code
stringlengths
114
1.05M
path
stringlengths
3
312
quality_prob
float64
0.5
0.99
learning_prob
float64
0.2
1
filename
stringlengths
3
168
kind
stringclasses
1 value
%% rufus_tokenize inserts expression terminators into a stream of tokens %% produced by rufus_raw_tokenize, based on the set of rules defined in the %% _Semicolons_ section of the language spec: %% %% 1. When the input is broken into tokens, a semicolon is automatically %% inserted into the token stream immediately after a line's final token if %% that token is: %% - An identifier. %% - An `atom`, `bool`, `float`, `int` or `string` literal. %% - One of the punctuation `)` or `}`. %% 2. To allow complex expressions to occupy a single line, a semicolon may be %% omitted before a closing `)` or `}`. -module(rufus_tokenize). -include_lib("rufus_type.hrl"). %% API exports -export([string/1]). %% API %% string tokenizes RufusText and inserts semicolon expression terminators based %% on a set of rules. Return values: %% - `{ok, Tokens}` with `Tokens` as a list of tokens, if tokenization is %% successful. %% - `{error, Reason}` if an error occurs. -spec string(rufus_text()) -> ok_tuple() | error_tuple(). string(RufusText) -> case rufus_raw_tokenize:string(RufusText) of {ok, Tokens, _Lines} -> insert_semicolons(discard_comments(Tokens)); {error, Reason, _LineNumber} -> {error, Reason} end. %% Private API make_semicolon_token(TokenLine) -> {';', TokenLine}. terminate(Acc = [{';', _TokenLine1} | _T], _TokenLine2) -> Acc; terminate(Acc, TokenLine) -> [make_semicolon_token(TokenLine) | Acc]. %% discard_comments discards comment tokens. -spec discard_comments(list(tuple())) -> list(tuple()). discard_comments(Tokens) -> discard_comments([], Tokens). discard_comments(Acc, [{comment, _TokenLine, _TokenChars} | T]) -> discard_comments(Acc, T); discard_comments(Acc, [Token | T]) -> discard_comments([Token | Acc], T); discard_comments(Acc, []) -> lists:reverse(Acc). %% insert_semicolons inserts `;` tokens after some `eol` tokens to terminate %% expressions. All `eol` tokens are discarded in the resulting list of tokens. -spec insert_semicolons(list(tuple())) -> {ok, list(tuple())}. insert_semicolons(Tokens) -> LastToken = undefined, SemicolonTerminatedTokens = insert_semicolons([], LastToken, Tokens), {ok, SemicolonTerminatedTokens}. %% insert_semicolons inserts a semicolon when the following tokens are the last %% on a line or the last in the source text. The `eol` token is always discarded. insert_semicolons(Acc, {identifier, TokenLine, _TokenChars}, [Token = {eol, _TokenLine} | T]) -> insert_semicolons(terminate(Acc, TokenLine), Token, T); insert_semicolons(Acc, {identifier, TokenLine, _TokenChars}, []) -> insert_semicolons(terminate(Acc, TokenLine), undefined, []); insert_semicolons(Acc, {atom_lit, TokenLine, _TokenChars}, [Token = {eol, _TokenLine} | T]) -> insert_semicolons(terminate(Acc, TokenLine), Token, T); insert_semicolons(Acc, {atom_lit, TokenLine, _TokenChars}, []) -> insert_semicolons(terminate(Acc, TokenLine), undefined, []); insert_semicolons(Acc, {bool_lit, TokenLine, _TokenChars}, [Token = {eol, _TokenLine} | T]) -> insert_semicolons(terminate(Acc, TokenLine), Token, T); insert_semicolons(Acc, {bool_lit, TokenLine, _TokenChars}, []) -> insert_semicolons(terminate(Acc, TokenLine), undefined, []); insert_semicolons(Acc, {float_lit, TokenLine, _TokenChars}, [Token = {eol, _TokenLine} | T]) -> insert_semicolons(terminate(Acc, TokenLine), Token, T); insert_semicolons(Acc, {float_lit, TokenLine, _TokenChars}, []) -> insert_semicolons(terminate(Acc, TokenLine), undefined, []); insert_semicolons(Acc, {int_lit, TokenLine, _TokenChars}, [Token = {eol, _TokenLine} | T]) -> insert_semicolons(terminate(Acc, TokenLine), Token, T); insert_semicolons(Acc, {int_lit, TokenLine, _TokenChars}, []) -> insert_semicolons(terminate(Acc, TokenLine), undefined, []); insert_semicolons(Acc, {string_lit, TokenLine, _TokenChars}, [Token = {eol, _TokenLine} | T]) -> insert_semicolons(terminate(Acc, TokenLine), Token, T); insert_semicolons(Acc, {string_lit, TokenLine, _TokenChars}, []) -> insert_semicolons(terminate(Acc, TokenLine), undefined, []); insert_semicolons(Acc, {')', TokenLine}, [Token = {eol, _TokenLine} | T]) -> insert_semicolons(terminate(Acc, TokenLine), Token, T); insert_semicolons(Acc, {')', TokenLine}, []) -> insert_semicolons(terminate(Acc, TokenLine), undefined, []); insert_semicolons(Acc, {'}', TokenLine}, [Token = {eol, _TokenLine} | T]) -> insert_semicolons(terminate(Acc, TokenLine), Token, T); insert_semicolons(Acc, {'}', TokenLine}, []) -> insert_semicolons(terminate(Acc, TokenLine), undefined, []); %% insert_semicolons discards `eol` tokens. insert_semicolons(Acc, _LastToken, [Token = {eol, _TokenLine} | T]) -> insert_semicolons(Acc, Token, T); %% insert_semicolons keeps all other tokens. insert_semicolons(Acc, _LastToken, [Token | T]) -> insert_semicolons([Token | Acc], Token, T); %% insert_semicolons terminates when all tokens have been processed. insert_semicolons(Acc, _LastToken, []) -> lists:reverse(Acc).
rf/src/rufus_tokenize.erl
0.610802
0.799951
rufus_tokenize.erl
starcoder
-module(morpheus_instrument). %% API exports -export([instrument_and_load/6, whitelist_func/3]). -include("morpheus_priv.hrl"). %%==================================================================== %% API functions %%==================================================================== %% Instrument a module with object code `ObjectCode` and module name `OriginalModule`. %% During the function, the instrumented module will be loaded, where: %% - The module is renamed to `NewModule` (NewModule is supposed to be globally unique) %% - call M:F(A...) is redirected to CtlMod:handle(OriginalModule, NewModule, call, M, F, A, _Ann) %% - apply(M, F, A) is redirected to CtlMod:handle(OriginalModule, NewModule, apply, M, F, A, _Ann) %% - T ! M is transformed to CtlMod:handle(OriginalModule, NewModule, call, erlang, send, [T, M], _Ann) %% - receive ... after ... end is transformed to case CtlMod:handle(OriginalModule, NewModule, 'receive', [PatFun, Timeout], _Ann) of ... end %% - Nif stubs in the new module are forwarding calls to the original stubs %% - _Ann is code annotation that is unused for now %% If success, {ok, Warnings} will be returned. %% Otherwise, failed will be returned instrument_and_load(CtlMod, CtlState, OriginalModule, NewModule, Filename, ObjectCode) -> ?DEBUG("instrument(~w, ~w, ~w, ~p, ...)", [CtlMod, OriginalModule, NewModule, Filename]), Core = get_core(ObjectCode), ?VERBOSE_DEBUG("~w: gonna instrument ~p", [?FUNCTION_NAME, Core]), _InstrResult = {{Original, Instrumented}, CtlState0, Nifs, Warnings} = instrument_core(CtlMod, CtlState, Core, OriginalModule, NewModule), case Filename of [] -> %% empty filename indicate dynamic module. So far there is no good way of handling that. ok; _ -> case Original of undefined -> {module, OriginalModule} = code:ensure_loaded(OriginalModule); _ -> ?DEBUG("Reload original module ~p", [OriginalModule]), load_module_from_core(OriginalModule, Filename, Original) end end, load_module_from_core(NewModule, Filename, Instrumented), {ok, CtlState0, Nifs, Warnings}. %%==================================================================== %% Internal functions %%==================================================================== load_module_from_core(Name, Filename, Core) -> ?DEBUG("load_module_from_core: ~p", [Name]), CompResult = compile:forms(Core, [from_core, report_errors, binary]), Binary = case CompResult of {ok, _, _Binary} -> _Binary; _Other -> ?WARNING("compile failed on module ~p:~n" " info: ~p~n" ,[Name, _Other]), error(load_module_failed) end, {module, Name} = code:load_binary(Name, Filename, Binary), ok. %% Some functions are inspired by Concuerror get_core(CodeOrFilename) -> {ok, {Module, [{abstract_code, ChunkInfo}]}} = beam_lib:chunks(CodeOrFilename, [abstract_code]), case ChunkInfo of {_, Chunk} -> {ok, Module, Core} = compile:forms(Chunk, [binary, to_core0]), Core; no_abstract_code -> {ok, {Module, [{compile_info, CompileInfo}]}} = beam_lib:chunks(CodeOrFilename, [compile_info]), {source, File} = proplists:lookup(source, CompileInfo), {options, CompileOptions} = proplists:lookup(options, CompileInfo), Filter = fun(Option) -> lists:member(element(1, Option), [d, i, parse_transform]) end, CleanOptions = lists:filter(Filter, CompileOptions), Options = [debug_info, report_errors, binary, to_core0|CleanOptions], {ok, Module, Core} = compile:file(File, Options), Core end. -record(ins_state, { nif_error_found :: boolean(), nifs :: [{atom(), integer()}], debug_counter :: integer() }). instrument_core(CtlMod, CtlState, Core, OriginalModule, NewModule) -> Opt = {CtlMod, OriginalModule, NewModule}, {R, {Opt, CtlState0, #ins_state{nifs = Nifs}, Warn}} = cerl_trees:mapfold(fun instrument_core_tree/2, {Opt, CtlState, #ins_state{nif_error_found = false, nifs = [], debug_counter = 0}, []}, Core), ToExport = case OriginalModule of erl_eval -> [{match_clause, 5} | Nifs]; _ -> Nifs end, case ToExport of [] -> {{undefined, R}, CtlState0, [], Warn}; _ -> %% XXX some module cannot be loaded twice due to on_load handler limitation ToReload = case code:is_sticky(OriginalModule) of false -> true; true -> code:unstick_mod(OriginalModule) end, case ToReload of true -> ?DEBUG("Reload original module of ~w with export list: ~p", [OriginalModule, ToExport]), {ModifiedOriginal, _} = cerl_trees:mapfold(fun export_fun/2, ToExport, Core), {{ModifiedOriginal, R}, CtlState0, %% Since NIFs is not handled specially anywhere except in the process entry (and it's rare!), simply pass [] %% IS#ins_state.nifs, [], Warn}; false -> ?INFO("Skip reloading module ~p", [OriginalModule]), {{undefined, R}, CtlState0, [], Warn} end end. export_fun(Tree, FunList) -> Type = cerl:type(Tree), case Type of module -> NewTree = cerl:update_c_module( Tree, cerl:module_name(Tree), lists:foldr(fun (V, Acc) -> [ cerl:c_var(V) | Acc ] end, [], FunList) ++ cerl:module_exports(Tree), cerl:module_attrs(Tree), cerl:module_defs(Tree)), {NewTree, FunList}; _ -> {Tree, FunList} end. %% Currently, we identify nif overrides by assuming that the nif stubs will call nif_error to generate error instrument_core_tree(Tree, {{CtlMod, OriginalModule, NewModule} = Opt, CtlState, IS, Warnings}) -> Type = cerl:type(Tree), {NewTreeAndMaybeWarn, CtlState0, NewIS} = case Type of module -> _OriginalModule = cerl:atom_val(cerl:module_name(Tree)), %% Ignore on_load property assuming it's only loading NIF binaries, %% and the original module already handled it. NewAttrs = lists:foldr( fun ({NameTree, ValueTree}, Acc) -> Name = cerl:concrete(NameTree), case Name of on_load -> Acc; _ -> [{NameTree, ValueTree} | Acc] end end, [], cerl:module_attrs(Tree)), {Nifs, ToExpose, DefsWithForwarding} = lists:foldr( fun ({NT, DT}, {NifsL, ToExposeL0, DefsL}) -> N = cerl:var_name(NT), {F, A} = N, Vars = cerl:fun_vars(DT), Body = cerl:fun_body(DT), IsNifStub = lists:member(nif_stub, cerl:get_ann(DT)), ToExposeL = case erlang:function_exported(CtlMod, to_expose, 4) andalso CtlMod:to_expose(CtlState, OriginalModule, F, A) of true -> [{F, A} | ToExposeL0]; false -> ToExposeL0 end, case N =/= {on_load, 0} andalso erlang:function_exported(CtlMod, to_override, 4) andalso CtlMod:to_override(CtlState, OriginalModule, F, A) of {true, Action} when Action =:= trace; Action =:= callback -> OrigName = case IsNifStub of true -> []; false -> list_to_atom(atom_to_list(F) ++ "$orig") end, InspectedBody = inspect( override , [ cerl:c_atom(Action) , cerl:c_atom(F) , cerl:c_atom(OrigName) , cerl:make_list(Vars) ] , Body, Opt), OverridedDef = cerl:update_c_fun(DT, Vars, InspectedBody), case IsNifStub of true -> {[N | NifsL], ToExposeL, [{NT, OverridedDef} | DefsL]}; false -> OrigNT = {OrigName, A}, {NifsL, [OrigNT | ToExposeL], [{NT, OverridedDef}, {cerl:c_var(OrigNT), DT} | DefsL]} end; false when N =:= {on_load, 0} -> ?INFO("ignored on_load/0", []), {NifsL, ToExposeL, DefsL}; false when IsNifStub -> UpdatedDef = case CtlMod:is_undet_nif(CtlState, OriginalModule, F, A) of true -> InspectedBody = inspect(undet_nif_stub, [cerl:c_atom(element(1, N)), cerl:make_list(Vars)], Body, Opt), cerl:update_c_fun(DT, Vars, InspectedBody); false -> cerl:update_c_fun( DT, Vars, cerl:update_tree(DT, call, [[cerl:c_atom(OriginalModule)], [cerl:c_atom(F)], Vars])) end, {[N | NifsL], ToExposeL, [{NT, UpdatedDef} | DefsL]}; false -> {NifsL, ToExposeL, [{NT, DT} | DefsL]} end end, {[], [], []}, cerl:module_defs(Tree)), NewExports = lists:usort(cerl:module_exports(Tree) ++ lists:foldr(fun (VN, Acc) -> [cerl:c_var(VN) | Acc] end, [], ToExpose)) -- [cerl:c_var({on_load, 0})] , {cerl:update_c_module( Tree, cerl:c_atom(NewModule), NewExports, NewAttrs, DefsWithForwarding), CtlState, IS#ins_state{nifs = Nifs}}; 'fun' -> case IS#ins_state.nif_error_found of true -> {cerl:add_ann([nif_stub], Tree), CtlState, IS#ins_state{nif_error_found = false}}; false -> {Tree, CtlState, IS} end; apply -> Op = cerl:apply_op(Tree), case cerl:is_c_fname(Op) of true -> case cerl:fname_id(Op) of %% Some modules (crypto I'm looking at you) wrap nif_error in their own function ... nif_stub_error -> {Tree, CtlState, IS#ins_state{nif_error_found = true}}; _ -> ?VERBOSE_DEBUG("~w: ignore c_fname apply ~p", [?FUNCTION_NAME, Op]), {Tree, CtlState, IS} end; false -> OldArgs = cerl:make_list(cerl:apply_args(Tree)), {inspect(apply, [Op, OldArgs], Tree, Opt), CtlState, IS} end; call -> Module = cerl:call_module(Tree), Name = cerl:call_name(Tree), Args = cerl:call_args(Tree), Arity = length(Args), ?DEBUG("~w: call ~p", [?FUNCTION_NAME, {Module, Name, Arity}]), {Tree1, ToInspect, _CtlState, IS0} = case cerl:is_literal(Module) andalso cerl:is_literal(Name) andalso {cerl:concrete(Module), cerl:concrete(Name), Arity} of false -> {Tree, true, CtlState, IS}; {erlang, nif_error, _} -> {Tree, false, CtlState, IS#ins_state{nif_error_found = true}}; {erlang, node, 0} -> %% HACK: prevent compiling erlang:node() into node vm instruction {cerl:update_tree(Tree, call, [[cerl:c_atom(erlang)], [cerl:c_atom(node)], [cerl:update_tree(Tree, call, [[cerl:c_atom(erlang)], [cerl:c_atom(self)], []])]]), false, CtlState, IS}; {M, F, A} -> case whitelist_func(M, F, A) of true -> {Tree, false, CtlState, IS}; false -> {_A, _B} = CtlMod:to_handle( CtlState, OriginalModule, NewModule, {call, erlang, F, A}), {Tree, _A, _B, IS} end end, ?VERBOSE_DEBUG("~w: to instrument? ~p", [?FUNCTION_NAME, ToInspect]), case ToInspect of true -> {inspect(call, [Module, Name, cerl:make_list(Args)], Tree1, Opt), _CtlState, IS0}; false -> {Tree1, _CtlState, IS0} end; 'receive' -> PatFun = receive_matching_fun(Tree), Clauses = cerl:receive_clauses(Tree), Timeout = cerl:receive_timeout(Tree), TimeoutAction = cerl:receive_action(Tree), InspectedMsgTree = inspect('receive', [PatFun, Timeout], Tree, Opt), TransformedClauses = transform_receive_clauses(Clauses, TimeoutAction), Tree1 = cerl:update_tree(Tree, 'case', [[InspectedMsgTree], TransformedClauses]), {Tree1, CtlState, IS}; _ -> {Tree, CtlState, IS} end, {NewTree, NewWarnings} = case NewTreeAndMaybeWarn of {warn, NT, W} -> {NT, [W|Warnings]}; _ -> {NewTreeAndMaybeWarn, Warnings} end, {NewTree, {Opt, CtlState0, NewIS, NewWarnings}}. inspect(Tag, Args, Tree, {CtlMod, OriginalModule, NewModule}) -> CTag = cerl:c_atom(Tag), CArgs = cerl:make_list(Args), cerl:update_tree(Tree, call, [[cerl:c_atom(CtlMod)], [cerl:c_atom(handle)], [cerl:c_atom(OriginalModule), cerl:c_atom(NewModule), CTag, CArgs, cerl:abstract(cerl:get_ann(Tree))]]). receive_matching_fun(Tree) -> Msg = cerl:c_var(message), case cerl:receive_clauses(Tree) of [] -> %% this could happen in receive after X -> Y end cerl:ann_c_atom(cerl:get_ann(Tree), undefined); Clauses -> PatClauses = extract_patterns(Clauses), Body = cerl:update_tree(Tree, 'case', [[Msg], PatClauses]), cerl:update_tree(Tree, 'fun', [[Msg], [Body]]) end. extract_patterns(Clauses) -> extract_patterns(Clauses, []). extract_patterns([], Acc) -> Pat = [cerl:c_var(message)], Guard = cerl:c_atom(true), Body = cerl:c_atom(false), lists:reverse([cerl:c_clause(Pat, Guard, Body)|Acc]); extract_patterns([Tree|Rest], Acc) -> Body = cerl:c_atom(true), Pats = cerl:clause_pats(Tree), Guard = cerl:clause_guard(Tree), extract_patterns(Rest, [cerl:update_c_clause(Tree, Pats, Guard, Body)|Acc]). transform_receive_clauses(Clauses, TimeoutAction) -> transform_receive_clauses(Clauses, [], TimeoutAction). transform_receive_clauses([], Acc, TimeoutAction) -> lists:reverse([cerl:update_c_clause(TimeoutAction, %%%% this wil trigger failure in OTP-22-dev %% [cerl:c_atom(timeout)], [cerl:c_var(anything)], cerl:c_atom(true), TimeoutAction) | Acc]); transform_receive_clauses([Tree | Rest], Acc, TimeoutAction) -> [Pat] = cerl:clause_pats(Tree), Guard = cerl:clause_guard(Tree), Body = cerl:clause_body(Tree), transform_receive_clauses( Rest, [cerl:update_c_clause( Tree, [cerl:c_cons(cerl:c_atom(message), Pat)], Guard, Body) | Acc], TimeoutAction). whitelist_func(lists, _, _) -> true; whitelist_func(queue, _, _) -> true; whitelist_func(proplists, _, _) -> true; whitelist_func(dict, _, _) -> true; whitelist_func(orddict, _, _) -> true; whitelist_func(math, _, _) -> true; whitelist_func(maps, _, _) -> true; whitelist_func(gb_sets, _, _) -> true; whitelist_func(gb_trees, _, _) -> true; whitelist_func(sets, _, _) -> true; whitelist_func(re, _, _) -> true; whitelist_func(binary, _, _) -> true; whitelist_func(string, _, _) -> true; whitelist_func(unicode, _, _) -> true; whitelist_func(crypto, _, _) -> true; whitelist_func(io_lib, _, _) -> true; %% XXX could not instrument zlib properly - seems nif stub related whitelist_func(zlib, _, _) -> true; whitelist_func(filename, _, _) -> true; whitelist_func(erl_internal, _, _) ->true; whitelist_func(erl_parse, _, _) -> true; whitelist_func(erl_lint, _, _) -> true; whitelist_func(erl_prim_loader, _, _) -> true; whitelist_func(erl_anno, _, _) -> true; whitelist_func(erl_scan, _, _) -> true; whitelist_func(erts_internal, _, _) -> true; whitelist_func(erlang, error, _) -> true; whitelist_func(erlang, F, A) -> (erl_internal:guard_bif(F, A) orelse erl_internal:arith_op(F, A) orelse erl_internal:bool_op(F, A) orelse erl_internal:comp_op(F, A)); whitelist_func(_, _, _) -> false.
src/morpheus_instrument.erl
0.570212
0.47025
morpheus_instrument.erl
starcoder
% Licensed under the Apache License, Version 2.0 (the "License"); you may not % use this file except in compliance with the License. You may obtain a copy of % the License at % % http://www.apache.org/licenses/LICENSE-2.0 % % Unless required by applicable law or agreed to in writing, software % distributed under the License is distributed on an "AS IS" BASIS, WITHOUT % WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the % License for the specific language governing permissions and limitations under % the License. -module(couch_task_status). -behaviour(gen_server). % This module is used to track the status of long running tasks. % Long running tasks register themselves, via a call to add_task/1, and then % update their status properties via update/1. The status of a task is a % list of properties. Each property is a tuple, with the first element being % either an atom or a binary and the second element must be an EJSON value. When % a task updates its status, it can override some or all of its properties. % The properties {started_on, UnitTimestamp}, {updated_on, UnixTimestamp} and % {pid, ErlangPid} are automatically added by this module. % When a tracked task dies, its status will be automatically removed from % memory. To get the tasks list, call the all/0 function. -export([start_link/0, stop/0]). -export([all/0, add_task/1, update/1, get/1, set_update_frequency/1]). -export([is_task_added/0]). -export([init/1, terminate/2, code_change/3]). -export([handle_call/3, handle_cast/2, handle_info/2]). -include("couch_db.hrl"). -define(set(L, K, V), lists:keystore(K, 1, L, {K, V})). start_link() -> gen_server:start_link({local, ?MODULE}, ?MODULE, [], []). stop() -> gen_server:cast(?MODULE, stop). all() -> gen_server:call(?MODULE, all, infinity). add_task(Props) -> put(task_status_update, {{0, 0, 0}, 0}), Ts = timestamp(), TaskProps = lists:ukeysort( 1, [{started_on, Ts}, {updated_on, Ts} | Props]), put(task_status_props, TaskProps), gen_server:call(?MODULE, {add_task, TaskProps}, infinity). is_task_added() -> undefined /= erlang:get(task_status_props). set_update_frequency(Msecs) -> put(task_status_update, {{0, 0, 0}, Msecs * 1000}). update(Props) -> MergeProps = lists:ukeysort(1, Props), TaskProps = lists:ukeymerge(1, MergeProps, erlang:get(task_status_props)), put(task_status_props, TaskProps), maybe_persist(TaskProps). get(Props) when is_list(Props) -> TaskProps = erlang:get(task_status_props), [couch_util:get_value(P, TaskProps) || P <- Props]; get(Prop) -> TaskProps = erlang:get(task_status_props), couch_util:get_value(Prop, TaskProps). maybe_persist(TaskProps0) -> {LastUpdateTime, Frequency} = erlang:get(task_status_update), case timer:now_diff(Now = now(), LastUpdateTime) >= Frequency of true -> put(task_status_update, {Now, Frequency}), TaskProps = ?set(TaskProps0, updated_on, timestamp(Now)), gen_server:cast(?MODULE, {update_status, self(), TaskProps}); false -> ok end. init([]) -> % read configuration settings and register for configuration changes ets:new(?MODULE, [ordered_set, protected, named_table]), {ok, nil}. terminate(_Reason,_State) -> ok. handle_call({add_task, TaskProps}, {From, _}, Server) -> case ets:lookup(?MODULE, From) of [] -> true = ets:insert(?MODULE, EtsTuple = {From, TaskProps}), erlang:monitor(process, From), gen_event:notify(couch_task_events, {created, EtsTuple}), {reply, ok, Server}; [_] -> {reply, {add_task_error, already_registered}, Server} end; handle_call(all, _, Server) -> All = [ [{pid, ?l2b(pid_to_list(Pid))} | TaskProps] || {Pid, TaskProps} <- ets:tab2list(?MODULE) ], {reply, All, Server}. handle_cast({update_status, Pid, NewProps}, Server) -> case ets:lookup(?MODULE, Pid) of [{Pid, _CurProps}] -> ?LOG_DEBUG("New task status for ~p: ~p", [Pid, NewProps]), true = ets:insert(?MODULE, {Pid, NewProps}); _ -> % Task finished/died in the meanwhile and we must have received % a monitor message before this call - ignore. ok end, {noreply, Server}; handle_cast(stop, State) -> {stop, normal, State}. handle_info({'DOWN', _MonitorRef, _Type, Pid, _Info}, Server) -> gen_event:notify(couch_task_events, {deleted, Pid}), ets:delete(?MODULE, Pid), {noreply, Server}. code_change(_OldVsn, State, _Extra) -> {ok, State}. timestamp() -> timestamp(now()). timestamp({Mega, Secs, _}) -> Mega * 1000000 + Secs.
src/couchdb/couch_task_status.erl
0.693058
0.442456
couch_task_status.erl
starcoder
%% This Source Code Form is subject to the terms of the Mozilla Public %% License, v. 2.0. If a copy of the MPL was not distributed with this %% file, You can obtain one at https://mozilla.org/MPL/2.0/. %% %% Copyright (c) 2021-2022 VMware, Inc. or its affiliates. All rights reserved. %% %% @doc Khepri API for transactional queries and updates. %% %% Transactions are anonymous functions which take no arguments, much like %% what Mnesia supports. However, unlike with Mnesia, transaction functions in %% Khepri are restricted: %% %% <ul> %% <li>Calls to BIFs and other functions is limited to a set of whitelisted %% APIs. See {@link is_remote_call_valid/3} for the complete list.</li> %% <li>Sending or receiving messages is denied.</li> %% </ul> %% %% The reason is that the transaction function must always have the exact same %% outcome given its inputs. Indeed, the transaction function is executed on %% every Ra cluster members participating in the consensus. The function must %% therefore modify the Khepri state (the database) identically on all Ra %% members. This is also true for Ra members joining the cluster later or %% catching up after a network partition. %% %% To achieve that: %% <ol> %% <li>The code of the transaction function is extracted from the its initial %% Erlang module. This way, the transaction function does not depend on the %% initial module availability and is not affected by a module reload. See %% {@link khepri_fun})</li> %% <li>The code is verified to make sure it does not perform any denied %% operations.</li> %% <li>The extracted transaction function is stored as a Khepri state machine %% command in the Ra journal to be replicated on all Ra members.</li> %% </ol> -module(khepri_tx). -include_lib("stdlib/include/assert.hrl"). -include("include/khepri.hrl"). -include("src/internal.hrl"). -include("src/khepri_machine.hrl"). %% IMPORTANT: When adding a new khepri_tx function to be used inside a %% transaction function: %% 1. The function must be added to the whitelist in %% `is_remote_call_valid()' in this file. %% 2. If the function modifies the tree, it must be handled in %% `is_standalone_fun_still_needed()' is this file too. -export([put/2, put/3, get/1, get/2, exists/1, has_data/1, list/1, find/2, delete/1, abort/1, is_transaction/0]). %% For internal user only. -export([to_standalone_fun/2, run/3]). -compile({no_auto_import, [get/1, put/2, erase/1]}). -type tx_fun_result() :: any() | no_return(). -type tx_fun() :: fun(() -> tx_fun_result()). -type tx_fun_bindings() :: #{Name :: atom() => Value :: any()}. -type tx_abort() :: {aborted, any()}. -type tx_props() :: #{allow_updates := boolean()}. -export_type([tx_fun/0, tx_fun_bindings/0, tx_fun_result/0, tx_abort/0]). -spec put(PathPattern, Payload) -> Result when PathPattern :: khepri_path:pattern(), Payload :: khepri_machine:payload(), Result :: khepri_machine:result(). %% @doc Creates or modifies a specific tree node in the tree structure. put(PathPattern, Payload) -> put(PathPattern, Payload, #{}). -spec put(PathPattern, Payload, Extra) -> Result when PathPattern :: khepri_path:pattern(), Payload :: khepri_machine:payload(), Extra :: #{keep_while => khepri_condition:keep_while()}, Result :: khepri_machine:result(). %% @doc Creates or modifies a specific tree node in the tree structure. put(PathPattern, Payload, Extra) when ?IS_KHEPRI_PAYLOAD(Payload) -> ensure_path_pattern_is_valid(PathPattern), ensure_updates_are_allowed(), {State, SideEffects} = get_tx_state(), Ret = khepri_machine:insert_or_update_node( State, PathPattern, Payload, Extra), case Ret of {NewState, Result, NewSideEffects} -> set_tx_state(NewState, SideEffects ++ NewSideEffects); {NewState, Result} -> set_tx_state(NewState, SideEffects) end, Result; put(PathPattern, Payload, _Extra) -> abort({invalid_payload, PathPattern, Payload}). get(PathPattern) -> get(PathPattern, #{}). get(PathPattern, Options) -> ensure_path_pattern_is_valid(PathPattern), {#khepri_machine{root = Root}, _SideEffects} = get_tx_state(), khepri_machine:find_matching_nodes(Root, PathPattern, Options). -spec exists(Path) -> Exists when Path :: khepri_path:pattern(), Exists :: boolean(). exists(Path) -> case get(Path, #{expect_specific_node => true}) of {ok, _} -> true; _ -> false end. -spec has_data(Path) -> HasData when Path :: khepri_path:pattern(), HasData :: boolean(). has_data(Path) -> case get(Path, #{expect_specific_node => true}) of {ok, Result} -> [NodeProps] = maps:values(Result), maps:is_key(data, NodeProps); _ -> false end. list(Path) -> Path1 = Path ++ [?STAR], get(Path1). find(Path, Condition) -> Condition1 = #if_all{conditions = [?STAR_STAR, Condition]}, Path1 = Path ++ [Condition1], get(Path1). delete(PathPattern) -> ensure_path_pattern_is_valid(PathPattern), ensure_updates_are_allowed(), {State, SideEffects} = get_tx_state(), Ret = khepri_machine:delete_matching_nodes(State, PathPattern), case Ret of {NewState, Result, NewSideEffects} -> set_tx_state(NewState, SideEffects ++ NewSideEffects); {NewState, Result} -> set_tx_state(NewState, SideEffects) end, Result. -spec abort(Reason) -> no_return() when Reason :: any(). abort(Reason) -> throw({aborted, Reason}). -spec is_transaction() -> boolean(). is_transaction() -> StateAndSideEffects = erlang:get(?TX_STATE_KEY), case StateAndSideEffects of {#khepri_machine{}, _SideEffects} -> true; _ -> false end. -spec to_standalone_fun(Fun, ReadWrite) -> StandaloneFun | no_return() when Fun :: fun(), ReadWrite :: ro | rw | auto, StandaloneFun :: khepri_fun:standalone_fun(). to_standalone_fun(Fun, ReadWrite) when is_function(Fun, 0) andalso (ReadWrite =:= auto orelse ReadWrite =:= rw) -> Options = #{ensure_instruction_is_permitted => fun ensure_instruction_is_permitted/1, should_process_function => fun should_process_function/4, is_standalone_fun_still_needed => fun(Params) -> is_standalone_fun_still_needed(Params, ReadWrite) end}, try khepri_fun:to_standalone_fun(Fun, Options) catch throw:Error -> throw({invalid_tx_fun, Error}) end; to_standalone_fun(Fun, ro) -> Fun. ensure_instruction_is_permitted({allocate, _, _}) -> ok; ensure_instruction_is_permitted({allocate_zero, _, _}) -> ok; ensure_instruction_is_permitted({allocate_heap, _, _, _}) -> ok; ensure_instruction_is_permitted({apply, _}) -> throw(dynamic_apply_denied); ensure_instruction_is_permitted({apply_last, _, _}) -> throw(dynamic_apply_denied); ensure_instruction_is_permitted({badmatch, _}) -> ok; ensure_instruction_is_permitted({bif, Bif, _, Args, _}) -> Arity = length(Args), ensure_bif_is_valid(Bif, Arity); ensure_instruction_is_permitted({bs_add, _, _, _}) -> ok; ensure_instruction_is_permitted({bs_append, _, _, _, _, _, _, _, _}) -> ok; ensure_instruction_is_permitted({bs_init2, _, _, _, _, _, _}) -> ok; ensure_instruction_is_permitted({BsPutSomething, _, _, _, _, _}) when BsPutSomething =:= bs_put_binary orelse BsPutSomething =:= bs_put_integer -> ok; ensure_instruction_is_permitted({bs_put_string, _, _}) -> ok; ensure_instruction_is_permitted({bs_get_position, _, _, _}) -> ok; ensure_instruction_is_permitted({bs_set_position, _, _}) -> ok; ensure_instruction_is_permitted({bs_get_tail, _, _, _}) -> ok; ensure_instruction_is_permitted({bs_start_match4, _, _, _, _}) -> ok; ensure_instruction_is_permitted({Call, _, _}) when Call =:= call orelse Call =:= call_only orelse Call =:= call_ext orelse Call =:= call_ext_only -> ok; ensure_instruction_is_permitted({Call, _, _, _}) when Call =:= call_last orelse Call =:= call_ext_last -> ok; ensure_instruction_is_permitted({call_fun, _}) -> ok; ensure_instruction_is_permitted({case_end, _}) -> ok; ensure_instruction_is_permitted({'catch', _, _}) -> ok; ensure_instruction_is_permitted({catch_end, _}) -> ok; ensure_instruction_is_permitted({deallocate, _}) -> ok; ensure_instruction_is_permitted({func_info, _, _, _}) -> ok; ensure_instruction_is_permitted({gc_bif, Bif, _, Arity, _, _}) -> ensure_bif_is_valid(Bif, Arity); ensure_instruction_is_permitted({get_hd, _, _}) -> ok; ensure_instruction_is_permitted({get_tl, _, _}) -> ok; ensure_instruction_is_permitted({get_tuple_element, _, _, _}) -> ok; ensure_instruction_is_permitted({get_map_elements, _, _, _}) -> ok; ensure_instruction_is_permitted({get_list, _, _, _}) -> ok; ensure_instruction_is_permitted(if_end) -> ok; ensure_instruction_is_permitted({init, _}) -> ok; ensure_instruction_is_permitted({init_yregs, _}) -> ok; ensure_instruction_is_permitted({jump, _}) -> ok; ensure_instruction_is_permitted({move, _, _}) -> ok; ensure_instruction_is_permitted({loop_rec, _, _}) -> throw(receiving_message_denied); ensure_instruction_is_permitted({loop_rec_env, _}) -> throw(receiving_message_denied); ensure_instruction_is_permitted({make_fun2, _, _, _, _}) -> ok; ensure_instruction_is_permitted({make_fun3, _, _, _, _, _}) -> ok; ensure_instruction_is_permitted({put_list, _, _, _}) -> ok; ensure_instruction_is_permitted({put_map_assoc, _, _, _, _, _}) -> ok; ensure_instruction_is_permitted({put_tuple2, _, _}) -> ok; ensure_instruction_is_permitted(raw_raise) -> ok; ensure_instruction_is_permitted(remove_message) -> throw(receiving_message_denied); ensure_instruction_is_permitted(return) -> ok; ensure_instruction_is_permitted(send) -> throw(sending_message_denied); ensure_instruction_is_permitted({select_tuple_arity, _, _, {list, _}}) -> ok; ensure_instruction_is_permitted({select_val, _, _, {list, _}}) -> ok; ensure_instruction_is_permitted({set_tuple_element, _, _, _}) -> ok; ensure_instruction_is_permitted({swap, _, _}) -> ok; ensure_instruction_is_permitted({test, _, _, _}) -> ok; ensure_instruction_is_permitted({test, _, _, _, _}) -> ok; ensure_instruction_is_permitted({test, _, _, _, _, _}) -> ok; ensure_instruction_is_permitted({test_heap, _, _}) -> ok; ensure_instruction_is_permitted({trim, _, _}) -> ok; ensure_instruction_is_permitted({'try', _, _}) -> ok; ensure_instruction_is_permitted({try_end, _}) -> ok; ensure_instruction_is_permitted({try_case, _}) -> ok; ensure_instruction_is_permitted(Unknown) -> throw({unknown_instruction, Unknown}). should_process_function(Module, Name, Arity, FromModule) -> ShouldCollect = khepri_utils:should_collect_code_for_module(Module), case ShouldCollect of true -> case Module of FromModule -> true; _ -> _ = code:ensure_loaded(Module), case erlang:function_exported(Module, Name, Arity) of true -> true; false -> throw({call_to_unexported_function, {Module, Name, Arity}}) end end; false -> ensure_call_is_valid(Module, Name, Arity), false end. ensure_call_is_valid(Module, Name, Arity) -> case is_remote_call_valid(Module, Name, Arity) of true -> ok; false -> throw({call_denied, {Module, Name, Arity}}) end. ensure_bif_is_valid(Bif, Arity) -> try ensure_call_is_valid(erlang, Bif, Arity) catch throw:{call_denied, {erlang, Bif, Arity}} -> throw({call_denied, {Bif, Arity}}) end. is_remote_call_valid(khepri, no_payload, 0) -> true; is_remote_call_valid(khepri, data_payload, 1) -> true; is_remote_call_valid(khepri_tx, put, _) -> true; is_remote_call_valid(khepri_tx, get, _) -> true; is_remote_call_valid(khepri_tx, exists, _) -> true; is_remote_call_valid(khepri_tx, has_data, _) -> true; is_remote_call_valid(khepri_tx, list, _) -> true; is_remote_call_valid(khepri_tx, find, _) -> true; is_remote_call_valid(khepri_tx, delete, _) -> true; is_remote_call_valid(khepri_tx, abort, _) -> true; is_remote_call_valid(khepri_tx, is_transaction, _) -> true; is_remote_call_valid(_, module_info, _) -> false; is_remote_call_valid(erlang, abs, _) -> true; is_remote_call_valid(erlang, adler32, _) -> true; is_remote_call_valid(erlang, adler32_combine, _) -> true; is_remote_call_valid(erlang, append_element, _) -> true; is_remote_call_valid(erlang, atom_to_binary, _) -> true; is_remote_call_valid(erlang, atom_to_list, _) -> true; is_remote_call_valid(erlang, binary_part, _) -> true; is_remote_call_valid(erlang, binary_to_atom, _) -> true; is_remote_call_valid(erlang, binary_to_float, _) -> true; is_remote_call_valid(erlang, binary_to_integer, _) -> true; is_remote_call_valid(erlang, binary_to_list, _) -> true; is_remote_call_valid(erlang, binary_to_term, _) -> true; is_remote_call_valid(erlang, bit_size, _) -> true; is_remote_call_valid(erlang, bitstring_to_list, _) -> true; is_remote_call_valid(erlang, byte_size, _) -> true; is_remote_call_valid(erlang, ceil, _) -> true; is_remote_call_valid(erlang, crc32, _) -> true; is_remote_call_valid(erlang, crc32_combine, _) -> true; is_remote_call_valid(erlang, delete_element, _) -> true; is_remote_call_valid(erlang, element, _) -> true; is_remote_call_valid(erlang, error, _) -> true; is_remote_call_valid(erlang, exit, _) -> true; is_remote_call_valid(erlang, external_size, _) -> true; is_remote_call_valid(erlang, float, _) -> true; is_remote_call_valid(erlang, float_to_binary, _) -> true; is_remote_call_valid(erlang, float_to_list, _) -> true; is_remote_call_valid(erlang, hd, _) -> true; is_remote_call_valid(erlang, insert_element, _) -> true; is_remote_call_valid(erlang, integer_to_binary, _) -> true; is_remote_call_valid(erlang, integer_to_list, _) -> true; is_remote_call_valid(erlang, iolist_size, _) -> true; is_remote_call_valid(erlang, iolist_to_binary, _) -> true; is_remote_call_valid(erlang, iolist_to_iovec, _) -> true; is_remote_call_valid(erlang, is_atom, _) -> true; is_remote_call_valid(erlang, is_binary, _) -> true; is_remote_call_valid(erlang, is_bitstring, _) -> true; is_remote_call_valid(erlang, is_boolean, _) -> true; is_remote_call_valid(erlang, is_float, _) -> true; is_remote_call_valid(erlang, is_integer, _) -> true; is_remote_call_valid(erlang, is_list, _) -> true; is_remote_call_valid(erlang, is_map, _) -> true; is_remote_call_valid(erlang, is_map_key, _) -> true; is_remote_call_valid(erlang, is_number, _) -> true; is_remote_call_valid(erlang, is_pid, _) -> true; is_remote_call_valid(erlang, is_record, _) -> true; is_remote_call_valid(erlang, is_reference, _) -> true; is_remote_call_valid(erlang, is_tuple, _) -> true; is_remote_call_valid(erlang, length, _) -> true; is_remote_call_valid(erlang, list_to_atom, _) -> true; is_remote_call_valid(erlang, list_to_binary, _) -> true; is_remote_call_valid(erlang, list_to_bitstring, _) -> true; is_remote_call_valid(erlang, list_to_float, _) -> true; is_remote_call_valid(erlang, list_to_integer, _) -> true; is_remote_call_valid(erlang, list_to_pid, _) -> true; is_remote_call_valid(erlang, list_to_tuple, _) -> true; is_remote_call_valid(erlang, make_tuple, _) -> true; is_remote_call_valid(erlang, map_get, _) -> true; is_remote_call_valid(erlang, map_size, _) -> true; is_remote_call_valid(erlang, max, _) -> true; is_remote_call_valid(erlang, md5, _) -> true; is_remote_call_valid(erlang, md5_final, _) -> true; is_remote_call_valid(erlang, md5_init, _) -> true; is_remote_call_valid(erlang, md5_update, _) -> true; is_remote_call_valid(erlang, min, _) -> true; is_remote_call_valid(erlang, 'not', _) -> true; is_remote_call_valid(erlang, phash2, _) -> true; is_remote_call_valid(erlang, pid_to_list, _) -> true; is_remote_call_valid(erlang, raise, _) -> true; is_remote_call_valid(erlang, round, _) -> true; is_remote_call_valid(erlang, setelement, _) -> true; is_remote_call_valid(erlang, size, _) -> true; is_remote_call_valid(erlang, split_binary, _) -> true; %% FIXME: What about changes to the marshalling code between versions of %% Erlang? is_remote_call_valid(erlang, term_to_binary, _) -> true; is_remote_call_valid(erlang, term_to_iovec, _) -> true; is_remote_call_valid(erlang, throw, _) -> true; is_remote_call_valid(erlang, tl, _) -> true; is_remote_call_valid(erlang, tuple_size, _) -> true; is_remote_call_valid(erlang, tuple_to_list, _) -> true; is_remote_call_valid(erlang, '++', _) -> true; is_remote_call_valid(erlang, '--', _) -> true; is_remote_call_valid(erlang, '+', _) -> true; is_remote_call_valid(erlang, '>=', _) -> true; is_remote_call_valid(erlang, '=<', _) -> true; is_remote_call_valid(erlang, '>', _) -> true; is_remote_call_valid(erlang, '<', _) -> true; is_remote_call_valid(erlang, '==', _) -> true; is_remote_call_valid(erlang, '/=', _) -> true; is_remote_call_valid(erlang, '=:=', _) -> true; is_remote_call_valid(erlang, '=/=', _) -> true; is_remote_call_valid(dict, _, _) -> true; is_remote_call_valid(io_lib, format, _) -> true; is_remote_call_valid(lists, _, _) -> true; is_remote_call_valid(logger, alert, _) -> true; is_remote_call_valid(logger, critical, _) -> true; is_remote_call_valid(logger, debug, _) -> true; is_remote_call_valid(logger, emergency, _) -> true; is_remote_call_valid(logger, error, _) -> true; is_remote_call_valid(logger, info, _) -> true; is_remote_call_valid(logger, notice, _) -> true; is_remote_call_valid(logger, warning, _) -> true; is_remote_call_valid(maps, _, _) -> true; is_remote_call_valid(orddict, _, _) -> true; is_remote_call_valid(ordsets, _, _) -> true; is_remote_call_valid(proplists, _, _) -> true; is_remote_call_valid(re, compile, _) -> true; is_remote_call_valid(re, inspect, _) -> true; is_remote_call_valid(re, replace, _) -> true; is_remote_call_valid(re, run, _) -> true; is_remote_call_valid(re, split, _) -> true; is_remote_call_valid(sets, _, _) -> true; is_remote_call_valid(string, _, _) -> true; is_remote_call_valid(unicode, _, _) -> true; is_remote_call_valid(_, _, _) -> false. is_standalone_fun_still_needed(_, rw) -> true; is_standalone_fun_still_needed(#{calls := Calls}, auto) -> ReadWrite = case Calls of #{{khepri_tx, put, 2} := _} -> rw; #{{khepri_tx, put, 3} := _} -> rw; #{{khepri_tx, delete, 1} := _} -> rw; _ -> ro end, ReadWrite =:= rw. -spec run(State, Fun, AllowUpdates) -> Ret when State :: khepri_machine:state(), Fun :: tx_fun(), AllowUpdates :: boolean(), Ret :: {State, tx_fun_result() | Exception, SideEffects}, Exception :: {exception, Class, Reason, Stacktrace}, Class :: error | exit | throw, Reason :: any(), Stacktrace :: list(), SideEffects :: ra_machine:effects(). %% @private run(State, Fun, AllowUpdates) -> SideEffects = [], TxProps = #{allow_updates => AllowUpdates}, NoState = erlang:put(?TX_STATE_KEY, {State, SideEffects}), NoProps = erlang:put(?TX_PROPS, TxProps), ?assertEqual(undefined, NoState), ?assertEqual(undefined, NoProps), try Ret = Fun(), {NewState, NewSideEffects} = erlang:erase(?TX_STATE_KEY), NewTxProps = erlang:erase(?TX_PROPS), ?assert(is_record(NewState, khepri_machine)), ?assertEqual(TxProps, NewTxProps), {NewState, Ret, NewSideEffects} catch Class:Reason:Stacktrace -> _ = erlang:erase(?TX_STATE_KEY), _ = erlang:erase(?TX_PROPS), Exception = {exception, Class, Reason, Stacktrace}, {State, Exception, []} end. -spec get_tx_state() -> {State, SideEffects} when State :: khepri_machine:state(), SideEffects :: ra_machine:effects(). %% @private get_tx_state() -> StateAndSideEffects = {#khepri_machine{}, _SideEffects} = erlang:get(?TX_STATE_KEY), StateAndSideEffects. -spec set_tx_state(State, SideEffects) -> ok when State :: khepri_machine:state(), SideEffects :: ra_machine:effects(). %% @private set_tx_state(#khepri_machine{} = NewState, SideEffects) -> _ = erlang:put(?TX_STATE_KEY, {NewState, SideEffects}), ok. -spec get_tx_props() -> TxProps when TxProps :: tx_props(). %% @private get_tx_props() -> erlang:get(?TX_PROPS). -spec ensure_path_pattern_is_valid(PathPattern) -> ok | no_return() when PathPattern :: khepri_path:pattern(). ensure_path_pattern_is_valid(PathPattern) -> case khepri_path:is_valid(PathPattern) of true -> ok; {false, Path} -> abort({invalid_path, Path}) end. -spec ensure_updates_are_allowed() -> ok | no_return(). %% @private ensure_updates_are_allowed() -> case get_tx_props() of #{allow_updates := true} -> ok; #{allow_updates := false} -> abort(store_update_denied) end.
src/khepri_tx.erl
0.691706
0.415373
khepri_tx.erl
starcoder
-module(coello_basic_spec). -include_lib("espec/include/espec.hrl"). -include_lib("hamcrest/include/hamcrest.hrl"). -include_lib("amqp_client/include/amqp_client.hrl"). spec() -> before_all(fun() -> meck:new([amqp_channel]), meck:expect(amqp_channel, cast, 2, ok), meck:expect(amqp_channel, cast, 3, ok) end), after_all(fun() -> meck:unload([amqp_channel]) end), describe("publish/4", fun() -> it("should publish a binary message", fun()-> Data = crypto:rand_bytes(30), Msg = #amqp_msg{payload = Data }, Method = #'basic.publish'{ exchange = exchange, routing_key = routing_key}, ok = coello_basic:publish(channel, Data, exchange,routing_key), assert_that(meck:called(amqp_channel, cast, [channel, Method, Msg]), is(true)) end), it("should publish a text message", fun()-> Msg = #amqp_msg{payload = <<"abc">> }, Method = #'basic.publish'{ exchange = exchange, routing_key = routing_key}, ok = coello_basic:publish(channel, "abc", exchange, routing_key), assert_that(meck:called(amqp_channel, cast, [channel, Method, Msg]), is(true)) end) end), describe("publish/5", fun() -> it("should publish a binary message", fun()-> ReplyTo = <<"respon<PASSWORD>">>, Data = crypto:rand_bytes(30), Msg = #amqp_msg{payload = Data, props = #'P_basic'{reply_to = ReplyTo} }, Method = #'basic.publish'{ exchange = exchange, routing_key = routing_key}, ok = coello_basic:publish(channel, Data, exchange, routing_key, ReplyTo), assert_that(meck:called(amqp_channel, cast, [channel, Method, Msg]), is(true)) end), it("should publish a text message", fun() -> Data = "abc", ReplyTo = <<"respondeaqui">>, Msg = #amqp_msg{payload = <<"abc">>, props = #'P_basic'{reply_to = ReplyTo} }, Method = #'basic.publish'{ exchange = exchange, routing_key = routing_key}, ok = coello_basic:publish(channel, Data, exchange, routing_key, ReplyTo), assert_that(meck:called(amqp_channel, cast, [channel, Method, Msg]), is(true)) end) end), describe("consume/3", fun() -> it("should consume messages and invoke the passed in callback", fun()-> meck:expect(amqp_channel, subscribe, 3, #'basic.consume_ok'{consumer_tag = 1234}), QueueName = <<"queue">>, Method = #'basic.consume'{ queue = QueueName}, Pid = self(), {ConsumerPid, _} = coello_basic:consume(channel, QueueName, fun(_, _) -> Pid ! on_message end), ConsumerPid ! {#'basic.deliver'{}, #amqp_msg{}}, assert_that( receive on_message -> true after 500 -> false end, is(true)), assert_that(meck:called(amqp_channel, subscribe, [channel, Method, ConsumerPid]), is(true)) end) end), describe("consume/4", fun() -> describe("when option 'no_ack' is present", fun() -> it("should set its value in the amqp method", fun() -> meck:new(coello_consumer), meck:expect(coello_consumer, start, 1, consumer), meck:expect(amqp_channel, subscribe, 3, #'basic.consume_ok'{consumer_tag = 1234}), Options = [{no_ack, true}], QueueName = <<"queue">>, Method = #'basic.consume'{ queue = QueueName, no_ack = true}, coello_basic:consume(channel, QueueName, fun(_, _) -> ok end, Options), assert_that(meck:called(amqp_channel, subscribe, [channel, Method, '_']), is(true)), meck:unload(coello_consumer) end) end) end), describe("cancel", fun() -> it("should cancel a consumer", fun() -> meck:new(coello_consumer), meck:expect(amqp_channel, call, 2, ok), meck:expect(coello_consumer, stop, 1, ok), Method = #'basic.cancel'{consumer_tag = consumer_tag}, ok = coello_basic:cancel(channel, {consumer, consumer_tag}), assert_that(meck:called(amqp_channel, call, [channel, Method]), is(true)), assert_that(meck:called(coello_consumer, stop, [consumer]), is(true)), meck:unload(coello_consumer) end) end), describe("ack", fun() -> it("should ack a message", fun() -> DeliveryTag = tag, Multiple = 0, Method = #'basic.ack'{delivery_tag = DeliveryTag, multiple = Multiple}, ok = coello_basic:ack(channel, DeliveryTag, Multiple), assert_that(meck:called(amqp_channel, cast, [channel, Method]), is(true)) end) end), describe("reject", fun() -> it("should reject a message", fun() -> DeliveryTag = tag, Requeue = true, Method = #'basic.reject'{delivery_tag = DeliveryTag, requeue = Requeue}, ok = coello_basic:reject(channel, DeliveryTag, Requeue), assert_that(meck:called(amqp_channel, cast, [channel, Method]), is(true)) end) end).
test/spec/coello_basic_spec.erl
0.668447
0.456349
coello_basic_spec.erl
starcoder
%% roman -- Roman Numerals in Erlang %% %% @doc Allows you to add and subtract Roman numerals in pure Erlang, with %% Roman numeral "literals": %% %% ``` %% roman:add(ii, ii) =:= iv. %% roman:subtract(v, iii) =:= ii. %% ''' %% %% Finally! A way to do calculations on movie copyright years! %% %% ``` %% roman:subtract(mmi, mcmlxviii) =:= xxxiii. %% ''' %% %% At no point does any of this code attempt to convert anything into an %% integer. In fact, apart from Erlang's function_notation/0, there are no %% digits in the source code whatsoever. All arithmetic is done %% "symbolically", and without using your computer's integer arithmetic %% hardware. Give those adders a break! %% %% Other fun facts: %% - This code has no concept of zero. %% - Since this code only uses "standardized" Roman numerals, the maximum %% value is ⅯⅯⅯⅭⅯⅩⅭⅠⅩ (three thousand, nine hundred ninety nine). %% %% This gives a total range of Ⅰ to ⅯⅯⅯⅭⅯⅩⅭⅠⅩ, inclusive. %% %% Also, I used Google Translate to name as many variables as I could in %% Latin, because... quidnī? %% (I don't remember enough from tenth grade Latin to know whether any of the %% translated names are GOOD, so they are most likely BAD). -module(roman). -export([add/2, subtract/2]). -ifdef(TEST). -include_lib("eunit/include/eunit.hrl"). -endif. %% ==================================================================== %% Public API %% ==================================================================== %% @doc Adds two Roman numerals together. add(A, B) -> Summa = add_internus(parse(A), parse(B)), scribe(Summa). %% @doc Subtracts the second Roman numeral from the first. subtract(A, B) -> Differentia = sub_internus(parse(A), parse(B)), scribe(Differentia). %% ==================================================================== %% Internal Functions %% ==================================================================== %% The internal representation of roman numerals is a list of decimal places. %% Decimal places are ordered in INCREASING magnitude. That is, the lowest %% magnitude item is first, and the highest magnitude item is last. %% %% This is **technically** not a place value system, since if a decimal place %% does not participate in a number, it is absent from the list, rather than %% having a value of zero. It is also **technically** annoying. %% %% Examples of valid internal representations: %% %% i ≡ [{unus, i}] %% x ≡ [{decem, x}] %% c ≡ [{centum, c}] %% m ≡ [{mille, m}] %% %% Values are in increasing magnitude, and intervening decimal places are %% absent: %% %% mdcvii ≡ [{unus, vii}, {decem, dc}, {mille, m}] %% %% Notice the lack of a {decem, _} term! -type numerum() ::[valorem(), ...]. %% A decimal place can be ones, tens, hundreds, thousands: -type valorem() :: unus() | decem() | centum() | mille(). %% And these enumerate all possible value within that decimal place: -type unus() :: {unus, i | ii | iii | iv | v | vi | vii | viii | ix}. -type decem() :: {decem, x | xx | xxx | xl | l | lx | lxx | lxxx | xc}. -type centum() :: {centum, c | cc | ccc | cd | d | dc | dcc | dccc | cm}. %% note bene: in "standardized" Roman numerals, there's no way to count from %% four thousand and beyond: -type mille() :: {mille, m | mm | mmm }. %% @doc Write a Roman numeral as an atom. -spec scribe(numerum()) -> atom(). scribe(Numerum) -> DecreasingInMagntiude = lists:reverse(Numerum), ListOfLists = [atom_to_list(I) || {_, I} <- DecreasingInMagntiude], list_to_atom(lists:concat(ListOfLists)). %% ==================================================================== %% The way adding works is by decomposing the addition into %% a bunch of adding ones. %% %% Say you want the sum x + v (ten plus five). We know: %% - how to increment a number by one %% - how to decrement a number by one %% %% Why, that's the same as the sum of one more than first and one less than %% the second: %% %% x + v =:= (x + i) + (v - i) %% x + v =:= xi + iv %% %% Continue this process; at each step, the first number gets one bigger, and %% the second number gets one smaller: %% %% x + v =:= (x + i) + (v - i) %% x + v =:= xi + iv %% xi + iv =:= (xi + i) + (iv - i) %% xi + iv =:= xii + iii %% xii + iii =:= (xii + i) + (iii - i) %% xii + iii =:= xiii + ii %% %% Eventually, the second number becomes one. Since we know how to add one to %% a number, we return the end result %% %% x + v =:= xiv + i %% x + v =:= xv %% %% @see increment/1 %% @see decrement/1 -spec add_internus(numerum(), numerum()) -> numerum(). add_internus(A, [{unus, i}]) -> increment(A); add_internus(A, B) -> ASucc = increment(A), BPred = decrement(B), add_internus(ASucc, BPred). %% Subtracting works similarly to adding. %% In this case, we are taking away from both numbers by the same amount. %% %% We know how to take away one from a number. So we repeat this process: %% %% x - v =:= (x - i) - (v - i) %% x - v =:= ix - iv %% %% ix - iv =:= (ix - i) - (iv - i) %% ix - iv =:= viii - iii %% %% viii - iii =:= (viii - i) - (iii - i) %% viii - iii =:= vii - ii %% %% Eventually, the second number becomes one. Since we know how to subtract %% one from a number, we return the result: %% %% x - v =:= vi - i %% x - v =:= v %% %% @see add_internus/2 %% @see decrement/1 -spec sub_internus(numerum(), numerum()) -> numerum(). sub_internus(A, [{unus, i}]) -> decrement(A); sub_internus(A, B) -> sub_internus(decrement(A), decrement(B)). %% ==================================================================== %% @doc add one to a number. -spec increment(numerum()) -> numerum(). increment(Numerum=[{unus, _}|_]) -> %% Incrementing when there is a ones place in a number is straightforward: incrementum_generalis(Numerum); increment(Etc) -> %% However, when the lowest decimal place is tens or larger, adding one %% means adding the term {unus, i} to entire number. [primus(unus)|Etc]. %% Increments any decimal place. incrementum_generalis([{Magnitudo, I}|Etc]) -> case succssorem(I) of %% We must carry over into the next decimal place: carry -> propagate_carry(Magnitudo, Etc); %% No carry required; we keep the same decimal place: II -> [{Magnitudo, II}|Etc] end. %% When handling carry, we need to know, what decimal initiated the carry. propagate_carry(ExMagnitudo, []) -> %% In the case that there is a carry, and there is not a decimal place %% high enough to absorb it, that means that we need to spontaneously %% create the immediate NEXT decimal place into existence: [primus(diende(ExMagnitudo))]; propagate_carry(ExMagnitudo, Numerum=[{Magnitudo, _}|_]) -> %% If there is something bigger, we need to check whether it's... case MagnitudoProximo = diende(ExMagnitudo) of %% ...the immediately next decimal place: %% e.g., xix + i ≡ [{unus, ix}, {decem, x}] + i %% ≡ [ {decem, xx}] Magnitudo -> incrementum_generalis(Numerum); %% ...or an even larger decimal place, so we need to vivify the %% decimal place in between: %% e.g., cix + i ≡ [{unus, ix}, {centum, c}] %% ≡ [ {decem, x}, {centum, c}] _ -> [primus(MagnitudoProximo)|Numerum] end. %% The next decimal place: diende(unus) -> decem; diende(decem) -> centum; diende(centum) -> mille; diende(mille) -> error(nimis_magna). %% The "first" of a decimal place. primus(unus) -> {unus, i}; primus(decem) -> {decem, x}; primus(centum) -> {centum, c}; primus(mille) -> {mille, m}. %% ==================================================================== %% @doc take one away from a number. -spec decrement(numerum()) -> numerum(). decrement([{Magnitudo, II}|Etc]) -> ante(Magnitudo) ++ case praedecessor(II) of % If we have to borrow, this term disappears from % existence: borrow -> Etc; I -> [{Magnitudo, I}|Etc] end. %% What number is immediately before a decimal place? ante(mille) -> ante(centum) ++ [{centum, cm}]; ante(centum) -> ante(decem) ++ [{decem, xc}]; ante(decem) -> [{unus, ix}]; ante(unus) -> []. %% @doc What numeral comes next for THIS decimal place? % Special cases: succssorem(m) -> mm; succssorem(mm) -> mmm; succssorem(mmm) -> error(nimis_magna); % General case: succssorem(Numerum) -> Partes = split_atom(Numerum), Simbola = simbola(magnitudo(Partes)), case succssorem(Simbola, Partes) of carry -> carry; PartesNovem -> fuse_atom(PartesNovem) end. %% The successor of any numeral looks the same, regardless of its decimal %% place. This table generalizes adding one to any numeral, with your %% choice of {I, V, X} := {i, v, x} | {x, l, c} | {c, d, m}. %% %% @see simbola/1 succssorem({I, _, _}, [I]) -> [I, I]; succssorem({I, _, _}, [I, I]) -> [I, I, I]; succssorem({I, V, _}, [I, I, I]) -> [I, V]; succssorem({I, V, _}, [I, V]) -> [V]; succssorem({I, V, _}, [V]) -> [V, I]; succssorem({I, V, _}, [V, I]) -> [V, I, I]; succssorem({I, V, _}, [V, I, I]) -> [V, I, I, I]; succssorem({I, V, X}, [V, I, I, I]) -> [I, X]; succssorem({I, _, X}, [I, X]) -> carry. %% @doc What numeral comes immediately before for THIS decimal place? % Special cases: praedecessor(m) -> borrow; praedecessor(mm) -> m; praedecessor(mmm) -> mm; % General case praedecessor(Numerum) -> Partes = split_atom(Numerum), Simbola = simbola(magnitudo(Partes)), case praedecessor(Simbola, Partes) of borrow -> borrow; PartesNovem -> fuse_atom(PartesNovem) end. %% The predecessor of any numeral looks the same, regardless of its decimal %% place. This table generalizes taking one away from any numeral, with your %% choice of {I, V, X} := {i, v, x} | {x, l, c} | {c, d, m}. %% %% @see simbola/1 praedecessor({I, _, _}, [I]) -> borrow; praedecessor({I, _, _}, [I, I]) -> [I]; praedecessor({I, _, _}, [I, I, I]) -> [I, I]; praedecessor({I, V, _}, [I, V]) -> [I, I, I]; praedecessor({I, V, _}, [V]) -> [I, V]; praedecessor({I, V, _}, [V, I]) -> [V]; praedecessor({I, V, _}, [V, I, I]) -> [V, I]; praedecessor({I, V, _}, [V, I, I, I]) -> [V, I, I]; praedecessor({I, V, X}, [I, X]) -> [V, I, I, I]. %% @doc What is the decimal place of a term? magnitudo([i|_]) -> unus; magnitudo([v|_]) -> unus; magnitudo([x|_]) -> decem; magnitudo([l|_]) -> decem; magnitudo([c|_]) -> centum; magnitudo([d|_]) -> centum; magnitudo([m|_]) -> mille. %% @doc What letters are used to represent this decimal place? simbola(unus) -> {i, v, x}; simbola(decem) -> {x, l, c}; simbola(centum) -> {c, d, m}; % Note: in practice, this is never used: simbola(mille) -> {m}. %% ==================================================================== %% @doc Parses a "Roman numeral literal" (really, an atom) into the internal %% representation. -spec parse(atom()) -> numerum(). parse(Term) -> AtomList = split_atom(Term), {Mille, EtceteraI} = parse_mille(AtomList), {Centum, EtceteraII} = parse_generalis(centum, {c, d, m}, EtceteraI), {Decem, EtceteraIII} = parse_generalis(decem, {x, l, c}, EtceteraII), {Unus, []} = parse_generalis(unus, {i, v, x}, EtceteraIII), Unus ++ Decem ++ Centum ++ Mille. %% @doc Returns a parsed numeral, if any, and a list of leftover characters. parse_generalis(Nomen, {I, V, X}, AtomList) -> {Parse, Etc} = parse_generalis_initium({I, V, X}, AtomList, []), Term = case Parse of % Did not parse anything (everything is leftover): [] -> []; % Parsed something! The accepted string is in reverse: Reversed -> [{Nomen, fuse_atom(lists:reverse(Reversed))}] end, {Term, Etc}. %% @doc Implements a "generic" finite state machine (FSM) that accepts Roman %% numerals at a given order of magnitude. The FSM is parameterized by the %% symbols used for its ones (I), its fives (V), and its tens (X). %% %% Note: accepted atoms are returned in REVERESED order! %% %% See the README for a diagram of this state machine. parse_generalis_initium({I, V, X}, [I|Etc], Acc) -> i({I, V, X}, Etc, [I|Acc]); parse_generalis_initium({I, V, X}, [V|Etc], Acc) -> v({I, V, X}, Etc, [V|Acc]); parse_generalis_initium({_, _, _}, Etc, []) -> {[], Etc}. i({I, V, X}, [I|Etc], Acc) -> ii({I, V, X}, Etc, [I|Acc]); i({I, V, X}, [V|Etc], Acc) -> iv({I, V, X}, Etc, [V|Acc]); i({I, V, X}, [X|Etc], Acc) -> ix({I, V, X}, Etc, [X|Acc]); i({_, _, _}, Etc, Acc) -> {Acc, Etc}. ii({I, V, X}, [I|Etc], Acc) -> iii({I, V, X}, Etc, [I|Acc]); ii({_, _, _}, Etc, Acc) -> {Acc, Etc}. iii({_, _, _}, Etc, Acc) -> {Acc, Etc}. iv({_, _, _}, Etc, Acc) -> {Acc, Etc}. v({I, V, X}, [I|Etc], Acc) -> vi({I, V, X}, Etc, [I|Acc]); v({_, _, _}, Etc, Acc) -> {Acc, Etc}. vi({I, V, X}, [I|Etc], Acc) -> ii({I, V, X}, Etc, [I|Acc]); vi({_, _, _}, Etc, Acc) -> {Acc, Etc}. ix({_, _, _}, Etc, Acc) -> {Acc, Etc}. %% Parse thousands. Special cased, because the Romans could not count beyond %% MMMCMXCIX. parse_mille([m,m,m|Etc]) -> {[{mille, mmm}], Etc}; parse_mille([m,m|Etc]) -> {[{mille, mm}], Etc}; parse_mille([m|Etc]) -> {[{mille, m}], Etc}; parse_mille(Etc) -> {[], Etc}. %% ==================================================================== %% @doc Returns a list of single character atoms taken from the given atom. -spec split_atom(atom()) -> [atom()]. split_atom(Atom) -> [list_to_atom([Char]) || Char <- atom_to_list(Atom)]. %% @doc Concatenates a list of atoms into a single atom. -spec fuse_atom(['i' | 'v' | 'x' | 'l' | 'c' | 'd' | 'm', ...]) -> atom(). fuse_atom(AtomList) -> ListOfLists = [atom_to_list(Char) || Char <- AtomList], list_to_atom([Char || [Char] <- ListOfLists]). %% ==================================================================== %% Unit Tests %% ==================================================================== -ifdef(EUNIT). % largest value: maximum() -> mmmcmxcix. % before the largest value: ante_maximum() -> mmmcmxcviii. -endif. -ifdef(EUNIT). add_test() -> [?assertEqual(ii, add(i, i)), ?assertEqual(iv, add(ii, ii)), ?assertEqual(mmxx, add(i, mmxix)), ?assertEqual(cxi, add(lxix, xlii)), ?assertEqual(cdlxxxix, add(lxix, cdxx)), ?assertEqual(lxix, add(lxviii, i)), % commendatus ?assertEqual(maximum(), add(ante_maximum(), i))]. -endif. -ifdef(EUNIT). sub_test() -> [?assertEqual(i, subtract(ii, i)), ?assertEqual(ii, subtract(iv, ii)), ?assertEqual(xxvii, subtract(lxix, xlii)), ?assertEqual(xxx, subtract(mmxxii, mcmxcii)), ?assertEqual(ante_maximum(), subtract(maximum(), i)), ?assertEqual(i, subtract(maximum(), ante_maximum()))]. -endif. -ifdef(EUNIT). decrement_test() -> [?assertEqual(parse(mmccxix), decrement(parse(mmccxx))), ?assertEqual(parse(mmcxcix), decrement(parse(mmcc)))]. -endif. -ifdef(EUNIT). parse_unum_test() -> [?assertEqual([{unus, i}], parse(i)), ?assertEqual([{unus, ii}], parse(ii)), ?assertEqual([{unus, iii}], parse(iii)), ?assertEqual([{unus, iv}], parse(iv)), ?assertEqual([{unus, v}], parse(v)), ?assertEqual([{unus, vi}], parse(vi)), ?assertEqual([{unus, vii}], parse(vii)), ?assertEqual([{unus, viii}], parse(viii)), ?assertEqual([{unus, ix}], parse(ix))]. -endif. -ifdef(EUNIT). parse_mixed_test() -> [?assertEqual([{unus, ix}, {decem, lx}], parse(lxix)), ?assertEqual([{decem, xx}, {centum, cd}], parse(cdxx))]. -endif. -ifdef(EUNIT). parse_invalid_test() -> [?assertError(_, parse(viv)), ?assertError(_, parse(vix))]. -endif. -ifdef(EUNIT). split_test() -> [?assertEqual([c, x, i], split_atom(cxi))]. -endif. -ifdef(EUNIT). join_test() -> [?assertEqual(cxi, fuse_atom([c, x, i]))]. -endif.
roman.erl
0.603932
0.533215
roman.erl
starcoder
%% ========================================================================================================== %% Ram - A distributed KV store for Erlang and Elixir. %% %% The MIT License (MIT) %% %% Copyright (c) 2021-2022 <NAME> <<EMAIL>>. %% %% Permission is hereby granted, free of charge, to any person obtaining a copy %% of this software and associated documentation files (the "Software"), to deal %% in the Software without restriction, including without limitation the rights %% to use, copy, modify, merge, publish, distribute, sublicense, and/or sell %% copies of the Software, and to permit persons to whom the Software is %% furnished to do so, subject to the following conditions: %% %% The above copyright notice and this permission notice shall be included in %% all copies or substantial portions of the Software. %% %% THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR %% IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, %% FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE %% AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER %% LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, %% OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN %% THE SOFTWARE. %% ========================================================================================================== %% =================================================================== %% @doc Exposes all of the Key Value store APIs. %% @end %% =================================================================== -module(ram). %% API -export([start/0, stop/0]). -export([start_cluster/1, stop_cluster/1]). -export([add_node/1, remove_node/1, nodes/0]). -export([restart_server/0]). -export([get/1, get/2, fetch/1]). -export([put/2]). -export([update/3]). -export([delete/1]). %% =================================================================== %% API %% =================================================================== %% @doc Starts Ram manually. %% %% In most cases Ram will be started as one of your application's dependencies, %% however you may use this helper method to start it manually. -spec start() -> ok. start() -> {ok, _} = application:ensure_all_started(ram), ok. %% @doc Stops Ram manually. -spec stop() -> ok | {error, Reason :: term()}. stop() -> application:stop(ram). %% @doc Starts the Ram cluster. -spec start_cluster([node()]) -> ok | {error, Reason :: term()}. start_cluster(Nodes) -> ram_backbone:start_cluster(Nodes). %% @doc Stops the Ram cluster. -spec stop_cluster([node()]) -> ok | {error, Reason :: term()}. stop_cluster(Nodes) -> ram_backbone:stop_cluster(Nodes). %% @doc Restart a previously stopped node of the cluster, so that it joins the cluster again. -spec restart_server() -> ok | {error, Reason :: term()}. restart_server() -> ram_backbone:restart_server(). %% @doc Adds Node to an existing Ram cluster. %% %% This method is to be called when: %% <ul> %% <li>Adding a new node to the cluster.</li> %% <li>Restarting a previously stopped node of the cluster, so that it joins the cluster again.</li> %% </ul> %% %% Note that when restarting a server it might be preferable to use {@link restart_server/0} instead. -spec add_node(Node :: node()) -> ok | {error, Reason :: term()}. add_node(Node) -> ram_backbone:add_node(Node). %% @doc Removes Node from the Ram cluster. -spec remove_node(Node :: node()) -> ok | {error, Reason :: term()}. remove_node(Node) -> ram_backbone:remove_node(Node). %% @doc Returns the nodes in the Ram cluster. -spec nodes() -> [node()]. nodes() -> ram_backbone:nodes(). %% @equiv get(Key, undefined) %% @end -spec get(Key :: term()) -> Value :: term(). get(Key) -> get(Key, undefined). %% @doc Returns the Key's Value or Default if the Key is not found. %% %% <h2>Examples</h2> %% <h3>Elixir</h3> %% ``` %% iex(1)> :ram.get("key") %% :undefined %% iex(2)> :ram.get("key", "default") %% "default" %% iex(3)> :ram.put("key", "value") %% :ok %% iex(4)> :ram.get("key") %% "value" %% ''' %% <h3>Erlang</h3> %% ``` %% 1> ram:get("key"). %% undefined %% 2> ram:get("key", "default"). %% "default" %% 3> ram:put("key", "value"). %% ok %% 4> ram:get("key"). %% "value" %% ''' -spec get(Key :: term(), Default :: term()) -> Value :: term(). get(Key, Default) -> ram_kv:get(Key, Default). %% @doc Looks up a Key. %% %% Returns `error' if the Key is not found. -spec fetch(Key :: term()) -> {ok, Value :: term()} | error. fetch(Key) -> ram_kv:fetch(Key). %% @doc Puts a Value for a Key. -spec put(Key :: term(), Value :: term()) -> ok. put(Key, Value) -> ram_kv:put(Key, Value). %% @doc Atomically updates a Key with the given function. %% %% If Key is found then the existing Value is passed to the fun and its result is used as the updated Value of Key. %% If Key is not found, Default is put as the Value of Key. The Default value will not be passed through the update function. %% %% <div style="font-size:30px;float:left;padding-right:10px;color:#ffcc00;">⚠</div> %% <i>Passing functions as arguments might not be compatible across different Erlang versions, %% so if your cluster is composed of nodes running different Erlang versions do not use this method.</i> %% %% <h2>Examples</h2> %% <h3>Elixir</h3> %% ``` %% iex(1)> update_fun = fn existing_value -> existing_value * 2 end %% #Function<44.65746770/1 in :erl_eval.expr/5> %% iex(2)> :ram.update("key", 10, update_fun) %% ok %% iex(3)> :ram.get("key") %% 10 %% iex(4)> :ram.update("key", 10, update_fun) %% ok %% iex(5)> :ram.get("key") %% 20 %% ''' %% <h3>Erlang</h3> %% ``` %% 1> UpdateFun = fun(ExistingValue) -> ExistingValue * 2 end. %% #Fun<erl_eval.44.65746770> %% 2> ram:update("key", 10, UpdateFun). %% ok %% 3> ram:get("key"). %% 10 %% 4> ram:update("key", 10, UpdateFun). %% ok %% 5> ram:get("key"). %% 20 %% ''' -spec update(Key :: term(), Default :: term(), function()) -> ok. update(Key, Default, Fun) -> ram_kv:update(Key, Default, Fun). %% @doc Deletes a Key. -spec delete(Key :: term()) -> ok. delete(Key) -> ram_kv:delete(Key).
src/ram.erl
0.532911
0.42931
ram.erl
starcoder
%% ------------------------------------------------------------------- %% %% Copyright (c) 2014 SyncFree Consortium. All Rights Reserved. %% %% This file is provided to you under the Apache License, %% Version 2.0 (the "License"); you may not use this file %% except in compliance with the License. You may obtain %% a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, %% software distributed under the License is distributed on an %% "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY %% KIND, either express or implied. See the License for the %% specific language governing permissions and limitations %% under the License. %% %% ------------------------------------------------------------------- -module(clocksi_materializer). -include("antidote.hrl"). -ifdef(TEST). -include_lib("eunit/include/eunit.hrl"). -endif. -export([new/1, materialize/4, materialize_eager/3]). %% @doc Creates an empty CRDT for a given type. -spec new(type()) -> snapshot(). new(Type) -> materializer:create_snapshot(Type). %% The materializer is given of tuple containing ordered update operations. %% Each update operation has an id number that is one larger than %% the previous. This function takes as input that tuple and returns the id number of the first update %% operation (i.e. the one with the largest id) -spec get_first_id([{non_neg_integer(), #clocksi_payload{}}] | tuple()) -> non_neg_integer(). get_first_id([]) -> 0; get_first_id([{Id, _Op}|_]) -> Id; get_first_id(Tuple) when is_tuple(Tuple) -> {Length, _ListLen} = element(2, Tuple), case Length of 0 -> 0; Length -> {Id, _Op} = element(?FIRST_OP+(Length-1), Tuple), Id end. %% @doc Applies the operation of a list to a previously created CRDT snapshot. Only the %% operations that are not already in the previous snapshot and %% with smaller timestamp than the specified %% are considered. Newer operations are discarded. %% Input: %% Type: The type of CRDT to create %% Snapshot: Current state of the CRDT %% SnapshotCommitTime: The time used to describe the state of the CRDT given in Snapshot %% MinSnapshotTime: The threshold time given by the reading transaction %% Ops: The list of operations to apply in causal order %% TxId: The Id of the transaction requesting the snapshot %% Output: A tuple. The first element is ok, the seond is the CRDT after appliying the operations, %% the third element 1 minus the number of the operation with the smallest id not included in the snapshot, %% the fourth element is the smallest vectorclock that describes this snapshot, %% the fifth element is a boolean, it it is true it means that the returned snapshot contains %% more operations than the one given as input, false otherwise. %% the sixth element is an integer the counts the number of operations applied to make the snapshot -spec materialize(type(), txid() | ignore, snapshot_time() | ignore, #snapshot_get_response{} ) -> {ok, snapshot(), integer(), snapshot_time() | ignore, boolean(), non_neg_integer()} | {error, reason()}. materialize(Type, TxId, MinSnapshotTime, #snapshot_get_response{snapshot_time = SnapshotCommitTime, ops_list = Ops, materialized_snapshot = #materialized_snapshot{last_op_id = LastOp, value = Snapshot}}) -> FirstId = get_first_id(Ops), {ok, OpList, NewLastOp, LastOpCt, IsNewSS} = materialize_intern(Type, [], LastOp, FirstId, SnapshotCommitTime, MinSnapshotTime, Ops, TxId, SnapshotCommitTime, false, 0), case apply_operations(Type, Snapshot, 0, OpList) of {ok, NewSS, Count} -> {ok, NewSS, NewLastOp, LastOpCt, IsNewSS, Count}; {error, Reason} -> {error, Reason} end. %% @doc Applies a list of operations to a snapshot %% Input: %% Type: The type of CRDT of the snapshot %% Snapshot: The initial snapshot to apply the operations to %% Count: Should be input as 0, this will count the number of ops applied %% OpList: The list of operations to apply %% Output: Either the snapshot with the operations applied to %% it, or an error. -spec apply_operations(type(), snapshot(), non_neg_integer(), [clocksi_payload()]) -> {ok, snapshot(), non_neg_integer()} | {error, reason()}. apply_operations(_Type, Snapshot, Count, []) -> {ok, Snapshot, Count}; apply_operations(Type, Snapshot, Count, [Op | Rest]) -> case materializer:update_snapshot(Type, Snapshot, Op#clocksi_payload.op_param) of {ok, NewSnapshot} -> apply_operations(Type, NewSnapshot, Count+1, Rest); {error, Reason} -> {error, Reason} end. %% @doc Internal function that goes through a list of operations and a snapshot %% time and returns which operations from the list should be applied for %% the given snapshot time. %% Input: %% Type: The type of the CRDT %% OpList: Should be given initially as an empty list, this will accumulate %% the operations to apply. %% LastOp: 1 minus the number of the operation with the smallest id not included in the initial snapshot %% FirstHole: The variable keeps track of 1 minus the number of the operation with the smallest id %% not included in the new snapshot that is currently being generated, it should be initialised to the %% id of the first op in OpList %% SnapshotCommitTime: The time used to describe the intitial state of the CRDT given in Snapshot %% MinSnapshotTime: The threshold time given by the reading transaction %% Ops: The list of operations to apply in causal order, the most recent op is on the left %% TxId: The Id of the transaction requesting the snapshot %% LastOpCommitTime: The snapshot time of the last operation in the list of operations to apply %% NewSS: Boolean that is true if any operations should be applied, fale otherwise. Should start as false. %% Output: A tuple with 4 elements or an error. The first element of the tuple is the atom ok. %% The second element is the list of operations that should be applied to the snapshot. %% The third element 1 minus the number of the operation with the smallest id not included in the snapshot. %% The fourth element is the snapshot time of the last operation in the list. %% The fifth element is a boolean, true if a new snapshot should be generated, false otherwise. -spec materialize_intern(type(), [clocksi_payload()], integer(), integer(), snapshot_time() | ignore, snapshot_time(), [{integer(), clocksi_payload()}] | tuple(), txid() | ignore, snapshot_time() | ignore, boolean(), non_neg_integer()) -> {ok, [clocksi_payload()], integer(), snapshot_time()|ignore, boolean()}. materialize_intern(_Type, OpList, _LastOp, FirstHole, _SnapshotCommitTime, _MinSnapshotTime, [], _TxId, LastOpCt, NewSS, _Location) -> {ok, OpList, FirstHole, LastOpCt, NewSS}; materialize_intern(Type, OpList, LastOp, FirstHole, SnapshotCommitTime, MinSnapshotTime, [{OpId, Op}|Rest], TxId, LastOpCt, NewSS, Location) -> materialize_intern_perform(Type, OpList, LastOp, FirstHole, SnapshotCommitTime, MinSnapshotTime, {OpId, Op}, Rest, TxId, LastOpCt, NewSS, Location + 1); materialize_intern(Type, OpList, LastOp, FirstHole, SnapshotCommitTime, MinSnapshotTime, TupleOps, TxId, LastOpCt, NewSS, Location) -> {Length, _ListLen} = element(2, TupleOps), case Length == Location of true -> {ok, OpList, FirstHole, LastOpCt, NewSS}; false -> materialize_intern_perform(Type, OpList, LastOp, FirstHole, SnapshotCommitTime, MinSnapshotTime, element((?FIRST_OP+Length-1) - Location, TupleOps), TupleOps, TxId, LastOpCt, NewSS, Location + 1) end. materialize_intern_perform(Type, OpList, LastOp, FirstHole, SnapshotCommitTime, MinSnapshotTime, {OpId, Op}, Rest, TxId, LastOpCt, NewSS, Location) -> Result = case Type == Op#clocksi_payload.type of true -> OpCom=Op#clocksi_payload.commit_time, OpSS=Op#clocksi_payload.snapshot_time, %% Check if the op is not in the previous snapshot and should be included in the new one case (is_op_in_snapshot(TxId, Op, OpCom, OpSS, MinSnapshotTime, SnapshotCommitTime, LastOpCt)) of {true, _, NewOpCt} -> %% Include the new op because it has a timestamp bigger than the snapshot being generated {ok, [Op | OpList], NewOpCt, false, true, FirstHole}; {false, false, _} -> %% Dont include the op {ok, OpList, LastOpCt, false, NewSS, OpId-1}; % no update {false, true, _} -> %% Dont Include the op, because it was already in the SS {ok, OpList, LastOpCt, true, NewSS, FirstHole} end; false -> %% Op is not for this {Key, Type} %% @todo THIS CASE PROBABLY SHOULD NOT HAPPEN?! {ok, OpList, LastOpCt, false, NewSS, FirstHole} %% no update end, case Result of {ok, NewOpList1, NewLastOpCt, false, NewSS1, NewHole} -> materialize_intern(Type, NewOpList1, LastOp, NewHole, SnapshotCommitTime, MinSnapshotTime, Rest, TxId, NewLastOpCt, NewSS1, Location); {ok, NewOpList1, NewLastOpCt, true, NewSS1, NewHole} -> case OpId - 1 =< LastOp of true -> %% can skip the rest of the ops because they are already included in the SS materialize_intern(Type, NewOpList1, LastOp, NewHole, SnapshotCommitTime, MinSnapshotTime, [], TxId, NewLastOpCt, NewSS1, Location); false -> materialize_intern(Type, NewOpList1, LastOp, NewHole, SnapshotCommitTime, MinSnapshotTime, Rest, TxId, NewLastOpCt, NewSS1, Location) end end. %% @doc Check whether an udpate is included in a snapshot and also %% if that update is newer than a snapshot's commit time %% Input: %% TxId: Descriptor of the transaction requesting the snapshot %% Op: The operation to check %% {OpDc, OpCommitTime}: The DC and commit time of the operation %% OperationSnapshotTime: The snapshot time of the operation %% SnapshotTime: The snapshot time to check if the operation is included in %% LastSnapshot: The previous snapshot that is being used to generate the new snapshot %% PrevTime: The snapshot time of the previous operation that was checked %% Outptut: A tuple of 3 elements. The first element is a boolean that is true %% if the operation should be included in the snapshot false otherwise, the second element %% is a boolean that is true if the operation was already included in the previous snapshot, %% false otherwise. The thrid element is the snapshot time of the last operation to %% be applied to the snapshot -spec is_op_in_snapshot(txid(), clocksi_payload(), dc_and_commit_time(), snapshot_time(), snapshot_time(), snapshot_time() | ignore, snapshot_time()) -> {boolean(), boolean(), snapshot_time()}. is_op_in_snapshot(TxId, Op, {OpDc, OpCommitTime}, OperationSnapshotTime, SnapshotTime, LastSnapshot, PrevTime) -> %% First check if the op was already included in the previous snapshot %% Is the "or TxId ==" part necessary and correct???? case materializer_vnode:belongs_to_snapshot_op( LastSnapshot, {OpDc, OpCommitTime}, OperationSnapshotTime) or (TxId == Op#clocksi_payload.txid) of true -> %% If not, check if it should be included in the new snapshot %% Replace the snapshot time of the dc where the transaction committed with the commit time OpSSCommit = dict:store(OpDc, OpCommitTime, OperationSnapshotTime), %% PrevTime2 is the time of the previous snapshot, if there was none, it usues the snapshot time %% of the new operation PrevTime2 = case PrevTime of ignore -> OpSSCommit; _ -> PrevTime end, %% Result is true if the op should be included in the snapshot %% NewTime is the vectorclock of the snapshot with the time of Op included {Result, NewTime} = dict:fold(fun(DcIdOp, TimeOp, {Acc, PrevTime3}) -> Res1 = case dict:find(DcIdOp, SnapshotTime) of {ok, TimeSS} -> case TimeSS < TimeOp of true -> false; false -> Acc end; error -> lager:error("Could not find DC in SS ~p", [SnapshotTime]), false end, Res2 = dict:update(DcIdOp, fun(Val) -> case TimeOp > Val of true -> TimeOp; false -> Val end end, TimeOp, PrevTime3), {Res1, Res2} end, {true, PrevTime2}, OpSSCommit), case Result of true -> {true, false, NewTime}; false -> {false, false, PrevTime} end; false-> %% was already in the prev ss, done searching ops {false, true, PrevTime} end. %% @doc Apply updates in given order without any checks. %% Careful: In contrast to materialize/6, it takes just operations, not clocksi_payloads! -spec materialize_eager(type(), snapshot(), [op()]) -> snapshot(). materialize_eager(Type, Snapshot, Ops) -> materializer:materialize_eager(Type, Snapshot, Ops). -ifdef(TEST). materializer_clocksi_test()-> Type = antidote_crdt_counter_pn, PNCounter = new(Type), ?assertEqual(0, Type:value(PNCounter)), %% need to add the snapshot time for these for the test to pass Op1 = #clocksi_payload{key = abc, type = Type, op_param = 2, commit_time = {1, 1}, txid = 1, snapshot_time=vectorclock:from_list([{1, 1}])}, Op2 = #clocksi_payload{key = abc, type = Type, op_param = 1, commit_time = {1, 2}, txid = 2, snapshot_time=vectorclock:from_list([{1, 2}])}, Op3 = #clocksi_payload{key = abc, type =Type, op_param = 1, commit_time = {1, 3}, txid = 3, snapshot_time=vectorclock:from_list([{1, 3}])}, Op4 = #clocksi_payload{key = abc, type = Type, op_param = 2, commit_time = {1, 4}, txid = 4, snapshot_time=vectorclock:from_list([{1, 4}])}, Ops = [{4, Op4}, {3, Op3}, {2, Op2}, {1, Op1}], SS = #snapshot_get_response{snapshot_time = ignore, ops_list = Ops, materialized_snapshot = #materialized_snapshot{last_op_id = 0, value = PNCounter}}, {ok, PNCounter2, 3, CommitTime2, _SsSave, _} = materialize(Type, ignore, vectorclock:from_list([{1, 3}]), SS), ?assertEqual({4, vectorclock:from_list([{1, 3}])}, {Type:value(PNCounter2), CommitTime2}), {ok, PNcounter3, 4, CommitTime3, _SsSave1, _} = materialize(Type, ignore, vectorclock:from_list([{1, 4}]), SS), ?assertEqual({6, vectorclock:from_list([{1, 4}])}, {Type:value(PNcounter3), CommitTime3}), {ok, PNcounter4, 4, CommitTime4, _SsSave2, _} = materialize(Type, ignore, vectorclock:from_list([{1, 7}]), SS), ?assertEqual({6, vectorclock:from_list([{1, 4}])}, {Type:value(PNcounter4), CommitTime4}). %% This test tests when a a snapshot is generated that does not include all of the updates in the %% list of operations, precisely in the case where an operation is not taken, but the operations to %% the left and right of it in the list are taken. When this snapshot is then used for a future %% read with a different timestamp, this missing value must be checked. materializer_missing_op_test() -> Type = antidote_crdt_counter_pn, PNCounter = new(Type), ?assertEqual(0, Type:value(PNCounter)), Op1 = #clocksi_payload{key = abc, type = Type, op_param = 1, commit_time = {1, 1}, txid = 1, snapshot_time=vectorclock:from_list([{1, 1}, {2, 1}])}, Op2 = #clocksi_payload{key = abc, type = Type, op_param = 1, commit_time = {1, 2}, txid = 2, snapshot_time=vectorclock:from_list([{1, 2}, {2, 1}])}, Op3 = #clocksi_payload{key = abc, type = Type, op_param = 1, commit_time = {2, 2}, txid = 3, snapshot_time=vectorclock:from_list([{1, 1}, {2, 1}])}, Op4 = #clocksi_payload{key = abc, type = Type, op_param = 1, commit_time = {1, 3}, txid = 2, snapshot_time=vectorclock:from_list([{1, 2}, {2, 1}])}, Ops = [{4, Op4}, {3, Op3}, {2, Op2}, {1, Op1}], SS = #snapshot_get_response{snapshot_time = ignore, ops_list = Ops, materialized_snapshot = #materialized_snapshot{last_op_id = 0, value = PNCounter}}, {ok, PNCounter2, LastOp, CommitTime2, _SsSave, _} = materialize(Type, ignore, vectorclock:from_list([{1, 3}, {2, 1}]), SS), ?assertEqual({3, vectorclock:from_list([{1, 3}, {2, 1}])}, {Type:value(PNCounter2), CommitTime2}), SS2 = #snapshot_get_response{snapshot_time = CommitTime2, ops_list = Ops, materialized_snapshot = #materialized_snapshot{last_op_id = LastOp, value = PNCounter2}}, {ok, PNCounter3, 4, CommitTime3, _SsSave, _} = materialize(Type, ignore, vectorclock:from_list([{1, 3}, {2, 2}]), SS2), ?assertEqual({4, vectorclock:from_list([{1, 3}, {2, 2}])}, {Type:value(PNCounter3), CommitTime3}). %% This test tests the case when there are updates that only snapshots that contain entries from one of the DCs. %% This can happen for example if an update is commited before the DCs have been connected. %% It ensures that when we read using a snapshot with and without all the DCs we still include the correct updates. materializer_missing_dc_test() -> Type = antidote_crdt_counter_pn, PNCounter = new(Type), ?assertEqual(0, Type:value(PNCounter)), Op1 = #clocksi_payload{key = abc, type = Type, op_param = 1, commit_time = {1, 1}, txid = 1, snapshot_time=vectorclock:from_list([{1, 1}])}, Op2 = #clocksi_payload{key = abc, type = Type, op_param = 1, commit_time = {1, 2}, txid = 2, snapshot_time=vectorclock:from_list([{1, 2}])}, Op3 = #clocksi_payload{key = abc, type = Type, op_param = 1, commit_time = {2, 2}, txid = 3, snapshot_time=vectorclock:from_list([{2, 1}])}, Op4 = #clocksi_payload{key = abc, type = Type, op_param = 1, commit_time = {1, 3}, txid = 2, snapshot_time=vectorclock:from_list([{1, 2}])}, Ops = [{4, Op4}, {3, Op3}, {2, Op2}, {1, Op1}], SS = #snapshot_get_response{snapshot_time = ignore, ops_list = Ops, materialized_snapshot = #materialized_snapshot{last_op_id = 0, value = PNCounter}}, {ok, PNCounterA, LastOpA, CommitTimeA, _SsSave, _} = materialize(Type, ignore, vectorclock:from_list([{1, 3}]), SS), ?assertEqual({3, vectorclock:from_list([{1, 3}])}, {Type:value(PNCounterA), CommitTimeA}), SS2 = #snapshot_get_response{snapshot_time = CommitTimeA, ops_list = Ops, materialized_snapshot = #materialized_snapshot{last_op_id = LastOpA, value = PNCounterA}}, {ok, PNCounterB, 4, CommitTimeB, _SsSave, _} = materialize(Type, ignore, vectorclock:from_list([{1, 3}, {2, 2}]), SS2), ?assertEqual({4, vectorclock:from_list([{1, 3}, {2, 2}])}, {Type:value(PNCounterB), CommitTimeB}), {ok, PNCounter2, LastOp, CommitTime2, _SsSave, _} = materialize(Type, ignore, vectorclock:from_list([{1, 3}, {2, 1}]), SS), ?assertEqual({3, vectorclock:from_list([{1, 3}])}, {Type:value(PNCounter2), CommitTime2}), SS3 = #snapshot_get_response{snapshot_time = CommitTime2, ops_list = Ops, materialized_snapshot = #materialized_snapshot{last_op_id = LastOp, value = PNCounter2}}, {ok, PNCounter3, 4, CommitTime3, _SsSave, _} = materialize(Type, ignore, vectorclock:from_list([{1, 3}, {2, 2}]), SS3), ?assertEqual({4, vectorclock:from_list([{1, 3}, {2, 2}])}, {Type:value(PNCounter3), CommitTime3}). materializer_clocksi_concurrent_test() -> Type = antidote_crdt_counter_pn, PNCounter = new(Type), ?assertEqual(0, Type:value(PNCounter)), Op1 = #clocksi_payload{key = abc, type = Type, op_param = 2, commit_time = {1, 1}, txid = 1, snapshot_time=vectorclock:from_list([{1, 1}, {2, 1}])}, Op2 = #clocksi_payload{key = abc, type = Type, op_param = 1, commit_time = {1, 2}, txid = 2, snapshot_time=vectorclock:from_list([{1, 2}, {2, 1}])}, Op3 = #clocksi_payload{key = abc, type = Type, op_param = 1, commit_time = {2, 2}, txid = 3, snapshot_time=vectorclock:from_list([{1, 1}, {2, 1}])}, Ops = [{3, Op2}, {2, Op3}, {1, Op1}], {ok, PNCounter2, 3, CommitTime2, _Keep} = materialize_intern(Type, [], 0, 3, ignore, vectorclock:from_list([{2, 2}, {1, 2}]), Ops, ignore, ignore, false, 0), {ok, PNCounter3, _} = apply_operations(Type, PNCounter, 0, PNCounter2), ?assertEqual({4, vectorclock:from_list([{1, 2}, {2, 2}])}, {Type:value(PNCounter3), CommitTime2}), Snapshot=new(Type), SS = #snapshot_get_response{snapshot_time = ignore, ops_list = Ops, materialized_snapshot = #materialized_snapshot{last_op_id = 0, value = Snapshot}}, {ok, PNcounter3, 1, CommitTime3, _SsSave1, _} = materialize(Type, ignore, vectorclock:from_list([{1, 2}, {2, 1}]), SS), ?assertEqual({3, vectorclock:from_list([{1, 2}, {2, 1}])}, {Type:value(PNcounter3), CommitTime3}), {ok, PNcounter4, 2, CommitTime4, _SsSave2, _} = materialize(Type, ignore, vectorclock:from_list([{1, 1}, {2, 2}]), SS), ?assertEqual({3, vectorclock:from_list([{1, 1}, {2, 2}])}, {Type:value(PNcounter4), CommitTime4}), {ok, PNcounter5, 1, CommitTime5, _SsSave3, _} = materialize(Type, ignore, vectorclock:from_list([{1, 1}, {2, 1}]), SS), ?assertEqual({2, vectorclock:from_list([{1, 1}, {2, 1}])}, {Type:value(PNcounter5), CommitTime5}). %% Testing gcounter with empty update log materializer_clocksi_noop_test() -> Type = antidote_crdt_counter_pn, PNCounter = new(Type), ?assertEqual(0, Type:value(PNCounter)), Ops = [], {ok, PNCounter2, 0, ignore, _SsSave} = materialize_intern(Type, [], 0, 0, ignore, vectorclock:from_list([{1, 1}]), Ops, ignore, ignore, false, 0), {ok, PNCounter3, _} = apply_operations(Type, PNCounter, 0, PNCounter2), ?assertEqual(0, Type:value(PNCounter3)). materializer_eager_clocksi_test()-> Type = antidote_crdt_counter_pn, PNCounter = new(Type), ?assertEqual(0, Type:value(PNCounter)), % test - no ops PNCounter2 = materialize_eager(Type, PNCounter, []), ?assertEqual(0, Type:value(PNCounter2)), % test - several ops Op1 = 1, Op2 = 2, Op3 = 3, Op4 = 4, Ops = [Op1, Op2, Op3, Op4], PNCounter3 = materialize_eager(Type, PNCounter, Ops), ?assertEqual(10, Type:value(PNCounter3)). is_op_in_snapshot_test() -> Type = antidote_crdt_counter_pn, Op1 = #clocksi_payload{key = abc, type = Type, op_param = {increment, 2}, commit_time = {dc1, 1}, txid = 1, snapshot_time=vectorclock:from_list([{dc1, 1}])}, OpCT1 = {dc1, 1}, OpCT1SS = vectorclock:from_list([OpCT1]), ST1 = vectorclock:from_list([{dc1, 2}]), ST2 = vectorclock:from_list([{dc1, 0}]), ?assertEqual({true, false, OpCT1SS}, is_op_in_snapshot(2, Op1, OpCT1, OpCT1SS, ST1, ignore, ignore)), ?assertEqual({false, false, ignore}, is_op_in_snapshot(2, Op1, OpCT1, OpCT1SS, ST2, ignore, ignore)). -endif.
src/clocksi_materializer.erl
0.651133
0.403185
clocksi_materializer.erl
starcoder
%% @doc This is the main API for the library. %% %% You can think of this module as an abstraction layer %% on the raw API which attempts to add some convenience %% to plumbing the entire thing by hand-coding maps with %% the appropriate AWS magic. %% %% The low level API calls are in `darcy_ddb_api.erl', but generally %% users shouldn't call those directly. -module(darcy). -include("darcy.hrl"). -define(TIMEOUT, 5000). % 5 seconds -define(BIG_TIMEOUT, 5*60*1000). % 5 minutes -define(BATCH_SIZE, 100). % how many items to split into a batch -define(BATCH_MAX, 10). % maximum number of batches to process in a single go -define(INITIAL_ERROR_DELAY, 500). % initial throughput retry delay in milliseconds -define(MAX_ERROR_DELAY, 32000). % maximum millseconds of delay before terminating operation -define(INITIAL_ERROR_STATE, #error_delay{}). %% This is a record because, later, I might try to extend this to reduce the sleep %% time based on the number of successful operations. So for now it only has one %% field, but it may have two or more in the future. -record(error_delay, { delay = 0 :: non_neg_integer() }). -export([ start/0, to_map/1, clean_map/1, to_ddb/1, default_decode/1, default_encode/1, make_attribute_defs/1, make_key_schema/1, make_provisioned_throughput/2, table_name/1, make_table_spec/3, make_table_spec/5, make_global_index_spec/3, make_global_index_spec/5, add_global_index/2, make_table_if_not_exists/2, make_global_table_if_not_exists/3, describe_table/2, describe_global_table/2, delete_table/2, get_item/3, batch_get_items/3, put_item/3, put_item/4, batch_write_items/3, query/3, query/4, scan/3, scan_all/3, scan_all/4, scan_parallel/5, scan_parallel/6, scan_parallel/7 ]). -type lookup_value() :: integer() | float() | binary() | {blob, binary()}. %-type set_value() :: {number_set, [ integer() | float() ] } | {string_set, [ binary() ]}. %-type list_value() :: {list, [ map() | set_value() | lookup_value() ]}. %% @doc Convenience function to start `darcy' and all %% of its dependent applications. start() -> application:ensure_all_started(darcy). %% @doc Return a map with the key of `AttributeDefinitions' %% suitable for using in a table or index specification. -spec make_attribute_defs( [ { AttributeName :: binary(), AttributeType :: binary() } ] ) -> AttributeDefinitions :: map(). make_attribute_defs(Attributes) when is_list(Attributes) -> #{ <<"AttributeDefinitions">> => [ #{ <<"AttributeName">> => N, <<"AttributeType">> => T } || {N, T} <- Attributes ] }. %% @doc Return a `KeySchema' map suitable for use in a %% table or index specification. %% %% If you pass one attribute, it will be assigned the `HASH' %% key type. If you pass two attributes, the first will be %% `HASH' and the second will be the `RANGE' type. %% %% You can <a href="http://docs.aws.amazon.com/amazondynamodb/latest/developerguide/HowItWorks.Partitions.html">read more about hash and range keys</a> %% in the official Dynamo documentation. -spec make_key_schema( [ Keys :: binary() ] ) -> KeySchema :: map(). make_key_schema([HashKey]) -> make_schema_impl([{HashKey, <<"HASH">>}]); make_key_schema([HashKey, RangeKey]) -> make_schema_impl([{HashKey, <<"HASH">>}, {RangeKey, <<"RANGE">>}]). make_schema_impl(Schema) -> #{ <<"KeySchema">> => [ #{ <<"AttributeName">> => N, <<"KeyType">> => T } || {N, T} <- Schema ] }. %% @doc Makes a `ProvisionedThroughput' map to indicate the number %% read and write units Dynamo should reserve for your index or %% table. -spec make_provisioned_throughput( ReadUnits :: pos_integer(), WriteUnits :: pos_integer() ) -> ProvisionedThroughput :: map(). make_provisioned_throughput(ReadUnits, WriteUnits) -> #{ <<"ProvisionedThroughput">> => #{ <<"ReadCapacityUnits">> => ReadUnits, <<"WriteCapacityUnits">> => WriteUnits } }. %% @doc Makes a `TableName' map suitable for use in a table %% or index specification. -spec table_name( Name :: binary() ) -> TableName :: map(). table_name(N) when is_binary(N) -> #{ <<"TableName">> => N }. %% @doc Convenience function which returns a complete %% table specification. This function uses the default %% number of read and write units (currently 5 each). -spec make_table_spec( TableName :: binary(), Attr :: [{ AttrName :: binary(), AttrType :: binary() }], Keys :: [ Keys :: binary() ] ) -> TableSpec :: map(). make_table_spec(TableName, Attributes, Keys) -> make_table_spec(TableName, Attributes, Keys, ?DEFAULT_READ_UNITS, ?DEFAULT_WRITE_UNITS). %% @doc Convenience function which returns a complete %% table specification. -spec make_table_spec( TableName :: binary(), Attr :: [{ AttrName :: binary(), AttrType :: binary() }], Keys :: [ binary() ], ReadUnits :: pos_integer(), WriteUnits :: pos_integer() ) -> TableSpec :: map(). make_table_spec(TableName, Attributes, Keys, Read, Write) -> lists:foldl(fun(M, Acc) -> maps:merge(M, Acc) end, #{}, [make_attribute_defs(Attributes), make_key_schema(Keys), make_provisioned_throughput(Read, Write), table_name(TableName)]). %% @doc Convenience function which returns a global index %% specification. This function uses the default read and %% write units (currently 5 each). -spec make_global_index_spec( IndexName :: binary(), Keys :: [ binary() ], ProjectionSpec :: {} | { [ binary() ], binary() } ) -> GlobalIndexSpec :: map(). make_global_index_spec(IndexName, Keys, ProjectionSpec) -> make_global_index_spec(IndexName, Keys, ProjectionSpec, ?DEFAULT_READ_UNITS, ?DEFAULT_WRITE_UNITS). %% @doc Convenience function which returns a global index %% specification. -spec make_global_index_spec( IndexName :: binary(), Keys :: [ binary() ], ProjectionSpec :: {} | { [ NonKeyAttribute :: binary() ], binary() }, ReadUnits :: pos_integer(), WriteUnits :: pos_integer() ) -> GlobalIndexSpec :: map(). make_global_index_spec(IndexName, Keys, ProjectionSpec, Read, Write) -> lists:foldl(fun(M, Acc) -> maps:merge(M, Acc) end, #{}, [make_key_schema(Keys), make_provisioned_throughput(Read, Write), make_projection(ProjectionSpec), index_name(IndexName)]). %% @doc Add a global index specification to an existing table %% specification. If a global index specification has already %% been added, this function will add the new one to the %% current one. %% %% Tables may not have more than two global indices. -spec add_global_index( TableSpec :: map(), GlobalIndexSpec :: map() ) -> NewTableSpec :: map(). add_global_index(#{ <<"GlobalSecondaryIndexes">> := CurrentGSI } = TableSpec, GSISpec) -> maps:put(<<"GlobalSecondaryIndexes">>, [ GSISpec | CurrentGSI ], TableSpec); add_global_index(TableSpec, GSISpec) -> maps:put(<<"GlobalSecondaryIndexes">>, [ GSISpec ], TableSpec). %% @doc Create an `IndexName' map for use in an index specification. -spec index_name( Name :: binary() ) -> IndexName :: map(). index_name(N) when is_binary(N) -> #{ <<"IndexName">> => N }. %% @doc This function returns a `Projection' map suitable for %% use in an index specification. It is expected that if you %% want the `ALL' or `KEYS_ONLY' projection type, your list of %% non-key attributes will be empty. %% %% If you want an empty projection map, pass in an empty tuple %% `{}'. -spec make_projection( ProjectionSpec :: {} | { [ NonKeyAttribute :: binary() ], binary() }) -> Projection :: map(). make_projection({}) -> #{ <<"Projection">> => #{} }; make_projection({Attr, <<"INCLUDE">>}) -> #{ <<"Projection">> => #{ <<"NonKeyAttributes">> => Attr, <<"ProjectionType">> => <<"INCLUDE">> } }; make_projection({[], T}) -> #{ <<"Projection">> => #{ <<"ProjectionType">> => T } }. %% @doc Make a table if it doesn't already exist. -spec make_table_if_not_exists( Client :: darcy_client:aws_client(), TableSpec :: map() ) -> ok | {error, Error :: term()}. make_table_if_not_exists(Client, #{ <<"TableName">> := TableName} = Spec) -> case darcy_ddb_api:describe_table(Client, #{ <<"TableName">> => TableName }) of {ok, _Result, _Details} -> ok; {error, _Error, {400, _Headers, _Client}} -> attempt_make_table(Client, Spec); {error, Error, {Status, _Headers, _NClient}} -> {error, {table_creation_error, {Status, Error}}} end. attempt_make_table(Client, Spec) -> case darcy_ddb_api:create_table(Client, Spec) of {ok, _Result, { 200, _Headers, _Client}} -> ok; {error, Error, {Status, _Headers, _NewClient}} -> {error, {table_creation_failed, {Status, Error}}} end. %% @doc Make a global table if it doesn't already exist. -spec make_global_table_if_not_exists(Client :: darcy_client:aws_client(), TableSpec :: map(), Regions :: [ binary() ]) -> ok | {error, Error :: term()}. make_global_table_if_not_exists(#{ region := Region } = Client, #{ <<"TableName">> := TableName } = Spec, Regions) -> case lists:member(Region, Regions) of false -> {error, {bad_region_spec, [Region, Regions]}}; true -> case describe_global_table(Client, TableName) of {ok, _Result} -> ok; {error, _} -> ok = global_table_setup(Client, Spec, Regions), attempt_make_global_table(Client, TableName, Regions) end end. all_ok(ok) -> true; all_ok(_) -> false. global_table_setup(Client, Spec, Regions) -> true = lists:all(fun all_ok/1, pmap(fun(R) -> do_table_creation(Client, Spec, R) end, Regions) ), ok. do_table_creation(Client, Spec, Region) -> NewClient = darcy_client:switch_region(Client, Region), NewSpec = enable_global_streams(Spec), make_table_if_not_exists(NewClient, NewSpec). enable_global_streams(Spec) -> Streams = #{ <<"StreamSpecification">> => #{ <<"StreamEnabled">> => true, <<"StreamViewType">> => <<"NEW_AND_OLD_IMAGES">> } }, maps:merge(Spec, Streams). attempt_make_global_table(Client, TableName, Regions) -> Req = #{ <<"GlobalTableName">> => TableName, <<"ReplicationGroup">> => [ #{ <<"RegionName">> => R } || R <- Regions ] }, %% this is a long-running operation, make the response receive timeout 60 seconds case darcy_ddb_api:create_global_table(Client, Req, [{recv_timeout, 60*1000}]) of {ok, _Result, {200, _Headers, _Client}} -> ok; {error, Error, {Status, _Headers, _Client}} -> {error, {global_table_creation_failed, {Status, Error}}} end. %% @doc Delete a Dynamo table with the given name. -spec delete_table( Client :: darcy_client:aws_client(), TableName :: binary() ) -> ok | {error, Error :: term()}. delete_table(Client, TableName) -> case darcy_ddb_api:delete_table(Client, table_name(TableName)) of {ok, #{ <<"TableDescription">> := Desc }, Details} -> ensure_deleting_state(Desc, Details); {error, Error, {Status, _Headers, _C}} -> {error, {table_deletion_error, {Status, Error}}} end. ensure_deleting_state( #{ <<"TableStatus">> := <<"DELETING">> }, _Details ) -> ok; ensure_deleting_state( Other , {Status, _Headers, _C} ) -> {error, {table_deletion_error, {Status, Other}}}. %% @doc This returns a map representing the current state of the %% given Dynamo table. -spec describe_table( Client :: darcy_client:aws_client(), TableName :: binary() ) -> {ok, TableDesc :: map()} | {error, Error :: term()}. describe_table(Client, TableName) -> case darcy_ddb_api:describe_table(Client, table_name(TableName)) of {ok, Result, _Details } -> {ok, Result}; {error, Error, {Status, _Headers, Client}} -> {error, {table_description_error, {Status, Error}}} end. %% @doc This returns a map representing the current state of the %% given Dynamo global table. -spec describe_global_table( Client :: darcy_client:aws_client(), TableName :: binary() ) -> {ok, TableDesc :: map()} | {error, Error :: term()}. describe_global_table(Client, TableName) -> case darcy_ddb_api:describe_global_table(Client, #{ <<"GlobalTableName">> => TableName }) of {ok, Result, _Details } -> {ok, Result}; {error, Error, {Status, _Headers, _Client}} -> {error, {table_description_error, {Status, Error}}} end. %% GET ITEM %% @doc Retrieve a single item from the given Dynamo table using %% the hash and if needed, range keys. -spec get_item( Client :: darcy_client:aws_client(), TableName :: binary(), Key :: #{ KeyName :: binary() => LookupValue :: lookup_value() } ) -> {ok, Item :: map()} | {error, not_found} | {error, Error :: term()}. get_item(Client, TableName, Key) -> Request = #{ <<"TableName">> => TableName, <<"Key">> => to_ddb(Key) }, case darcy_ddb_api:get_item(Client, Request) of {ok, Raw, _Details} -> return_value(Raw); {error, Error, {Code, Headers, _Client}} -> {error, {Error, [Code, Headers]}} end. %% @doc Retrieve a set of records given a list of keys. %% %% The underlying API supports a maximum of 100 keys per %% request, so this call batches keys into sets of up %% to 100 and folds across these sets into an %% accumulator for all keys. %% %% Items which are not found in the table will not be part %% of the result set but will consume provisioned read %% capacity. %% %% <B>N.B.</B>: Dynamo and this client will not return your items %% in any particular order! -spec batch_get_items( Client :: darcy_client:aws_client(), TableName :: binary(), Items :: [ map() ] ) -> {ok, [ Result :: term() ]} | {error, Error :: term() }. batch_get_items(Client, TableName, Keys) -> make_batch_get(Client, TableName, Keys, []). make_batch_get(_Client, _TableName, [], Acc) -> {ok, lists:flatten(Acc)}; make_batch_get(Client, TableName, Keys, Acc) -> {Request, Rest} = format_batch_get(TableName, Keys), Result = execute_batch_get(Client, TableName, Request, ?RETRIES, []), make_batch_get(Client, TableName, Rest, [ Result | Acc ]). format_batch_get(TableName, Keys) -> {Current, Rest} = maybe_split_keys(Keys), {format_batch_get_request(TableName, Current), Rest}. maybe_split_keys(Keys) when length(Keys) =< 100 -> {Keys, []}; maybe_split_keys(Keys) -> lists:split(100, Keys). format_batch_get_request(TableName, Keys) -> #{ <<"RequestItems">> => #{ TableName => #{ <<"Keys">> => [ to_ddb(K) || K <- Keys ] } } }. execute_batch_get(_Client, _TableName, Request, 0, Acc) -> {error, {retries_exceeded, Request, Acc}}; execute_batch_get(Client, TableName, Request, Retries, Acc) -> case darcy_ddb_api:batch_get_item(Client, Request) of {ok, Raw, _Details} -> return_batch_get_results(Raw, Client, TableName, Retries, Acc); {error, Error, {Code, Headers, _Client}} -> {error, {Error, [Code, Headers]}} end. %% happy path: no unprocessed keys, no accumulator return_batch_get_results(#{ <<"Responses">> := R, <<"UnprocessedKeys">> := U }, _Client, TableName, _Retries, []) when map_size(U) == 0 -> batch_get_results(TableName, R); %% not as happy path: no unprocessed keys, accumulator has partial results return_batch_get_results(#{ <<"Responses">> := R, <<"UnprocessedKeys">> := U }, _Client, TableName, _Retries, Acc) when map_size(U) == 0 -> lists:flatten([ batch_get_results(TableName, R) | Acc ]); %% not very happy path: unprocessed keys return_batch_get_results(#{ <<"Responses">> := R, <<"UnprocessedKeys">> := U }, Client, TableName, Retries, Acc) -> retry_sleep(Retries), execute_batch_get(Client, TableName, U, Retries - 1, [ batch_get_results(TableName, R) | Acc ]). batch_get_results(TableName, Responses) -> [ clean_map(to_map(I)) || I <- maps:get(TableName, Responses) ]. %% PUT ITEM %% @doc Put a single item into the given dynamo table, with conditions! -spec put_item( Client :: darcy_client:aws_client(), TableName :: binary(), Item :: map(), ConditionMap :: map() ) -> ok | {error, Error :: term()}. put_item(Client, TableName, Item, ConditionMap) -> ConditionExpression = maps:get(condition_expression, ConditionMap, undefined), ExpressionAttributeNames = maps:get(expression_attribute_names, ConditionMap, undefined), ExpressionAttributeValues = maps:get(expression_attribute_values, ConditionMap, undefined), Request = make_put_request(TableName, Item, ConditionExpression, ExpressionAttributeNames, ExpressionAttributeValues), case darcy_ddb_api:put_item(Client, Request) of {ok, #{}, {200, _Headers, _Client}} -> ok; {error, Error, {Code, Headers, _Client}} -> {error, {Error, [Code, Headers]}} end. make_put_request(TableName, Item, undefined, undefined, undefined) -> #{ <<"TableName">> => TableName, <<"Item">> => to_ddb(Item) }; make_put_request(TableName, Item, ConditionExpression, ExpressionAttributeNames, ExpressionAttributeValues) when is_binary(ConditionExpression) and is_map(ExpressionAttributeNames) and is_map(ExpressionAttributeValues) -> #{ <<"TableName">> => TableName, <<"ConditionExpression">> => ConditionExpression, <<"ExpressionAttributeNames">> => ExpressionAttributeNames, <<"ExpressionAttributeValues">> => to_ddb(ExpressionAttributeValues), <<"Item">> => to_ddb(Item) }. %% @doc Put a single item into the given dynamo table. -spec put_item( Client :: darcy_client:aws_client(), TableName :: binary(), Item :: map() ) -> ok | {error, Error :: term()}. put_item(Client, TableName, Item) -> put_item(Client, TableName, Item, #{}). %% @doc Put a list of items into the given Dynamo table. %% %% This function currently does not support deleting %% items (although the underlying API supports this.) %% %% Items are automatically batched into groups of 25 or %% less as required by AWS. Unprocessed keys are automatically %% retried up to 5 times. -spec batch_write_items( Client :: darcy_client:aws_client(), TableName :: binary(), Items :: [ map() ] ) -> ok | {error, Error :: term()}. batch_write_items(Client, TableName, Items) when length(Items) =< 25 -> Request = make_batch_put(TableName, Items), Result = darcy_ddb_api:batch_write_item(Client, Request), handle_batch_write_result(Client, ?RETRIES, Result); batch_write_items(Client, TableName, Items) -> {Part, Tail} = lists:split(25, Items), ok = batch_write_items(Client, TableName, Part), batch_write_items(Client, TableName, Tail). make_batch_put(TableName, Items) when length(Items) =< 25 -> #{ <<"RequestItems">> => #{ TableName => [ #{ <<"PutRequest">> => #{ <<"Item">> => to_ddb(I) } } || I <- Items ] } }. handle_batch_write_result(_Client, _N, {ok, #{ <<"UnprocessedItems">> := U }, _Details}) when map_size(U) == 0 -> ok; handle_batch_write_result(Client, N, {ok, #{ <<"UnprocessedItems">> := U }, _Details}) -> reprocess_batch_write(Client, N, U); handle_batch_write_result(_Client, _N, {error, Error, {Status, Headers, _Ref}}) -> {error, {Error, [Status, Headers]}}. reprocess_batch_write(_Client, 0, RetryItems) -> {error, {retries_exceeded, RetryItems}}; reprocess_batch_write(Client, N, RetryItems) -> retry_sleep(N), Results = darcy_ddb_api:batch_write_item(Client, #{ <<"RequestItems">> => RetryItems }), handle_batch_write_result(Client, N-1, Results). return_value(#{ <<"Item">> := Item }) -> {ok, clean_map(to_map(Item))}; return_value(#{} = M) when map_size(M) == 0 -> {error, not_found}. %% QUERY %% @doc Lookup records using the partition and range keys from %% a table or an index. %% %% Unfortunately this call requires quite a bit of understanding %% of both the Dynamo data model and the table and/or index %% structures. %% %% The query expression should take the form of a map which follows %% the <a href="http://docs.aws.amazon.com/amazondynamodb/latest/APIReference/API_Query.html">query guidelines</a> laid out in the official AWS documentation. %% %% The return value also punts on the issue of result pagination. -spec query( Client :: darcy_client:aws_client(), TableName :: binary(), QueryExpression :: map() ) -> {ok, Results :: map() } | {error, Error :: term()}. query(Client, TableName, Expr) -> query_impl(Client, [table_name(TableName), Expr]). %% @doc A query that operates on an index instead of a table. -spec query( Client :: darcy_client:aws_client(), TableName :: binary(), IndexName :: binary(), QueryExpression :: map() ) -> {ok, Results :: map() } | {error, Error :: term()}. query(Client, TableName, IndexName, Expr) -> query_impl(Client, [table_name(TableName), index_name(IndexName), Expr]). query_impl(Client, Ops) -> Request = lists:foldl(fun(M, Acc) -> maps:merge(M, Acc) end, #{}, Ops), case darcy_ddb_api:query(Client, Request) of {ok, Result, _Details } -> process_result_set(Result); {error, Error, {Status, Headers, _Client}} -> {error, {Error, {Status, Headers}}} end. process_result_set(#{ <<"Count">> := C }) when C == 0 -> {error, not_found}; process_result_set(#{ <<"Items">> := Items, <<"Count">> := C }) -> {ok, #{ <<"Count">> => C, <<"Items">> => [ clean_map(to_map(I)) || I <- Items ] } }; process_result_set(Other) -> {error, {query_error, Other}}. %% SCAN %% @doc This function executes a sequential table scan (no parallelism here) %% and returns results synchronously to the caller. If you want to scan a table %% with parallel workers, look at `scan_parallel'. %% %% %% Scans return automatically when 1 MB of data has accumulated. If more data %% is available, the atom `partial' will be returned instead of `ok'. %% %% If you want to continue your scanning activities, you must add the %% `LastEvaluatedKey' as the `ExclusiveStartKey' in the next call to %% this function's expression map. (See the official API docs for %% further details about continuing scans: %% %% https://docs.aws.amazon.com/amazondynamodb/latest/APIReference/API_Scan.html#API_Scan_ResponseSyntax.) %% %% If a table has no data, the atom `empty_table' will be returned. %% %% If a table has data, but a filter expression has filtered all results, %% the atom `no_results' will be returned. -spec scan( Client :: darcy_request:aws_client(), TableName :: binary(), Expr :: map() ) -> {ok, Result :: map()} | {ok, empty_table} | {ok, no_results} | {partial, Result :: map()} | {error, Reason :: term()}. scan(Client, TableName, Expr) -> execute_scan(Client, make_scan_request([Expr, table_name(TableName)])). %% @doc This function executes a sequential scan over an entire table. In other %% words, it will continue to make new calls until no `LastEvaluatedKey' field %% is returned from Dynamo. %% %% The items will be accumulated into a list and returned. This call is %% syntactic sugar for `scan_all/4' with a function of `same(X) -> X' passed %% in. scan_all(Client, TableName, Expr) -> scan_all(Client, TableName, Expr, fun same/1). same(X) -> X. %% @doc This function executes a sequential scan over an entire table. In other %% words, it will continue to make new calls until no `LastEvaluatedKey' field %% is returned from Dynamo. %% %% For each item returned, the function `Fun' will be executed and the results %% accumulated and returned when all valid rows from the scan query expression %% have been processed. (You may or may not care about these results if you're %% doing something in your function for the side effect.) -spec scan_all( Client :: darcy_request:aws_client(), TableName :: binary(), Expr :: map(), Fun :: function() ) -> {ok, [ term() ]} | {error, { Reason :: term(), Acc :: [ map() ]}}. scan_all(Client, TableName, Expr, Fun) -> Request = make_scan_request([Expr, table_name(TableName)]), do_scan_all(Client, Request, execute_scan(Client, Request), Fun, [], ?INITIAL_ERROR_STATE). do_scan_all(_Client, _Req, {ok, empty_table}, _Fun, _Acc, _ErrDelay) -> {ok, []}; do_scan_all(_Client, _Req, {ok, no_results}, _Fun, Acc, _ErrDelay) -> {ok, flatten(Acc)}; do_scan_all(_Client, _Req, {ok, #{ <<"Items">> := I }}, Fun, Acc, _ErrDelay) -> {ok, flatten([ batch_pmap(Fun, I) | Acc ])}; do_scan_all(Client, Req, {error, {EType, _ETxt}=Error}, Fun, Acc, EDelay) -> case maybe_retry_scan_op(get_error_type(EType), EDelay) of {true, NewErrDelay} -> error_logger:warning_msg("Got error ~p. Will retry request.", [Error]), maybe_sleep(NewErrDelay), do_scan_all(Client, Req, execute_scan(Client, Req), Fun, Acc, NewErrDelay); _ -> error_logger:error_msg("Error executing scan_all: ~p request: ~p", [Error, Req]), {error, {Error, Acc}} end; %% TODO? Maybe throttle back up if we get "a lot" of successes. do_scan_all(Client, Req, {partial, #{ <<"LastEvaluatedKey">> := LEK, <<"Items">> := I }}, Fun, Acc, ErrDelay) -> NewRequest = make_scan_request([Req, #{ <<"ExclusiveStartKey">> => LEK }]), maybe_sleep(ErrDelay), do_scan_all(Client, NewRequest, execute_scan(Client, NewRequest), Fun, [ batch_pmap(Fun, I) | Acc], ErrDelay). maybe_sleep(#error_delay{ delay = 0 }) -> ok; maybe_sleep(#error_delay{ delay = D }) -> error_logger:warning_msg("THROTTLE: Pid ~p sleeping for ~p ms before next request...", [self(), D]), timer:sleep(D). maybe_retry_scan_op(<<"ProvisionedThroughputExceededException">>, #error_delay{ delay = 0 } = ErrState) -> {true, ErrState#error_delay{ delay = ?INITIAL_ERROR_DELAY }}; maybe_retry_scan_op(<<"ProvisionedThroughputExceededException">>, #error_delay{ delay = D } = ErrState) when D > ?MAX_ERROR_DELAY -> {false, ErrState}; maybe_retry_scan_op(<<"ProvisionedThroughputExceededException">>, #error_delay{ delay = D } = ErrState) -> {true, ErrState#error_delay{ delay = D*2 }}; maybe_retry_scan_op(_Type, ErrorState) -> {false, ErrorState}. get_error_type(Type) -> [_, Exception] = binary:split(<<"#">>, Type), Exception. %% @doc Scan a table in parallel using the given expression. This %% function is equivalent to `scan_parallel/7' with a timeout value of 60000 %% milliseconds. %% %% Instead of returning the coordinator pid to the caller, this %% function blocks and waits for the return values from the %% coordinator. scan_parallel(Client, TableName, Expr, Fun, SegmentCount) -> scan_parallel(Client, TableName, Expr, Fun, ?BIG_TIMEOUT, SegmentCount). scan_parallel(Client, TableName, Expr, Fun, Timeout, SegmentCount) -> {Ref, _Pid} = scan_parallel(Client, TableName, Expr, Fun, Timeout, SegmentCount, self()), receive {Ref, {Results, []}} -> {ok, Results}; {Ref, {Partial, Errors}} -> {error, {Errors, Partial}}; Other -> Other after Timeout -> {error, scan_timeout} end. %% @doc Scan a table in parallel using the given expression. %% %% DynamoDB supports parallel scans using a partitioning technique described in the Developer Guide. %% %% https://docs.aws.amazon.com/amazondynamodb/latest/developerguide/Scan.html %% %% This function spawns a coordinator process which in turn spawns a set of %% workers, one per segment count which executes the `scan_all/4' function. %% %% <B>N.B.</B>: This operation can consume a lot of read capacity. It is a good %% idea to limit the number of segments used in a scan operation. %% %% The coordinator that is spawned here is <B>linked</B> to the caller. If you %% want more robust error handling, you should trap exit messages by the %% caller's process. %% %% Since this function returns the "raw" results, you will have to %% handle them appropriately within your own receive block. %% %% They will be in the form of `{Ref, {Results, Errors}}' where %% `Ref' matches the reference returned by this call. The coordinator %% pid is the process which is coordinating the parallel workers. %% (You may or may not care about those.) -spec scan_parallel( Client :: darcy_request:aws_client(), TableName :: binary(), Expr :: map(), Fun :: function(), Timeout :: pos_integer(), SegmentCount :: pos_integer(), ReplyPid :: pid()) -> {Ref :: reference(), Coordinator :: pid()}. scan_parallel(Client, TableName, Expr, Fun, Timeout, SegmentCount, ReplyPid) -> Reqs = [ make_scan_request([Expr, table_name(TableName), make_segments(N, SegmentCount)]) || N <- lists:seq(0, SegmentCount - 1) ], Ref = make_ref(), Pid = spawn_link(fun() -> start_coordinator(Client, Ref, Timeout, Fun, Reqs, ReplyPid) end), {Ref, Pid}. results_ok({ok, _}) -> true; results_ok(_) -> false. start_coordinator(Client, Ref, Timeout, Fun, Reqs, Reply) -> L = pmap(fun(Req) -> worker_scan_all(Client, Req, Fun) end, Reqs, Timeout, {Ref, {[], scan_timeout}}), ReplyMsg = make_reply_msg(Ref, L), Reply ! {Ref, ReplyMsg}. make_reply_msg(Ref, {Ref, {[], scan_timeout}}) -> {[], {error, scan_timeout}}; make_reply_msg(_Ref, Results) -> {Res, Err} = lists:partition(fun results_ok/1, Results), {lists:flatten([ R || {ok, R} <- Res ]), Err}. worker_scan_all(Client, Request, Fun) -> do_scan_all(Client, Request, execute_scan(Client, Request), Fun, [], ?INITIAL_ERROR_STATE). make_segments(N, Count) -> #{ <<"Segment">> => N, <<"TotalSegments">> => Count }. make_scan_request(Ops) -> lists:foldl(fun(M, Acc) -> maps:merge(M, Acc) end, #{}, Ops). execute_scan(Client, Request) -> case darcy_ddb_api:scan(Client, Request) of {ok, Result, _Details } -> process_scan_result(Result); {error, Error, {Status, Headers, _Client}} -> {error, {Error, {Status, Headers}}} end. process_scan_result(#{ <<"Count">> := 0, <<"ScannedCount">> := 0 }) -> {ok, empty_table}; process_scan_result(#{ <<"Count">> := 0, <<"ScannedCount">> := _SC }) -> {ok, no_results}; process_scan_result(#{ <<"Items">> := Items, <<"LastEvaluatedKey">> := _LEK } = M) -> NewItems = [ clean_map(to_map(I)) || I <- Items ], {partial, maps:put(<<"Items">>, NewItems, M)}; process_scan_result(#{ <<"Items">> := Items } = M) -> NewItems = [ clean_map(to_map(I)) || I <- Items ], {ok, maps:put(<<"Items">>, NewItems, M)}; process_scan_result(Other) -> {error, {scan_error, Other}}. flatten(L) when is_list(L) -> lists:flatten(lists:reverse(L)). %% @doc This function returns a map without any Dynamo specific type tuples, %% which is useful for passing around internally in an application that doesn't %% care or understand Dynamo data types. clean_map(M) when is_map(M) -> maps:map(fun(_K, {_, V}) -> V; (_K, V) -> V end, M). %% @doc This is the default decoding function for binary data. It base64 %% decodes the binary, and decompresses it. default_decode(Blob) -> zlib:uncompress(base64:decode(Blob)). %% @doc This is the default encoding function for binary data. It compresses %% the data and base64 encodes it. default_encode(Data) -> base64:encode(zlib:compress(Data)). %% @doc Translate from a "raw" JSON map representation of a Dynamo %% data item to an Erlang data item. Uses the following tuples %% to remove ambiguities in Erlang JSON encoding: %% <ul> %% <li>`NULL' values are returned as `undefined'</li> %% <li>`{blob, Binary}'</li> %% <li>`{list, List}'</li> %% <li>`{string_set, Set}' (internally stored as an ordset)</li> %% <li>`{number_set, Set}' (internally stored as an ordset)</li> %% </ul> to_map(M) when is_map(M) -> maps:map(fun(_K, V) when is_map(V) -> unddt(V); (_K, V) -> V end, M). %% @private unddt(#{ <<"B">> := V }) -> {M, F, A} = application:get_env(darcy, blob_decode_fun, {darcy, default_decode, []}), {blob, erlang:apply(M, F, [V | A])}; unddt(#{ <<"N">> := V }) -> %% could be an integer or a float. Try integer conversion %% first. try binary_to_integer(V) catch error:badarg -> binary_to_float(V) end; unddt(#{ <<"S">> := V }) -> V; unddt(#{ <<"BOOL">> := <<"true">> }) -> true; unddt(#{ <<"BOOL">> := true }) -> true; unddt(#{ <<"BOOL">> := <<"false">> }) -> false; unddt(#{ <<"BOOL">> := false }) -> false; unddt(#{ <<"L">> := V }) -> {list, [ unddt(E) || E <- V ]}; unddt(#{ <<"M">> := V }) -> maps:map(fun(_K, Val) -> unddt(Val) end, V); unddt(#{ <<"SS">> := V }) -> {string_set, ?SET:from_list([ E || E <- V ])}; unddt(#{ <<"NS">> := V }) -> {number_set, ?SET:from_list([ binary_to_integer(E) || E <- V ])}; unddt(#{ <<"NULL">> := _V }) -> undefined; unddt(Other) -> erlang:error({error, badarg}, [Other]). %% @doc This function takes an Erlang map and attempts to encode it using Dynamo %% data type annotations. Because there are ambiguities in how Erlang internally %% represents things like strings, lists and sets, tagged tuples are used to %% remove ambiguity. They are the same tagged tuples as above: %% <ul> %% <li>`undefined' is stored as a `NULL' data type</li> %% <li>`{blob, Binary}'</li> %% <li>`{list, List}'</li> %% <li>`{string_set, Set}' (internally stored as an ordset)</li> %% <li>`{number_set, Set}' (internally stored as an ordset)</li> %% </ul> %% %% Generally, you should try to modify your internal data representation values %% to remove these ambiguities <i>before</i> you pass them into this function. to_ddb(M) when is_map(M) -> maps:map(fun(_K, V) -> ddt(V) end, M); to_ddb(Other) -> erlang:error({error, badarg}, [Other]). %% @private %% see http://boto3.readthedocs.io/en/latest/_modules/boto3/dynamodb/types.html ddt(undefined) -> #{ <<"NULL">> => true }; ddt(null) -> #{ <<"NULL">> => true }; ddt({blob, Data}) -> {M, F, A} = application:get_env(darcy, blob_encode_fun, {darcy, default_encode, []}), #{ <<"B">> => erlang:apply(M, F, [ Data | A ]) }; ddt({list, L}) -> #{ <<"L">> => [ ddt(E) || E <- L ] }; ddt({string_set, S}) -> #{ <<"SS">> => [ ddt(E) || E <- ?SET:to_list(S) ] }; ddt({number_set, S}) -> #{ <<"NS">> => [ ddt(E) || E <- ?SET:to_list(S) ] }; ddt(V) when is_integer(V) -> #{ <<"N">> => number_to_binary(V) }; ddt(V) when is_float(V) -> #{ <<"N">> => number_to_binary(V) }; ddt(V) when is_binary(V) -> #{ <<"S">> => V }; ddt(V) when is_boolean(V) -> #{ <<"BOOL">> => V }; ddt(V) when is_map(V) -> #{ <<"M">> => maps:map(fun(_K, Val) -> ddt(Val) end, V) }; ddt(V) when is_list(V) -> try #{ <<"S">> => list_to_binary(V) } catch _:_ -> #{ <<"L">> => [ ddt(E) || E <- V ] } end; ddt(Other) -> erlang:error({error, badarg}, [Other]). number_to_binary(V) when is_integer(V) -> integer_to_binary(V); number_to_binary(V) when is_float(V) -> float_to_binary(V, [{decimals, 20}, compact]). %% sleep for RETRIES - N * 1000 milliseconds before retrying an operation. %% Current RETRIES value is 5. Always sleeps for <i>at least</i> 1000 %% milliseconds. %% N = 5 => 1000 ms sleep %% N = 4 => 1000 ms sleep %% N = 3 => 2000 ms sleep %% N = 2 => 3000 ms sleep %% N = 1 => 4000 ms sleep retry_sleep(N) -> S = max(1000, (?RETRIES-N) * 1000), timer:sleep(S). %% split the big list into smaller batches and execute them in parallel. batch_pmap(F, List) when length(List) =< ?BATCH_SIZE -> pmap(F, List); batch_pmap(F, BigList) -> Len = length(BigList), I = items_per_batch(Len), PC = lists:seq(1, Len, I), Batches = make_batches(BigList, I, PC, []), pmap(fun(E) -> lists:map(F, E) end, Batches). items_per_batch(Len) -> case Len div ?BATCH_SIZE of I when I =< ?BATCH_MAX -> ?BATCH_SIZE; _ -> Len div ?BATCH_MAX end. make_batches(_, _, [], Acc) -> lists:reverse(Acc); make_batches(L, Len, [H|T], Acc) -> make_batches(L, Len, T, [ lists:sublist(L, H, Len) | Acc ]). %% parallel map %% http://erlang.org/pipermail/erlang-questions/2009-January/041214.html %% %% TODO: Maybe we do not care about the order messages are received pmap(F, Arglist) -> pmap(F, Arglist, ?TIMEOUT, pmap_timeout). pmap(F, Arglist, Timeout, TimeoutError) -> S = self(), TaskID = make_ref(), Workers = lists:map( fun(X) -> spawn_link(fun() -> do_F(S, TaskID, F, X) end) end, Arglist), gather(Workers, TaskID, Timeout, TimeoutError). do_F(Caller, TaskID, F, X) -> Caller ! {self(), TaskID, catch(F(X))}. gather([], _, _, _) -> []; gather([W|R], TaskID, Timeout, TimeoutError) -> receive {W, TaskID, Val} -> [Val | gather(R, TaskID, Timeout, TimeoutError)] after Timeout -> TimeoutError end. %% Tests -ifdef(TEST). -include_lib("eunit/include/eunit.hrl"). make_put_request_test() -> TableName = <<"foo">>, Item = #{ <<"Student">> => <<"Foo">>, <<"Version">> => <<"totally_a_uuid">> }, ConditionExpression = <<"#version = :old_version OR attribute_not_exists(#version)">>, ExpressionAttributeNames = #{<<"#version">> => <<"Version">>}, ExpressionAttributeValues = #{<<":old_version">> => <<"totally_a_uuid">>}, Request = make_put_request(TableName, Item, ConditionExpression, ExpressionAttributeNames, ExpressionAttributeValues), Expected = #{<<"ConditionExpression">> => <<"#version = :old_version OR attribute_not_exists(#version)">>, <<"ExpressionAttributeNames">> => #{<<"#version">> => <<"Version">>}, <<"ExpressionAttributeValues">> => #{<<":old_version">> => #{<<"S">> => <<"totally_a_uuid">>}}, <<"Item">> => #{<<"Student">> => #{<<"S">> => <<"Foo">>}, <<"Version">> => #{<<"S">> => <<"totally_a_uuid">>}}, <<"TableName">> => <<"foo">>}, ?assertEqual(Expected, Request). to_ddb_test() -> Raw = #{ <<"Grades">> => {list, [17,39,76,27]}, <<"Average">> => 39.75, <<"Student">> => <<"Quentin">>, <<"Subject">> => <<"Science">> }, Expected = #{<<"Grades">> => #{<<"L">> => [#{<<"N">> => <<"17">>}, #{<<"N">> => <<"39">>}, #{<<"N">> => <<"76">>}, #{<<"N">> => <<"27">>}]}, <<"Average">> => #{<<"N">> => <<"39.75">>}, <<"Student">> => #{<<"S">> => <<"Quentin">>}, <<"Subject">> => #{<<"S">> => <<"Science">>}}, ?assertEqual(Expected, to_ddb(Raw)). to_map_test() -> Raw = #{<<"Grades">> => #{<<"L">> => [#{<<"N">> => <<"17">>}, #{<<"N">> => <<"39">>}, #{<<"N">> => <<"76">>}, #{<<"N">> => <<"27">>}]}, <<"Student">> => #{<<"S">> => <<"Quentin">>}, <<"Subject">> => #{<<"S">> => <<"Science">>}}, Expected = #{ <<"Grades">> => {list, [17, 39, 76, 27]}, <<"Student">> => <<"Quentin">>, <<"Subject">> => <<"Science">> }, ?assertEqual(Expected, to_map(Raw)). clean_map_test() -> Raw = #{<<"Grades">> => #{<<"L">> => [#{<<"N">> => <<"17">>}, #{<<"N">> => <<"39">>}, #{<<"N">> => <<"76">>}, #{<<"N">> => <<"27">>}]}, <<"Student">> => #{<<"S">> => <<"Quentin">>}, <<"Subject">> => #{<<"S">> => <<"Science">>}}, Expected = #{ <<"Grades">> => [17, 39, 76, 27], <<"Student">> => <<"Quentin">>, <<"Subject">> => <<"Science">> }, ?assertEqual(Expected, clean_map(to_map(Raw))). -endif.
src/darcy.erl
0.558568
0.444866
darcy.erl
starcoder
-module(interval_table). -export([new/0, gap_size/1, span/2, assign/2, shrink/2, lookup/2]). -export_type([table/0]). -type key() :: any(). -record(table, {gaps = [] :: list(interval:interval()), intervals = [] :: list({interval:interval(), key()})}). -opaque table() :: #table{}. -spec new() -> table(). new() -> #table{gaps = [interval:new(rational:new(0, 1), rational:new(1, 1))]}. %% @doc Get the span covered by all gaps in the table. -spec gap_size(table()) -> rational:rational(). gap_size(#table{gaps=Gaps}) -> lists:foldl( fun(Interval, Size) -> rational:add(Size, interval:length(Interval)) end, rational:new(0, 1), Gaps). %% @doc Assign all keys in `NewKeys' to gaps in `Table'. Uses the %% greedy strategy (largest span/largest gap first strategy). -spec assign(NewKeys::#{key() => rational:rational()}, Table::table()) -> table(). assign(NewKeys, Table) when map_size(NewKeys) =:= 0 -> Table; assign(NewKeys, Table=#table{gaps=[]}) -> NewNewKeys = maps:filter(fun(_, Span) -> Span =/= rational:new(0) end, NewKeys), case maps:size(NewNewKeys) of 0 -> Table; _ -> throw({badarg, "Span too large for available gaps."}) end; assign(NewKeys, Table=#table{gaps=Gaps, intervals=Intervals}) -> {LargestKey, LargestSpan} = maps:fold( fun(Key, Span, Acc={_, LargestSpan}) -> case rational:compare(Span, LargestSpan) of eq -> {Key, Span}; lt -> Acc; gt -> {Key, Span} end end, {'$no_key$', rational:new(0)}, NewKeys), LargestGap = lists:foldl( fun(Gap, LargestGap) -> case rational:compare(interval:length(Gap), interval:length(LargestGap)) of eq -> Gap; lt -> LargestGap; gt -> Gap end end, interval:new(rational:new(0), rational:new(0)), Gaps), RemainingGaps = lists:delete(LargestGap, Gaps), case interval:split(LargestSpan, LargestGap, left) of {empty, _} -> % if the key with the largest span has an empty span the % assignment is done Table; {Interval, empty} -> assign( maps:update_with( LargestKey, fun(Span) -> rational:subtract(Span, interval:length(Interval)) end, NewKeys), Table#table{ gaps = RemainingGaps, intervals = insert_inorder({Interval, LargestKey}, Intervals)}); {Interval, Gap} -> assign( maps:remove(LargestKey, NewKeys), Table#table{ gaps = [Gap | RemainingGaps], intervals = insert_inorder({Interval, LargestKey}, Intervals)}) end. insert_inorder({Gap, Key}, []) -> [{Gap, Key}]; insert_inorder({Gap, Key}, [Assigned={I, _}|Intervals]) -> % no case for eq because we should never insert an interval that % is equal to an interval that is already assigned. case interval:preceeds(Gap, I) of true -> [{Gap, Key},Assigned|Intervals]; false -> [Assigned|insert_inorder({Gap, Key}, Intervals)] end. %% @doc Get the total span assigned to `Key'. -spec span(key(), table()) -> rational:rational(). span(Key, IntervalTable) -> lists:foldl( fun({Interval, AssignedTo}, Span) -> if AssignedTo =:= Key -> rational:add(interval:length(Interval), Span); AssignedTo =/= Key -> Span end end, rational:new(0), IntervalTable#table.intervals). %% @doc Create gaps by shrinking the span assigned to each key. %% @param Changes a map ``#{Key => ShrinkAmount}'' giving the amount %% by which to shrink the span assigned to `Key'. %% @reutrns the table with gaps. -spec shrink(#{key() => rational:rational()}, table()) -> table(). shrink(Changes, Table=#table{gaps=Gaps, intervals=Intervals}) -> {NewGaps, NewIntervals} = cut_shift(Changes, Intervals), Table#table{gaps = merge_gaps(NewGaps, Gaps), intervals = NewIntervals}. %% merge two lists of gaps combining adjacent gaps into single larger %% gaps. merge_gaps(Gaps1, Gaps2) -> SortedGaps = lists:merge(fun interval:preceeds/2, lists:sort(fun interval:preceeds/2, Gaps1), lists:sort(fun interval:preceeds/2, Gaps2)), lists:foldr(fun(Gap, []) -> [Gap]; (Gap, [PreceedingGap|Gaps]) -> case interval:adjacent(Gap, PreceedingGap) of true -> [interval:merge(Gap, PreceedingGap) | Gaps]; false -> [Gap, PreceedingGap | Gaps] end end, [], SortedGaps). cut_shift(AssignmentChanges, Intervals) -> SubtractLength = fun(I) -> fun(X) -> rational:subtract(X, interval:length(I)) end end, {_, Gaps, Remaining, _} = lists:foldl( fun({I, Key}, {Changes, Gaps, Remaining, left}) -> Span = maps:get(Key, Changes, rational:new(0)), case interval:split(Span, I, left) of {empty, _} -> {Changes, Gaps, [{I, Key}|Remaining], right}; {Gap, empty} -> {maps:update_with( Key, SubtractLength(Gap), rational:new(0), Changes), [Gap|Gaps], Remaining, left}; {Gap, Rem} -> {maps:update_with( Key, SubtractLength(Gap), rational:new(0), Changes), [Gap|Gaps], [{Rem, Key}|Remaining], right} end; ({I, Key}, {Changes, Gaps, Remaining, right}) -> Span = maps:get(Key, Changes, rational:new(0)), case interval:split(Span, I, right) of {_, empty} -> {Changes, Gaps, [{I, Key}|Remaining], right}; {empty, Gap} -> {maps:update_with( Key, SubtractLength(Gap), rational:new(0), Changes), [Gap|Gaps], Remaining, left}; {Rem, Gap} -> {maps:update_with( Key, SubtractLength(Gap), rational:new(0), Changes), [Gap|Gaps], [{Rem, Key}|Remaining], left} end end, {AssignmentChanges, [], [], right}, Intervals), {Gaps, Remaining}. %% @doc Look up the key assigned `X'. %% %% @returns `{ok, Key}' if `X' falls in the interval assigned to `Key' %% or `unassigned' if `X' falls in a gap. -spec lookup(rational:rational(), table()) -> {ok, key()} | unassigned. lookup(X, #table{intervals=Intervals}) -> case lists:filter(fun({I, _}) -> interval:contains(X, I) end, Intervals) of [{_, Key}|_] -> {ok, Key}; [] -> unassigned end.
src/interval_table.erl
0.676834
0.488649
interval_table.erl
starcoder
% Licensed under the Apache License, Version 2.0 (the "License"); you may not % use this file except in compliance with the License. You may obtain a copy of % the License at % % http://www.apache.org/licenses/LICENSE-2.0 % % Unless required by applicable law or agreed to in writing, software % distributed under the License is distributed on an "AS IS" BASIS, WITHOUT % WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the % License for the specific language governing permissions and limitations under % the License. -module(vtree_util). -include("vtree.hrl"). -export([calc_mbb/2, nodes_mbb/2, min/2, max/2, calc_perimeter/1, calc_volume/1, intersect_mbb/3, find_min_value/2, within_mbb/3]). -spec min(Tuple::{any(), any()} | [any()], Less::fun()) -> Min::any(). min({A, B}, Less) -> case Less(A, B) of true -> A; false -> B end; min([H|T], Less) -> min(T, Less, H). -spec min(List::[any()], Less::fun(), Min::any()) -> Min::any(). min([], _Less, Min) -> Min; min([H|T], Less, Min) -> Min2 = case Less(H, Min) of true -> H; false -> Min end, min(T, Less, Min2). -spec max(Tuple::{any(), any()} | [any()], Less::fun()) -> Max::any(). max({A, B}, Less) -> case Less(A, B) of true -> B; false -> A end; max([H|T], Less) -> max(T, Less, H). -spec max(List::[any()], Less::fun(), Max::any()) -> Max::any(). max([], _Less, Max) -> Max; max([H|T], Less, Max) -> Max2 = case Less(H, Max) of true -> Max; false -> H end, max(T, Less, Max2). % Calculate the perimeter of list of 2-tuples that contain a min and a max % value -spec calc_perimeter(mbb()) -> number(). calc_perimeter(Values) -> lists:foldl(fun({Min, Max}, Acc) -> Acc + (Max-Min) end, 0, Values). % Calculate the volume of list of 2-tuples that contain a min and a max % value -spec calc_volume(mbb()) -> number(). calc_volume(Values) -> lists:foldl(fun({Min, Max}, Acc) -> Acc * (Max-Min) end, 1, Values). % Calculate the enclosing bounding box from a list of bounding boxes -spec calc_mbb(List::[[{any(), any()}]], Less::fun()) -> [{any(), any()}]. calc_mbb([H|T], Less) -> calc_mbb(T, Less, H). -spec calc_mbb(List::[[{any(), any()}]], Less::fun(), Mbb::[{any(), any()}]) -> Mbb::[{any(), any()}]. calc_mbb([], _Less, Mbb) -> Mbb; calc_mbb([H|T], Less, Mbb) -> Mbb2 = lists:map( fun({{Min, Max}, {MinMbb, MaxMbb}}) -> {?MODULE:min({Min, MinMbb}, Less), ?MODULE:max({Max, MaxMbb}, Less)} end, lists:zip(H, Mbb)), calc_mbb(T, Less, Mbb2). % Calculate the enclosing MBB from a list of nodes -spec nodes_mbb(Nodes :: [#kv_node{} | #kp_node{} | split_node()], Less :: lessfun()) -> mbb(). nodes_mbb([#kv_node{}|_]=Nodes, Less) -> Mbbs = [Node#kv_node.key || Node <- Nodes], vtree_util:calc_mbb(Mbbs, Less); nodes_mbb([#kp_node{}|_]=Nodes, Less) -> Mbbs = [Node#kp_node.key || Node <- Nodes], vtree_util:calc_mbb(Mbbs, Less); nodes_mbb(Nodes, Less) -> {Mbbs, _} = lists:unzip(Nodes), vtree_util:calc_mbb(Mbbs, Less). % Returns the intersection of two MBBs. Touching also counts as intersection. -spec intersect_mbb(A :: mbb(), B :: mbb(), Less :: lessfun()) -> mbb() | overlapfree. intersect_mbb(A, B, Less) -> intersect_mbb0(lists:zip(A, B), Less, []). -spec intersect_mbb0([{{keyval(), keyval()}, {keyval(), keyval()}}], Less :: lessfun(), Acc :: mbb()) -> mbb() | overlapfree. intersect_mbb0([], _Less, Acc) -> lists:reverse(Acc); % If both values are `nil`, it is like a wildcard, it covers the other % range completely intersect_mbb0([{{nil, nil}, MinMax}|T], Less, Acc) -> intersect_mbb0(T, Less, [MinMax|Acc]); intersect_mbb0([{MinMax, {nil, nil}}|T], Less, Acc) -> intersect_mbb0(T, Less, [MinMax|Acc]); % If one end of the range is `nil` it's an open range and the intersection % will be the one of the other given range. The guards are needed to prevent % endless loops. intersect_mbb0([{{nil, MaxA}, {MinB, MaxB}}|T], Less, Acc) when MinB =/= nil -> intersect_mbb0([{{MinB, MaxA}, {MinB, MaxB}}|T], Less, Acc); intersect_mbb0([{{MinA, nil}, {MinB, MaxB}}|T], Less, Acc) when MaxB =/= nil -> intersect_mbb0([{{MinA, MaxB}, {MinB, MaxB}}|T], Less, Acc); intersect_mbb0([{{MinA, MaxA}, {nil, MaxB}}|T], Less, Acc) when MinA =/= nil -> intersect_mbb0([{{MinA, MaxA}, {MinA, MaxB}}|T], Less, Acc); intersect_mbb0([{{MinA, MaxA}, {MinB, nil}}|T], Less, Acc) when MaxA =/= nil -> intersect_mbb0([{{MinA, MaxA}, {MinB, MaxA}}|T], Less, Acc); % All `nil` cases are resolved, do the actual work intersect_mbb0([{{MinA, MaxA}, {MinB, MaxB}}|T], Less, Acc) -> Min = vtree_util:max({MinA, MinB}, Less), Max = vtree_util:min({MaxA, MaxB}, Less), case Less(Max, Min) of true -> overlapfree; false -> intersect_mbb0(T, Less, [{Min, Max}|Acc]) end. % Find the minumum value from a list of things. One item from the % list will be passed into the the suppplied `MinFun`, which returns % some numner. The lowest number together with the corresponding item % will be returned. -spec find_min_value(MinFun :: fun(), Data :: [any()]) -> {number(), any()}. find_min_value(MinFun, [_|_]=Data) -> lists:foldl( fun(Item, {MinVal, _}=Acc) -> Val = MinFun(Item), case (Val < MinVal) orelse (MinVal =:= nil) of true -> {Val, Item}; false -> Acc end end, {nil, []}, Data). % Returns true if MBB `A` is completely within `B`, i.E. merging `A` with `B` % wouldn't expand `B` and just return `B`. -spec within_mbb(A :: mbb(), B :: mbb(), Less :: lessfun()) -> true | false. within_mbb(A, B, Less) -> within_mbb0(lists:zip(A, B), Less). -spec within_mbb0([{{keyval(), keyval()}, {keyval(), keyval()}}], Less :: lessfun()) -> true | false. within_mbb0([], _Less) -> true; within_mbb0([{{MinA, MaxA}, {MinB, MaxB}}|T], Less) -> case Less(MinA, MinB) orelse Less(MaxB, MaxA) of true -> false; _ -> within_mbb0(T, Less) end.
vtree/src/vtree_util.erl
0.818845
0.660638
vtree_util.erl
starcoder
%% -------- Overview --------- %% %% There are two primary types of exchange sorted %% - a full exchange aimed at implementations with cached trees, where the %% cached trees represent all the data in the location, and the comparion is %% between two complete data sets %% - a partial exchange where it is expected that trees will be dynamically %% created covering a subset of data within the location %% %% The full exchange assumes access to cached trees, with a low cost of %% repeated access, and a relatively high proportion fo the overall cost in %% network bandwitdh. These exchanges go through the following process: %% %% - Root Compare (x n) %% - Branch Compare (x n) %% - Clock Compare %% - Repair %% %% The partial, dynamic tree exchange is based on dynamically produced trees, %% where a relatively high proportion of the cost is in the production of the %% trees. In a tree exchange, whole trees are compared (potentially reduced by %% use of a segment filter), until the delta stops decreasing at a significant %% rate and a Clock Compare is run. So these exchanges for through the %% following process: %% %% - Tree Compare (x n) %% - Clock Compare %% - Repair %% %% Each exchange has a 'blue' list and a 'pink' list. Each list (blue and %% pink) is a set of partitions pertinent to this exchange, with the state %% to be compared being the merging of all the trees referenced by the list. %% %% The lists can be a single item each (for a pairwise exchange), or a %% ring-size number of partitions for a coverage query exchange. %% %% -------- Root Compare --------- %% %% This allows the comparison between the roots of trees. Each root (with a %% tree size of large and 4-byte hashes), will be 4KB in size. The outcome of %% the comparison should be a set of BranchIDs where the (merged) roots are %% showing differences. %% %% The Exchange can terminate if the set of differences is empty. A timeout %% should trigger the commencement of the next stage (to provide a pause %% between vnode requests). %% %% -------- Root Confirm --------- %% %% In the next stage the roots are again requested, received and compared. %% Again a set of branchIDs which differ is created - and the set of %% confirmed deltas is the intersection of the sets generated from both root %% exchanges. %% %% The purpose of the confirm stage is to rule out false negative results %% related to timing differences in the result of PUTs. %% %% The Exchange can terminate if the set of differences is empty. A timeout %% should trigger the commencement of the next stage (to provide a pause %% between vnode requests). %% %% -------- Branch Compare / Confirm --------- %% %% The set of branch ID differences should now be fetched (Compare), and then %% re-fetched following a timeout (Confirm) to produce a set of SegmentIDs (or %% tree leaves) that represent differences ebwteen blue and pink, eliminating %% false negatives related to timing as with the Root Compare and Confirm. %% %% Each Branch is 1KB in size. So if there are more than 16 branches which %% have differrences, only 16 should be chosen for the Compare and Confirm to %% control the volume of network traffic prompted by the exchange. %% %% The Exchange can terminate if the set of differences is empty. A timeout %% should trigger the commencement of the next stage (to provide a pause %% between vnode requests). %% %% -------- Clock Compare --------- %% %% The final stage is clock compare. The clock compare can be done on up to %% 128 segments across a maximum of 8 BranchIDs. This is to control the %% potential overhead of the comparison and subsequent repairs. This may mean %% for empty vnodes o(1000) exchanges may be required to fully recover the %% store. However, in these cases it is likely that handoff and read repair %% is already recovering the data so overly-aggressive read repair is %% unnecessary. %% -module(aae_exchange). -behaviour(gen_fsm). -ifdef(fsm_deprecated). -compile({nowarn_deprecated_function, [{gen_fsm, start, 3}, {gen_fsm, send_event, 2}]}). -endif. -include("include/aae.hrl"). -define(TRANSITION_PAUSE_MS, 500). % A pause between phases - allow queue lengths to change, and avoid % generating an excess workload for AAE -define(CACHE_TIMEOUT_MS, 60000). % 60 seconds (used in fetch root/branches) -define(SCAN_TIMEOUT_MS, 600000). % 10 minutes (used in fetch clocks) -define(UNFILTERED_SCAN_TIMEOUT_MS, 14400000). % 4 hours (used in fetch trees with no filters) -define(MAX_RESULTS, 128). % Maximum number of results to request in one round of -define(WORTHWHILE_REDUCTION, 0.3). % If the last comparison of trees has reduced the size of the dirty leaves % by 30%, probably worth comparing again before a clock fetch is run. % Number a suck-teeth estimate, not even a fag-packet calculation involved. -define(WORTHWHILE_REDUCTION_CACHED, 0). % When checking a cached tree - then even a small reduction is worth % another check, as the cost per check is so small. This changed after % seeing cost of false negative results in large stores. -define(WORTHWHILE_FILTER, 256). % If the number of segment IDs to pass into a filter is too large, the % filter is probably not worthwhile - more effort checking the filter, than % time saved in the accumulator. Another suck-teeth estimate here as to % what this value is, at this level with a small tree it will save opening % all but one block in most slots (with the sst file). I suspect the % optimal number is more likely to be higher than lower. -export([init/1, handle_sync_event/4, handle_event/3, handle_info/3, terminate/3, code_change/4]). -export([waiting_all_results/2, prepare_full_exchange/2, prepare_partial_exchange/2, root_compare/2, branch_compare/2, clock_compare/2, tree_compare/2, merge_root/2, merge_branches/2]). -export([compare_roots/2, compare_branches/2, compare_clocks/2, compare_trees/2]). -export([insync_responses/0]). -export([start/4, start/7, reply/3]). -include_lib("eunit/include/eunit.hrl"). -record(state, {root_compare_deltas = [] :: list(), branch_compare_deltas = [] :: list(), tree_compare_deltas = [] :: list(), key_deltas = [] :: list(), repair_fun :: repair_fun()|undefined, reply_fun :: reply_fun()|undefined, blue_list = [] :: input_list(), pink_list = [] :: input_list(), exchange_id = "not_set" :: list(), blue_returns = {0, 0} :: {integer(), integer()}, pink_returns = {0, 0} :: {integer(), integer()}, pink_acc, blue_acc, merge_fun, start_time = os:timestamp() :: erlang:timestamp(), pending_state :: atom(), reply_timeout = 0 :: integer(), exchange_type :: exchange_type(), exchange_filters = none :: filters(), last_tree_compare = none :: list(non_neg_integer())|none, last_root_compare = none :: list(non_neg_integer())|none, last_branch_compare = none :: list(non_neg_integer())|none, tree_compares = 0 :: integer(), root_compares = 0 :: integer(), branch_compares = 0 :: integer(), prethrottle_branches = 0 :: non_neg_integer(), prethrottle_leaves = 0 :: non_neg_integer(), transition_pause_ms = ?TRANSITION_PAUSE_MS :: pos_integer(), log_levels :: aae_util:log_levels()|undefined, scan_timeout = ?SCAN_TIMEOUT_MS :: non_neg_integer(), max_results = ?MAX_RESULTS :: pos_integer(), purpose :: atom()|undefined }). -type branch_results() :: list({integer(), binary()}). % Results to branch queries are a list mapping Branch ID to the binary for % that branch -type exchange_state() :: #state{}. -type exchange_type() :: full|partial. -type compare_state() :: root_compare|tree_compare|branch_compare|clock_compare. -type closing_state() :: compare_state()|timeout|error|not_supported. -type bucket_range() :: {binary(), binary()}|binary()|all. -type key_range() :: {binary(), binary()}|all. -type modified_range() :: {non_neg_integer(), non_neg_integer()}|all. -type segment_filter() :: {segments, list(non_neg_integer()), leveled_tictac:tree_size()}|all. -type hash_method() :: pre_hash|{rehash, non_neg_integer()}. -type filters() :: {filter, bucket_range(), key_range(), leveled_tictac:tree_size(), segment_filter(), modified_range(), hash_method()}|none. % filter to be used in partial exchanges -type option_item() :: {transition_pause_ms, pos_integer()}| {scan_timeout, non_neg_integer()}| {log_levels, aae_util:log_levels()}| {max_results, non_neg_integer()}| {purpose, atom()}. -type options() :: list(option_item()). -type send_message() :: fetch_root | {fetch_branches, list(non_neg_integer())} | {fetch_clocks, list(non_neg_integer())} | {merge_tree_range, filters()} | {fetch_clocks_range, filters()}. -type send_fun() :: fun((send_message(), list(tuple())|all, blue|pink) -> ok). -type input_list() :: [{send_fun(), list(tuple())|all}]. % The Blue List and the Pink List are made up of: % - a SendFun, which should be a 3-arity function, taking a preflist, % a message and a colour to be used to flag the reply; % - a list of preflists, to be used in the SendFun to be filtered by the % target. The Preflist might be {Index, Node} for remote requests or % {Index, Pid} for local requests % For partial exchanges only, the preflist can and must be set to 'all' -type repair_input() :: {{any(), any()}, {any(), any()}}. % {{Bucket, Key}, {BlueClock, PinkClock}}. -type repair_fun() :: fun((list(repair_input())) -> ok). % Input will be Bucket, Key, Clock -type reply_fun() :: fun(({closing_state(), non_neg_integer()}) -> ok). -define(FILTERIDX_SEG, 5). -define(FILTERIDX_TRS, 4). -export_type([send_fun/0, repair_fun/0, reply_fun/0, filters/0]). %%%============================================================================ %%% API %%%============================================================================ start(BlueList, PinkList, RepairFun, ReplyFun) -> % API for backwards compatability start(full, BlueList, PinkList, RepairFun, ReplyFun, none, []). -spec start(exchange_type(), input_list(), input_list(), repair_fun(), reply_fun(), filters(), options()) -> {ok, pid(), list()}. %% @doc %% Start an FSM to manage an exchange and compare the preflsist in the %% BlueList with those in the PinkList, using the RepairFun to repair any %% keys discovered to have inconsistent clocks. ReplyFun used to reply back %% to calling client the StateName at termination. %% %% The ReplyFun should be a 1 arity function that expects a tuple with the %% closing state and the cout of deltas. start(Type, BlueList, PinkList, RepairFun, ReplyFun, Filters, Opts) -> ExchangeID = leveled_util:generate_uuid(), {ok, ExPID} = gen_fsm:start(?MODULE, [{Type, Filters}, BlueList, PinkList, RepairFun, ReplyFun, ExchangeID, Opts], []), {ok, ExPID, ExchangeID}. -spec reply(pid(), any(), pink|blue) -> ok. %% @doc %% Support events to be sent back to the FSM reply(Exchange, {error, Error}, _Colour) -> gen_fsm:send_event(Exchange, {error, Error}); reply(Exchange, Result, Colour) -> gen_fsm:send_event(Exchange, {reply, Result, Colour}). %%%============================================================================ %%% gen_fsm callbacks %%%============================================================================ init([{Type, Filters}, BlueList, PinkList, RepairFun, ReplyFun, ExChID, Opts]) -> leveled_rand:seed(), PinkTarget = length(PinkList), BlueTarget = length(BlueList), State = #state{blue_list = BlueList, pink_list = PinkList, repair_fun = RepairFun, reply_fun = ReplyFun, exchange_id = ExChID, pink_returns = {PinkTarget, PinkTarget}, blue_returns = {BlueTarget, BlueTarget}, exchange_type = Type, exchange_filters = Filters}, State0 = process_options(Opts, State), aae_util:log("EX001", [ExChID, PinkTarget + BlueTarget, State0#state.purpose], logs(), State0#state.log_levels), InitState = case Type of full -> prepare_full_exchange; partial -> prepare_partial_exchange end, {ok, InitState, State0, 0}. prepare_full_exchange(timeout, State) -> aae_util:log("EX006", [prepare_tree_exchange, State#state.exchange_id], logs(), State#state.log_levels), trigger_next(fetch_root, root_compare, fun merge_root/2, <<>>, false, ?CACHE_TIMEOUT_MS, State). prepare_partial_exchange(timeout, State) -> aae_util:log("EX006", [prepare_partial_exchange, State#state.exchange_id], logs(), State#state.log_levels), Filters = State#state.exchange_filters, ScanTimeout = filtered_timeout(Filters, State#state.scan_timeout), TreeSize = element(?FILTERIDX_TRS, Filters), trigger_next({merge_tree_range, Filters}, tree_compare, fun merge_tree/2, leveled_tictac:new_tree(empty_tree, TreeSize), false, ScanTimeout, State). tree_compare(timeout, State) -> aae_util:log("EX006", [root_compare, State#state.exchange_id], logs(), State#state.log_levels), DirtyLeaves = compare_trees(State#state.blue_acc, State#state.pink_acc), TreeCompares = State#state.tree_compares + 1, {StillDirtyLeaves, Reduction} = case State#state.last_tree_compare of none -> {DirtyLeaves, 1.0}; PreviouslyDirtyLeaves -> SDL = intersect_ids(PreviouslyDirtyLeaves, DirtyLeaves), {SDL, 1.0 - length(SDL) / length(PreviouslyDirtyLeaves)} end, % We want to keep comparing trees until the number of deltas stops reducing % significantly. Then there should be a clock comparison. % It is expected there will be natural deltas with tree compare because of % timing differences. Ideally the natural deltas will be small enough so % that there should be no more than 2 tree compares before a segment filter % can be applied to accelerate the process. Filters = State#state.exchange_filters, TreeSize = element(?FILTERIDX_TRS, Filters), case ((length(StillDirtyLeaves) > 0) and (Reduction > ?WORTHWHILE_REDUCTION)) of true -> % Keep comparing trees, this is reducing the segments we will % eventually need to compare Filters0 = case length(StillDirtyLeaves) < ?WORTHWHILE_FILTER of true -> Segments = {segments, StillDirtyLeaves, TreeSize}, setelement(?FILTERIDX_SEG, Filters, Segments); false -> Filters end, ScanTimeout = filtered_timeout(Filters0, State#state.scan_timeout), trigger_next({merge_tree_range, Filters0}, tree_compare, fun merge_tree/2, leveled_tictac:new_tree(empty_tree, TreeSize), false, ScanTimeout, State#state{last_tree_compare = StillDirtyLeaves, tree_compares = TreeCompares}); false -> % Compare clocks. Note if there are no Mismatched segment IDs the % stop condition in trigger_next will be met SegmentIDs = select_ids(StillDirtyLeaves, State#state.max_results, tree_compare, State#state.exchange_id, State#state.log_levels), % TODO - select_ids doesn't account for TreeSize Filters0 = setelement(?FILTERIDX_SEG, Filters, {segments, SegmentIDs, TreeSize}), trigger_next({fetch_clocks_range, Filters0}, clock_compare, fun merge_clocks/2, [], length(SegmentIDs) == 0, State#state.scan_timeout, State#state{tree_compare_deltas = StillDirtyLeaves, tree_compares = TreeCompares, prethrottle_leaves = length(StillDirtyLeaves)}) end. root_compare(timeout, State) -> aae_util:log("EX006", [root_compare, State#state.exchange_id], logs(), State#state.log_levels), DirtyBranches = compare_roots(State#state.blue_acc, State#state.pink_acc), RootCompares = State#state.root_compares + 1, {BranchIDs, Reduction} = case State#state.last_root_compare of none -> {DirtyBranches, DirtyBranches}; PreviouslyDirtyBranches -> BDL = intersect_ids(PreviouslyDirtyBranches, DirtyBranches), {BDL, length(BDL) - length(PreviouslyDirtyBranches)} end, % Should we loop again on root_compare? As longs as root_compare is % reducing the result set sufficiently, keep doing it until we switch to % branch_compare case ((length(BranchIDs) > 0) and (Reduction > ?WORTHWHILE_REDUCTION_CACHED)) of true -> trigger_next(fetch_root, root_compare, fun merge_root/2, <<>>, false, ?CACHE_TIMEOUT_MS, State#state{last_root_compare = BranchIDs, root_compares = RootCompares}); false -> BranchesToFetch = select_ids(BranchIDs, State#state.max_results, root_confirm, State#state.exchange_id, State#state.log_levels), trigger_next({fetch_branches, BranchesToFetch}, branch_compare, fun merge_branches/2, [], length(BranchIDs) == 0, ?CACHE_TIMEOUT_MS, State#state{root_compare_deltas = BranchesToFetch, root_compares = RootCompares, prethrottle_branches = length(BranchIDs)}) end. branch_compare(timeout, State) -> aae_util:log("EX006", [branch_compare, State#state.exchange_id], logs(), State#state.log_levels), DirtySegments = compare_branches(State#state.blue_acc, State#state.pink_acc), BranchCompares = State#state.branch_compares + 1, {SegmentIDs, Reduction} = case State#state.last_branch_compare of none -> {DirtySegments, DirtySegments}; PreviouslyDirtySegments -> SDL = intersect_ids(PreviouslyDirtySegments, DirtySegments), {SDL, length(SDL) - length(PreviouslyDirtySegments)} end, % Should we loop again on root_compare? As longs as root_compare is % reducing the result set sufficiently, keep doing it until we switch to % branch_compare case ((length(SegmentIDs) > 0) and (Reduction > ?WORTHWHILE_REDUCTION_CACHED)) of true -> trigger_next({fetch_branches, State#state.root_compare_deltas}, branch_compare, fun merge_branches/2, [], false, ?CACHE_TIMEOUT_MS, State#state{last_branch_compare = SegmentIDs, branch_compares = BranchCompares}); false -> SegstoFetch = select_ids(SegmentIDs, State#state.max_results, branch_confirm, State#state.exchange_id, State#state.log_levels), trigger_next({fetch_clocks, SegstoFetch, State#state.exchange_filters}, clock_compare, fun merge_clocks/2, [], length(SegmentIDs) == 0, State#state.scan_timeout, State#state{branch_compare_deltas = SegstoFetch, branch_compares = BranchCompares, prethrottle_leaves = length(SegmentIDs)}) end. clock_compare(timeout, State) -> aae_util:log("EX006", [clock_compare, State#state.exchange_id], logs(), State#state.log_levels), aae_util:log("EX008", [State#state.blue_acc, State#state.pink_acc], logs(), State#state.log_levels), RepairKeys = compare_clocks(State#state.blue_acc, State#state.pink_acc), RepairFun = State#state.repair_fun, aae_util:log("EX004", [State#state.exchange_id, State#state.purpose, length(RepairKeys)], logs(), State#state.log_levels), RepairFun(RepairKeys), {stop, normal, State#state{key_deltas = RepairKeys}}. waiting_all_results({reply, not_supported, Colour}, State) -> aae_util:log("EX010", [State#state.exchange_id, Colour, State#state.purpose], logs(), State#state.log_levels), {stop, normal, State#state{pending_state = not_supported}}; waiting_all_results({reply, {error, Reason}, _Colour}, State) -> waiting_all_results({error, Reason}, State); waiting_all_results({reply, Result, Colour}, State) -> aae_util:log("EX007", [Colour, State#state.exchange_id], logs(), State#state.log_levels), {PC, PT} = State#state.pink_returns, {BC, BT} = State#state.blue_returns, MergeFun = State#state.merge_fun, {State0, AllPink, AllBlue} = case Colour of pink -> PinkAcc = MergeFun(Result, State#state.pink_acc), {State#state{pink_returns = {PC + 1, PT}, pink_acc = PinkAcc}, PC + 1 == PT, BC == BT}; blue -> BlueAcc = MergeFun(Result, State#state.blue_acc), {State#state{blue_returns = {BC + 1, BT}, blue_acc = BlueAcc}, PC == PT, BC + 1 == BT} end, case AllBlue and AllPink of true -> {next_state, State0#state.pending_state, State0, jitter_pause(State#state.transition_pause_ms)}; false -> {next_state, waiting_all_results, State0, set_timeout(State0#state.start_time, State0#state.reply_timeout)} end; waiting_all_results(UnexpectedResponse, State) -> % timeout expected here, but also may get errors from vnode - such as % {error, mailbox_overload} when vnode has entered overload state. Not % possible to complete exchange so stop {PC, PT} = State#state.pink_returns, {BC, BT} = State#state.blue_returns, MissingCount = PT + BT - (PC + BC), aae_util:log("EX002", [UnexpectedResponse, State#state.pending_state, MissingCount, State#state.exchange_id, State#state.purpose], logs(), State#state.log_levels), ReplyState = case UnexpectedResponse of timeout -> timeout; _ -> error end, {stop, normal, State#state{pending_state = ReplyState}}. handle_sync_event(_msg, _From, StateName, State) -> {reply, ok, StateName, State}. handle_event(_Msg, StateName, State) -> {next_state, StateName, State}. handle_info(_Msg, StateName, State) -> {next_state, StateName, State}. terminate(normal, StateName, State) -> case State#state.exchange_type of full -> case StateName of StateName when StateName == root_compare; StateName == branch_compare -> aae_util:log("EX003", [State#state.purpose, true, StateName, State#state.exchange_id, 0, State#state.root_compares, State#state.branch_compares, length(State#state.key_deltas)], logs(), State#state.log_levels); BrokenState -> EstDamage = estimated_damage(State#state.prethrottle_branches, State#state.prethrottle_leaves, State#state.max_results), aae_util:log("EX003", [State#state.purpose, false, BrokenState, State#state.exchange_id, EstDamage, State#state.root_compares, State#state.branch_compares, length(State#state.key_deltas)], logs(), State#state.log_levels) end; partial -> case StateName of tree_compare -> aae_util:log("EX009", [State#state.purpose, true, tree_compare, State#state.exchange_id, 0, State#state.tree_compares, length(State#state.key_deltas)], logs(), State#state.log_levels); BrokenState -> aae_util:log("EX009", [State#state.purpose, false, BrokenState, State#state.exchange_id, State#state.prethrottle_leaves, State#state.tree_compares, length(State#state.key_deltas)], logs(), State#state.log_levels) end end, ReplyFun = State#state.reply_fun, ReplyFun({State#state.pending_state, length(State#state.key_deltas)}). code_change(_OldVsn, StateName, State, _Extra) -> {ok, StateName, State}. %%%============================================================================ %%% External Functions %%%============================================================================ -spec insync_responses() -> list(compare_state()). %% @doc %% To help external applications understand the states returned in replies, %% a list of those response that imply that the systems are in sync. Note, %% that with branch_compare this is possibly not true if not all branch deltas %% were checked (e.g. branch IDs > max_results). insync_responses() -> [root_compare, branch_compare]. -spec merge_binary(binary(), binary()) -> binary(). %% @doc %% Merge two binaries - where one might be empty (as nothing has been seen for %% that preflist, or the accumulator is the initial one) merge_binary(<<>>, AccBin) -> AccBin; merge_binary(ResultBin, <<>>) -> ResultBin; merge_binary(ResultBin, AccBin) -> leveled_tictac:merge_binaries(ResultBin, AccBin). -spec merge_branches(branch_results(), branch_results()) -> branch_results(). %% @doc %% Branches should be returned as a list of {BranchID, BranchBin} pairs. For %% each branch in a result, merge into the accumulator. merge_branches([], BranchAccL) -> BranchAccL; merge_branches([{BranchID, BranchBin}|Rest], BranchAccL) -> case lists:keyfind(BranchID, 1, BranchAccL) of false -> % First response has an empty accumulator merge_branches(Rest, [{BranchID, BranchBin}|BranchAccL]); {BranchID, BinAcc} -> BinAcc0 = merge_binary(BranchBin, BinAcc), merge_branches(Rest, lists:keyreplace(BranchID, 1, BranchAccL, {BranchID, BinAcc0})) end. -spec merge_root(binary(), binary()) -> binary(). %% @doc %% Merge an individual result for a set of preflists into the accumulated %% binary for the tree root merge_root(Root, RootAcc) -> merge_binary(Root, RootAcc). -spec merge_tree(leveled_tictac:tictactree(), leveled_tictac:tictactree()) -> leveled_tictac:tictactree(). %% @doc %% Merge two trees into an XOR'd tree representing the total result set merge_tree(Tree0, Tree1) -> leveled_tictac:merge_trees(Tree0, Tree1). %%%============================================================================ %%% Internal Functions %%%============================================================================ -spec estimated_damage(pos_integer(), pos_integer(), pos_integer()) -> non_neg_integer(). estimated_damage(BrokenBranches, BrokenLeaves, MaxResults) -> Mult = max(1.0, BrokenBranches / MaxResults), round(Mult * BrokenLeaves). -spec process_options(options(), exchange_state()) -> exchange_state(). %% @doc %% Alter state reflecting any passed in options process_options([], State) -> State; process_options([{transition_pause_ms, PauseMS}|Tail], State) when is_integer(PauseMS) -> process_options(Tail, State#state{transition_pause_ms = PauseMS}); process_options([{log_levels, LogLevels}|Tail], State) when is_list(LogLevels) -> process_options(Tail, State#state{log_levels = LogLevels}); process_options([{scan_timeout, Timeout}|Tail], State) when is_integer(Timeout) -> process_options(Tail, State#state{scan_timeout = Timeout}); process_options([{max_results, MaxResults}|Tail], State) when is_integer(MaxResults) -> process_options(Tail, State#state{max_results = MaxResults}); process_options([{purpose, Purpose}|Tail], State) when is_atom(Purpose) -> process_options(Tail, State#state{purpose = Purpose}). -spec trigger_next(any(), atom(), fun(), any(), boolean(), integer(), exchange_state()) -> any(). %% @doc %% Trigger the next request trigger_next(NextRequest, PendingStateName, MergeFun, InitAcc, StopTest, Timeout, LoopState) -> case StopTest of true -> {stop, normal, LoopState}; false -> ok = send_requests(NextRequest, LoopState#state.blue_list, LoopState#state.pink_list, always_blue), {next_state, waiting_all_results, LoopState#state{start_time = os:timestamp(), pending_state = PendingStateName, pink_acc = InitAcc, blue_acc = InitAcc, merge_fun = MergeFun, pink_returns = reset(LoopState#state.pink_returns), blue_returns = reset(LoopState#state.blue_returns), reply_timeout = Timeout}, Timeout} end. -spec set_timeout(erlang:timestamp(), pos_integer()) -> integer(). %% @doc %% Set the timeout in a given state based on the time the state was commenced set_timeout(StartTime, Timeout) -> max(0, Timeout - timer:now_diff(os:timestamp(), StartTime) div 1000). -spec send_requests(any(), list(tuple()), list(tuple()), always_blue|always_pink) -> ok. %% @doc %% Alternate between sending requests to items on the blue and pink list send_requests({merge_tree_range, {filter, B, KR, TS, SF, MR, HM}}, BlueList, PinkList, Always) -> % unpack the filter into a single tuple msg or merge_tree_range send_requests({merge_tree_range, B, KR, TS, SF, MR, HM}, BlueList, PinkList, Always); send_requests({fetch_clocks_range, {filter, B, KR, _TS, SF, MR, _HM}}, BlueList, PinkList, Always) -> % unpack the filter into a single tuple msg or merge_tree_range send_requests({fetch_clocks_range, B, KR, SF, MR}, BlueList, PinkList, Always); send_requests({fetch_clocks, SegIDs, none}, BlueList, PinkList, Always) -> send_requests({fetch_clocks, SegIDs}, BlueList, PinkList, Always); send_requests({fetch_clocks, SegIDs, {filter, all, all, large, all, MR, pre_hash}}, BlueList, PinkList, Always) -> send_requests({fetch_clocks, SegIDs, MR}, BlueList, PinkList, Always); send_requests({fetch_clocks, SegIDs, {filter, B, KR, large, all, MR, pre_hash}}, BlueList, PinkList, Always) -> F0 = {filter, B, KR, large, {segments, SegIDs, large}, MR, pre_hash}, send_requests({fetch_clocks_range, F0}, BlueList, PinkList, Always); send_requests(_Msg, [], [], _Always) -> ok; send_requests(Msg, [{SendFun, Preflists}|Rest], PinkList, always_blue) -> SendFun(Msg, Preflists, blue), case length(PinkList) > 0 of true -> send_requests(Msg, Rest, PinkList, always_pink); false -> send_requests(Msg, Rest, PinkList, always_blue) end; send_requests(Msg, BlueList, [{SendFun, Preflists}|Rest], always_pink) -> SendFun(Msg, Preflists, pink), case length(BlueList) > 0 of true -> send_requests(Msg, BlueList, Rest, always_blue); false -> send_requests(Msg, BlueList, Rest, always_pink) end. -spec merge_clocks(list(tuple()), list(tuple())) -> list(tuple()). %% @doc %% Accumulate keys and clocks returned in the segment query, outputting a %% sorted list of keys and clocks. merge_clocks(KeyClockL, KeyClockLAcc) -> lists:merge(lists:usort(KeyClockL), KeyClockLAcc). -spec compare_roots(binary(), binary()) -> list(integer()). %% @doc %% Compare the roots of two trees (i.e. the Pink and Blue root), and return a %% list of branch IDs which are mismatched. compare_roots(BlueRoot, PinkRoot) -> leveled_tictac:find_dirtysegments(BlueRoot, PinkRoot). -spec compare_branches(branch_results(), branch_results()) -> list(integer()). %% @doc %% Compare two sets of branches , and return a list of segment IDs which are %% mismatched compare_branches(BlueBranches, PinkBranches) -> FoldFun = fun(Idx, Acc) -> {BranchID, BlueBranch} = lists:nth(Idx, BlueBranches), {BranchID, PinkBranch} = lists:keyfind(BranchID, 1, PinkBranches), DirtySegs = leveled_tictac:find_dirtysegments(BlueBranch, PinkBranch), lists:map(fun(S) -> leveled_tictac:join_segment(BranchID, S) end, DirtySegs) ++ Acc end, lists:foldl(FoldFun, [], lists:seq(1, length(BlueBranches))). -spec compare_clocks(list(tuple()), list(tuple())) -> list(repair_input()). %% @doc %% Find the differences between the lists - and return a list of %% {{B, K}, {blue-side VC | none, pink-side VC | none}} %% If the blue-side or pink-side does not contain the key, then none is used %% in place of the clock compare_clocks(BlueList, PinkList) -> % Two lists of {B, K, VC} want to remove everything where {B, K, VC} is % the same in both lists SortClockFun = fun({B, K, VC}) -> {B, K, refine_clock(VC)} end, BlueSet = ordsets:from_list(lists:map(SortClockFun, BlueList)), PinkSet = ordsets:from_list(lists:map(SortClockFun, PinkList)), BlueDelta = ordsets:subtract(BlueSet, PinkSet), PinkDelta = ordsets:subtract(PinkSet, BlueSet), % Want to subtract out from the Pink and Blue Sets any example where % both pink and blue are the same % % This should speed up the folding and key finding to provide the % joined list BlueDeltaList = lists:reverse( ordsets:fold(fun({B, K, VCB}, Acc) -> % Assume for now that element may be only % blue [{{B, K}, {VCB, none}}|Acc] end, [], BlueDelta)), % BlueDeltaList is the output of compare clocks, assuming the item % is only on the Blue side (so it compares the blue vector clock with % none) PinkEnrichFun = fun({B, K, VCP}, Acc) -> case lists:keyfind({B, K}, 1, Acc) of {{B, K}, {VCB, none}} -> ElementWithClockDiff = {{B, K}, {VCB, VCP}}, lists:keyreplace({B, K}, 1, Acc, ElementWithClockDiff); false -> ElementOnlyPink = {{B, K}, {none, VCP}}, lists:keysort(1, [ElementOnlyPink|Acc]) end end, % The Foldfun to be used on the PinkDelta, will now fill in the Pink % vector clock if the element also exists in Pink AllDeltaList = ordsets:fold(PinkEnrichFun, BlueDeltaList, PinkDelta), % The accumulator starts with the Blue side only perspective, and % either adds to it or enriches it by folding over the Pink side % view AllDeltaList. -spec compare_trees(leveled_tictac:tictactree(), leveled_tictac:tictactree()) -> list(non_neg_integer()). %% @doc %% Compare the trees - get list of dirty leaves (Segment IDs) compare_trees(Tree0, Tree1) -> leveled_tictac:find_dirtyleaves(Tree0, Tree1). -spec intersect_ids(list(integer()), list(integer())) -> list(integer()). %% @doc %% Provide the intersection of two lists of integer IDs intersect_ids(IDs0, IDs1) -> lists:filter(fun(ID) -> lists:member(ID, IDs1) end, IDs0). -spec select_ids(list(integer()), pos_integer(), atom(), list(), aae_util:log_levels()|undefined) -> list(integer()). %% @doc %% Select a cluster of IDs if the list of IDs is smaller than the maximum %% output size. The lookup based on these IDs will be segment based, so it %% is expected that the tightest clustering will yield the most efficient %% results. However, if we always get the same list, then concurrent exchanges %% will wastefully correct the same data - so randomly chose one of the better %% lists select_ids(IDList, MaxOutput, StateName, ExchangeID, LogLevels) when length(IDList) > MaxOutput -> IDList0 = lists:sort(IDList), aae_util:log("EX005", [ExchangeID, length(IDList0), StateName], logs(), LogLevels), IDList1 = lists:sublist(IDList0, 1 + length(IDList0) - MaxOutput), IDList2 = lists:sublist(IDList0, MaxOutput, 1 + length(IDList0) - MaxOutput), FoldFun = fun({Start, End}, {Idx, Acc}) -> {Idx + 1, [{End - Start, Idx}|Acc]} end, {_EndIdx, SpaceIdxL} = lists:foldl(FoldFun, {1, []}, lists:zip(IDList1, IDList2)), Selections = lists:sublist(lists:sort(SpaceIdxL), MaxOutput), {_ChosenSpace, ChosenIdx} = lists:nth(leveled_rand:uniform(length(Selections)), Selections), lists:sublist(IDList0, ChosenIdx, MaxOutput); select_ids(IDList, _MaxOutput, _StateName, _ExchangeID, _LogLevels) -> lists:sort(IDList). -spec jitter_pause(pos_integer()) -> pos_integer(). %% @doc %% Jitter a pause, so if multiple FSMs started at once, they don't all use %% the network at the same time jitter_pause(Timeout) -> leveled_rand:uniform(Timeout) + Timeout div 2. -spec reset({pos_integer(), pos_integer()}) -> {non_neg_integer(), pos_integer()}. %% @doc %% Rest the count back to 0 reset({Target, Target}) -> {0, Target}. -spec filtered_timeout(filters(), pos_integer()) -> pos_integer(). %% @doc %% Has a filter been applied to the scan (true), or are we scanning the whole %% bucket (false) filtered_timeout({filter, _B, KeyRange, _TS, SegFilter, ModRange, _HM}, ScanTimeout) -> case ((KeyRange == all) and (SegFilter == all) and (ModRange == all)) of true -> ?UNFILTERED_SCAN_TIMEOUT_MS; false -> ScanTimeout end. -spec refine_clock(list()|binary()) -> list()|binary(). %% @doc %% When the lock is a list, always sort the list so as not to confuse clocks %% differentiated only by sorting refine_clock(Clock) when is_list(Clock) -> lists:sort(Clock); refine_clock(Clock) -> Clock. %%%============================================================================ %%% log definitions %%%============================================================================ -spec logs() -> list(tuple()). %% @doc %% Define log lines for this module logs() -> [{"EX001", {info, "Exchange id=~s with target_count=~w expected purpose=~w"}}, {"EX002", {error, "~w with pending_state=~w and missing_count=~w" ++ " for exchange id=~s purpose=~w"}}, {"EX003", {info, "Normal exit for full exchange purpose=~w in_sync=~w " ++ " pending_state=~w for exchange id=~s" ++ " scope of mismatched_segments=~w" ++ " root_compare_loops=~w " ++ " branch_compare_loops=~w " ++ " keys_passed_for_repair=~w"}}, {"EX004", {info, "Exchange id=~s purpose=~w led to prompting" ++ " of repair_count=~w"}}, {"EX005", {info, "Exchange id=~s throttled count=~w at state=~w"}}, {"EX006", {debug, "State change to ~w for exchange id=~s"}}, {"EX007", {debug, "Reply received for colour=~w in exchange id=~s"}}, {"EX008", {debug, "Comparison between BlueList ~w and PinkList ~w"}}, {"EX009", {info, "Normal exit for full exchange purpose=~w in_sync=~w" ++ " pending_state=~w for exchange id=~s" ++ " scope of mismatched_segments=~w" ++ " tree_compare_loops=~w " ++ " keys_passed_for_repair=~w"}}, {"EX010", {warn, "Exchange not_supported in exchange id=~s" ++ " for colour=~w purpose=~w"}} ]. %%%============================================================================ %%% Test %%%============================================================================ -ifdef(TEST). select_id_test() -> L0 = [1, 2, 3], ?assertMatch(L0, select_ids(L0, 3, root_confirm, "t0", undefined)), L1 = [3, 2, 1], ?assertMatch(L0, select_ids(L1, 3, root_confirm, "t0", undefined)), ?assertMatch(2, length(select_ids(L1, 2, root_confirm, "t0", undefined))). select_best_id_rand_test() -> L2 = [1, 2, 3, 5, 16, 17, 18], F = fun(_N, {S1, S2, S3}) -> case {S1, S2, S3} of {true, true, true} -> {true, true, true}; _ -> case select_ids(L2, 3, root_confirm, "r3", undefined) of [1, 2, 3] -> {true, S2, S3}; [2, 3, 5] -> {S1, true, S3}; [16, 17, 18] -> {S1, S2, true} end end end, ?assertMatch({true, true, true}, lists:foldl(F, {false, false, false}, lists:seq(1, 1000))). compare_clocks_test() -> KV1 = {<<"B1">>, <<"K1">>, [{a, 1}]}, KV2 = {<<"B1">>, <<"K2">>, [{b, 1}]}, KV3 = {<<"B1">>, <<"K3">>, [{a, 2}]}, KV4 = {<<"B1">>, <<"K1">>, [{a, 1}, {b, 2}]}, KV5 = {<<"B1">>, <<"K2">>, [{b, 1}, {c, 1}]}, BL1 = [KV1, KV2, KV3], PL1 = [KV1, KV2, KV3], ?assertMatch([], compare_clocks(BL1, PL1)), BL2 = [KV2, KV3, KV4], ?assertMatch([{{<<"B1">>, <<"K1">>}, {[{a, 1}, {b, 2}], [{a, 1}]}}], compare_clocks(BL2, PL1)), ?assertMatch([{{<<"B1">>, <<"K1">>}, {[{a, 1}], [{a, 1}, {b, 2}]}}], compare_clocks(PL1, BL2)), PL2 = [KV4, KV5], ?assertMatch([{{<<"B1">>, <<"K1">>}, {[{a, 1}], [{a, 1}, {b, 2}]}}, {{<<"B1">>, <<"K2">>}, {[{b, 1}], [{b, 1}, {c, 1}]}}, {{<<"B1">>, <<"K3">>}, {[{a, 2}], none}}], compare_clocks(BL1, PL2)). compare_unsorted_clocks_test()-> KV1 = {<<"B1">>, <<"K1">>, [{a, 1}, {b, 2}]}, KV2 = {<<"B1">>, <<"K1">>, [{b, 2}, {a, 1}]}, KV1b = {<<"B1">>, <<"K1">>, term_to_binary([{a, 1}, {b, 2}])}, KV2b = {<<"B1">>, <<"K1">>, term_to_binary([{b, 2}, {a, 1}])}, ?assertMatch([], compare_clocks([KV1], [KV2])), KL = compare_clocks([KV1b], [KV2b]), ?assertMatch(1, length(KL)). clean_exit_ontimeout_test() -> State0 = #state{pink_returns={4, 5}, blue_returns={8, 8}, exchange_type = full}, State1 = State0#state{pending_state = timeout}, {stop, normal, State1} = waiting_all_results(timeout, State0). connect_error_test() -> SendFun = fun(_Msg, _PLs, Colour) -> Exchange = self(), reply(Exchange, {error, disconnected}, Colour) end, BlueList = [{SendFun, [{0, 1}]}], PinkList = [{SendFun, [{0, 1}]}], RepairFun = fun(_RL) -> ok end, ReceiveReply = spawn(fun() -> receive {error, 0} -> ok end end), ReplyFun = fun(R) -> ReceiveReply ! R end, {ok, Test, _ExID} = start(BlueList, PinkList, RepairFun, ReplyFun), ?assertMatch(true, lists:foldl(fun(X, Acc) -> case Acc of true -> true; false -> timer:sleep(X), not is_process_alive(ReceiveReply) end end, false, [1000, 1000, 1000])), ?assertMatch(false, is_process_alive(Test)). waiting_for_error_test() -> {stop, normal, _S0} = waiting_all_results({reply, {error, query_backlog}, blue}, #state{exchange_type = full, merge_fun = fun merge_clocks/2}). coverage_cheat_test() -> {next_state, prepare, _State0} = handle_event(null, prepare, #state{exchange_type = full}), {reply, ok, prepare, _State1} = handle_sync_event(null, nobody, prepare, #state{exchange_type = full}), {next_state, prepare, _State2} = handle_info(null, prepare, #state{exchange_type = full}), {ok, prepare, _State3} = code_change(null, prepare, #state{exchange_type = full}, null), [root_compare, branch_compare] = insync_responses(). -endif.
src/aae_exchange.erl
0.650245
0.871311
aae_exchange.erl
starcoder
%%-------------------------------------------------------------------- %% Copyright (c) 2013-2018 EMQ Enterprise, Inc. (http://emqtt.io) %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %%-------------------------------------------------------------------- -module(emqttd_hooks). -behaviour(gen_server). -author("<NAME> <<EMAIL>>"). %% Start -export([start_link/0]). %% Hooks API -export([add/3, add/4, delete/2, run/2, run/3, lookup/1]). %% gen_server Function Exports -export([init/1, handle_call/3, handle_cast/2, handle_info/2, terminate/2, code_change/3]). -record(state, {}). -type(hooktag() :: atom() | string() | binary()). -export_type([hooktag/0]). -record(callback, {tag :: hooktag(), function :: function(), init_args = [] :: list(any()), priority = 0 :: integer()}). -record(hook, {name :: atom(), callbacks = [] :: list(#callback{})}). -define(HOOK_TAB, mqtt_hook). start_link() -> gen_server:start_link({local, ?MODULE}, ?MODULE, [], []). %%-------------------------------------------------------------------- %% Hooks API %%-------------------------------------------------------------------- -spec(add(atom(), function() | {hooktag(), function()}, list(any())) -> ok). add(HookPoint, Function, InitArgs) when is_function(Function) -> add(HookPoint, {undefined, Function}, InitArgs, 0); add(HookPoint, {Tag, Function}, InitArgs) when is_function(Function) -> add(HookPoint, {Tag, Function}, InitArgs, 0). -spec(add(atom(), function() | {hooktag(), function()}, list(any()), integer()) -> ok). add(HookPoint, Function, InitArgs, Priority) when is_function(Function) -> add(HookPoint, {undefined, Function}, InitArgs, Priority); add(HookPoint, {Tag, Function}, InitArgs, Priority) when is_function(Function) -> gen_server:call(?MODULE, {add, HookPoint, {Tag, Function}, InitArgs, Priority}). -spec(delete(atom(), function() | {hooktag(), function()}) -> ok). delete(HookPoint, Function) when is_function(Function) -> delete(HookPoint, {undefined, Function}); delete(HookPoint, {Tag, Function}) when is_function(Function) -> gen_server:call(?MODULE, {delete, HookPoint, {Tag, Function}}). %% @doc Run hooks without Acc. -spec(run(atom(), list(Arg :: any())) -> ok | stop). run(HookPoint, Args) -> run_(lookup(HookPoint), Args). -spec(run(atom(), list(Arg :: any()), any()) -> any()). run(HookPoint, Args, Acc) -> run_(lookup(HookPoint), Args, Acc). %% @private run_([#callback{function = Fun, init_args = InitArgs} | Callbacks], Args) -> case apply(Fun, lists:append([Args, InitArgs])) of ok -> run_(Callbacks, Args); stop -> stop; _Any -> run_(Callbacks, Args) end; run_([], _Args) -> ok. %% @private run_([#callback{function = Fun, init_args = InitArgs} | Callbacks], Args, Acc) -> case apply(Fun, lists:append([Args, [Acc], InitArgs])) of ok -> run_(Callbacks, Args, Acc); {ok, NewAcc} -> run_(Callbacks, Args, NewAcc); stop -> {stop, Acc}; {stop, NewAcc} -> {stop, NewAcc}; _Any -> run_(Callbacks, Args, Acc) end; run_([], _Args, Acc) -> {ok, Acc}. -spec(lookup(atom()) -> [#callback{}]). lookup(HookPoint) -> case ets:lookup(?HOOK_TAB, HookPoint) of [#hook{callbacks = Callbacks}] -> Callbacks; [] -> [] end. %%-------------------------------------------------------------------- %% gen_server Callbacks %%-------------------------------------------------------------------- init([]) -> ets:new(?HOOK_TAB, [set, protected, named_table, {keypos, #hook.name}]), {ok, #state{}}. handle_call({add, HookPoint, {Tag, Function}, InitArgs, Priority}, _From, State) -> Callback = #callback{tag = Tag, function = Function, init_args = InitArgs, priority = Priority}, {reply, case ets:lookup(?HOOK_TAB, HookPoint) of [#hook{callbacks = Callbacks}] -> case contain_(Tag, Function, Callbacks) of false -> insert_hook_(HookPoint, add_callback_(Callback, Callbacks)); true -> {error, already_hooked} end; [] -> insert_hook_(HookPoint, [Callback]) end, State}; handle_call({delete, HookPoint, {Tag, Function}}, _From, State) -> {reply, case ets:lookup(?HOOK_TAB, HookPoint) of [#hook{callbacks = Callbacks}] -> case contain_(Tag, Function, Callbacks) of true -> insert_hook_(HookPoint, del_callback_(Tag, Function, Callbacks)); false -> {error, not_found} end; [] -> {error, not_found} end, State}; handle_call(Req, _From, State) -> {reply, {error, {unexpected_request, Req}}, State}. handle_cast(_Msg, State) -> {noreply, State}. handle_info(_Info, State) -> {noreply, State}. terminate(_Reason, _State) -> ok. code_change(_OldVsn, State, _Extra) -> {ok, State}. %%-------------------------------------------------------------------- %% Internal functions %%-------------------------------------------------------------------- insert_hook_(HookPoint, Callbacks) -> ets:insert(?HOOK_TAB, #hook{name = HookPoint, callbacks = Callbacks}), ok. add_callback_(Callback, Callbacks) -> lists:keymerge(#callback.priority, Callbacks, [Callback]). del_callback_(Tag, Function, Callbacks) -> lists:filter( fun(#callback{tag = Tag1, function = Func1}) -> not ((Tag =:= Tag1) andalso (Function =:= Func1)) end, Callbacks). contain_(_Tag, _Function, []) -> false; contain_(Tag, Function, [#callback{tag = Tag, function = Function}|_Callbacks]) -> true; contain_(Tag, Function, [_Callback | Callbacks]) -> contain_(Tag, Function, Callbacks).
src/emqttd_hooks.erl
0.587943
0.404507
emqttd_hooks.erl
starcoder
% @doc % <a href="https://reference.digilentinc.com/reference/pmod/pmodgps/reference-manual"> % PmodGPS</a> % module. % % The PmodGPS sends the GPS data over UART. % % Start the driver with % ``` % 1> grisp:add_device(uart, pmod_gps). % ''' % @end -module(pmod_gps). -behaviour(gen_server). %--- Exports ------------------------------------------------------------------- % API -export([start_link/2]). -export([get/1]). % Callbacks -export([init/1]). -export([handle_call/3]). -export([handle_cast/2]). -export([handle_info/2]). -export([code_change/3]). -export([terminate/2]). %--- Includes ------------------------------------------------------------------ -include("grisp.hrl"). %--- Macros -------------------------------------------------------------------- -define(MAX_CONSECUTIVE_ERRORS, 10). %--- Records ------------------------------------------------------------------- -record(state, {port, last_sentences, error_count = 0}). %--- API ----------------------------------------------------------------------- % @private start_link(Slot, _Opts) -> gen_server:start_link({local, ?MODULE}, ?MODULE, Slot, []). % @doc Get the GPS data. % % The input parameter specifies which type of sentence to get. % For a description of the sentences see the % <a href="https://reference.digilentinc.com/_media/reference/pmod/pmodgps/pmodgps_rm.pdf"> % PmodGPS Reference Manual % The sentence CRC is checked for all the sentence types, but for now only the % GGA values are parsed. If other sentences are needed, grisp_nmea needs to be % extended to support more types. % </a>. % % === Example === % ``` % 2> pmod_gps:get(gga). % {gps,gga,#{alt => 61.9,fixed => true,lat => 52.122661666666666,long => 11.594928333333334,time => 53912000}}} % 3> pmod_gps:get(gsa). % {gps,gsa,<<"A,3,17,06,19,02,24,,,,,,,,2.69,2.51,0.97">>} % 4> pmod_gps:get(gsv). % {gps,gsv,<<"3,3,12,14,22,317,17,17,10,040,35,29,09,203,,22,02,351">>} % 5> pmod_gps:get(rmc). % {gps,rmc,<<"150007.000,A,5207.3592,N,01135.6895,E,0.46,255.74,120220,,,A">>} % 6> pmod_gps:get(vtg). % {gps,vtg,<<"297.56,T,,M,0.65,N,1.21,K,A">>} % ''' -spec get(grisp_nmea:message_id()) -> {grisp_nmea:talker_id(), grisp_nmea:message_id(), map() | binary()} | undefined. get(MessageId) -> call({get, MessageId}). %--- Callbacks ----------------------------------------------------------------- % @private init(Slot = uart) -> Port = open_port({spawn_driver, "grisp_termios_drv"}, [binary]), grisp_devices:register(Slot, ?MODULE), Sentences = maps:from_list([{T, undefined} || T <- [ dtm, gbq, gbs, gga, gll, glq, gnq, gns, gpq, grs, gsa, gst, gsv, rmc, txt, vlw, vtg, zda ]]), {ok, #state{port = Port, last_sentences = Sentences}}. % @private handle_call(Call, _From, State) -> try execute_call(Call, State) catch throw:Reason -> {reply, {error, Reason}, State} end. % @private handle_cast(Request, _State) -> error({unknown_cast, Request}). % @private % We need to support at least one message failing parsing, because % when starting to read randomly in the stream of sentences, the first % one may be truncated. For now, we fail after a maximum number of % consecutive errors. handle_info({Port, {data, Data}}, #state{port = Port, last_sentences = LastSentences, error_count = ErrorCount} = State) when is_binary(Data) -> case {ErrorCount, grisp_nmea:parse(Data)} of {Count, {error, Reason}} when Count > ?MAX_CONSECUTIVE_ERRORS -> erlang:error({gps_nmea_parsing_error, Reason}); {Count, {error, _Reason}} -> {noreply, State#state{error_count = Count + 1}}; {_, {ok, {_TalkerId, MessageType, _Values} = Sentence}} -> LastSentences2 = LastSentences#{MessageType => Sentence}, {noreply, State#state{last_sentences = LastSentences2, error_count = 0}} end; handle_info(_Any, State) -> {noreply, State}. % @private code_change(_OldVsn, State, _Extra) -> {ok, State}. % @private terminate(_Reason, _State) -> ok. %--- Internal ----------------------------------------------------------------- call(Call) -> Dev = grisp_devices:default(?MODULE), case gen_server:call(Dev#device.pid, Call) of {error, Reason} -> error(Reason); {ok, Result} -> Result end. execute_call({get, MessageId}, #state{last_sentences = Sentences} = State) -> case maps:find(MessageId, Sentences) of error -> {reply, {error, {unknown_sentence, MessageId}}, State}; {ok, Sentence} -> {reply, {ok, Sentence}, State} end; execute_call(Request, _State) -> error({unknown_call, Request}).
src/pmod_gps.erl
0.587943
0.481271
pmod_gps.erl
starcoder
%% @author <NAME> <<EMAIL>> %% @copyright 2012 <NAME> %% @doc Calculate a quadtile code from a lat/long location. (http://wiki.openstreetmap.org/wiki/QuadTiles) %% Copyright 2012 <NAME> %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. -module(geomap_quadtile). -author('<NAME> <<EMAIL>>'). %% Default 31 bits, as 62 bits fit in a PostgreSQL 'bigint' -define(BITS, 31). %% Rounding precision when decoding (ca. 10cm) -define(PREC, 1000000). -export([ encode/2, encode/3, decode/1, decode/2, test/0 ]). -include_lib("zotonic_core/include/zotonic.hrl"). %% @doc Calculate a quadtile number in 64 bits (32 bits per coordinate) -spec encode(Latitude :: float(), Longitude :: float()) -> integer(). encode(Latitude, Longitude) -> encode(Latitude, Longitude, ?BITS). %% @doc Calculate a quadtile number in a 2*Bits integer (Bits precision per coordinate) -spec encode(Latitude :: float(), Longitude :: float(), Bits :: integer()) -> integer(). encode(Latitude, Longitude, Bits) -> interleave(to_int(Longitude, 180.0, Bits), to_int(Latitude, 90.0, Bits), Bits - 1, 0). %% @doc Decode the long/lat from a quadtile -spec decode(integer()) -> {Latitude :: float(), Longitude :: float()}. decode(Quadtile) -> decode(Quadtile, ?BITS). -spec decode(Quadtile :: integer(), Bits :: integer()) -> {Latitude :: float(), Longitude :: float()}. decode(Quadtile, Bits) -> {LongInt, LatInt} = deinterleave(0, 0, 1 bsl (Bits+Bits-1), Bits, Quadtile), {to_float(LatInt, 90.0, Bits), to_float(LongInt, 180.0, Bits)}. %% @doc Interleave the longitude and latitude bit by bit, start with a longitude bit. interleave(_, _, -1, Acc) -> Acc; interleave(Long, Lat, N, Acc) -> Acc1 = (Acc bsl 1) bor ((Long bsr N) band 1), Acc2 = (Acc1 bsl 1) bor ((Lat bsr N) band 1), interleave(Long, Lat, N-1, Acc2). %% @doc Get the longitude and latitude by de-interleaving the bits deinterleave(Long, Lat, 0, _N, _Q) -> {Long, Lat}; deinterleave(Long, Lat, Mask, N, Q) -> Long1 = Long bor ((Q band Mask) bsr N), Lat1 = Lat bor ((Q band (Mask bsr 1)) bsr (N-1)), deinterleave(Long1, Lat1, Mask bsr 2, N-1, Q). %% @doc Map a longitude/latitude to an integer in the range 0..(2**Bits-1) to_int(Angle, MaxAngle, Bits) -> round((fmod(Angle, MaxAngle)+MaxAngle)/(2.0*MaxAngle) * ((1 bsl Bits)-1)). %% @doc Map a N bits integer to a float longitude/latitude to_float(N, MaxAngle, Bits) -> round(((N / ((1 bsl Bits) -1))*2.0*MaxAngle - MaxAngle) * ?PREC) / ?PREC. %% @doc Ensure that the angle is between [-Max, Max> fmod(Angle, Max) when Angle < -Max -> fmod(Angle+2.0*Max, Max); fmod(Angle, Max) when Angle >= Max -> fmod(Angle-2.0*Max, Max); fmod(Angle, _Max) -> Angle. test() -> 0 = geomap_quadtile:encode(-90, -180), 0 = geomap_quadtile:encode(90, 180), {-90.0,-180.0} = geomap_quadtile:decode(0), {1.0, 2.0} = geomap_quadtile:decode(geomap_quadtile:encode(1,2)), % Below is close to (1 bsl (?BITS*2))-1 = 4611686018427387903 4611686018427387900 = geomap_quadtile:encode(89.9999999, 179.9999999), {90.0, 180.0} = geomap_quadtile:decode(4611686018427387900), ok.
src/support/geomap_quadtile.erl
0.753467
0.58516
geomap_quadtile.erl
starcoder
%%%------------------------------------------------------------------- %%% @author yangcancai %%% Copyright (c) 2021 by yangcancai(<EMAIL>), All Rights Reserved. %%% %%% Licensed under the Apache License, Version 2.0 (the "License"); %%% you may not use this file except in compliance with the License. %%% You may obtain a copy of the License at %%% %%% https://www.apache.org/licenses/LICENSE-2.0 %%% %%% Unless required by applicable law or agreed to in writing, software %%% distributed under the License is distributed on an "AS IS" BASIS, %%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %%% See the License for the specific language governing permissions and %%% limitations under the License. %%% %%% @doc %%% %%% @end %%% Created : 2021-05-19T07:19:22+00:00 %%%------------------------------------------------------------------- -module(bd_bench). -author("yangcancai"). -include("big_data.hrl"). -export([run/0, run/2, overview/0]). -define(TotalCmdIx, 1). -define(TotalTime, 2). -define(FIELDS, [total_command, sum_time]). -define(SIZE, erlang:length(?FIELDS)). run() -> run(100, 100). run(N, Degree) when is_integer(N) -> run(#{bucket_counter => N, degree => Degree, min_row => 1, max_row => 300, mix_row_bytes => 100, max_row_bytes => 500}). run(#{bucket_counter := BucketCounter, degree := Degree} = Config) -> spawn_link(fun() -> Each = BucketCounter div Degree, Pid = self(), Size = ?SIZE, Ref = counters:new(Size, []), persistent_term:put(bd_bench_s, erlang:system_time(1000)), persistent_term:put(counters, Ref), Pids = [spawn_link(fun() -> loop(Pid, Each, Config) end) || _ <- lists:seq(1, Degree)], wait_loop(Pids), Rs = overview(), io:format("Overview = ~p~n", [Rs]) end). overview() -> Ref = persistent_term:get(counters), L = [counters:get(Ref, I) || I <- lists:seq(1, ?SIZE)], R = #{total_command := Cmd, sum_time := Time} = maps:from_list( lists:zip(?FIELDS, L)), Tps = Time div Cmd, TotalTime = persistent_term:get(bd_bench_e) - persistent_term:get(bd_bench_s), S = TotalTime div 1000, Sec = case S of 0 -> 1; S -> S end, R#{aver => Tps div 1000, tps => Cmd div Sec, total_time => TotalTime}. wait_loop([]) -> persistent_term:put(bd_bench_e, erlang:system_time(1000)), ok; wait_loop(Pids) -> receive {Pid, done} -> ?DEBUG("Bech done pid = ~p", [Pid]), wait_loop(lists:delete(Pid, Pids)); {'EXIT', Pid, _} -> ?DEBUG("Bech exit pid = ~p", [Pid]), wait_loop(lists:delete(Pid, Pids)) end. loop(Parent, Each, #{min_row := MixRow, max_row := MaxRow, mix_row_bytes := MixRowBytes, max_row_bytes := MaxRowBytes}) -> Row = gen_row(MixRow, MaxRow), Ref = persistent_term:get(counters), %% insert [begin Bucket = gen_bucket(), [begin RowBytes = gen_data(MixRowBytes, MaxRowBytes), counters:add(Ref, ?TotalCmdIx, 1), {T, _Rs} = timer:tc(fun() -> big_data:command(#bd_wal{action = insert, args = [Bucket, erlang:integer_to_binary(RowID), erlang:system_time(1000), RowBytes]}) end), counters:add(Ref, ?TotalTime, T) end || RowID <- lists:seq(1, Row)] end || _ <- lists:seq(1, Each)], Parent ! {self(), done}. gen_row(MixRow, MaxRow) -> rand(MixRow, MaxRow). gen_bucket() -> rand_bytes(32). gen_data(MixRowBytes, MaxRowBytes) -> RowBytes = rand(MixRowBytes, MaxRowBytes), rand_bytes(RowBytes). rand(Mix, Max) -> rand:uniform(Max - Mix) + Mix. rand_bytes(N) -> ?BD_RAND_BYTES(N).
src/bd_bench.erl
0.554109
0.455078
bd_bench.erl
starcoder
%%%------------------------------------------------------------------------ %% Copyright 2018, OpenCensus Authors %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %% %% @doc %% Measure represents a type of metric to be tracked and recorded. %% For example, latency, request Mb/s, and response Mb/s are measures %% to collect from a server. %% %% Measure is a generic interface for recording values in aggregations %% via subscribed views. %% When recording a value, we have to obtain the list of all subscribed views %% and call respective aggregations. We use code generation to optimize this. %% When a view subscribed or unsubscribed we regenerate unrolled loop in a %% special module (one for each measure). Module names generated from measurement %% names (1-to-1). If we know a measure name at the compile time, we can eliminate %% the module name lookup and inject remote call directly, replacing `oc_stat:record' %% with `<GENERATED_MEASURE_MODULE>:record'. %% For that {parse_transform, oc_stat_measure} option must be used. %% @end %%%----------------------------------------------------------------------- -module(oc_stat_measure). %% user api -export([new/3, exists/1, unit/1]). %% codegen -export([measure_module/1, module_name/1, maybe_module_name/1, regen_record/2, delete_measure/1]). %% unsafe api, needs snychronization -export([register_/1, add_subscription_/2, remove_subscription_/2, terminate_/0]). -export(['__init_backend__'/0]). -export_type([name/0, description/0, unit/0, measure/0]). -record(measure, {name :: name(), module :: module(), description :: description(), unit :: unit()}). -type name() :: atom() | binary() | string(). -type description() :: binary() | string(). -type unit() :: atom(). -type measure() :: #measure{}. -define(MEASURES_TABLE, ?MODULE). -dialyzer({nowarn_function, regen_module/3}). -dialyzer({nowarn_function, insert_measure_/1}). -dialyzer({nowarn_function, regen_record/2}). -dialyzer({nowarn_function, delete_measure/1}). %% @doc %% Creates and registers a measure. If a measure with the same name %% already exists, old measure returned. %% @end -spec new(name(), description(), unit()) -> oc_stat_view:measure(). new(Name, Description, Unit) -> gen_server:call(oc_stat, {measure_register, #measure{name=Name, module=oc_stat_measure:module_name(Name), description=Description, unit=Unit}}). %% @doc %% Returns a measure with the `Name' or `false'.. %% @end -spec exists(name() | measure()) -> measure() | false. exists(#measure{name=Name}) -> exists(Name); exists(Name) -> case ets:lookup(?MEASURES_TABLE, Name) of [Measure] -> Measure; _ -> false end. -spec unit(measure()) -> unit(). unit(#measure{unit=Unit}) -> Unit. %% ============================================================================= %% internal %% ============================================================================= %% @private register_(#measure{name=Name}=Measure) -> case exists(Name) of false -> insert_measure_(Measure); OldMeasure -> OldMeasure end. %% @private insert_measure_(#measure{module=Module}=Measure) -> ets:insert(?MEASURES_TABLE, Measure), regen_record(Module, []), Measure. %% @private add_subscription_(Measure, VS) -> case exists(Measure) of false -> {error, {unknown_measure, Measure}}; #measure{module=Module} -> Subs = Module:subs(), regen_record(Module, [VS | Subs]), ok end. %% @private remove_subscription_(Name, VS) -> case exists(Name) of false -> ok; #measure{module=Module} -> Subs = Module:subs(), regen_record(Module, lists:delete(VS, Subs)), ok end. %% @private terminate_() -> [delete_measure(M) || M <- ets:tab2list(?MEASURES_TABLE)]. %% @private '__init_backend__'() -> ?MEASURES_TABLE = ets:new(?MEASURES_TABLE, [set, named_table, public, {keypos, 2}]), ok. %% ============================================================================= %% codegen %% ============================================================================= %% @private measure_module(Name) -> case ets:lookup(?MEASURES_TABLE, Name) of [#measure{module=Module}] -> Module; _ -> erlang:error({unknown_measure, Name}) end. %% @private -spec module_name(name()) -> module(). module_name(Name) -> list_to_atom(module_name_str(Name)). module_name_str(Name) when is_atom(Name) -> name_template(atom_to_list(Name)); module_name_str(Name) when is_binary(Name) -> name_template(binary_to_list(Name)); module_name_str(Name) when is_list(Name) -> name_template(binary_to_list(iolist_to_binary(Name))). name_template(Name) -> lists:flatten(["$_MEASURE_", Name]). %% @private maybe_module_name(Name) -> list_to_existing_atom(module_name_str(Name)). %% @private regen_record(ModuleName, VSs) -> regen_module(ModuleName, gen_add_sample_calls(VSs), erl_parse:abstract(VSs)). %% @private delete_measure(#measure{name=Name, module=Module}) -> ErrorA = erl_parse:abstract({unknown_measure, Name}), regen_module(Module, gen_add_sample_calls([]) ++ [{call, 1, {remote, 1, {atom, 1, erlang}, {atom, 1, error}}, [ErrorA]}], {call, 1, {remote, 1, {atom, 1, erlang}, {atom, 1, error}}, [ErrorA]}). %% @private regen_module(ModuleName, RecordBody, Subs) -> ModuleNameStr = atom_to_list(ModuleName), {ok, Module, Binary} = compile:forms( [{attribute, 1, file, {ModuleNameStr, 1}}, {attribute, 1, module, ModuleName}, {attribute, 1, export, [{record, 2}]}, {attribute, 1, export, [{subs, 0}]}, {function, 1, record, 2, [{clause, 1, [{var, 1, 'ContextTags'}, {var, 1, 'Value'}], [], RecordBody ++ [{atom, 1, ok}] }]}, {function, 1, subs, 0, [{clause, 1, [], [], [Subs] }]}, {eof, 2}]), {module, Module} = code:load_binary(Module, ModuleNameStr, Binary). gen_add_sample_calls([]) -> [{match, 1, {var, 1, '_'}, {var, 1, 'ContextTags'}}, {match, 1, {var, 1, '_'}, {var, 1, 'Value'}}]; gen_add_sample_calls(VSs) -> lists:map(fun oc_stat_view:gen_add_sample_/1, VSs).
src/oc_stat_measure.erl
0.750461
0.496338
oc_stat_measure.erl
starcoder
%% ===================================================================== %% @doc Hugin is a framework to simplify the process of defining your %% own web crawlers. It tries to do so while maintaining maximum %% flexibility and efficiency. %% %% The name Hugin comes from one the two ravens owned by Odin in %% Norse mythology, Hugin and Munin. Odin sent out the two ravens every %% morning to scout the world and bring the entire world's accumulated %% information back to him. %% %% Similar to the case for Odin, the Hugin framework helps you to crawl %% the internet in the pursuit of gathering the World Wide Web's %% collective information and bringing it back to you. %% @copyright 2015 <NAME> %% @author <NAME> <<EMAIL>> %% @version {@version} %% @end %% ===================================================================== -module(hugin). -export([start/0, start/1]). -export([url/1, url/2]). -export([set_option/2, set_options/2, max_freq/3, max_freq/4, max_par/2]). -type url() :: binary(). -type server_ref() :: atom() | pid(). %%%========================================================================= %%% API %%%========================================================================= -spec start() -> ok | {error, any()}. %% @doc Start the application with no active crawlers. start() -> hackney:start(), application:ensure_started(hugin). -spec start(Callback :: module() | fun()) -> {ok, pid()}. %% @doc Start the application together with one crawler. %% The callback can be either a atom with the module or a function. %% See {@link raven} for more information about the callback. start(M) when is_atom(M) -> start1(M); %% Start application with one crawler and callback function F. start(F) when is_function(F) -> start1(F). start1(Cback) -> start(), Id = supervisor_id(Cback), supervisor:start_child( hugin_sup, {Id, {supervisor, start_link, [{local, Id}, hugin_server_sup, [Cback, Id]]}, transient, 5000, supervisor, [ hugin_server_sup ]}). %% @doc Return the pool of URLs waiting to be fetched. -spec url(ServerRef :: server_ref()) -> [ url() ]. url(ServerRef) -> hugin_server:url(ServerRef). -spec url(ServerRef :: server_ref(), [url()]) -> ok | {error, any()}. %% @doc Add more URLs to be fetched. url(ServerRef, Urls) -> hugin_server:url(ServerRef, Urls). %%%======================================================================== %%% Functions to edit the behavior of the server %%%======================================================================== -spec max_freq(Ref :: server_ref(), Amount :: integer(), U :: hugin_opts:time_unit()) -> ok | {error, Reason :: any()}. %% @doc An option to limit the frequency of calls that Hugin makes. %% The default option is to have no frequency limits. %% However, Hugin still limits the amount of simultaneous outgoing connections %% to parallel sessions by default. %% See {@link max_par/2} for more information. %% @equiv max_freq(ServerRef, Amount, 1, U) %% @see hugin_opts:max_freq/3 max_freq(ServerRef, Amount, Unit) -> max_freq(ServerRef, Amount, 1, Unit). %% @equiv set_option(ServerRef, hugin_opts:max_freq(ServerRef, Amount, 1, U)) %% @see hugin_opts:max_freq/4 max_freq(ServerRef, Amount, N, Unit) -> set_option(ServerRef, hugin_opts:max_freq(Amount, N, Unit)). %% @doc An option to limit the amount of simultaneous parallel outgoing %% connections that Hugin will keep. The default is 5. max_par(ServerRef, N) -> set_option(ServerRef, hugin_opts:max_par(N)). -spec set_option(ServerRef :: server_ref(), Option :: hugin_opts:opt()) -> ok | {error, Reason :: any()}. %% @doc Immediately update the behavior of the Hugin server. %% @equiv set_options(ServerRef, [Option]) set_option(ServerRef, O) -> set_options(ServerRef, [O]). -spec set_options(ServerRef :: server_ref(), Options :: [ hugin_opts:opt() ]) -> ok | {error, Reason :: any()}. %% @doc Immediately update the behavior of the Hugin server. set_options(ServerRef, Options) -> %% Verify that ServerRef is of the right type. The options will be checked %% in hugin_server and return {error, badarg} if any of them are not. case is_atom(ServerRef) orelse is_pid(ServerRef) of false -> exit(badarg); true -> hugin_server:set_options(ServerRef, Options) end. %%%========================================================================= %%% Private Functions %%%========================================================================= supervisor_id(F) when is_function(F) -> supervisor_id(erlang:ref_to_list(make_ref())); supervisor_id(Module) -> list_to_atom( lists:concat([Module, "_sup"])).
src/hugin.erl
0.544075
0.570511
hugin.erl
starcoder
%% @author <NAME> <<EMAIL>> %% @doc This is AOC2020 <em>day11</em> solution %% @reference <a href="https://adventofcode.com/2020/day/11">AOC 2020 day 11</a> for %% @since 2020-12-18 %% @version 0.5.0 -module(day11). -export([solve_part1/1, solve_part2/1]). % could be shielded with -ifdef(TEST) macro -export([parse/1, adjacent/2, at/2, visible/2, next/2, next_adj/2, next_vis/2, count_occupied/1]). %%% solution solve_part1(Input) -> part1(parse(Input)). solve_part2(Input) -> part2(parse(Input)). -type tile() :: floor | empty | occupied. -type grid() :: [[tile(), ...], ...]. -spec parse(GridStr) -> Grid when GridStr :: string(), Grid :: grid(). %% @doc Convert Grid string to Grid. %% @param GridStr string in the format of task input. %% @returns is a double-list of atoms floor, empty, %% @throws shadows. %% and occupied. parse(GridStr) -> Lines = string:lexemes(GridStr, "\n"), [parse_line(Line) || Line <- Lines]. parse_line(Line) -> [parse_symbol(Symbol) || Symbol <- Line]. parse_symbol($.) -> floor; parse_symbol($L) -> empty; parse_symbol($#) -> occupied. %% @doc Lists sorted adjacent tiles' coordinates adjacent({MaxX, MaxY}, {X, Y}) -> lists:sort([ {AdjX, AdjY} || AdjX <- [X - 1, X, X + 1], AdjY <- [Y - 1, Y, Y + 1], {AdjX, AdjY} =/= {X, Y}, % self is not adjacent AdjX > 0, AdjY > 0, AdjX =< MaxX, AdjY =< MaxY ]). -spec at(Grid, {X, Y}) -> Tile when Grid :: grid(), X :: integer(), Y :: integer(), Tile :: tile(). %% @doc Shows what is at coordinates {X, Y}. at(Grid, {X, Y}) -> lists:nth(Y, lists:nth(X, Grid)). -spec visible(Grid, {X, Y}) -> Tiles when Grid :: grid(), X :: integer(), Y :: integer(), Tiles :: [tile()]. %% @doc Lists non-floor tiles visible from {X, Y}. visible(Grid, Origin) -> WithWalls = [arrow(Grid, Origin, Vector) || Vector <- [{1, 0}, {1, 1}, {0, 1}, {-1, 1}, {-1, 0}, {-1, -1}, {0, -1}, {1, -1}] ], WithoutWalls = [Tile || Tile <- WithWalls, Tile =/= wall], WithoutWalls. %% @doc Shoots an arrow from {X, Y} in the direction of {Dx, Dy} %% returns either `empty`, `occupied`, or `wall`. arrow(Grid, _Origin={X, Y}, Vector={Dx, Dy}) -> {MaxX, MaxY} = dimensions(Grid), NewOrigin = {NewX, NewY} = {X + Dx, Y + Dy}, case lists:member(NewX, [0, MaxX + 1]) or lists:member(NewY, [0, MaxY + 1]) of true -> wall; false -> case at(Grid, {NewX, NewY}) of empty -> empty; occupied -> occupied; floor -> arrow(Grid, NewOrigin, Vector) end end. %% @doc Measures X, Y dimensions of the Grid. dimensions(Grid) -> {length(Grid), length(lists:nth(1, Grid))}. %% @doc Advances grid to the next generation using strategy fun. next(Grid, Strategy) -> {MaxX, MaxY} = dimensions(Grid), [ [Strategy(Grid, {Row, Col}) || Col <- lists:seq(1, MaxY)] || Row <- lists:seq(1, MaxX) ]. %%% Counting Strategies %% @doc Produces the next generation tile at {X, Y}. next_adj(Grid, {X, Y}) -> {MaxX, MaxY} = dimensions(Grid), Tile = at(Grid, {X, Y}), Adjacent = [at(Grid, {AdjX, AdjY}) || {AdjX, AdjY} <- adjacent({MaxX, MaxY}, {X, Y})], next_adj2(Tile, Adjacent). %% @doc Produces the next generation tile given current and %% adjacent tiles. next_adj2(floor, _Adjacent) -> floor; next_adj2(empty, Adjacent) -> case lists:member(occupied, Adjacent) of false -> occupied; true -> empty end; next_adj2(occupied, Adjacent) -> case count(occupied, Adjacent) >= 4 of true -> empty; false -> occupied end. %% @doc Produces the next generation tile using visibility rules next_vis(Grid, Origin) -> Tile = at(Grid, Origin), Visible = visible(Grid, Origin), case Tile of floor -> floor; occupied -> case count(occupied, Visible) >= 5 of true -> empty; false -> occupied end; empty -> case count(occupied, Visible) =:= 0 of true -> occupied; false -> empty end end. %% @doc Counts the number of occupied seats. count_occupied(Grid) -> count(occupied, lists:flatten(Grid)). %% @doc Counts X's occurence in list. count(_, []) -> 0; count(X, [X | XS]) -> 1 + count(X, XS); count(X, [_|XS]) -> count(X, XS). %% @doc Counts occupied seats after equilibrium using next_adj. part1(Grid) -> NewGrid = next(Grid, fun next_adj/2), case NewGrid =:= Grid of true -> count_occupied(Grid); false -> part1(NewGrid) end. %% @doc Counts occupied seats after equilibrium using next_vis. part2(Grid) -> NewGrid = next(Grid, fun next_vis/2), case NewGrid =:= Grid of true -> count_occupied(Grid); false -> part2(NewGrid) end.
src/day11.erl
0.685107
0.586641
day11.erl
starcoder
% Licensed under the Apache License, Version 2.0 (the "License"); you may not % use this file except in compliance with the License. You may obtain a copy of % the License at % % http://www.apache.org/licenses/LICENSE-2.0 % % Unless required by applicable law or agreed to in writing, software % distributed under the License is distributed on an "AS IS" BASIS, WITHOUT % WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the % License for the specific language governing permissions and limitations under % the License. -module(couch_stats). -export([ start/0, stop/0, fetch/0, reload/0, sample/1, new/2, delete/1, list/0, increment_counter/1, increment_counter/2, decrement_counter/1, decrement_counter/2, update_histogram/2, update_gauge/2 ]). -include("couch_stats.hrl"). -type response() :: ok | {error, unknown_metric}. -type stat() :: {any(), [{atom(), any()}]}. start() -> application:start(couch_stats). stop() -> application:stop(couch_stats). fetch() -> couch_stats_aggregator:fetch(). reload() -> couch_stats_aggregator:reload(). -spec sample(any()) -> stat(). sample(Name) -> [{Name, Info}] = folsom_metrics:get_metric_info(Name), sample_type(Name, proplists:get_value(type, Info)). -spec new(atom(), any()) -> ok | {error, metric_exists | unsupported_type}. new(counter, Name) -> case folsom_metrics:new_counter(Name) of ok -> ok; {error, Name, metric_already_exists} -> {error, metric_exists} end; new(histogram, Name) -> Time = config:get_integer("stats", "interval", ?DEFAULT_INTERVAL), case folsom_metrics:new_histogram(Name, slide_uniform, {Time, 1024}) of ok -> ok; {error, Name, metric_already_exists} -> {error, metric_exists} end; new(gauge, Name) -> case folsom_metrics:new_gauge(Name) of ok -> ok; {error, Name, metric_already_exists} -> {error, metric_exists} end; new(_, _) -> {error, unsupported_type}. delete(Name) -> folsom_metrics:delete_metric(Name). list() -> folsom_metrics:get_metrics_info(). -spec increment_counter(any()) -> response(). increment_counter(Name) -> notify_existing_metric(Name, {inc, 1}, counter). -spec increment_counter(any(), pos_integer()) -> response(). increment_counter(Name, Value) -> notify_existing_metric(Name, {inc, Value}, counter). -spec decrement_counter(any()) -> response(). decrement_counter(Name) -> notify_existing_metric(Name, {dec, 1}, counter). -spec decrement_counter(any(), pos_integer()) -> response(). decrement_counter(Name, Value) -> notify_existing_metric(Name, {dec, Value}, counter). -spec update_histogram(any(), number()) -> response(); (any(), function()) -> any(). update_histogram(Name, Fun) when is_function(Fun, 0) -> Begin = os:timestamp(), Result = Fun(), Duration = timer:now_diff(os:timestamp(), Begin) div 1000, case notify_existing_metric(Name, Duration, histogram) of ok -> Result; {error, unknown_metric} -> throw({unknown_metric, Name}) end; update_histogram(Name, Value) when is_number(Value) -> notify_existing_metric(Name, Value, histogram). -spec update_gauge(any(), number()) -> response(). update_gauge(Name, Value) -> notify_existing_metric(Name, Value, gauge). -spec notify_existing_metric(any(), any(), any()) -> response(). notify_existing_metric(Name, Op, Type) -> try ok = folsom_metrics:notify_existing_metric(Name, Op, Type) catch _:_ -> couch_log:notice("unknown metric: ~p", [Name]), {error, unknown_metric} end. -spec sample_type(any(), atom()) -> stat(). sample_type(Name, histogram) -> folsom_metrics:get_histogram_statistics(Name); sample_type(Name, _) -> folsom_metrics:get_metric_value(Name).
src/couch_stats/src/couch_stats.erl
0.802981
0.491212
couch_stats.erl
starcoder
%% %% %CopyrightBegin% %% %% Copyright Ericsson AB 2022. All Rights Reserved. %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %% %% %CopyrightEnd% %% %% Purpose: Calculate tight bounds for integer operations. %% %% Reference: %% %% <NAME>, <NAME> (2 ed). <NAME> - %% Pearson Education, Inc. Chapter 4. Arithmetic Bounds. %% %% -module(beam_bounds). -export([bounds/2, bounds/3, relop/3, infer_relop_types/3]). -export_type([range/0]). -type range() :: {integer(), integer()} | {'-inf', integer()} | {integer(), '+inf'} | 'any'. -type range_result() :: range() | 'any'. -type relop() :: '<' | '=<' | '>' | '>='. -type bool_result() :: 'true' | 'false' | 'maybe'. -type op() :: atom(). %% Maximum size of integers in bits to keep ranges for. -define(NUM_BITS, 128). -spec bounds(op(), range()) -> range_result(). bounds(abs, R) -> case R of {A,B} when is_integer(A), is_integer(B) -> Min = 0, Max = max(abs(A), abs(B)), {Min,Max}; _ -> {0,'+inf'} end. -spec bounds(op(), range(), range()) -> range_result(). bounds('+', R1, R2) -> case {R1,R2} of {{A,B}, {C,D}} when abs(A) bsr ?NUM_BITS =:= 0, abs(B) bsr ?NUM_BITS =:= 0, abs(C) bsr ?NUM_BITS =:= 0, abs(D) bsr ?NUM_BITS =:= 0 -> normalize({A+C,B+D}); {{'-inf',B}, {_C,D}} when abs(B) bsr ?NUM_BITS =:= 0, abs(D) bsr ?NUM_BITS =:= 0 -> normalize({'-inf',B+D}); {{_A,B}, {'-inf',D}} when abs(B) bsr ?NUM_BITS =:= 0, abs(D) bsr ?NUM_BITS =:= 0 -> normalize({'-inf',B+D}); {{A,'+inf'}, {C,_D}} when abs(A) bsr ?NUM_BITS =:= 0, abs(C) bsr ?NUM_BITS =:= 0 -> normalize({A+C,'+inf'}); {{A,_B}, {C,'+inf'}} when abs(A) bsr ?NUM_BITS =:= 0, abs(C) bsr ?NUM_BITS =:= 0 -> normalize({A+C,'+inf'}); {_, _} -> any end; bounds('-', R1, R2) -> case {R1,R2} of {{A,B}, {C,D}} when abs(A) bsr ?NUM_BITS =:= 0, abs(B) bsr ?NUM_BITS =:= 0, abs(C) bsr ?NUM_BITS =:= 0, abs(D) bsr ?NUM_BITS =:= 0 -> normalize({A-D,B-C}); {{A,'+inf'}, {_C,D}} when abs(A) bsr ?NUM_BITS =:= 0, abs(D) bsr ?NUM_BITS =:= 0 -> normalize({A-D,'+inf'}); {{_A,B}, {C,'+inf'}} when abs(B) bsr ?NUM_BITS =:= 0, abs(C) bsr ?NUM_BITS =:= 0 -> normalize({'-inf',B-C}); {{'-inf',B}, {C,_D}} when abs(B) bsr ?NUM_BITS =:= 0, abs(C) bsr ?NUM_BITS =:= 0 -> normalize({'-inf',B-C}); {{A,_B}, {'-inf',D}} when abs(A) bsr ?NUM_BITS =:= 0, abs(D) bsr ?NUM_BITS =:= 0 -> normalize({A-D,'+inf'}); {_, _} -> any end; bounds('*', R1, R2) -> case {R1,R2} of {{A,B}, {C,D}} when abs(A) bsr ?NUM_BITS =:= 0, abs(B) bsr ?NUM_BITS =:= 0, abs(C) bsr ?NUM_BITS =:= 0, abs(D) bsr ?NUM_BITS =:= 0 -> All = [X * Y || X <- [A,B], Y <- [C,D]], Min = lists:min(All), Max = lists:max(All), normalize({Min,Max}); {{A,'+inf'}, {C,D}} when abs(A) bsr ?NUM_BITS =:= 0, abs(C) bsr ?NUM_BITS =:= 0, abs(D) bsr ?NUM_BITS =:= 0, C >= 0 -> {min(A*C, A*D),'+inf'}; {{'-inf',B}, {C,D}} when abs(B) bsr ?NUM_BITS =:= 0, abs(C) bsr ?NUM_BITS =:= 0, abs(D) bsr ?NUM_BITS =:= 0, C >= 0 -> {'-inf',max(B*C, B*D)}; {{A,B}, {'-inf',_}} when is_integer(A), is_integer(B) -> bounds('*', R2, R1); {{A,B}, {_,'+inf'}} when is_integer(A), is_integer(B) -> bounds('*', R2, R1); {_, _} -> any end; bounds('div', R1, R2) -> div_bounds(R1, R2); bounds('rem', R1, R2) -> rem_bounds(R1, R2); bounds('band', R1, R2) -> case {R1,R2} of {{A,B}, {C,D}} when A bsr ?NUM_BITS =:= 0, A >= 0, C bsr ?NUM_BITS =:= 0, C >= 0, is_integer(B), is_integer(D) -> Min = min_band(A, B, C, D), Max = max_band(A, B, C, D), {Min,Max}; {_, {C,D}} when is_integer(C), C >= 0 -> {0,D}; {{A,B}, _} when is_integer(A), A >= 0 -> {0,B}; {_, _} -> any end; bounds('bor', R1, R2) -> case {R1,R2} of {{A,B}, {C,D}} when A bsr ?NUM_BITS =:= 0, A >= 0, C bsr ?NUM_BITS =:= 0, C >= 0, is_integer(B), is_integer(D) -> Min = min_bor(A, B, C, D), Max = max_bor(A, B, C, D), {Min,Max}; {_, _} -> any end; bounds('bxor', R1, R2) -> case {R1,R2} of {{A,B}, {C,D}} when A bsr ?NUM_BITS =:= 0, A >= 0, C bsr ?NUM_BITS =:= 0, C >= 0, is_integer(B), is_integer(D) -> Max = max_bxor(A, B, C, D), {0,Max}; {_, _} -> any end; bounds('bsr', R1, R2) -> case {R1,R2} of {{A,B}, {C,D}} when is_integer(C), C >= 0 -> Min = inf_min(inf_bsr(A, C), inf_bsr(A, D)), Max = inf_max(inf_bsr(B, C), inf_bsr(B, D)), normalize({Min,Max}); {_, _} -> any end; bounds('bsl', R1, R2) -> case {R1,R2} of {{A,B}, {C,D}} when abs(A) bsr ?NUM_BITS =:= 0, abs(B) bsr ?NUM_BITS =:= 0 -> Min = inf_min(inf_bsl(A, C), inf_bsl(A, D)), Max = inf_max(inf_bsl(B, C), inf_bsl(B, D)), normalize({Min,Max}); {_, _} -> any end; bounds(max, R1, R2) -> case {R1,R2} of {{A,B},{C,D}} -> normalize({inf_max(A, C),inf_max(B, D)}); {_,_} -> any end; bounds(min, R1, R2) -> case {R1,R2} of {{A,B},{C,D}} -> normalize({inf_min(A, C),inf_min(B, D)}); {_,_} -> any end. -spec relop(relop(), range(), range()) -> bool_result(). relop('<', {A,B}, {C,D}) -> case {inf_lt(B, C),inf_lt(A, D)} of {Bool,Bool} -> Bool; {_,_} -> 'maybe' end; relop('=<', {A,B}, {C,D}) -> case {inf_le(B, C),inf_le(A, D)} of {Bool,Bool} -> Bool; {_,_} -> 'maybe' end; relop('>=', {A,B}, {C,D}) -> case {inf_ge(B, C),inf_ge(A, D)} of {Bool,Bool} -> Bool; {_,_} -> 'maybe' end; relop('>', {A,B}, {C,D}) -> case {inf_gt(B, C),inf_gt(A, D)} of {Bool,Bool} -> Bool; {_,_} -> 'maybe' end; relop(_, _, _) -> 'maybe'. -spec infer_relop_types(relop(), range(), range()) -> any(). infer_relop_types(Op, {_,_}=Range1, {_,_}=Range2) -> case relop(Op, Range1, Range2) of 'maybe' -> infer_relop_types_1(Op, Range1, Range2); _ -> any end; infer_relop_types('<', {A,_}=R1, any) -> {R1, normalize({inf_add(A, 1), '+inf'})}; infer_relop_types('<', any, {_,D}=R2) -> {normalize({'-inf', inf_add(D, -1)}), R2}; infer_relop_types('=<', {A,_}=R1, any) -> {R1, normalize({A, '+inf'})}; infer_relop_types('=<', any, {_,D}=R2) -> {normalize({'-inf', D}), R2}; infer_relop_types('>=', {_,B}=R1, any) -> {R1, normalize({'-inf', B})}; infer_relop_types('>=', any, {C,_}=R2) -> {normalize({C, '+inf'}), R2}; infer_relop_types('>', {_,B}=R1, any) -> {R1, normalize({'-inf', inf_add(B, -1)})}; infer_relop_types('>', any, {C,_}=R2) -> {normalize({inf_add(C, 1), '+inf'}), R2}; infer_relop_types(_Op, _R1, _R2) -> any. %%% %%% Internal functions. %%% div_bounds({A,B}, {C,D}) when is_integer(A), is_integer(B), is_integer(C), is_integer(D) -> Denominators = [min(C, D),max(C, D)| %% Handle zero crossing for the denominator. if C < 0, 0 < D -> [-1, 1]; C =:= 0 -> [1]; D =:= 0 -> [-1]; true -> [] end], All = [X div Y || X <- [A,B], Y <- Denominators, Y =/= 0], Min = lists:min(All), Max = lists:max(All), normalize({Min,Max}); div_bounds({A,'+inf'}, {C,D}) when is_integer(C), C > 0, is_integer(D) -> Min = min(A div C, A div D), Max = '+inf', normalize({Min,Max}); div_bounds({'-inf',B}, {C,D}) when is_integer(C), C > 0, is_integer(D) -> Min = '-inf', Max = max(B div C, B div D), normalize({Min,Max}); div_bounds(_, _) -> any. rem_bounds({A,_}, {C,D}) when is_integer(C), is_integer(D), C > 0 -> Max = inf_add(D, -1), Min = if A =:= '-inf' -> -Max; A >= 0 -> 0; true -> -Max end, normalize({Min,Max}); rem_bounds(_, {C,D}) when is_integer(C), is_integer(D), C =/= 0 orelse D =/= 0 -> Max = max(abs(C), abs(D)) - 1, Min = -Max, normalize({Min,Max}); rem_bounds(_, _) -> any. min_band(A, B, C, D) -> M = 1 bsl (upper_bit(A bor C) + 1), min_band(A, B, C, D, M). min_band(A, _B, C, _D, 0) -> A band C; min_band(A, B, C, D, M) -> if (bnot A) band (bnot C) band M =/= 0 -> case (A bor M) band -M of NewA when NewA =< B -> min_band(NewA, B, C, D, 0); _ -> case (C bor M) band -M of NewC when NewC =< D -> min_band(A, B, NewC, D, 0); _ -> min_band(A, B, C, D, M bsr 1) end end; true -> min_band(A, B, C, D, M bsr 1) end. max_band(A, B, C, D) -> M = 1 bsl upper_bit(B bxor D), max_band(A, B, C, D, M). max_band(_A, B, _C, D, 0) -> B band D; max_band(A, B, C, D, M) -> if B band (bnot D) band M =/= 0 -> case (B band (bnot M)) bor (M - 1) of NewB when NewB >= A -> max_band(A, NewB, C, D, 0); _ -> max_band(A, B, C, D, M bsr 1) end; (bnot B) band D band M =/= 0 -> case (D band (bnot M)) bor (M - 1) of NewD when NewD >= C -> max_band(A, B, C, NewD, 0); _ -> max_band(A, B, C, D, M bsr 1) end; true -> max_band(A, B, C, D, M bsr 1) end. min_bor(A, B, C, D) -> M = 1 bsl upper_bit(A bxor C), min_bor(A, B, C, D, M). min_bor(A, _B, C, _D, 0) -> A bor C; min_bor(A, B, C, D, M) -> if (bnot A) band C band M =/= 0 -> case (A bor M) band -M of NewA when NewA =< B -> min_bor(NewA, B, C, D, 0); _ -> min_bor(A, B, C, D, M bsr 1) end; A band (bnot C) band M =/= 0 -> case (C bor M) band -M of NewC when NewC =< D -> min_bor(A, B, NewC, D, 0); _ -> min_bor(A, B, C, D, M bsr 1) end; true -> min_bor(A, B, C, D, M bsr 1) end. max_bor(A, B, C, D) -> Intersection = B band D, M = 1 bsl upper_bit(Intersection), max_bor(Intersection, A, B, C, D, M). max_bor(_Intersection, _A, B, _C, D, 0) -> B bor D; max_bor(Intersection, A, B, C, D, M) -> if Intersection band M =/= 0 -> case (B - M) bor (M - 1) of NewB when NewB >= A -> max_bor(Intersection, A, NewB, C, D, 0); _ -> case (D - M) bor (M - 1) of NewD when NewD >= C -> max_bor(Intersection, A, B, C, NewD, 0); _ -> max_bor(Intersection, A, B, C, D, M bsr 1) end end; true -> max_bor(Intersection, A, B, C, D, M bsr 1) end. max_bxor(A, B, C, D) -> M = 1 bsl upper_bit(B band D), max_bxor(A, B, C, D, M). max_bxor(_A, B, _C, D, 0) -> B bxor D; max_bxor(A, B, C, D, M) -> if B band D band M =/= 0 -> case (B - M) bor (M - 1) of NewB when NewB >= A -> max_bxor(A, NewB, C, D, M bsr 1); _ -> case (D - M) bor (M - 1) of NewD when NewD >= C -> max_bxor(A, B, C, NewD, M bsr 1); _ -> max_bxor(A, B, C, D, M bsr 1) end end; true -> max_bxor(A, B, C, D, M bsr 1) end. upper_bit(Val) -> upper_bit_1(Val, 0). upper_bit_1(Val0, N) -> case Val0 bsr 1 of 0 -> N; Val -> upper_bit_1(Val, N + 1) end. infer_relop_types_1('<', {A,B}, {C,D}) -> Left = normalize({A, clamp(inf_add(D, -1), A, B)}), Right = normalize({clamp(inf_add(A, 1), C, D), D}), {Left,Right}; infer_relop_types_1('=<', {A,B}, {C,D}) -> Left = normalize({A, clamp(D, A, B)}), Right = normalize({clamp(A, C, D), D}), {Left,Right}; infer_relop_types_1('>=', {A,B}, {C,D}) -> Left = normalize({clamp(C, A, B), B}), Right = normalize({C, clamp(B, C, D)}), {Left,Right}; infer_relop_types_1('>', {A,B}, {C,D}) -> Left = normalize({clamp(inf_add(C, 1), A, B), B}), Right = normalize({C,clamp(inf_add(B, -1), C, D)}), {Left,Right}. %%% %%% Handling of ranges. %%% %%% A range can begin with '-inf' OR end with '+inf'. %%% %%% Atoms are greater than all integers. Therefore, we don't %%% need any special handling of '+inf'. %%% normalize({'-inf','-inf'}) -> {'-inf',-1}; normalize({'-inf','+inf'}) -> any; normalize({'+inf','+inf'}) -> {0,'+inf'}; normalize({Min,Max}=T) -> true = inf_ge(Max, Min), T. clamp(V, A, B) -> inf_min(inf_max(V, A), B). inf_min(A, B) when A =:= '-inf'; B =:= '-inf' -> '-inf'; inf_min(A, B) when A =< B -> A; inf_min(A, B) when A > B -> B. inf_max('-inf', B) -> B; inf_max(A, '-inf') -> A; inf_max(A, B) when A >= B -> A; inf_max(A, B) when A < B -> B. inf_neg('-inf') -> '+inf'; inf_neg('+inf') -> '-inf'; inf_neg(N) -> -N. inf_add(Int, N) when is_integer(Int) -> Int + N; inf_add(Inf, _N) -> Inf. inf_bsr('-inf', _S) -> '-inf'; inf_bsr('+inf', _S) -> '+inf'; inf_bsr(N, S0) when S0 =:= '-inf'; S0 < 0 -> S = inf_neg(S0), if S >= ?NUM_BITS, N < 0 -> '-inf'; S >= ?NUM_BITS, N >= 0 -> '+inf'; true -> N bsl S end; inf_bsr(N, '+inf') -> if N < 0 -> -1; N >= 0 -> 0 end; inf_bsr(N, S) when S >= 0 -> N bsr S. inf_bsl(N, S) -> inf_bsr(N, inf_neg(S)). inf_lt(_, '-inf') -> false; inf_lt('-inf', _) -> true; inf_lt(A, B) -> A < B. inf_ge(_, '-inf') -> true; inf_ge('-inf', _) -> false; inf_ge(A, B) -> A >= B. inf_le(A, B) -> inf_ge(B, A). inf_gt(A, B) -> inf_lt(B, A).
lib/compiler/src/beam_bounds.erl
0.675444
0.433442
beam_bounds.erl
starcoder
%% This Source Code Form is subject to the terms of the Mozilla Public %% License, v. 2.0. If a copy of the MPL was not distributed with this %% file, You can obtain one at https://mozilla.org/MPL/2.0/. %% %% Copyright (c) 2021-2022 VMware, Inc. or its affiliates. All rights reserved. %% %% @doc Condition support. %% %% Conditions can be used in path patterns and `keep_while' conditions. They %% allow to point to a specific node only if conditions are met, or to match %% several tree nodes with a single path pattern. %% %% A condition is an Erlang record defining a specific property. Some of them %% have arguments to further define the condition. %% %% All supported conditions are described in the <a href="#types">Data Types %% section</a>. -module(khepri_condition). -include("include/khepri.hrl"). -include("src/internal.hrl"). -type comparison_op(Type) :: {eq, Type} | {ne, Type} | {lt, Type} | {le, Type} | {gt, Type} | {ge, Type}. %% Comparison operator in some {@link condition()}. -type if_name_matches() :: #if_name_matches{}. %% Condition. Evaluates to true if the name of the tested node matches the %% condition pattern. %% %% Record fields: %% <ul> %% <li>`regex': defines the condition pattern. It can be either: %% <ul> %% <li>a regular expression</li> %% <li>the atom `any' to match any node names; the equivalent of the `".*"' %% regular expression but more efficient</li> %% </ul></li> %% </ul> %% %% Example: %% ``` %% #if_name_matches{regex = "^user_"}. %% #if_name_matches{regex = any}. %% ''' -type if_path_matches() :: #if_path_matches{}. %% Condition. Evaluates to true if the name of the tested node matches the %% condition pattern. If it does not match, child node names are tested %% recursively. %% %% Record fields: %% <ul> %% <li>`regex': defines the condition pattern. It can be either: %% <ul> %% <li>a regular expression</li> %% <li>the atom `any' to match any node names; the equivalent of the `".*"' %% regular expression but more efficient</li> %% </ul></li> %% </ul> %% %% Example: %% ``` %% #if_path_matches{regex = "^user_"}. %% #if_path_matches{regex = any}. %% ''' -type if_has_data() :: #if_has_data{}. %% Condition. Evaluates to true if the tested node's data payload presence %% corresponds to the expected state. %% %% Record fields: %% <ul> %% <li>`has_data': boolean set to the expected presence of a data %% payload.</li> %% </ul> %% %% Data absence is either no payload or a non-data type of payload. %% %% Example: %% ``` %% #if_has_data{has_data = false}. %% ''' -type if_data_matches() :: #if_data_matches{}. %% Condition. Evaluates to true if the tested node has a data payload and the %% data payload term matches the given pattern. %% %% Record fields: %% <ul> %% <li>`pattern': an ETS-like match pattern.</li> %% </ul> %% %% Example: %% ``` %% #if_data_matches{pattern = {user, '_'}}. %% ''' -type if_node_exists() :: #if_node_exists{}. %% Condition. Evaluates to true if the tested node existence corresponds to %% the expected state. %% %% Record fields: %% <ul> %% <li>`exists': boolean set to the expected presence of the node.</li> %% </ul> %% %% Example: %% ``` %% #if_node_exists{exists = false}. %% ''' -type if_payload_version() :: #if_payload_version{}. %% Condition. Evaluates to true if the tested node's payload version %% corresponds to the expected value. %% %% Record fields: %% <ul> %% <li>`version': integer or {@link comparison_op()} to compare to the actual %% payload version.</li> %% </ul> %% %% Example: %% ``` %% #if_payload_version{version = 1}. %% #if_payload_version{version = {gt, 10}}. %% ''' -type if_child_list_version() :: #if_child_list_version{}. %% Condition. Evaluates to true if the tested node's child list version %% corresponds to the expected value. %% %% Record fields: %% <ul> %% <li>`version': integer or {@link comparison_op()} to compare to the actual %% child list version.</li> %% </ul> %% %% Example: %% ``` %% #if_child_list_version{version = 1}. %% #if_child_list_version{version = {gt, 10}}. %% ''' -type if_child_list_length() :: #if_child_list_length{}. %% Condition. Evaluates to true if the tested node's child list size %% corresponds to the expected value. %% %% Record fields: %% <ul> %% <li>`version': integer or {@link comparison_op()} to compare to the actual %% child list child.</li> %% </ul> %% %% Example: %% ``` %% #if_child_list_length{count = 1}. %% #if_child_list_length{count = {gt, 10}}. %% ''' -type if_not() :: #if_not{}. %% Condition. Evaluates to true if the inner condition evalutes to false. %% %% Record fields: %% <ul> %% <li>`condition': the inner condition to evaluate.</li> %% </ul> %% %% Example: %% ``` %% #if_not{condition = #if_name_matches{regex = "^a"}}. %% ''' -type if_all() :: #if_all{}. %% Condition. Evaluates to true if all inner conditions evalute to true. %% %% Record fields: %% <ul> %% <li>`conditions': a list of inner conditions to evaluate.</li> %% </ul> %% %% Example: %% ``` %% #if_all{conditions = [#if_name_matches{regex = "^a"}, %% #if_has_data{has_data = true}]}. %% ''' -type if_any() :: #if_any{}. %% Condition. Evaluates to true if any of the inner conditions evalute to %% true. %% %% Record fields: %% <ul> %% <li>`conditions': a list of inner conditions to evaluate.</li> %% </ul> %% %% Example: %% ``` %% #if_any{conditions = [#if_name_matches{regex = "^a"}, %% #if_has_data{has_data = true}]}. %% ''' -type condition() :: if_name_matches() | if_path_matches() | if_has_data() | if_data_matches() | if_node_exists() | if_payload_version() | if_child_list_version() | if_child_list_length() | if_not() | if_all() | if_any(). %% All supported conditions. -type condition_using_regex() :: if_name_matches() | if_path_matches(). -type condition_using_comparison_op() :: if_payload_version() | if_child_list_version() | if_child_list_length(). -type keep_while() :: #{khepri_path:path() => condition()}. %% An association between a path and a condition. As long as the condition %% evalutes to true, the tree node is kept. Once the condition evalutes to %% false, the tree node is deleted. -export([compile/1, applies_to_grandchildren/1, is_met/3, is_valid/1]). -ifdef(TEST). -export([eval_regex/4, compare_numerical_values/2]). -endif. -export_type([condition/0, comparison_op/1, keep_while/0]). -spec compile(Condition) -> Condition when Condition :: khepri_path:pattern_component(). %% @doc Preprocess properties inside some conditions to make them more %% efficient. %% %% An example is the regular expression inside an {@link if_name_matches()} %% condition. %% %% Conditions are also optimized if possible. An example is the replacement of %% an {@link if_all()} condition and all its sub-conditions if one of them is %% a specific node name. In this case, it is replaced by the node name %% directly. %% %% @param Condition the condition to compile. %% %% @returns the same condition with all its properties preprocessed. %% %% @private compile(#if_name_matches{regex = any} = Cond) -> Cond; compile(#if_name_matches{regex = Re, compiled = undefined} = Cond) -> Compiled = re:compile(Re), Cond#if_name_matches{compiled = Compiled}; compile(#if_data_matches{pattern = Pattern, compiled = undefined} = Cond) -> Compiled = ets:match_spec_compile([{Pattern, [], [match]}]), Cond#if_data_matches{compiled = Compiled}; compile(#if_not{condition = InnerCond} = Cond) -> InnerCond1 = compile(InnerCond), Cond#if_not{condition = InnerCond1}; compile(#if_all{conditions = InnerConds} = Cond) -> InnerConds1 = lists:map(fun compile/1, InnerConds), case optimize_if_all_conditions(InnerConds1) of [InnerCond] -> InnerCond; InnerConds2 -> Cond#if_all{conditions = InnerConds2} end; compile(#if_any{conditions = InnerConds} = Cond) -> InnerConds1 = lists:map(fun compile/1, InnerConds), case optimize_if_any_conditions(InnerConds1) of [InnerCond] -> InnerCond; InnerConds2 -> Cond#if_any{conditions = InnerConds2} end; compile(Cond) -> Cond. -spec optimize_if_all_conditions([condition()]) -> [condition()]. %% @private %% @hidden optimize_if_all_conditions(Conds) -> optimize_if_all_conditions(Conds, []). %% @private %% @hidden optimize_if_all_conditions([ChildName | Rest], Result) when ?IS_PATH_COMPONENT(ChildName) -> %% The path component exact match condition will become the first one %% tested. Result1 = Result ++ [ChildName], optimize_if_all_conditions(Rest, Result1); optimize_if_all_conditions([Cond | Rest], Result) -> Result1 = [Cond | Result], optimize_if_all_conditions(Rest, Result1); optimize_if_all_conditions([], Result) -> lists:reverse(Result). -spec optimize_if_any_conditions([condition()]) -> [condition()]. %% @private %% @hidden optimize_if_any_conditions(Conds) -> Conds. -spec applies_to_grandchildren(condition()) -> boolean(). %% @doc Returns true if a condition should be evaluted against child nodes in %% addition to the current node. %% %% An example is the {@link if_path_matches()} condition. %% %% @private applies_to_grandchildren(#if_path_matches{}) -> true; applies_to_grandchildren(#if_not{condition = Cond}) -> applies_to_grandchildren(Cond); applies_to_grandchildren(#if_all{conditions = Conds}) -> lists:any(fun applies_to_grandchildren/1, Conds); applies_to_grandchildren(#if_any{conditions = Conds}) -> lists:any(fun applies_to_grandchildren/1, Conds); applies_to_grandchildren(_) -> false. -spec is_met(Condition, PathOrChildName, Child) -> IsMet when Condition :: khepri_path:pattern_component(), PathOrChildName :: khepri_path:path() | khepri_path:component(), Child :: khepri_machine:tree_node() | khepri_machine:node_props(), IsMet :: true | IsNotMet1 | IsNotMet2, IsNotMet1 :: {false, khepri_path:pattern_component()}, IsNotMet2 :: {false, {condition(), any()}}. %% @doc Returns true if the given condition is met when evaluated against the %% given tree node name and properties. %% %% @param Condition the condition to evaluate. %% @param PathOrChildName the path or child name to consider. %% @param Child the properties or the tree node. %% %% @returns true if the condition is met, false otherwise. %% %% @private is_met(Condition, Path, Child) when ?IS_PATH(Path) -> ChildName = case Path of [] -> ''; _ -> lists:last(Path) end, is_met(Condition, ChildName, Child); is_met(ChildName, ChildName, _Child) when ?IS_PATH_COMPONENT(ChildName) -> true; is_met(?THIS_NODE, _ChildNameB, _Child) -> true; is_met(ChildNameA, _ChildNameB, _Child) when ?IS_PATH_COMPONENT(ChildNameA) -> {false, ChildNameA}; is_met(#if_node_exists{exists = true}, _ChildName, _Child) -> true; is_met(#if_node_exists{exists = false} = Cond, _ChildName, _Child) -> {false, Cond}; is_met( #if_name_matches{regex = SourceRegex, compiled = CompiledRegex} = Cond, ChildName, _Child) -> eval_regex(Cond, SourceRegex, CompiledRegex, ChildName); is_met( #if_path_matches{regex = SourceRegex, compiled = CompiledRegex} = Cond, ChildName, _Child) -> eval_regex(Cond, SourceRegex, CompiledRegex, ChildName); is_met(#if_has_data{has_data = true}, _ChildName, #node{payload = #kpayload_data{data = _}}) -> true; is_met(#if_has_data{has_data = false} = Cond, _ChildName, #node{payload = #kpayload_data{data = _}}) -> {false, Cond}; is_met(#if_has_data{has_data = true} = Cond, _ChildName, _Child) -> {false, Cond}; is_met(#if_has_data{has_data = false}, _ChildName, _Child) -> true; is_met(#if_data_matches{compiled = CompMatchSpec} = Cond, _ChildName, #node{payload = #kpayload_data{data = Data}}) -> case term_matches(Data, CompMatchSpec) of true -> true; false -> {false, Cond} end; is_met(#if_data_matches{compiled = CompMatchSpec} = Cond, _ChildName, #{data := Data}) -> case term_matches(Data, CompMatchSpec) of true -> true; false -> {false, Cond} end; is_met(#if_payload_version{version = DVersionB} = Cond, _ChildName, #node{stat = #{payload_version := DVersionA}}) -> compare_numerical_values(Cond, DVersionA, DVersionB); is_met(#if_payload_version{version = DVersionB} = Cond, _ChildName, #{payload_version := DVersionA}) -> compare_numerical_values(Cond, DVersionA, DVersionB); is_met(#if_child_list_version{version = CVersionB} = Cond, _ChildName, #node{stat = #{child_list_version := CVersionA}}) -> compare_numerical_values(Cond, CVersionA, CVersionB); is_met(#if_child_list_version{version = CVersionB} = Cond, _ChildName, #{child_list_version := CVersionA}) -> compare_numerical_values(Cond, CVersionA, CVersionB); is_met(#if_child_list_length{count = ExpectedCount} = Cond, _ChildName, #node{child_nodes = Children}) -> Count = maps:size(Children), compare_numerical_values(Cond, Count, ExpectedCount); is_met(#if_child_list_length{count = ExpectedCount} = Cond, _ChildName, #{child_list_length := Count}) -> compare_numerical_values(Cond, Count, ExpectedCount); is_met(#if_not{condition = InnerCond} = Cond, ChildName, Child) -> case is_met(InnerCond, ChildName, Child) of true -> {false, Cond}; {false, _} -> true end; is_met(#if_all{conditions = []}, _ChildName, _Child) -> true; is_met(#if_all{conditions = Conds}, ChildName, Child) -> lists:foldl( fun (_, {false, _} = False) -> False; (Cond, _) -> is_met(Cond, ChildName, Child) end, true, Conds); is_met(#if_any{conditions = []} = Cond, _ChildName, _Child) -> {false, Cond}; is_met(#if_any{conditions = Conds} = IfAnyCond, ChildName, Child) -> Ret = lists:foldl( fun (_, true) -> true; (Cond, _) -> is_met(Cond, ChildName, Child) end, {false, undefined}, Conds), case Ret of true -> true; {false, _} -> {false, IfAnyCond} end; is_met(Cond, _, _) -> {false, Cond}. -spec term_matches(Term, MatchSpec) -> Matches when Term :: khepri_machine:data(), MatchSpec :: ets:comp_match_spec(), Matches :: boolean(). %% @doc Returns true if the given match spec matches the given match term. %% %% @private %% @hidden term_matches(Term, MatchSpec) -> case ets:match_spec_run([Term], MatchSpec) of [match] -> true; _ -> false end. -spec eval_regex(Condition, SourceRegex, CompiledRegex, Value) -> Ret when Condition :: condition_using_regex(), SourceRegex :: any | iodata() | unicode:charlist(), CompiledRegex :: {ok, re_mp()} | {error, {string(), non_neg_integer()}} | undefined, Value :: atom() | iodata() | unicode:charlist(), Ret :: true | {false, condition_using_regex()} | {false, {condition_using_regex(), {error, match_limit | match_limit_recursion | {string(), non_neg_integer()}}}}. %% @doc Returns true if the given regular expression matches the given string. %% %% @private %% @hidden eval_regex(_Cond, any, _CompiledRegex, _Value) -> true; eval_regex(Cond, _SourceRegex, {ok, Regex}, Value) when not is_atom(Value) -> case re:run(Value, Regex, [{capture, none}]) of match -> true; nomatch -> {false, Cond}; Error -> {false, {Cond, Error}} end; eval_regex(Cond, _SourceRegex, {error, _} = Error, Value) when not is_atom(Value) -> {false, {Cond, Error}}; eval_regex(Cond, SourceRegex, CompiledRegex, Value) when is_atom(Value) -> eval_regex(Cond, SourceRegex, CompiledRegex, atom_to_list(Value)); eval_regex(Cond, SourceRegex, undefined, Value) -> Compiled = re:compile(SourceRegex), eval_regex(Cond, SourceRegex, Compiled, Value). -spec compare_numerical_values(Cond, ValueA, ValueB) -> Equal when Cond :: condition_using_comparison_op(), ValueA :: non_neg_integer(), ValueB :: non_neg_integer() | comparison_op(non_neg_integer()), Equal :: true | {false, condition_using_comparison_op()}. %% @private %% @hidden compare_numerical_values(Cond, ValueA, ValueB) -> case compare_numerical_values(ValueA, ValueB) of true -> true; false -> {false, Cond} end. -spec compare_numerical_values(ValueA, ValueB) -> Equal when ValueA :: non_neg_integer(), ValueB :: non_neg_integer() | comparison_op(non_neg_integer()), Equal :: boolean(). %% @private %% @hidden compare_numerical_values(Value, Value) -> true; compare_numerical_values(Value, {eq, Value}) -> true; compare_numerical_values(ValueA, {ne, ValueB}) -> ValueA =/= ValueB; compare_numerical_values(ValueA, {lt, ValueB}) -> ValueA < ValueB; compare_numerical_values(ValueA, {le, ValueB}) -> ValueA =< ValueB; compare_numerical_values(ValueA, {gt, ValueB}) -> ValueA > ValueB; compare_numerical_values(ValueA, {ge, ValueB}) -> ValueA >= ValueB; compare_numerical_values(_, _) -> false. -spec is_valid(Condition) -> IsValid when Condition :: khepri_path:pattern_component(), IsValid :: true | {false, khepri_path:pattern_component()}. %% @doc Returns true if the condition's properties are valid. %% %% For instance, the function verifies that {@link if_node_exists()} takes a %% boolean(). %% %% @param Condition the condition to verify. %% %% @returns true if the condition's properties are valid, false otherwise. %% %% @private is_valid(Component) when ?IS_PATH_COMPONENT(Component) -> true; is_valid(?THIS_NODE) -> true; is_valid(#if_node_exists{exists = Exists}) -> is_boolean(Exists); is_valid(#if_name_matches{}) -> true; is_valid(#if_path_matches{}) -> true; is_valid(#if_has_data{has_data = HasData}) -> is_boolean(HasData); is_valid(#if_data_matches{}) -> true; is_valid(#if_payload_version{}) -> true; is_valid(#if_child_list_version{}) -> true; is_valid(#if_child_list_length{}) -> true; is_valid(#if_not{condition = InnerCond}) -> is_valid(InnerCond); is_valid(#if_all{conditions = Conds}) when is_list(Conds) -> lists:foldl( fun (_, {false, _} = False) -> False; (Cond, _) -> is_valid(Cond) end, true, Conds); is_valid(#if_any{conditions = Conds}) when is_list(Conds) -> lists:foldl( fun (_, {false, _} = False) -> False; (Cond, _) -> is_valid(Cond) end, true, Conds); is_valid(Cond) -> {false, Cond}.
src/khepri_condition.erl
0.770853
0.458288
khepri_condition.erl
starcoder
%% @doc Observing a metric involves periodically recording a snapshot of %% a scalar metric in the form of a histogram. These snapshots are recorded %% over a pre-configured interval. Each histogram will be registered %% automatically with the `metrics_reader'. %% Currently, only the folsom backend is supported but it is possible to extend %% the idea to any metrics backend using the erlang-metrics library %% interface here: %% https://github.com/benoitc/erlang-metrics %% Warning: observing a metric will clear it's current value whenenver a %% snapshot is recorded. %% @end -module(metrics_observer). -behaviour(gen_server). -include("metrics_reader.hrl"). %% API -export([start_link/0, observe/2, unobserve/1, observed/0]). %% gen_server callbacks -export([init/1, handle_call/3, handle_cast/2, handle_info/2, terminate/2, code_change/3]). -define(SERVER, ?MODULE). -record(state, {running = false :: boolean(), observations = sets:new(), slide_interval :: pos_integer(), acc_interval :: pos_integer(), timer_ref :: reference()}). -type metric_name() :: any(). -type histogram_name() :: any(). -type state() :: #state{}. %%%=================================================================== %%% API %%%=================================================================== -spec start_link() -> gen_server_startlink_ret(). start_link() -> gen_server:start_link({local, ?SERVER}, ?MODULE, [], []). -spec observe(metric_name(), histogram_name()) -> ok. observe(Name, HistogramName) -> gen_server:call(?SERVER, {observe, Name, HistogramName}). -spec unobserve(metric_name()) -> ok. unobserve(Name) -> gen_server:call(?SERVER, {unobserve, Name}). -spec observed() -> list(). observed() -> gen_server:call(?SERVER, observed). %%%=================================================================== %%% gen_server callbacks %%%=================================================================== -spec init([]) -> {ok, state()}. init([]) -> %% We want a high priority to ensure reporting accuracy process_flag(priority, high), SInterval = metrics_reader_helper:opt(histogram_slide_interval_sec, 60), AInterval = metrics_reader_helper:opt(histogram_acc_interval_sec, 1), {ok, #state{slide_interval = SInterval, acc_interval = AInterval * 1000}}. -spec handle_call(any(), any(), state()) -> {reply, term(), state()}. handle_call({observe, Name, HistogramName}, _From, State = #state{running = false, acc_interval = AccInterval}) -> State1 = do_observe({Name, HistogramName}, State), TRef = erlang:send_after(AccInterval, self(), tick), Reply = ok, {reply, Reply, State1#state{running = true, timer_ref = TRef}}; handle_call({observe, Name, HistogramName}, _From, State = #state{running = true}) -> Reply = ok, State1 = do_observe({Name, HistogramName}, State), {reply, Reply, State1}; handle_call({unobserve, Name}, _From, State = #state{running = true, timer_ref = TRef, observations = OSet}) -> Empty = sets:new(), [{Name, Histogram}] = [{N, H} || {N, H} <- sets:to_list(OSet), N =:= Name], State1 = do_unobserve({Name, Histogram}, State), #state{observations = OSet1} = State1, case OSet1 of Empty -> erlang:cancel_timer(TRef), {reply, ok, State1#state{running = false, timer_ref = undefined}}; _ -> {reply, ok, State1} end; handle_call(observed, _From, State = #state{observations = OSet}) -> Reply = [Name || {Name, _HistogramName} <- sets:to_list(OSet)], {reply, Reply, State}; handle_call(_Request, _From, State) -> Reply = ok, {reply, Reply, State}. -spec handle_cast(any(), state()) -> {noreply, state()}. handle_cast(_Msg, State) -> {noreply, State}. -spec handle_info(any(), state()) -> {noreply, state()} | {stop, any(), state()}. handle_info(tick, State = #state{running = true, observations = OSet, acc_interval = AccInterval}) -> lists:foreach(fun ({Name, HistogramName}) -> Value = folsom_metrics_counter:get_value(Name), folsom_metrics:notify({HistogramName, Value}), folsom_metrics_counter:clear(Name) end, sets:to_list(OSet)), TRef = erlang:send_after(AccInterval, self(), tick), {noreply, State#state{timer_ref = TRef}}; handle_info(_Info, State) -> {noreply, State}. -spec terminate(any(), any()) -> ok. terminate(_Reason, State = #state{timer_ref = TRef, observations = OSet}) -> Empty = sets:new(), #state{observations = Empty} = lists:foldl( fun do_unobserve/2, State, sets:to_list(OSet)), erlang:cancel_timer(TRef), ok. -spec code_change(any(), state(), any()) -> {ok, state()}. code_change(_OldVsn, State, _Extra) -> {ok, State}. %%%=================================================================== %%% Internal functions %%%=================================================================== do_observe({Name, HistogramName}, State = #state{slide_interval = SlideInterval, observations = OSet}) -> case folsom_metrics:get_metric_info(Name) of [{_, [{type, histogram}]}] -> ok; _ -> folsom_metrics:new_histogram(HistogramName, slide, SlideInterval) end, metrics_reader:register(HistogramName), OSet1 = sets:add_element({Name, HistogramName}, OSet), State#state{observations = OSet1}. do_unobserve({Name, HistogramName}, State = #state{observations = OSet}) -> metrics_reader:deregister(HistogramName), folsom_metrics:delete_metric(HistogramName), OSet1 = sets:del_element({Name, HistogramName}, OSet), State#state{observations = OSet1}.
src/metrics_observer.erl
0.708011
0.558748
metrics_observer.erl
starcoder
-module(pvc_ring). -include("pvc.hrl"). %% API -export([grb_replica_info/3, random_indexnode/1, get_key_indexnode/3, size/1]). %% @doc Raw ring structure returned from antidote %% %% Nodes are in erlang format, i.e. node_name@ip_address -type raw_ring() :: list({partition_id(), node()}). %% @doc Fixed ring structured used to route protocol requests %% %% Uses a tuple-based structure to enable index accesses %% in constant time. %% -type fixed_ring() :: tuple(). -record(ring, { size :: non_neg_integer(), fixed_ring :: fixed_ring() }). -opaque ring() :: #ring{}. -export_type([ring/0]). %% @doc Given an address and port, get the replica info from that node %% %% Returns the layout of the ring where the given node lives, %% as well as the replica identifier from the cluster. %% -spec grb_replica_info(Address :: node_ip(), Port :: inet:port_number(), LenBits :: non_neg_integer()) -> {ok, inet:ip_address(), term(), ring(), unique_nodes()} | socket_error(). grb_replica_info(Address, Port, LenBits) -> case gen_tcp:connect(Address, Port, ?UTIL_CONN_OPTS) of {error, Reason} -> {error, Reason}; {ok, Sock} -> {ok, {LocalIP, _}} = inet:sockname(Sock), ok = gen_tcp:send(Sock, <<0:LenBits, (ppb_grb_driver:connect())/binary>>), Reply = case gen_tcp:recv(Sock, 0) of {error, Reason} -> {error, Reason}; {ok, <<0:LenBits, RawReply/binary>>} -> {ok, ReplicaID, RingSize, RawRing} = pvc_proto:decode_serv_reply(RawReply), UniqueNodes = unique_ring_nodes(RawRing), FixedRing = make_fixed_ring(RingSize, RawRing), {ok, LocalIP, ReplicaID, #ring{size=RingSize, fixed_ring=FixedRing}, UniqueNodes} end, ok = gen_tcp:close(Sock), Reply end. -spec random_indexnode(ring()) -> index_node(). random_indexnode(#ring{size=Size, fixed_ring=Layout}) -> Pos = rand:uniform(Size - 1), erlang:element(Pos, Layout). -spec get_key_indexnode(ring(), term(), term()) -> index_node(). get_key_indexnode(#ring{size=Size, fixed_ring=Layout}, Key, Bucket) -> Pos = convert_key(Key, Bucket) rem Size + 1, erlang:element(Pos, Layout). -spec size(ring()) -> non_neg_integer(). size(#ring{size=Size}) -> Size. %%==================================================================== %% Routing Internal functions %%==================================================================== -spec convert_key(term(), term()) -> non_neg_integer(). convert_key(Key, Bucket) -> if is_integer(Key) -> convert_key_int(Key); is_binary(Key) -> convert_key_binary(Key, Bucket); is_tuple(Key) -> convert_key(element(1, Key), Bucket); true -> convert_key_hash(Key, Bucket) end. -spec convert_key_int(integer()) -> non_neg_integer(). convert_key_int(Int) -> abs(Int). -spec convert_key_binary(binary(), term()) -> non_neg_integer(). convert_key_binary(Bin, Bucket) -> AsInt = (catch list_to_integer(binary_to_list(Bin))), if is_integer(AsInt) -> convert_key_int(AsInt); true -> convert_key_hash(Bin, Bucket) end. -spec convert_key_hash(term(), term()) -> non_neg_integer(). convert_key_hash(Key, Bucket) -> %% Looked into the internals of riak_core for this HashKey = crypto:hash(sha, term_to_binary({Bucket, Key})), abs(crypto:bytes_to_integer(HashKey)). %%==================================================================== %% Partition Internal functions %%==================================================================== %% @doc Get an unique list of the ring owning IP addresses -spec unique_ring_nodes(raw_ring()) -> unique_nodes(). unique_ring_nodes(Ring) -> ordsets:from_list(lists:foldl(fun({_, Node}, Acc) -> [erlang_node_to_ip(Node) | Acc] end, [], Ring)). %% @doc Get IP address from an erlang node name -spec erlang_node_to_ip(atom()) -> node_ip(). erlang_node_to_ip(Node) -> [_, Ip] = binary:split(atom_to_binary(Node, latin1), <<"@">>), binary_to_atom(Ip, latin1). %% @doc Convert a raw riak ring into a fixed tuple structure -spec make_fixed_ring(non_neg_integer(), raw_ring()) -> fixed_ring(). make_fixed_ring(Size, RawRing) -> erlang:make_tuple(Size, ignore, index_ring(RawRing)). %% @doc Converts a raw Antidote ring into an indexed structure %% %% Adds a 1-based index to each entry, plus converts Erlang %% nodes to an IP address, for easier matching with connection %% sockets %% -spec index_ring( RawRing :: raw_ring() ) -> [{non_neg_integer(), {partition_id(), node_ip()}}]. index_ring(RawRing) -> index_ring(RawRing, 1, []). index_ring([], _, Acc) -> lists:reverse(Acc); index_ring([{Partition, ErlangNode} | Rest], N, Acc) -> Converted = {N, {Partition, erlang_node_to_ip(ErlangNode)}}, index_ring(Rest, N + 1, [Converted | Acc]).
src/pvc_ring.erl
0.507568
0.418994
pvc_ring.erl
starcoder
% Licensed under the Apache License, Version 2.0 (the "License"); you may not % use this file except in compliance with the License. You may obtain a copy of % the License at % % http://www.apache.org/licenses/LICENSE-2.0 % % Unless required by applicable law or agreed to in writing, software % distributed under the License is distributed on an "AS IS" BASIS, WITHOUT % WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the % License for the specific language governing permissions and limitations under % the License. -module(kai_hash_SUITE). -compile(export_all). -include("kai.hrl"). -include("kai_test.hrl"). all() -> [test_hash]. % test_hash_efficiency % Hash ring of test1: % % 0 bucket-0 [3,2,1] -> [3,1,4] % 98,830,170 item-4 % % 536,870,912 bucket-1 [3,2,1] -> [3,1,4] % 953,414,533 item-2 % % 1,073,741,824 bucket-2 [3,2,1] -> [3,1,4] % 1,258,937,939 NODE3(2) % % 1,610,612,736 bucket-3 [2,1,4] -> [1,4,3] % 1,704,004,111 item-3 % 1,981,805,867 item-1 % % 2,147,483,648 bucket-4 [2,1,4] -> [1,4,3] % 2,203,089,259 NODE2(1) % 2,311,136,591 NODE1(2) % 2,365,722,681 NODE4(2) % % 2,684,354,560 bucket-5 [2,4,1] -> [4,1,3] % 2,772,605,746 NODE2(2) % 2,978,498,268 NODE4(1) % % 3,221,225,472 bucket-6 [1,3,2] -> [1,3,4] % 3,495,790,055 NODE1(1) % % 3,758,096,384 bucket-7 [3,2,1] -> [3,1,4] % 4,264,647,116 NODE3(1) test_hash() -> []. test_hash(_Conf) -> kai_config:start_link([ {hostname, "localhost"}, {rpc_port, 11011}, {n, 3}, {number_of_buckets, 8}, {number_of_virtual_nodes, 2} ]), kai_hash:start_link(), {node_info, ?NODE1, Info1} = kai_hash:node_info(), ?assertEqual(?INFO, Info1), {node_info, ?NODE1, Info1} = kai_hash:node_info(?NODE1), ?assertEqual(?INFO, Info1), {node_list, NodeList1} = kai_hash:node_list(), ?assertEqual([?NODE1], NodeList1), {virtual_node_list, VirtualNodeList1} = kai_hash:virtual_node_list(), ?assertEqual( [{2311136591, ?NODE1}, {3495790055, ?NODE1}], VirtualNodeList1 ), {bucket_list, BucketList1} = kai_hash:bucket_list(), ?assertEqual( [{0, [?NODE1]}, {1, [?NODE1]}, {2, [?NODE1]}, {3, [?NODE1]}, {4, [?NODE1]}, {5, [?NODE1]}, {6, [?NODE1]}, {7, [?NODE1]}], BucketList1 ), {buckets, Buckets1} = kai_hash:buckets(), ?assertEqual([0,1,2,3,4,5,6,7], Buckets1), {replaced_buckets, ReplacedBuckets2} = kai_hash:update_nodes([{?NODE2, ?INFO}, {?NODE3, ?INFO}, {?NODE4, ?INFO}], []), ?assertEqual( [{0,3,1}, {1,3,1}, {2,3,1}, {3,2,1}, {4,2,1}, {5,3,1}, {7,3,1}], ReplacedBuckets2 ), {node_list, NodeList2} = kai_hash:node_list(), ?assertEqual(4, length(NodeList2)), ?assert(lists:member(?NODE2, NodeList2)), ?assert(lists:member(?NODE3, NodeList2)), ?assert(lists:member(?NODE4, NodeList2)), {virtual_node_list, VirtualNodeList2} = kai_hash:virtual_node_list(), ?assertEqual( [{1258937939, ?NODE3}, {2203089259, ?NODE2}, {2311136591, ?NODE1}, {2365722681, ?NODE4}, {2772605746, ?NODE2}, {2978498268, ?NODE4}, {3495790055, ?NODE1}, {4264647116, ?NODE3}], VirtualNodeList2 ), {bucket_list, BucketList2} = kai_hash:bucket_list(), ?assertEqual( [{0, [?NODE3, ?NODE2, ?NODE1]}, {1, [?NODE3, ?NODE2, ?NODE1]}, {2, [?NODE3, ?NODE2, ?NODE1]}, {3, [?NODE2, ?NODE1, ?NODE4]}, {4, [?NODE2, ?NODE1, ?NODE4]}, {5, [?NODE2, ?NODE4, ?NODE1]}, {6, [?NODE1, ?NODE3, ?NODE2]}, {7, [?NODE3, ?NODE2, ?NODE1]}], BucketList2 ), {buckets, Buckets2} = kai_hash:buckets(), ?assertEqual([0,1,2,3,4,5,6,7], Buckets2), {bucket, Bucket1} = kai_hash:find_bucket("item-1"), ?assertEqual(3, Bucket1), {replica, Replica1} = kai_hash:find_replica(Bucket1), ?assertEqual(2, Replica1), {nodes, Nodes1} = kai_hash:find_nodes(Bucket1), ?assertEqual([?NODE2, ?NODE1, ?NODE4], Nodes1), {nodes, Nodes2} = kai_hash:find_nodes("item-1"), ?assertEqual([?NODE2, ?NODE1, ?NODE4], Nodes2), {nodes, Nodes3} = kai_hash:find_nodes("item-2"), ?assertEqual([?NODE3, ?NODE2, ?NODE1], Nodes3), {node, Node1} = kai_hash:choose_node_randomly(), ?assertNot(Node1 == ?NODE1), {bucket, Bucket2} = kai_hash:choose_bucket_randomly(), ?assert((Bucket2 >= 0) or (Bucket2 < 8)), % TODO: choose it from my buckets {replaced_buckets, ReplacedBuckets3} = kai_hash:update_nodes([], [?NODE2]), ?assertEqual( [{0,2,3}, {1,2,3}, {2,2,3}, {3,1,2}, {4,1,2}, {5,2,3}, {7,2,3}], ReplacedBuckets3 ), {node_list, NodeList3} = kai_hash:node_list(), ?assertEqual(3, length(NodeList3)), ?assertNot(lists:member(?NODE2, NodeList3)), {virtual_node_list, VirtualNodeList3} = kai_hash:virtual_node_list(), ?assertEqual( [{1258937939, ?NODE3}, {2311136591, ?NODE1}, {2365722681, ?NODE4}, {2978498268, ?NODE4}, {3495790055, ?NODE1}, {4264647116, ?NODE3}], VirtualNodeList3 ), {bucket_list, BucketList3} = kai_hash:bucket_list(), ?assertEqual( [{0, [?NODE3, ?NODE1, ?NODE4]}, {1, [?NODE3, ?NODE1, ?NODE4]}, {2, [?NODE3, ?NODE1, ?NODE4]}, {3, [?NODE1, ?NODE4, ?NODE3]}, {4, [?NODE1, ?NODE4, ?NODE3]}, {5, [?NODE4, ?NODE1, ?NODE3]}, {6, [?NODE1, ?NODE3, ?NODE4]}, {7, [?NODE3, ?NODE1, ?NODE4]}], BucketList3 ), {buckets, Buckets3} = kai_hash:buckets(), ?assertEqual([0,1,2,3,4,5,6,7], Buckets3), {nodes, Nodes4} = kai_hash:find_nodes("item-1"), ?assertEqual([?NODE1, ?NODE4, ?NODE3], Nodes4), {nodes, Nodes5} = kai_hash:find_nodes("item-2"), ?assertEqual([?NODE3, ?NODE1, ?NODE4], Nodes5), kai_hash:stop(), kai_config:stop(). test_hash_efficiency() -> []. test_hash_efficiency(_Conf) -> io:format("simulate network of 64 nodes"), kai_config:start_link([ {hostname, "localhost"}, {rpc_port, 1}, {n, 3}, {number_of_buckets, 16384}, % 16,384 = 128*64*2 {number_of_virtual_nodes, 128} ]), kai_hash:start_link(), Nodes = lists:map( fun(Port) -> {{{127,0,0,1}, Port}, ?INFO} end, lists:seq(2, 63) ), kai_hash:update_nodes(Nodes, []), Args = [[{{{127,0,0,1}, 64}, [{number_of_virtual_nodes, 128}]}], []], {Usec, _} = timer:tc(kai_hash, update, Args), io:format("time to add a node: ~p [usec]", [Usec]), ?assert(Usec < 300000), {Usec2, _} = timer:tc(kai_hash, find, ["item-1", 1]), io:format("time to find a node: ~p [usec]", [Usec2]), ?assert(Usec2 < 1000), {Usec3, _} = timer:tc(kai_hash, choose_node_randomly, []), io:format("time to choose a node randomly: ~p [usec]", [Usec3]), ?assert(Usec3 < 1000), {Usec4, _} = timer:tc(kai_hash, choose_bucket_randomly, []), io:format("time to choose a bucket randomly: ~p [usec]", [Usec4]), ?assert(Usec4 < 300000), {Usec5, _} = timer:tc(kai_hash, update, [[], [{{127,0,0,1}, 1}]]), io:format("time to remove a node: ~p [usec]", [Usec5]), ?assert(Usec5 < 300000), kai_hash:stop(), kai_config:stop().
test/kai_hash_SUITE.erl
0.62498
0.443118
kai_hash_SUITE.erl
starcoder
%% Copyright 2019 Octavo Labs AG Zurich Switzerland (http://octavolabs.com) %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. -module(vmq_storage_engine_dets). -export([open/2, close/1, write/2, read/2, fold/3, fold/4]). -record(state, {ref}). open(DataRoot, _Opts) -> %% Get the data root directory File = filename:join(DataRoot, "bucket.dets"), filelib:ensure_dir(File), case dets:open_file(File, []) of {ok, Ref} -> {ok, #state{ref=Ref}}; Error -> Error end. close(#state{ref=Ref}) -> dets:close(Ref). write(#state{ref=Ref}, WriteOps) -> lists:foreach(fun({put, Key, Val}) -> ok = dets:insert(Ref, {Key, Val}); ({delete, Key}) -> ok = dets:delete(Ref, Key) end, WriteOps). read(#state{ref=Ref}, Key) -> case dets:lookup(Ref, Key) of [{Key, Val}] -> {ok, Val}; [] -> not_found end. fold(#state{ref=Ref}, Fun, Acc) -> % we use a ets table to snapshot (and order) the dets content Tab = ets:new(?MODULE, [ordered_set]), Tab = dets:to_ets(Ref, Tab), fold_iterate(ets:first(Tab), Tab, Fun, Acc). fold(#state{ref=Ref}, Fun, Acc, FirstKey) -> % we use a ets table to snapshot (and order) the dets content Tab = ets:new(?MODULE, [ordered_set]), Tab = dets:to_ets(Ref, Tab), fold_iterate(ets:next(Tab, FirstKey), Tab, Fun, Acc). fold_iterate('$end_of_table', Tab, _Fun, Acc) -> ets:delete(Tab), Acc; fold_iterate(Key, Tab, Fun, Acc0) -> [{Key, Value}] = ets:lookup(Tab, Key), try Fun(Key, Value, Acc0) of Acc1 -> fold_iterate(ets:next(Tab, Key), Tab, Fun, Acc1) catch throw:_Throw -> ets:delete(Tab), Acc0 end.
apps/vmq_generic_msg_store/src/engines/vmq_storage_engine_dets.erl
0.546738
0.429011
vmq_storage_engine_dets.erl
starcoder
%% ------------------------------------------------------------------- %% %% Copyright (c) 2016 <NAME>. All Rights Reserved. %% %% This file is provided to you under the Apache License, %% Version 2.0 (the "License"); you may not use this file %% except in compliance with the License. You may obtain %% a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, %% software distributed under the License is distributed on an %% "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY %% KIND, either express or implied. See the License for the %% specific language governing permissions and limitations %% under the License. %% %% ------------------------------------------------------------------- -module(state_orset_ext). -author("<NAME> <<EMAIL>>"). -define(TYPE, state_orset). -export([intersect/2, map/2, union/2, product/2, filter/2]). -export([ delta_operations/2 ]). union(LValue, RValue) -> state_orset:merge(LValue, RValue). product({?TYPE, LValue}, {?TYPE, RValue}) -> FolderFun = fun({X, XCausality}, {?TYPE, Acc}) -> {?TYPE, Acc ++ [{{X, Y}, causal_product(XCausality, YCausality)} || {Y, YCausality} <- RValue]} end, lists:foldl(FolderFun, new(), LValue). intersect({?TYPE, LValue}, RValue) -> lists:foldl(intersect_folder(RValue), new(), LValue). %% @private intersect_folder({?TYPE, RValue}) -> fun({X, XCausality}, {?TYPE, Acc}) -> Values = case lists:keyfind(X, 1, RValue) of {_Y, YCausality} -> [{X, causal_union(XCausality, YCausality)}]; false -> [] end, {?TYPE, Acc ++ Values} end. map(Function, {?TYPE, V}) -> FolderFun = fun({X, Causality}, {?TYPE, Acc}) -> {?TYPE, Acc ++ [{Function(X), Causality}]} end, lists:foldl(FolderFun, new(), V). filter(Function, {?TYPE, V}) -> FolderFun = fun({X, Causality}, {?TYPE, Acc}) -> case Function(X) of true -> {?TYPE, Acc ++ [{X, Causality}]}; false -> {?TYPE, Acc} end end, lists:foldl(FolderFun, new(), V). %% @private new() -> state_orset:new(). %% @private causal_product(Xs, Ys) -> lists:foldl(fun({X, XActive}, XAcc) -> lists:foldl(fun({Y, YActive}, YAcc) -> [{[X, Y], XActive andalso YActive}] ++ YAcc end, [], Ys) ++ XAcc end, [], Xs). %% @private causal_union(Xs, Ys) -> Xs ++ Ys. %% @private get_operations(Orddict, Value, Token, true) -> case orddict:find(Value, Orddict) of error -> [{add, Value}]; {ok, Tokens} -> case orddict:find(Token, Tokens) of error -> [{add, Value}]; {ok, true} -> []; {ok, false} -> erlang:error("Not a strict inflation") end end; get_operations(Orddict, Value, Token, false) -> case orddict:find(Value, Orddict) of error -> [{add, Value}, {rmv, Value}]; {ok, Tokens} -> case orddict:find(Token, Tokens) of error -> [{add, Value}, {rmv, Value}]; {ok, false} -> []; {ok, true} -> [{rmv, Value}] end end. delta_operations({?TYPE, A}=ORSet1, {?TYPE, B}=ORSet2) -> case state_orset:is_strict_inflation(ORSet1, ORSet2) of true -> orddict:fold(fun(Value, Tokens, Acc1) -> Acc1 ++ orddict:fold(fun(Token, Flag, Acc2) -> Acc2 ++ get_operations(A, Value, Token, Flag) end, [], Tokens) end, [], B); false -> [] end.
src/state_orset_ext.erl
0.541894
0.404272
state_orset_ext.erl
starcoder
%% @author <NAME> <<EMAIL>> %% @copyright 2017 <NAME> %% @doc Vector operation functions. %% %% Vector is a 1-dimensional array, which elements stored sequentially into Erlang binary object of %% some IEEE758 floating point data type: single float, double float, single complex or double complex. %% %% Vector contents uses regular Erlang binary storage and relies on it's garbage collection. In some %% cases it is possible to refer the same binary contents. %% %% Note that vector indexes starts from 0, but not from 1 as for lists. %% @end -module(nvector). -include_lib("eunit/include/eunit.hrl"). -include("erlynum.hrl"). -export([full/2, full/3, zeros/1, zeros/2, ones/1, ones/2]). -export([range/1, range/2, range/3, range/4]). -export([linspace/3, linspace/4, logspace/3, logspace/4, logspace/5, geomspace/3, geomspace/4]). -export([from_list/1, from_list/2, to_list/1, to_list/2]). -export([get/2, get/3]). -export([ copy/1, copy/2, sum/2, sum/3, scale/2, scale/3, axpy/3, axpy/4, dot/3, dot/2, asum/2, asum/1, iamax/1, iamin/1, nrm2/2, nrm2/1, euclidean_norm/1, euclidean_norm/2 ]). -define(WE, erlynum_p:wrap_error). -spec zeros(non_neg_integer()) -> erlynum:nvector(). %% @equiv zeros(Size, [{dtype, d}]) zeros(Size) -> zeros(Size, [{dtype, d}]). -spec zeros(non_neg_integer(), [erlynum:create_option()]) -> erlynum:nvector(). %% @doc Returns a vector of the given size filled by zeroes. zeros(Size, Options) -> full(Size, 0, Options). -spec ones(non_neg_integer()) -> erlynum:nvector(). %% @equiv ones(Size, [{dtype, d}]) ones(Size) -> ones(Size, [{dtype, d}]). -spec ones(non_neg_integer(), [erlynum:create_option()]) -> erlynum:nvector(). %% @doc Returns a vector of the given size filled by ones. ones(Size, Options) -> ?WE(erlynum_nif:nvector_full(Size, 1, Options)). -spec full(non_neg_integer(), erlynum:nscalar()) -> erlynum:nvector(). %% @equiv full(Size, InitialValue, [{dtype, auto}]) full(Size, InitialValue) -> ?WE(erlynum_nif:nvector_full(Size, InitialValue, [{dtype, auto}])). -spec full(non_neg_integer(), erlynum:nscalar(), [erlynum:create_option()]) -> erlynum:nvector(). %% @doc Returns a vector of the given size filled by the specified scalar value `InitialValue'. full(Size, InitialValue, Options) -> ?WE(erlynum_nif:nvector_full(Size, InitialValue, Options)). -spec from_list([erlynum:nscalar()]) -> erlynum:nvector(). %% @equiv from_list(List, [{dtype, auto}]) from_list(List) -> from_list(List, [{dtype, auto}]). -spec from_list([erlynum:nscalar()], [erlynum:create_option()]) -> erlynum:nvector(). %% @doc Returns a vector from the given list items. from_list(List, Options) -> ?WE(erlynum_nif:nvector_from_list(List, Options)). -spec to_list(NVector :: erlynum:nvector()) -> [ erlynum:nscalar() ]. %% @equiv to_list(NVector, noconvert) to_list(NVector) -> to_list(NVector, noconvert). -spec to_list( NVector :: erlynum:nvector(), Convert :: erlynum:convert_option() ) -> [ erlynum:nscalar() ]. %% @doc Returns an Erlang list from the given vector value. %% @param Convert specifies target data type. to_list(NVector, Convert) -> ?WE(erlynum_nif:nvector_to_list(NVector, Convert)). -spec get( NVector :: erlynum:nvector(), Index :: non_neg_integer() ) -> erlynum:nscalar(). %% @equiv get(NVector, Index, noconvert) get(NVector, Index) -> get(NVector, Index, noconvert). -spec get( NVector :: erlynum:nvector(), Index :: non_neg_integer(), Convert :: erlynum:convert_option() ) -> erlynum:nscalar(). %% @doc Returns a scalar element value from the `Vector'. %% Note that `Index' starts from 0, but not 1. %% It is possible to specify `Convert' option to cast scalar value into integer, real of complex value. get(NVector, Index, Convert) -> ?WE(erlynum_nif:nvector_get(NVector, Index, Convert)). -spec range(erlynum:nscalar()) -> erlynum:nvector(). %% @equiv range(0, Stop, 1, []) range(Stop) -> range(0, Stop, 1, []). -spec range( Start :: erlynum:nscalar(), Options_Stop :: [ erlynum:create_option() | erlynum:range_option() ] | erlynum:nscalar() ) -> erlynum:nvector(). %% @doc Returns a vector filled by values withing from `Start' and step `1'. %% %% If last argument is a {@link erlynum:scalar()} value, it specifies range `Stop'. %% If last argument is a list, it specifies create options. %% %% @see range/4 range(Stop, Options) when is_list(Options) -> range(0, Stop, 1, Options); range(Start, Stop) -> range(Start, Stop, 1, []). -spec range( Start :: erlynum:nscalar(), Stop :: erlynum:nscalar(), Step_Options :: [ erlynum:create_option() | erlynum:range_option() ] | erlynum:nscalar() ) -> erlynum:nvector(). %% @doc Returns a vector filled by values withing the given range `[Start, Stop)'. %% %% If last argument is a {@link erlynum:scalar()} value, it specifies increment on next value. %% If last argument is a list, it specifies create options. %% %% @see range/4 range(Start, Stop, Step_Options) when is_list(Step_Options) -> range(Start, Stop, 1, Step_Options); range(Start, Stop, Step_Options) -> range(Start, Stop, Step_Options, []). -spec range( Start :: erlynum:nscalar(), Stop :: erlynum:nscalar(), Step :: erlynum:nscalar(), Options :: [ erlynum:create_option() | erlynum:range_option() ] ) -> erlynum:nvector(). %% @doc Returns a vector filled by values withing the given range `[Start, Stop)' and increment `Step'. range(Start, Stop, Step, Options) -> ?WE(erlynum_nif:nvector_range(Start, Stop, Step, Options)). -spec linspace( Start :: erlynum:nscalar(), Stop :: erlynum:nscalar(), Count :: non_neg_integer() ) -> erlynum:nvector(). %% @equiv linspace(Start, Stop, Count, []) linspace(Start, Stop, Count) -> linspace(Start, Stop, Count, []). -spec linspace( Start :: erlynum:nscalar(), Stop :: erlynum:nscalar(), Count :: non_neg_integer(), Options :: [ erlynum:create_option() | erlynum:range_option() ] ) -> erlynum:nvector(). %% @doc Returns a vector filled by the specified `Count' of values withing the given range `[Start, Stop)'. linspace(Start, Stop, Count, Options) -> ?WE(erlynum_nif:nvector_linspace(Start, Stop, Count, Options)). -spec logspace( Start :: erlynum:nscalar(), Stop :: erlynum:nscalar(), Count :: non_neg_integer() ) -> erlynum:nvector(). %% @equiv logspace(Stat, Stop, Count, 10, []) logspace(Start, Stop, Count) -> logspace(Start, Stop, Count, 10.0). -spec logspace( Start :: erlynum:nscalar(), Stop :: erlynum:nscalar(), Count :: erlynum:nscalar(), Base_Options :: [ erlynum:create_option() | erlynum:range_option() ] | erlynum:nscalar() ) -> erlynum:nvector(). %% @doc Returns a vector filled by the specified `Count' of powers withing the given range `[Start, Stop)'. %% %% If last argument is a {@link erlynum:scalar()} value, it specifies `Base'. %% If last argument is a list, it specifies create options. %% %% @see logspace/5 logspace(Start, Stop, Count, Base_Options) when is_number(Base_Options) or is_tuple(Base_Options) -> logspace(Start, Stop, Count, Base_Options, []); logspace(Start, Stop, Count, Base_Options) when is_list(Base_Options) -> logspace(Start, Stop, Count, 10.0, Base_Options). -spec logspace( Start :: erlynum:nscalar(), Stop :: erlynum:nscalar(), Count :: non_neg_integer(), Base :: erlynum:nscalar(), Options :: [ erlynum:create_option() | erlynum:range_option() ] ) -> erlynum:nvector(). %% @doc Returns a vector filled by the specified `Count' of powers of `Base' withing the given range `[Start, Stop)'. logspace(Start, Stop, Count, Base, Options) -> ?WE(erlynum_nif:nvector_logspace(Start, Stop, Count, Base, Options)). -spec geomspace( Start :: erlynum:nscalar(), Stop :: erlynum:nscalar(), Count :: non_neg_integer() ) -> erlynum:nvector(). %% @equiv geomspace(Start, Stop, Count, []) geomspace(Start, Stop, Count) -> geomspace(Start, Stop, Count, []). -spec geomspace( Start :: erlynum:nscalar(), Stop :: erlynum:nscalar(), Count :: non_neg_integer(), Options :: [ erlynum:create_option() | erlynum:range_option() ] ) -> erlynum:nvector(). %% @doc Returns a vector filled by specified `Count' of geometric progression values. geomspace(Start, Stop, Count, Options) -> ?WE(erlynum_nif:nvector_geomspace(Start, Stop, Count, Options)). -spec copy(X :: erlynum:nvector()) -> erlynum:nvector(). %% @equiv copy(X, []) copy(X) -> copy(X, []). -spec copy(X :: erlynum:nvector(), Options :: [erlynum:create_option()]) -> erlynum:nvector(). %% @doc Returns a packed copy of vector view `Y ← αX'. %% This allow Erlang garbage collector to release referenced binary. %% The function is useless in case of vector view source (some big matrix, for example) %% stored in the same process stack, due to referenced source binary still alive and will not be collected. %% %% Generic cases to use this function are: %% * to pass return value to another process or node if source vector generated from another big array-like object; %% * to convert internal data type while copying. copy(X, Options) -> ?WE(erlynum_nif:nvector_copy(X, Options)). -spec scale( Y :: erlynum:nvector(), Alpha :: erlynum:nscalar() ) -> erlynum:nvector(). %% @equiv scale(Y, Alpha, []) scale(Y, Alpha) -> scale(Y, Alpha, []). -spec scale( Y :: erlynum:nvector(), Alpha :: erlynum:nscalar(), Options :: [ erlynum:create_option() ] ) -> erlynum:nvector(). %% @doc Returns a scaled copy `Y ← αY' of the vector of the same size and the specified data type. scale(Y, Alpha, Options) -> ?WE(erlynum_nif:nvector_scale(Y, Alpha, Options)). -spec sum(Y :: erlynum:nvector(), X :: erlynum:nvector()) -> erlynum:nvector(). %% @equiv sum(Y, X, []) sum(NVectorY, NVectorX) -> axpy(NVectorX, NVectorY, 1). -spec sum( Y :: erlynum:nvector(), X :: erlynum:nvector(), Options :: [ erlynum:create_option() ] ) -> erlynum:nvector(). %% @doc Returns an element-wise sum of two vectors `Y ← X + Y'. %% %% If vector sizes differs, the result is a vector of greater size, leaving %% rest elements unchanged. %% %% If data types differs, the `Y' vector (first argument) data type is used until options provided. sum(Y, X, Options) -> axpy(X, Y, 1, Options). -spec axpy( Y :: erlynum:nvector(), X :: erlynum:nvector(), Alpha :: erlynum:nscalar() ) -> erlynum:nvector(). %% @equiv axpy(Y, X, Alpha, []) axpy(Y, X, Alpha) -> axpy(Y, X, Alpha, []). -spec axpy( Y :: erlynum:nvector(), X :: erlynum:nvector(), Alpha :: erlynum:nscalar(), Options :: [ erlynum:create_option() ] ) -> erlynum:nvector(). %% @doc Returns a vector, each element calculated in form `Y ← αX + Y' axpy(Y, X, Alpha, Options) -> ?WE(erlynum_nif:nvector_axpy(Y, X, Alpha, Options)). -spec dot( Y :: erlynum:nvector(), X :: erlynum:nvector() ) -> erlynum:nscalar(). %% @equiv dot(Y, X, []) dot(Y, X) -> dot(Y, X, []). -spec dot( Y :: erlynum:nvector(), X :: erlynum:nvector(), Options :: [ erlynum:create_option() | {conjuated, boolean()}] ) -> erlynum:nscalar(). %% @doc Returns the inner product of two vectors. dot(Y, X, Options) -> ?WE(erlynum_nif:nvector_dot(Y, X, Options)). -spec asum(X :: erlynum:nvector()) -> erlynum:nscalar(). %% @equiv asum(X, []) asum(X) -> asum(X, []). -spec asum( X :: erlynum:nvector(), Options :: [ erlynum:create_option() ] ) -> erlynum:nscalar(). %% @doc Returns the sum of magnitudes of elements of a real vector, or the sum of magnitudes %% of the real and imaginary parts of elements of a complex vector. asum(X, Options) -> ?WE(erlynum_nif:nvector_asum(X, Options)). -spec iamax(X :: erlynum:nvector()) -> non_neg_integer() | undefined. %% @doc Returns the lowest index of vector element that has the largest absolute value. iamax(X) -> ?WE(erlynum_nif:nvector_iamax_iamin(X, iamax)). -spec iamin(X :: erlynum:nvector()) -> non_neg_integer() | undefined. %% @doc Returns the lowest index of vector element that has the lowest absolute value. iamin(X) -> ?WE(erlynum_nif:nvector_iamax_iamin(X, iamin)). -spec nrm2(X :: erlynum:nvector()) -> erlynum:nscalar(). %% @equiv nrm2(X, []) nrm2(X) -> nrm2(X, []). -spec nrm2( X :: erlynum:nvector(), Options :: [ erlynum:create_option() ] ) -> erlynum:nscalar(). %% @doc Returns the Euclidean norm of vector. %% %% The function is named `nrm2' to match corresponding function name in BLAST. %% It is better to use `euclidean_norm' function for better readability. nrm2(X, Options) -> ?WE(erlynum_nif:nvector_nrm2(X, Options)). -spec euclidean_norm(X :: erlynum:nvector()) -> erlynum:nscalar(). %% @doc equiv nrm2(X, []) euclidean_norm(X) -> nrm2(X). -spec euclidean_norm( X :: erlynum:nvector(), Options :: [ erlynum:create_option() ] ) -> erlynum:nscalar(). %% @equiv nrm2(X, Options) euclidean_norm(X, Options) -> nrm2(X, Options).
src/nvector.erl
0.733643
0.517876
nvector.erl
starcoder
%% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% %% INSTRUCTION SCHEDULER %% %% This is a basic ILP cycle scheduler: %% * set cycle = 0 %% * while ready[cycle] nonempty do %% - take x with greatest priority from ready[cycle] %% - try to schedule x; %% * if scheduling x was possible, %% - reserve resources %% - add x to schedule and delete x from dag %% - update earliest-time for all successor nodes %% as max[earliest[y],cycle+latency[x]] %% - if some node y now has no predecessors, %% add y to ready[earliest[y]] %% * if it was impossible, put x in ready[cycle+1] %% (= try again) %% %% We use the following data structures: %% 1. all nodes are numbered and indices used as array keys %% 2. priority per node can be computed statically or dynamically %% * statically: before scheduling, each node gets a priority value %% * dynamically: at each cycle, compute priorities for all ready nodes %% 3. earliest: earliest cycle of issue, starts at 0 %% and is updated as predecessors issue %% 4. predecessors: number of predecessors (0 = ready to issue) %% 5. successors: list of {Latency,NodeID} %% 6. ready: an array indexed by cycle-time (integer), where %% ready nodes are kept. %% 7. resources: a resource representation (ADT) that answers %% certain queries, e.g., "can x be scheduled this cycle" %% and "reserve resources for x". %% 8. schedule: list of scheduled instructions {Instr,Cycle} %% in the order of issue %% 9. instructions: maps IDs back to instructions %% %% Inputs: %% - a list of {ID,Node} pairs (where ID is a unique key) %% - a dependence list {ID0,Latency,ID1}, which is used to %% build the DAG. %% %% Note that there is some leeway in how things are represented %% from here. %% %% MODIFICATIONS: %% - Some basic blocks are not worth scheduling (e.g., GC save/restore code) %% yet are pretty voluminous. How do we skip them? %% - Scheduling should be done at finalization time: when basic block is %% linearized and is definitely at Sparc assembly level, THEN reorder %% stuff. -module(hipe_schedule). -export([cfg/1, est_cfg/1, delete_node/5]). -include("../sparc/hipe_sparc.hrl"). %%-define(debug1,true). -define(debug2(Str,Args),ok). %%-define(debug2(Str,Args),io:format(Str,Args)). -define(debug3(Str,Args),ok). %%-define(debug3(Str,Args),io:format(Str,Args)). -define(debug4(Str,Args),ok). %%-define(debug4(Str,Args),io:format(Str,Args)). -define(debug5(Str,Args),ok). %%-define(debug5(Str,Args),io:format(Str,Args)). -define(debug(Str,Args),ok). %%-define(debug(Str,Args),io:format(Str,Args)). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : cfg %% Argument : CFG - the control flow graph %% Returns : CFG - A new cfg with scheduled blocks %% Description : Takes each basic block and schedules them one by one. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% cfg(CFG) -> ?debug3("CFG: ~n~p", [CFG]), update_all( [ {L, hipe_bb:mk_bb( block(L,hipe_bb:code(hipe_sparc_cfg:bb(CFG,L))) )} || L <- hipe_sparc_cfg:labels(CFG) ], CFG). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : update_all %% Argument : Blocks - [{Label, Block}] , a list with labels and new code %% used for updating the old CFG. %% CFG - The old controlflow graph %% Returns : An updated controlflow graph. %% Description : Just swappes the basic blocks in the CFG to the scheduled one. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% update_all([],CFG) -> CFG; update_all([{L,NewB}|Ls],CFG) -> update_all(Ls,hipe_sparc_cfg:bb_add(CFG,L,NewB)). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% est_cfg(CFG) -> update_all([ {L, hipe_bb:mk_bb(est_block(hipe_bb:code(hipe_sparc_cfg:bb(CFG,L))))} || L <- hipe_sparc_cfg:labels(CFG) ], CFG). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% %% Provides an estimation of how quickly a block will execute. %% This is done by chaining all instructions in sequential order %% by 0-cycle dependences (which means they will never be reordered), %% then scheduling the mess. est_block([]) -> []; est_block([I]) -> [I]; est_block(Blk) -> {IxBlk,DAG} = est_deps(Blk), Sch = bb(IxBlk,DAG), separate_block(Sch,IxBlk). est_deps(Blk) -> IxBlk = indexed_bb(Blk), DAG = deps(IxBlk), {IxBlk, chain_instrs(IxBlk,DAG)}. chain_instrs([{N,_}|Xs],DAG) -> chain_i(N,Xs,DAG). chain_i(_,[],DAG) -> DAG; chain_i(N,[{M,_}|Xs],DAG) -> NewDAG = dep_arc(N,zero_latency(),M,DAG), chain_i(M,Xs,NewDAG). zero_latency() -> 0. lookup_instr([{N,I}|_], N) -> I; lookup_instr([_|Xs], N) -> lookup_instr(Xs, N). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : block %% Argument : Instrs - [Instr], list of all the instructions in a basic %% block. %% Returns : A new scheduled block %% Description : Schedule a basic block %% %% Note: does not consider delay slots! %% (another argument for using only annulled delay slots?) %% * how do we add delay slots? somewhat tricky to %% reconcile with the sort of scheduling we consider. %% (as-early-as-possible) %% => rewrite scheduler into as-late-as-possible? %% (=> just reverse the dependence arcs??) %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Don't fire up the scheduler if there's no work to do. block(_, []) -> []; block(_L, [I]) -> case hipe_sparc:is_any_branch(I) of true -> [hipe_sparc:nop_create(), I]; false -> [I] end; block(_L, Blk) -> IxBlk = indexed_bb(Blk), case IxBlk of [{_N, I}] -> % comments and nops may have been removed. case hipe_sparc:is_any_branch(I) of true -> [hipe_sparc:nop_create(), I]; false -> [I] end; _ -> Sch = bb(IxBlk, {DAG, _Preds} = deps(IxBlk)), {NewSch, NewIxBlk} = fill_delays(Sch, IxBlk, DAG), X = finalize_block(NewSch, NewIxBlk), debug1_stuff(Blk, DAG, IxBlk, Sch, X), X end. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : fill_delays %% Argument : Sch - List of {{cycle, C}, {node, N}} : C = current cycle %% N = node index %% IxBlk - Indexed block [{N, Instr}] %% DAG - Dependence graph %% Returns : {NewSch, NewIxBlk} - vector with new schedule and vector %% with {N, Instr} %% Description : Goes through the schedule from back to front looking for %% branches/jumps. If one is found fill_del tries to find %% an instr to fill the delayslot. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% fill_delays(Sch, IxBlk, DAG) -> NewIxBlk = hipe_vectors:list_to_vector(IxBlk), %% NewSch = hipe_vectors:list_to_vector(Sch), NewSch = fill_del(length(Sch), hipe_vectors:list_to_vector(Sch), NewIxBlk, DAG), {NewSch, NewIxBlk}. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : fill_del %% Argument : N - current index in the schedule %% Sch - schedule %% IxBlk - indexed block %% DAG - dependence graph %% Returns : Sch - New schedule with possibly a delay instr in the last %% position. %% Description : If a call/jump is found fill_branch_delay/fill_call_delay %% is called to find a delay-filler. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% fill_del(N, Sch, _IxBlk, _DAG) when N < 1 -> Sch; fill_del(N, Sch, IxBlk, DAG) -> Index = get_index(Sch, N), ?debug2("Index for ~p: ~p~nInstr: ~p~n", [N, Index, get_instr(IxBlk, Index)]), NewSch = case get_instr(IxBlk, Index) of #call_link{} -> fill_branch_delay(N - 1, N, Sch, IxBlk, DAG); #jmp_link{} -> fill_call_delay(N - 1, N, Sch, IxBlk, DAG); #jmp{} -> fill_call_delay(N - 1, N, Sch, IxBlk, DAG); #b{} -> fill_branch_delay(N - 1, N, Sch, IxBlk, DAG); #br{} -> fill_branch_delay(N - 1, N, Sch, IxBlk, DAG); #goto{} -> fill_branch_delay(N - 1, N, Sch, IxBlk, DAG); _Other -> Sch end, NewSch. %% fill_del(N - 1, NewSch, IxBlk, DAG). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : fill_call_delay %% Argument : Cand - index in schedule of delay-candidate %% Call - index in schedule of call %% Sch - schedule vector: < {{cycle,Ci},{node,Nj}}, ... > %% IxBlk - block vector: < {N, Instr1}, {N+1, Instr2} ... > %% DAG - dependence graph %% Returns : Sch - new updated schedule. %% Description : Searches backwards through the schedule trying to find an %% instr without conflicts with the Call-instr. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% fill_call_delay(Cand, _Call, Sch, _IxBlk, _DAG) when Cand < 1 -> Sch; fill_call_delay(Cand, Call, Sch, IxBlk, DAG) -> CandIndex = get_index(Sch, Cand), CallIndex = get_index(Sch, Call), CandI = get_instr(IxBlk, CandIndex), case move_or_alu(CandI) of true -> case single_depend(CandIndex, CallIndex, DAG) of false -> % Other instrs depends on Cand ... fill_call_delay(Cand - 1, Call, Sch, IxBlk, DAG); true -> CallI = get_instr(IxBlk, CallIndex), CandDefs = ordsets:from_list(hipe_sparc:defines(CandI)), %% CandUses = ordsets:from_list(hipe_sparc:uses(CandI)), %% CallDefs = ordsets:from_list(hipe_sparc:defines(CallI)), CallUses = ordsets:from_list(hipe_sparc:uses(CallI)), Args = case CallI of #jmp_link{} -> ordsets:from_list( hipe_sparc:jmp_link_args(CallI)); #jmp{} -> ordsets:from_list(hipe_sparc:jmp_args(CallI)); #call_link{} -> ordsets:from_list( hipe_sparc:call_link_args(CallI)) end, CallUses2 = ordsets:subtract(CallUses, Args), Conflict = ordsets:intersection(CandDefs, CallUses2), %% io:format("single_depend -> true:~n ~p~n, ~p~n,~p~n",[CandI,CallI,DAG]), %% io:format("Cand = ~p~nCall = ~p~n",[CandI,CallI]), %% io:format("CandDefs = ~p~nCallDefs = ~p~n",[CandDefs,CallDefs]), %% io:format("CandUses = ~p~nCallUses = ~p~n",[CandUses,CallUses]), %% io:format("Args = ~p~nCallUses2 = ~p~n",[Args,CallUses2]), %% io:format("Conflict = ~p~n",[Conflict]), case Conflict of [] -> % No conflicts ==> Cand can fill delayslot after Call update_schedule(Cand, Call, Sch); _ -> % Conflict: try with preceeding instrs fill_call_delay(Cand - 1, Call, Sch, IxBlk, DAG) end end; false -> fill_call_delay(Cand - 1, Call, Sch, IxBlk, DAG) end. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : fill_branch_delay %% Argument : Cand - index in schedule of delay-candidate %% Branch - index in schedule of branch %% Sch - schedule %% IxBlk - indexed block %% DAG - dependence graph %% Returns : Sch - new updated schedule. %% Description : Searches backwards through the schedule trying to find an %% instr without conflicts with the Branch-instr. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% fill_branch_delay(Cand, _Br, Sch, _IxBlk, _DAG) when Cand < 1 -> Sch; fill_branch_delay(Cand, Br, Sch, IxBlk, DAG) -> CandIndex = get_index(Sch, Cand), BrIndex = get_index(Sch, Br), CandI = get_instr(IxBlk, CandIndex), case move_or_alu(CandI) of true -> case single_depend(CandIndex, BrIndex, DAG) of false -> % Other instrs depends on Cand ... fill_branch_delay(Cand - 1, Br, Sch, IxBlk, DAG); true -> BrI = get_instr(IxBlk, BrIndex), CandDefs = ordsets:from_list(hipe_sparc:defines(CandI)), %% CandUses = ordsets:from_list(hipe_sparc:uses(CandI)), %% BrDefs = ordsets:from_list(hipe_sparc:defines(BrI)), BrUses = ordsets:from_list(hipe_sparc:uses(BrI)), Conflict = ordsets:intersection(CandDefs, BrUses), %% io:format("single_depend -> true: ~p~n, ~p~n,~p~n", [CandI, BrI, DAG]), %% io:format("Cand = ~p~nBr = ~p~n",[CandI,BrI]), %% io:format("CandDefs = ~p~nBrDefs = ~p~n",[CandDefs,BrDefs]), %% io:format("CandUses = ~p~nBrUses = ~p~n",[CandUses,BrUses]), %% io:format("Conflict = ~p~n",[Conflict]); case Conflict of [] -> % No conflicts ==> % Cand can fill delayslot after Branch update_schedule(Cand, Br, Sch); _ -> % Conflict: try with preceeding instrs fill_branch_delay(Cand - 1, Br, Sch, IxBlk, DAG) end end; false -> fill_branch_delay(Cand - 1, Br, Sch, IxBlk, DAG) end. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : update_schedule %% Argument : From - the position from where to switch indexes in Sch %% To - the position to where to switch indexes in Sch %% Sch - schedule %% Returns : Sch - an updated schedule %% Description : If From is the delay-filler and To is the Call/jump, the %% schedule is updated so From gets index To, To gets index %% To - 1, and the nodes between From and To gets old_index - 1. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% update_schedule(To, To, Sch) -> {{cycle, C}, {node, _N} = Node} = hipe_vectors:get(Sch, To-1), hipe_vectors:set(Sch, To-1, {{cycle, C+1}, Node}); update_schedule(From, To, Sch) -> Temp = hipe_vectors:get(Sch, From-1), Sch1 = hipe_vectors:set(Sch, From-1, hipe_vectors:get(Sch, From)), update_schedule(From + 1, To, hipe_vectors:set(Sch1, From, Temp)). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : single_depend %% Argument : N - Index of the delayslot candidate %% M - Index of the node that N possibly has a single %% depend to. %% DAG - The dependence graph %% Returns : true if no other nodes than N os depending on N %% Description : Checks that no other nodes than M depends on N and that the %% latency between them is zero or 1. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% single_depend(N, M, DAG) -> Deps = hipe_vectors:get(DAG, N-1), single_depend(M, Deps). single_depend(_N, []) -> true; single_depend(N, [{0, N}]) -> true; single_depend(N, [{1, N}]) -> true; single_depend(_N, [{_Lat, _}|_]) -> false. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : get_index %% Argument : Sch - schedule %% N - index in schedule %% Returns : Index - index of the node %% Description : Returns the index of the node on position N in the schedule. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% get_index(Sch, N) -> {{cycle, _C}, {node, Index}} = hipe_vectors:get(Sch,N-1), Index. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : get_instr %% Argument : IxBlk - indexed block %% N - index in block %% Returns : Instr %% Description : Returns the instr on position N in the indexed block. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% get_instr(IxBlk, N) -> {_, Instr} = hipe_vectors:get(IxBlk, N-1), Instr. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : get_instr %% Argument : Sch - schedule %% IxBlk - indexed block %% N - index in schedule %% Returns : Instr %% Description : Returns the instr on position N in the schedule. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% get_instr(Sch, IxBlk, N) -> {{cycle, _C}, {node, Index}} = hipe_vectors:get(Sch, N-1), {_, Instr} = hipe_vectors:get(IxBlk, Index-1), Instr. separate_block(Sch,IxBlk) -> sep_comments([{C,lookup_instr(IxBlk,N)} || {{cycle,C},{node,N}} <- Sch]). sep_comments([]) -> []; sep_comments([{C,I}|Xs]) -> [hipe_sparc:comment_create({cycle,C}), I | sep_comments(Xs,C)]. sep_comments([], _) -> []; sep_comments([{C1,I}|Xs], C0) -> if C1 > C0 -> [hipe_sparc:comment_create({cycle,C1}),I|sep_comments(Xs,C1)]; true -> [I|sep_comments(Xs, C0)] end. finalize_block(Sch, IxBlk) -> ?debug5("Sch: ~p~nIxBlk: ~p~n",[Sch,IxBlk]), finalize_block(1, hipe_vectors:size(Sch), 1, Sch, IxBlk, []). finalize_block(N, End, _C, Sch, IxBlk, _Instrs) when N =:= End - 1 -> NextLast = get_instr(Sch, IxBlk, N), Last = get_instr(Sch, IxBlk, End), ?debug5("NextLast: ~p~nLast: ~p~n",[NextLast,Last]), case hipe_sparc:is_any_branch(Last) of true -> % Couldn't fill delayslot ==> add NOP [NextLast , hipe_sparc:nop_create(), Last]; false -> % Last is a delayslot-filler ==> change order... [Last, NextLast] end; finalize_block(N, End, C0, Sch, IxBlk, Instrs) -> {{cycle, _C1}, {node, _M}} = hipe_vectors:get(Sch, N-1), Instr = get_instr(Sch, IxBlk, N), ?debug5("Instr: ~p~n~n",[Instr]), [Instr | finalize_block(N + 1, End, C0, Sch, IxBlk, Instrs)]. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : bb %% Argument : IxBlk - indexed block %% DAG - {Dag, Preds} where Dag is dependence graph and %% Preds is number of predecessors for each node. %% Returns : Sch %% Description : Initializes earliest-list, ready-list, priorities, resources %% and so on, and calls the cycle_sched which does the scheduling %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% bb(IxBlk,DAG) -> bb(length(IxBlk), IxBlk, DAG). bb(N,IxBlk,{DAG, Preds}) -> Earliest = init_earliest(N), BigArray = N*10, % "nothing" is this big :-) Ready = hipe_schedule_prio:init_ready(BigArray,Preds), I_res = init_instr_resources(N, IxBlk), Prio = hipe_schedule_prio:init_instr_prio(N,DAG), Rsrc = init_resources(BigArray), ?debug4("I_res: ~n~p~nPrio: ~n~p~nRsrc: ~n~p~n", [I_res,Prio,Rsrc]), ?debug('cycle 1~n',[]), Sch = empty_schedule(), cycle_sched(1,Ready,DAG,Preds,Earliest,Rsrc,I_res,Prio,Sch,N,IxBlk). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : cycle_sched %% Argument : - C is current cycle, 1 or more. %% - Ready is an array (Cycle -> [Node]) %% yielding the collection of nodes ready to be %% scheduled in a cycle. %% - DAG is an array (Instr -> [{Latency,Instr}]) %% represents the dependence DAG. %% - Preds is an array (Instr -> NumPreds) %% counts the number of predecessors %% (0 preds = ready to be scheduled). %% - Earl is an array (Instr -> EarliestCycle) %% holds the earliest cycle an instruction can be scheduled. %% - Rsrc is a 'resource ADT' that handles scheduler resource %% management checks whether instruction can be scheduled %% this cycle without a stall. %% - I_res is an array (Instr -> Required_resources) %% holds the resources required to schedule an instruction. %% - Sch is the representation of the schedule current schedule. %% - N is the number of nodes remaining to be scheduled %% tells us when to stop the scheduler. %% - IxBlk is the indexed block with instrs %% Returns : present schedule %% Description : Scheduler main loop. %% Pick next ready node in priority order for cycle C until %% none remain. %% * check each node if it can be scheduled w/o stalling %% * if so, schedule it %% * otherwise, bump the node to the next cycle %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% cycle_sched(C,Ready,DAG,Preds,Earl,Rsrc,I_res,Prio,Sch,N,IxBlk) -> case hipe_schedule_prio:next_ready(C,Ready,Prio,IxBlk,DAG,Preds,Earl) of % case hipe_schedule_prio:next_ready(C,Ready,Prio,IxBlk) of {next,I,Ready1} -> ?debug('try ~p~n==> ready = ~p~n',[I, Ready1]), case resources_available(C,I,Rsrc,I_res) of {yes,NewRsrc} -> ?debug(' scheduled~n==> Rscrs = ~p~n',[NewRsrc]), NewSch = add_to_schedule(I,C,Sch), {ReadyNs,NewDAG,NewPreds,NewEarl} = delete_node(C,I,DAG,Preds,Earl), ?debug("NewPreds : ~p~n",[Preds]), ?debug(' ReadyNs: ~p~n',[ReadyNs]), NewReady = hipe_schedule_prio:add_ready_nodes(ReadyNs, Ready1), ?debug(' New ready: ~p~n',[NewReady]), cycle_sched(C,NewReady,NewDAG,NewPreds,NewEarl, NewRsrc,I_res,Prio,NewSch,N-1, IxBlk); no -> ?debug(' resource conflict~n',[]), NewReady = hipe_schedule_prio:insert_node(C+1,I,Ready1), cycle_sched(C,NewReady,DAG,Preds,Earl,Rsrc, I_res,Prio,Sch,N,IxBlk) end; none -> % schedule next cycle if some node remains if N > 0 -> ?debug('cycle ~p~n',[C+1]), cycle_sched(C+1,Ready,DAG,Preds,Earl, advance_cycle(Rsrc), I_res,Prio,Sch,N, IxBlk); true -> present_schedule(Sch) end end. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : init_earliest %% Argument : N - number of instrs %% Returns : %% Description : %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% init_earliest(N) -> hipe_vectors:new(N,1). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% %% Schedule is kept reversed until the end. -define(present_node(I,Cycle),{{cycle,Cycle},{node,I}}). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : empty_schedule %% Description : Returns an empty schedule. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% empty_schedule() -> []. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : add_to_schedule %% Argument : I - instr %% Cycle - cycle when I was placed %% Sch - schedule %% Description : Adds instr to schedule %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% add_to_schedule(I,Cycle,Sch) -> [?present_node(I,Cycle)|Sch]. present_schedule(Sch) -> lists:reverse(Sch). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% %% Interface to resource manager: %% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : init_resources %% Description : Yields a 'big enough' array mapping (Cycle -> Resources); %% this array is called Rsrc below. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% init_resources(S) -> hipe_target_machine:init_resources(S). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : init_instr_resources %% Argument : Nodes - a list of the instructions %% N - is the number of nodes %% Description : return a vector (NodeID -> Resource_requirements) %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% init_instr_resources(N,Nodes) -> hipe_target_machine:init_instr_resources(N,Nodes). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : resources_available %% Argument : Cycle - the current cycle %% I - the current instruction (index = NodeID) %% Rsrc - a map (Cycle -> Resources) %% I_res - maps (NodeID -> Resource_requirements) %% Description : returns {yes,NewResTab} | no %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% resources_available(Cycle,I,Rsrc,I_res) -> hipe_target_machine:resources_available(Cycle,I,Rsrc,I_res). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : advance_cycle %% Argument : Rsrc - resources %% Description : Returns an empty resources-state %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% advance_cycle(Rsrc) -> hipe_target_machine:advance_cycle(Rsrc). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : delete_node %% Argument : Cycle - current cycle %% I - index of instr %% DAG - dependence dag %% Preds - array with number of predecessors for nodes %% Earl - array with earliest-times for nodes %% Returns : {ReadyNs,NewDAG,NewPreds,NewEarl} %% Description : Deletes node I and updates earliest times for the rest. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% delete_node(Cycle,I,DAG,Preds,Earl) -> Succ = hipe_vectors:get(DAG,I-1), NewDAG = hipe_vectors:set(DAG,I-1,scheduled), % provides debug 'support' {ReadyNs,NewPreds,NewEarl} = update_earliest(Succ,Cycle,Preds,Earl,[]), ?debug('earliest after ~p: ~p~n',[I,[{Ix+1,V} || {Ix,V} <- hipe_vectors:list(NewEarl)]]), {ReadyNs,NewDAG,NewPreds,NewEarl}. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : update_earliest %% Argument : Succ - successor list %% Cycle - current cycle %% Preds - predecessors %% Earl - earliest times for nodes %% Ready - array with readynodes for cycles %% Returns : {Ready,Preds,Earl} %% Description : Updates the earliest times for nodes and updates number of %% predecessors for nodes %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% update_earliest([],_Cycle,Preds,Earl,Ready) -> {Ready,Preds,Earl}; update_earliest([{Lat,N}|Xs],Cycle,Preds,Earl,Ready) -> Old_earl = hipe_vectors:get(Earl,N-1), New_earl = erlang:max(Old_earl,Cycle+Lat), NewEarl = hipe_vectors:set(Earl,N-1,New_earl), Num_preds = hipe_vectors:get(Preds,N-1), NewPreds = hipe_vectors:set(Preds,N-1,Num_preds-1), if Num_preds =:= 0 -> ?debug('inconsistent DAG~n',[]), exit({update_earliest,N}); Num_preds =:= 1 -> NewReady = [{New_earl,N}|Ready], NewPreds2 = hipe_vectors:set(NewPreds,N-1,0), update_earliest(Xs,Cycle,NewPreds2,NewEarl,NewReady); is_integer(Num_preds), Num_preds > 1 -> update_earliest(Xs,Cycle,NewPreds,NewEarl,Ready) end. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% %% Collect instruction dependences. %% %% Three forms: %% - data/register %% * insert RAW, WAR, WAW dependences %% - memory %% * stores serialize memory references %% * alias analysis may allow loads to bypass stores %% - control %% * unsafe operations are 'trapped' between branches %% * branches are ordered %% %% returns { [{Index,Instr}], DepDAG } %% DepDAG is defined below. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : deps %% Argument : BB - Basic block %% Returns : {IxBB,DAG} - indexed block and dependence graph. DAG consists %% of both Dag and Preds, where Preds is number %% of predecessors for nodes. %% Description : Collect instruction dependences. %% %% Three forms: %% - data/register %% * insert RAW, WAR, WAW dependences %% - memory %% * stores serialize memory references %% * alias analysis may allow loads to bypass stores %% - control %% * unsafe operations are 'trapped' between branches %% * branches are ordered %% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% deps(IxBB) -> N = length(IxBB), DAG = empty_dag(N), % The DAG contains both dependence-arcs and % number of predeccessors... {_DepTab,DAG1} = dd(IxBB, DAG), DAG2 = md(IxBB, DAG1), cd(IxBB, DAG2). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : empty_dag %% Argument : N - number of nodes %% Returns : empty DAG %% Description : DAG consists of dependence graph and predeccessors %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% empty_dag(N) -> {hipe_vectors:new(N, []), hipe_vectors:new(N, 0)}. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : indexed_bb %% Argument : BB - basic block %% Returns : [{N, Instr}] %% Description : Puts indexes to all instrs of a block, removes comments. %% NOP's are also removed because if both sparc_schedule and %% sparc_post_schedule options are used, the first pass will %% add nop's before the branch if necessary, and these are %% removed before scheduling the second pass. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% indexed_bb(BB) -> indexed_bb(BB,1). indexed_bb([],_N) -> []; indexed_bb([X|Xs],N) -> case X of #comment{} -> indexed_bb(Xs,N); #nop{} -> indexed_bb(Xs,N); _Other -> [{N,X}|indexed_bb(Xs,N+1)] end. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : dep_arc %% Argument : N - Current node %% Lat - Latency from current node to M %% M - The dependent node %% DAG - The dependence graph. Consists of both DAG and %% predeccessors %% Returns : A new DAG with the arc added and number of predeccessors for %% M increased. %% Description : Adds a new arc to the graph, if an older arc goes from N to M %% it will be replaced with a new arc {max(OldLat, NewLat), M}. %% Number of predeccessors for node M is increased. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% dep_arc(N, Lat, M, {Dag,Preds}) -> OldDeps = hipe_vectors:get(Dag, N-1), %% io:format("{OldDeps} = {~p}~n",[OldDeps]), {NewDeps, Status} = add_arc(Lat, M, OldDeps), %% io:format("{NewDeps, Status} = {~p, ~p}~n",[NewDeps, Status]), NewDag = hipe_vectors:set(Dag, N-1, NewDeps), NewPreds = case Status of added -> % just increase preds if new arc was added OldPreds = hipe_vectors:get(Preds, M-1), hipe_vectors:set(Preds, M-1, OldPreds + 1); non_added -> Preds end, {NewDag, NewPreds}. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : add_arc %% Argument : Lat - The latency from current node to To. %% To - The instr-id of the node which the dependence goes to %% Arcs - The dependecies that are already in the dep-graph %% Returns : A dependence graph sorted by To. %% Description : A new arc that is added is sorted in the right place, and if %% there is already an arc between nodes A and B, the one with %% the greatest latency is chosen. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% add_arc(Lat,To, []) -> {[{Lat, To}], added}; add_arc(Lat1, To, [{Lat2, To} | Arcs]) -> {[{erlang:max(Lat1, Lat2), To} | Arcs], non_added}; add_arc(Lat1,To1, [{Lat2, To2} | Arcs]) when To1 < To2 -> {[{Lat1, To1}, {Lat2, To2} | Arcs], added}; add_arc(Lat1 ,To1, [{Lat2, To2} | Arcs]) -> {Arcs1, Status} = add_arc(Lat1, To1, Arcs), {[{Lat2, To2} | Arcs1], Status}. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% %% The register/data dependence DAG of a block is represented %% as a mapping (Variable -> {NextWriter,NextReaders}) %% where NextWriter is a pair {Ix,Type} %% and NextReaders is a list of pairs {Ix,Type}. %% %% Type is used to determine latencies of operations; on the UltraSparc, %% latencies of arcs (n -> m) are determined by both n and m. (E.g., if %% n is an integer op and m is a store, then latency is 0; if m is an %% integer op, it's 1.) dd([],DAG) -> { empty_deptab(), DAG }; dd([{N,I}|Is],DAG0) -> {DepTab,DAG1} = dd(Is,DAG0), add_deps(N,I,DepTab,DAG1). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : add_deps %% Argument : N - current node %% Instr - current instr %% DepTab - hashtable with {next-writer, next-readers} for reg %% DAG - dependence graph %% Returns : {DepTab, BlockInfo, DAG} - with new values %% Description : Adds dependencies for node N to the graph. The registers that %% node N defines and uses are used for computing the %% dependencies to the following nodes. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% add_deps(N,Instr,DepTab,DAG) -> {Ds,Us} = def_use(Instr), Type = dd_type(Instr), {DepTab1,DAG1} = add_write_deps(Ds,N,Type,DepTab,DAG), add_read_deps(Us,N,Type,DepTab1,DAG1). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Instructions are classified into symbolic categories, %% which are subsequently used to determine operation latencies %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% dd_type(Instr) -> case Instr of #b{} -> branch; %% #br{} -> branch; #call_link{} -> branch; #jmp_link{} -> branch; #jmp{} -> branch; #goto{} -> branch; #load{} -> load; #store{} -> store; #alu{} -> alu; #move{} -> alu; #multimove{} -> Src = hipe_sparc:multimove_src(Instr), Lat = round(length(Src)/2), {mmove,Lat}; #sethi{} -> alu; #alu_cc{} -> alu_cc; %% #cmov_cc{} -> cmov_cc; %% #cmov_r{} -> alu; #load_atom{} -> alu; #load_address{} -> alu; #pseudo_enter{} -> pseudo; #pseudo_pop{} -> pseudo; #pseudo_return{} -> pseudo; #pseudo_spill{} -> pseudo; #pseudo_unspill{} -> pseudo end. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : add_write_deps %% Argument : Defs - registers that node N defines. %% N - current node %% Ty - the type of current instr %% DepTab - Dependence-table %% DAG - The dependence graph. %% Returns : {DepTab,DAG} - with new values %% Description : Adds dependencies to the graph for nodes that depends on the %% registers that N defines. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% add_write_deps([],_N,_Ty,DepTab,DAG) -> {DepTab,DAG}; add_write_deps([D|Ds],N,Ty,DepTab,DAG) -> {NewDepTab,NewDAG} = add_write_dep(D,N,Ty,DepTab,DAG), add_write_deps(Ds,N,Ty,NewDepTab,NewDAG). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : add_write_dep %% Description : Updates the dependence table with N as next writer, and %% updates the DAG with the dependencies from N to subsequent %% nodes. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% add_write_dep(X,N,Ty,DepTab,DAG) -> {NxtWriter,NxtReaders} = lookup(X,DepTab), NewDepTab = writer(X,N,Ty,DepTab), NewDAG = write_deps(N,Ty,NxtWriter,NxtReaders,DAG), {NewDepTab, NewDAG}. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : write_deps %% Argument : Instr - Current instr %% Ty - Type of current instr %% NxtWriter - The node that is the next writer of the ragister %% that Instr defines. %% NxtReaders - The nodes that are subsequent readers of the %% register that N defines. %% DAG - The dependence graph %% Returns : Calls raw_deps that finally returns a new DAG with the new %% dependence arcs added. %% Description : If a next writer exists a dependence arc for this node is %% added, and after this raw_deps is called to compute the %% arcs for read-after-write dependencies. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% write_deps(Instr,Ty,NxtWriter,NxtReaders,DAG) -> DAG1 = case NxtWriter of none -> DAG; {Instr,_} -> DAG; {Wr,WrTy} -> dep_arc(Instr, hipe_target_machine:waw_latency(Ty,WrTy), Wr, DAG) end, raw_deps(Instr,Ty,NxtReaders,DAG1). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : raw_deps %% Argument : Instr - current instr %% Type - type of instr %% Readers - subsequent readers %% DAG - dependence graph %% Returns : DAG - A new DAG with read-after-write dependencies added %% Description : Updates the DAG with the dependence-arcs from Instr to the %% subsequent readers, with the appropriate latencies. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% raw_deps(_Instr,_Type,[],DAG) -> DAG; raw_deps(Instr,Ty,[{Rd,RdTy}|Xs],DAG) -> raw_deps(Instr,Ty,Xs, dep_arc(Instr,hipe_target_machine:raw_latency(Ty,RdTy), Rd,DAG)). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : add_read_deps %% Argument : Uses - The registers that node N uses. %% N - Index of the current node. %% Ty - Type of current node. %% DepTab - Dependence table %% DAG - Dependence graph %% Returns : {DepTab, DAG} - with updated values. %% Description : Adds the read dependencies from node N to subsequent ones, %% according to the registers that N uses. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% add_read_deps([],_N,_Ty,DepTab,DAG) -> {DepTab,DAG}; add_read_deps([U|Us],N,Ty,DepTab,DAG) -> {NewDepTab,NewDAG} = add_read_dep(U,N,Ty,DepTab,DAG), add_read_deps(Us,N,Ty,NewDepTab,NewDAG). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : add_read_dep %% Argument : X - Used register %% N - Index of checked instr %% Ty - Type of checked instr %% DepTab - Hashtable with {next-writer, next-readers} %% DAG - Dependence graph %% Returns : {DepTab, DAG} - with updated values %% Description : Looks up what the next-writer/next-readers are, and adjusts %% the table with current node as new reader. Finally %% read-dependencies are added to the DAG. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% add_read_dep(X,N,Ty,DepTab,DAG) -> {NxtWriter,_NxtReaders} = lookup(X,DepTab), NewDepTab = reader(X,N,Ty,DepTab), NewDAG = read_deps(N,Ty,NxtWriter,DAG), {NewDepTab, NewDAG}. % If NxtWriter is 'none', then this var is not written subsequently % Add WAR from Instr to NxtWriter (if it exists) % *** UNFINISHED *** %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : read_deps %% Argument : N - Index of current node %% Ty - Type of current node %% Writer - tuple {NextWriter, WrType} where NextWriter is the %% subsequent instr that writes this register next time, %% and WrType is the type of that instr. %% DAG - The dependence graph %% Returns : DAG %% Description : Returns a new DAG if a next-writer exists, otherwise the old %% DAG is returned. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% read_deps(_Instr,_Ty,none,DAG) -> DAG; read_deps(_Instr,_Ty,{_Instr,_},DAG) -> DAG; read_deps(Instr,Ty,{NxtWr,NxtWrTy},DAG) -> dep_arc(Instr,hipe_target_machine:war_latency(Ty,NxtWrTy),NxtWr, DAG). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : empty_deptab %% Description : Creates an empty dependence table (hash-table) %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% empty_deptab() -> gb_trees:empty(). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : lookup %% Argument : X - key (register) %% DepTab - dependence table %% Returns : {NextWriter, NextReaders} %% Description : Returns next writer and a list of following readers on %% register X. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% lookup(X, DepTab) -> case gb_trees:lookup(X, DepTab) of none -> {none, []}; {value, {W, Rs} = Val} -> Val end. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : writer %% Argument : X - key (register) %% N - index of writer %% Ty - type of writer %% DepTab - dependence table to be updated %% Returns : DepTab - new dependence table %% Description : Sets N tobe next writer on X %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% writer(X, N, Ty, DepTab) -> gb_trees:enter(X, {{N, Ty}, []}, DepTab). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : reader %% Argument : X - key (register) %% N - index of reader %% Ty - type of reader %% DepTab - dependence table to be updated %% Returns : DepTab - new dependence table %% Description : Adds N to the dependence table as a reader. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% reader(X,N,Ty,DepTab) -> {W,Rs} = lookup(X,DepTab), gb_trees:enter(X,{W,[{N,Ty}|Rs]},DepTab). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% %% The following version of md/2 separates heap- and stack operations, %% which allows for greater reordering. %% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : md %% Argument : IxBB - indexed block %% DAG - dependence graph %% Returns : DAG - new dependence graph %% Description : Adds arcs for load/store dependencies to the DAG. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% md(IxBB, DAG) -> md(IxBB,empty_md_state(),DAG). md([],_,DAG) -> DAG; md([{N,I}|Is],St,DAG) -> case md_type(I) of other -> md(Is,St,DAG); {st,T} -> { WAW_nodes, WAR_nodes, NewSt } = st_overlap(N,T,St), md(Is,NewSt, md_war_deps(WAR_nodes,N,md_waw_deps(WAW_nodes,N,DAG))); {ld,T} -> { RAW_nodes, NewSt } = ld_overlap(N,T,St), md(Is,NewSt, md_raw_deps(RAW_nodes,N,DAG)) end. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : md_war_deps %% Argument : WAR_nodes - write-after-read nodes depending on N %% N - index of current instr %% DAG - dependence graph %% Returns : DAG - updated DAG %% Description : Adds arcs for write-after-read dependencies for N %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% md_war_deps([],_,DAG) -> DAG; md_war_deps([M|Ms],N,DAG) -> md_war_deps(Ms,N,dep_arc(M,hipe_target_machine:m_war_latency(),N,DAG)). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : md_waw_deps %% Argument : WAW_nodes - write-after-write nodes depending on N %% N - index of current instr %% DAG - dependence graph %% Returns : DAG - updated DAG %% Description : Adds arcs for write-after-write dependencies for N %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% md_waw_deps([],_,DAG) -> DAG; md_waw_deps([M|Ms],N,DAG) -> md_waw_deps(Ms,N,dep_arc(M,hipe_target_machine:m_waw_latency(),N,DAG)). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : md_raw_deps %% Argument : RAW_nodes - read-after-write nodes depending on N %% N - index of current instr %% DAG - dependence graph %% Returns : DAG - updated DAG %% Description : Adds arcs for read-after-write dependencies for N %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% md_raw_deps([],_,DAG) -> DAG; md_raw_deps([M|Ms],N,DAG) -> md_raw_deps(Ms,N,dep_arc(M,hipe_target_machine:m_raw_latency(),N,DAG)). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : empty_md_state %% Description : Returns an empty memorydependence state, eg. 4 lists %% representing {StackStores, HeapStores, StackLoads, HeapLoads} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% empty_md_state() -> {[], [], [], []}. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : md_type %% Argument : I - instr %% Description : Maps the instr-type to a simplified type, telling if it's %% store/load resp. heap or stack. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% md_type(I) -> case I of #load{} -> Sp = hipe_sparc_registers:stack_pointer(), Src = hipe_sparc:load_src(I), N = hipe_sparc:reg_nr(Src), Off = hipe_sparc:load_off(I), if N =:= Sp -> % operation on stack {ld,{sp,Off}}; true -> {ld,{hp,Src,Off}} end; #store{} -> Sp = hipe_sparc_registers:stack_pointer(), Dst = hipe_sparc:store_dest(I), N = hipe_sparc:reg_nr(Dst), Off = hipe_sparc:store_off(I), if N =:= Sp -> {st,{sp,Off}}; true -> {st,{hp,Dst,Off}} end; _ -> other end. %% Given a memory operation and a 'memory op state', %% overlap(N,MemOp,State) returns { Preceding_Dependent_Ops, NewState }. %% which are either a tuple { WAW_deps, WAR_deps } or a list RAW_deps. %% %% NOTES: %% Note that Erlang's semantics ("heap stores never overwrite existing data") %% means we can be quite free in reordering stores to the heap. %% Ld/St to the stack are simply handled by their offsets; since we do not %% rename the stack pointer, this is sufficient. %% *** We assume all memory ops have uniform size = 4 *** %% %% NOTES: %% The method mentioned above has now been changed because the assumption that %% "heap stores never overwrite existing data" caused a bug when the %% process-pointer was treated the same way as the heap. We were also told %% that the semantics can possibly change in the future, so it would be more %% safe to treat the heap store/loads as the stack. %% A future improvement can be to do an alias analysis to give more freedom %% in reordering stuff... %% %% Alias state: %% { [StackOp], [HeapOp], [StackOp], [HeapOp] } %% where StackOp = {InstrID, Offset} %% HeapOp = {InstrID, Reg, Offset} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : st_overlap %% Argument : N - Index of current node %% Type - {sp,Off} or {hp,Dst,Off}, store on stack or heap %% State - { [StackStrs], [HeapStrs], [StackLds], [HeapLds] } %% where StackStrs/StackLds = {InstrID, Offset} %% and HeapStrs/HeapLds = {InstrID, Reg, Offset} %% Returns : { DepStrs, DepLds, State } - %% where DepStrs/DepLds = [NodeId] %% and State is the new state %% Description : Adds dependencies for overlapping stores. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% st_overlap(N, {sp, Off}, {St_Sp, St_Hp, Ld_Sp, Ld_Hp}) -> {DepSt, IndepSt_Sp} = st_sp_dep(St_Sp, Off), {DepLd, IndepLd_Sp} = ld_sp_dep(Ld_Sp, Off), {DepSt, DepLd, {[{N, Off}|IndepSt_Sp], St_Hp, IndepLd_Sp, Ld_Hp}}; st_overlap(N, {hp, Dst, Off}, {St_Sp, St_Hp, Ld_Sp, Ld_Hp}) -> DstOff = {Dst, Off}, {DepSt,_IndepSt_Hp} = st_hp_dep(St_Hp, DstOff), {DepLd, IndepLd_Hp} = ld_hp_dep(Ld_Hp, DstOff), {DepSt, DepLd, {St_Sp, [{N, Dst, Off}|St_Hp], Ld_Sp, IndepLd_Hp}}. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : ld_overlap %% Argument : N - Index of current node %% Type - {sp,Off} or {hp,Dst,Off}, store on stack or heap %% State - { [StackStrs], [HeapStrs], [StackLds], [HeapLds] } %% where StackStrs/StackLds = {InstrID, Offset} %% and HeapStrs/HeapLds = {InstrID, Reg, Offset} %% Returns : { DepStrs, State } %% Description : Adds dependencies for overlapping laods %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ld_overlap(N, {sp, Off}, {St_Sp, St_Hp, Ld_Sp, Ld_Hp}) -> DepSt = sp_dep_only(St_Sp, Off), {DepSt, {St_Sp, St_Hp, [{N, Off}|Ld_Sp], Ld_Hp}}; ld_overlap(N, {hp, Src, Off}, {St_Sp, St_Hp, Ld_Sp, Ld_Hp}) -> DepSt = hp_dep_only(St_Hp, Src, Off), {DepSt, {St_Sp, St_Hp, Ld_Sp, [{N, Src, Off}|Ld_Hp]}}. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : st_sp_dep %% Description : Adds dependencies that are depending on a stack store %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% st_sp_dep(Stores, Off) -> sp_dep(Stores, Off, [], []). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : ld_sp_dep %% Description : Adds dependencies that are depending on a stack load %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ld_sp_dep(Loads, Off) -> sp_dep(Loads, Off, [], []). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : st_hp_dep %% Description : Adds dependencies that are depending on a heap store %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% st_hp_dep(Stores, {_Reg, _Off} = RegOff) -> hp_dep(Stores, RegOff, [], []). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : ld_hp_dep %% Description : Adds dependencies that are depending on a heap load %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ld_hp_dep(Loads, {_Reg, _Off} = RegOff) -> hp_dep(Loads, RegOff, [], []). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : sp_dep %% Description : Returns {Dependent, Independent} which are lists of nodes %% that depends or not on a stack load/store %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% sp_dep([], _Off, Dep, Indep) -> {Dep, Indep}; sp_dep([{N,Off}|Xs], Off, Dep, Indep) -> sp_dep(Xs, Off, [N|Dep], Indep); sp_dep([X|Xs], Off, Dep, Indep) -> sp_dep(Xs, Off, Dep, [X|Indep]). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : hp_dep %% Description : Returns {Dependent, Independent} which are lists of nodes %% that depends or not on a heap load/store %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% hp_dep([], {_Reg,_Off}, Dep, Indep) -> {Dep,Indep}; hp_dep([{N,Reg,Off1}|Xs], {Reg,Off}, Dep, Indep) when Off1 =/= Off -> hp_dep(Xs, {Reg,Off}, Dep, [{N,Reg,Off1}|Indep]); hp_dep([{N,_,_}|Xs], {Reg,Off}, Dep, Indep) -> hp_dep(Xs, {Reg,Off}, [N|Dep], Indep). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : sp_dep_only %% Description : Returns a list of nodes that are depending on a stack store %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% sp_dep_only(Stores, Off) -> [N || {N,Off0} <- Stores, Off =:= Off0]. %% Dependences from heap stores to heap loads. %% *** UNFINISHED *** %% - but works %% This is somewhat subtle: %% - a heap load can only bypass a heap store if we KNOW it won't %% load the stored value %% - unfortunately, we do not know the relationships between registers %% at this point, so we can't say that store(p+4) is independent of %% load(q+0). %% (OR CAN WE? A bit closer reasoning might show that it's possible?) %% - We can ONLY say that st(p+c) and ld(p+c') are independent when c /= c' %% %% (As said before, it might be possible to lighten this restriction?) hp_dep_only([], _Reg, _Off) -> []; hp_dep_only([{_N,Reg,Off_1}|Xs], Reg, Off) when Off_1 =/= Off -> hp_dep_only(Xs, Reg, Off); hp_dep_only([{N,_,_}|Xs], Reg, Off) -> [N|hp_dep_only(Xs, Reg, Off)]. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Control dependences: %% - add dependences so that %% * branches are performed in order %% * unsafe operations are 'fenced in' by surrounding branches %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : cd %% Argument : IxBB - indexed block %% DAG - dependence graph %% Returns : DAG - new dependence graph %% Description : Adds conditional dependencies to the DAG %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% cd(IxBB,DAG) -> cd(IxBB, DAG, none, [], []). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : cd %% Argument : IxBB - indexed block %% DAG - dependence graph %% PrevBr - previous branch %% PrevUnsafe - previous unsafe instr (mem-op) %% PrevOthers - previous other instrs, used to "fix" preceeding %% instrs so they don't bypass a branch. %% Returns : DAG - new dependence graph %% Description : Adds conditional dependencies to the graph. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% cd([], DAG, _PrevBr, _PrevUnsafe, _PrevOthers) -> DAG; cd([{N,I}|Xs], DAG, PrevBr, PrevUnsafe, PrevOthers) -> case cd_type(I) of {branch,Ty} -> DAG1 = cd_branch_to_other_deps(N, PrevOthers, DAG), NewDAG = cd_branch_deps(PrevBr, PrevUnsafe, N, Ty, DAG1), cd(Xs,NewDAG,{N,Ty},[],[]); {unsafe,Ty} -> NewDAG = cd_unsafe_deps(PrevBr,N,Ty,DAG), cd(Xs, NewDAG, PrevBr, [{N,Ty}|PrevUnsafe], PrevOthers); {other,_Ty} -> cd(Xs, DAG, PrevBr, PrevUnsafe, [N|PrevOthers]) end. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : cd_branch_to_other_deps %% Argument : N - index of branch %% Ms - list of indexes of "others" preceding instrs %% DAG - dependence graph %% Returns : DAG - new graph %% Description : Makes preceding instrs fixed so they don't bypass a branch %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% cd_branch_to_other_deps(_, [], DAG) -> DAG; cd_branch_to_other_deps(N, [M | Ms], DAG) -> cd_branch_to_other_deps(N, Ms, dep_arc(M, zero_latency(), N, DAG)). %% Is the operation a branch, an unspeculable op or something else? %% Returns %% {branch,BranchType} %% {unsafe,OpType} %% {other,OpType} %% *** UNFINISHED *** %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : cd_type %% Argument : I - instr %% Description : Maps instrs to a simpler type. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% cd_type(I) -> case I of #goto{} -> {branch,uncond}; #br{} -> {branch,'cond'}; #b{} -> {branch,'cond'}; #call_link{} -> {branch,call}; #jmp_link{} -> {branch,call}; #jmp{} -> {branch,call}; #load{} -> {unsafe,load}; #store{} -> {unsafe,load}; T -> {other,T} end. %% add dependences to keep order of branches + unspeculable ops: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : cd_branch_deps %% Argument : PrevBr - preceeding branch %% PrevUnsafe - preceeding unsafe ops, eg, mem-ops %% N - current id. %% Ty - type of current instr %% DAG - dependence graph %% Returns : DAG - new DAG %% Description : Adds arcs between branches and calls deps_to_unsafe that adds %% arcs between branches and unsafe ops. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% cd_branch_deps(PrevBr, PrevUnsafe, N, Ty, DAG) -> DAG1 = case PrevBr of none -> DAG; {Br,BrTy} -> dep_arc(Br, hipe_target_machine:br_br_latency(BrTy,Ty), N, DAG) end, deps_to_unsafe(PrevUnsafe, N, Ty, DAG1). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : deps_to_unsafe %% Description : Adds dependencies between unsafe's and branches %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% deps_to_unsafe([], _, _, DAG) -> DAG; deps_to_unsafe([{M,UTy}|Us], N, Ty, DAG) -> deps_to_unsafe(Us,N,Ty, dep_arc(M, hipe_target_machine:unsafe_to_br_latency(UTy,Ty), N, DAG)). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : cd_unsafe_deps %% Description : Adds dependencies between branches and unsafe's %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% cd_unsafe_deps(none, _, _, DAG) -> DAG; cd_unsafe_deps({Br,BrTy}, N, Ty, DAG) -> dep_arc(Br, hipe_target_machine:br_to_unsafe_latency(BrTy, Ty), N, DAG). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : def_use %% Argument : Instr %% Description : Returns the registers that Instr defines resp. uses as 2 lists %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% def_use(Instr) -> {hipe_sparc:defines(Instr), hipe_sparc:uses(Instr)}. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Function : move_or_alu %% Description : True if the instruction is a move or an alu; false otherwise %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% move_or_alu(#move{}) -> true; move_or_alu(#alu{}) -> true; move_or_alu(_) -> false. %% Debugging stuff below %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% -ifdef(debug1). debug1_stuff(Blk, DAG, IxBlk, Sch, X) -> io:format("Blk: ~p~n",[Blk]), io:format("DAG: ~n~p~n~p",[DAG,IxBlk]), io:format("~n"), print_instrs(IxBlk), print_sch(Sch, IxBlk), print_instrs2(X). print_instrs([]) -> io:format("~n"); print_instrs([{N,Instr} | Instrs]) -> io:format("(~p): ",[N]), hipe_sparc_pp:pp_instr(Instr), io:format("~p~n",[element(1,Instr)]), print_instrs(Instrs). print_instrs2([]) -> io:format("~n"); print_instrs2([Instr | Instrs]) -> hipe_sparc_pp:pp_instr(Instr), print_instrs2(Instrs). print_sch([],_) -> io:format("~n"); print_sch([{{cycle,Cycle},{node,I}} | Rest], IxBlk) -> io:format("{C~p, N~p} ",[Cycle,I]), print_node(I, IxBlk), print_sch(Rest, IxBlk). print_node(_, []) -> io:format("~n"); print_node(I, [{I, Instr} | _]) -> hipe_sparc_pp:pp_instr(Instr); print_node(I, [_ | IxBlk]) -> print_node(I, IxBlk). -else. debug1_stuff(_Blk, _DAG, _IxBlk, _Sch, _X) -> ok. -endif.
lib/hipe/opt/hipe_schedule.erl
0.524882
0.503723
hipe_schedule.erl
starcoder
%% ------------------------------------------------------------------- %% %% riak_core_coverage_plan: Create a plan to cover a minimal set of VNodes. %% %% Copyright (c) 2007-2015 Basho Technologies, Inc. All Rights Reserved. %% %% This file is provided to you under the Apache License, %% Version 2.0 (the "License"); you may not use this file %% except in compliance with the License. You may obtain %% a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, %% software distributed under the License is distributed on an %% "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY %% KIND, either express or implied. See the License for the %% specific language governing permissions and limitations %% under the License. %% %% ------------------------------------------------------------------- %% @doc A module to calculate plans to cover a set of VNodes. %% %% There are two types of plans available: a "traditional" %% coverage plan which minimizes the set of VNodes to be %% contacted. This is used internally for functionality such as %% 2i and Riak Pipe. %% %% There is also a new "subpartition" coverage plan designed to %% achieve the opposite: allow clients to make parallel requests %% across as much of the cluster as is desired, designed for use %% with Basho Data Platform. %% %% Both plan types are now available through the protocol buffers %% API, delivered as discrete opaque binary chunks to be sent %% with queries that support the functionality (primarily 2i). %% Example traditional coverage plan for a two node, 8 vnode cluster %% at nval=3 %% { %% %% First component is a list of {vnode hash, node name} tuples %% [ %% {0, 'dev1@127.0.0.1'}, %% {548063113999088594326381812268606132370974703616, 'dev2@127.0.0.1'}, %% {913438523331814323877303020447676887284957839360, 'dev2@127.0.0.1'} %% ], %% %% Second component is a list of {vnode hash, [partition list]} %% %% tuples representing filters when not all partitions managed by a %% %% vnode are required to complete the coverage plan %% [ %% {913438523331814323877303020447676887284957839360, %% [730750818665451459101842416358141509827966271488, %% 913438523331814323877303020447676887284957839360] %% } %% ] %% } %% Snippet from a subpartition coverage plan for a two node, 8 vnode %% cluster at nval=3, with each partition represented twice for up to %% 16 parallel queries. %% The nested tuple in the third position is a representation of the %% mask against the cluster's full keyspace and bits necessary to %% shift keys right to compare against the mask (or shift the mask %% left to represent its actual value). %% [ %% %% Second vnode, first half of first partition %% {182687704666362864775460604089535377456991567872, %% 'dev2@127.0.0.1', {0, 156} %% }, %% %% Second vnode, second half of first partition %% {182687704666362864775460604089535377456991567872, %% 'dev2@127.0.0.1', {1, 156}}, %% %% Third vnode, first half of second partition %% {365375409332725729550921208179070754913983135744, %% 'dev1@127.0.0.1', {2, 156}}, %% %% Third vnode, second half of second partition %% {365375409332725729550921208179070754913983135744, %% 'dev1@127.0.0.1', {3, 156}}, %% ... %% ] -module(riak_core_coverage_plan). -include("riak_core_vnode.hrl"). -ifdef(TEST). -compile(export_all). -include_lib("eunit/include/eunit.hrl"). -endif. %% API -export([create_plan/5, create_plan/6, create_subpartition_plan/6]). -export([replace_subpartition_chunk/7, replace_traditional_chunk/7]). %% For other riak_core applications' coverage plan modules -export([identify_unavailable_vnodes/3, identify_unavailable_vnodes/4]). -export([add_offset/3]). -type ring_size() :: pos_integer(). -type index() :: chash:index_as_int(). %% This module is awash with lists of integers with the same range but %% and similar but distinct meanings (vnodes and partitions), or %% different ranges but same meanings (partition short id vs partition %% hash), or different values depending on context (partitions are %% incremented by a full RingIndexInc when used as traditional cover %% filters). %% %% Thus, tagged tuples are used in many places to make it easier for %% maintainers to keep track of what data is flowing where. %% IDs (vnode, partition) are integers in the [0, RingSize) space %% (and trivially map to indexes). Private functions deal with IDs %% instead of indexes as much as possible. %% ID, ring size, ring index increment -type vnode_id() :: {'vnode_id', non_neg_integer(), ring_size()}. -type partition_id() :: {'partition_id', non_neg_integer(), ring_size()}. %% Not a tagged tuple. ID + bits necessary to shift the ID to the left %% to create the subpartition index. -type subpartition_id() :: {non_neg_integer(), pos_integer()}. -type req_id() :: non_neg_integer(). -type coverage_vnodes() :: [{index(), node()}]. -type vnode_filters() :: [{index(), [index()]}]. -type coverage_plan() :: {coverage_vnodes(), vnode_filters()}. %% Vnode index, node, { Subpartition id, BSL } -type subp_plan() :: [{index(), node(), subpartition_id()}]. -export_type([coverage_plan/0, coverage_vnodes/0, vnode_filters/0]). %% Function to determine nodes currently available. This can be %% swapped out for testing to avoid using meck. The argument ignored %% here is the chashbin, potentially useful for testing. -define(AVAIL_NODE_FUN, fun(Svc, _) -> riak_core_node_watcher:nodes(Svc) end). %% =================================================================== %% Public API %% =================================================================== %% @doc Create Riak's traditional coverage plan to distribute work to %% a minimal set of covering VNodes around the ring. -spec create_plan(vnode_selector(), pos_integer(), pos_integer(), req_id(), atom()) -> {error, term()} | coverage_plan(). create_plan(VNodeTarget, NVal, PVC, ReqId, Service) -> %% jdaily added a sixth parameter, _Request, which is ignored %% here, apparently used in coverage API work. See %% riak_kv_qry_coverage_plan:create_plan for more clues. create_plan(VNodeTarget, NVal, PVC, ReqId, Service, undefined). -spec create_plan(vnode_selector(), pos_integer(), pos_integer(), req_id(), atom(), term()) -> {error, term()} | coverage_plan(). create_plan(VNodeTarget, NVal, PVC, ReqId, Service, _Request) -> {ok, CHBin} = riak_core_ring_manager:get_chash_bin(), create_traditional_plan(VNodeTarget, NVal, PVC, ReqId, Service, CHBin, ?AVAIL_NODE_FUN). %% @doc Create a "mini" traditional coverage plan to replace %% components of a previously-generated plan that are not useful %% to the client because the node is unavailable -spec replace_traditional_chunk(VnodeIdx :: index(), Node :: node(), Filters :: list(index()), NVal :: pos_integer(), ReqId :: req_id(), DownNodes :: list(node()), Service :: atom()) -> {error, term()} | coverage_plan(). replace_traditional_chunk(VnodeIdx, Node, Filters, NVal, ReqId, DownNodes, Service) -> {ok, CHBin} = riak_core_ring_manager:get_chash_bin(), replace_traditional_chunk(VnodeIdx, Node, Filters, NVal, ReqId, DownNodes, Service, CHBin, ?AVAIL_NODE_FUN). %% @doc Create a coverage plan with at least one slice per partition, %% originally designed for parallel extraction of data via 2i. -spec create_subpartition_plan('all'|'allup', pos_integer(), pos_integer() | {pos_integer(), pos_integer()}, pos_integer(), req_id(), atom()) -> {error, term()} | subp_plan(). create_subpartition_plan(VNodeTarget, NVal, {MinPar, RingSize}, PVC, ReqId, Service) -> Count = if MinPar =< RingSize -> RingSize; true -> next_power_of_two(MinPar) end, create_subpartition_plan(VNodeTarget, NVal, Count, PVC, ReqId, Service); create_subpartition_plan(VNodeTarget, NVal, Count, PVC, ReqId, Service) -> %% IMPORTANT: `Count' is assumed to be a power of 2. Anything else %% will behave badly. {ok, ChashBin} = riak_core_ring_manager:get_chash_bin(), create_subpartition_plan(VNodeTarget, NVal, Count, PVC, ReqId, Service, ChashBin, ?AVAIL_NODE_FUN). %% @doc Create a "mini" traditional coverage plan to replace %% components of a previously-generated plan that are not useful %% to the client because the node is unavailable -spec replace_subpartition_chunk(VnodeIdx :: index(), Node :: node(), {Mask :: non_neg_integer(), Bits :: non_neg_integer()}, NVal :: pos_integer(), ReqId :: req_id(), DownNodes :: list(node()), Service :: atom()) -> {error, term()} | subp_plan(). replace_subpartition_chunk(VnodeIdx, Node, {Mask, Bits}, NVal, ReqId, DownNodes, Service) -> {ok, CHBin} = riak_core_ring_manager:get_chash_bin(), replace_subpartition_chunk(VnodeIdx, Node, {Mask, Bits}, NVal, ReqId, DownNodes, Service, CHBin, ?AVAIL_NODE_FUN). %% ==================================================================== %% Internal functions that are useful for other riak_core applications %% that need to generate custom coverage plans %% ==================================================================== -spec identify_unavailable_vnodes(chashbin:chashbin(), pos_integer(), atom()) -> list(vnode_id()). identify_unavailable_vnodes(CHBin, RingSize, Service) -> identify_unavailable_vnodes(CHBin, RingSize, Service, ?AVAIL_NODE_FUN). -spec identify_unavailable_vnodes(chashbin:chashbin(), pos_integer(), atom(), fun((atom(), binary()) -> list(node()))) -> list(vnode_id()). identify_unavailable_vnodes(CHBin, RingSize, Service, AvailNodeFun) -> %% Get a list of the VNodes owned by any unavailable nodes [{vnode_id, index_to_id(Index, RingSize), RingSize} || {Index, _Node} <- riak_core_apl:offline_owners( AvailNodeFun(Service, CHBin), CHBin)]. %% Adding an offset while keeping the result inside [0, Top). Adding %% Top to the left of `rem' allows offset to be negative without %% violating the lower bound of zero add_offset(Position, Offset, Top) -> (Position + (Top + Offset)) rem Top. %% ==================================================================== %% Internal functions %% ==================================================================== %% @private create_traditional_plan(VNodeTarget, NVal, PVC, ReqId, Service, CHBin, AvailNodeFun) -> RingSize = chashbin:num_partitions(CHBin), %% Calculate an offset based on the request id to offer the %% possibility of different sets of VNodes being used even when %% all nodes are available. Used in compare_vnode_keyspaces as a %% tiebreaker. Offset = ReqId rem NVal, AllVnodes = list_all_vnode_ids(RingSize), %% Older versions of this call chain used the same list of %% integers for both vnode IDs and partition IDs, which made for %% confusing reading. We can cheat a little less obnoxiously AllPartitions = lists:map(fun({vnode_id, Id, RS}) -> {partition_id, Id, RS} end, AllVnodes), UnavailableVnodes = identify_unavailable_vnodes(CHBin, RingSize, Service, AvailNodeFun), %% Create function to map coverage keyspaces to %% actual VNode indexes and determine which VNode %% indexes should be filtered. CoverageVNodeFun = fun({VNodeID, KeySpaces}, Acc) -> %% Calculate the VNode index using the %% ring position and the increment of %% ring index values. VNodeIndex = convert(VNodeID, vnode_index), Node = chashbin:index_owner(VNodeIndex, CHBin), CoverageVNode = {VNodeIndex, Node}, case length(KeySpaces) < NVal of true -> KeySpaceIndexes = [convert(PartitionID, keyspace_filter) || PartitionID <- KeySpaces], {CoverageVNode, [{VNodeIndex, KeySpaceIndexes} | Acc]}; false -> {CoverageVNode, Acc} end end, %% The offset value serves as a tiebreaker in the %% compare_vnode_keyspaces function and is used to distribute work %% to different sets of VNodes. CoverageResult = find_minimal_coverage(AllPartitions, AllVnodes -- UnavailableVnodes, Offset, NVal, lists:min([PVC, NVal]), []), case CoverageResult of {ok, CoveragePlan} -> %% Assemble the data structures required for %% executing the coverage operation. lists:mapfoldl(CoverageVNodeFun, [], CoveragePlan); {insufficient_vnodes_available, _KeySpace, PartialCoverage} -> case VNodeTarget of allup -> %% The allup indicator means generate a coverage plan %% for any available VNodes. lists:mapfoldl(CoverageVNodeFun, [], PartialCoverage); all -> {error, insufficient_vnodes_available} end end. %% @private replace_traditional_chunk(VnodeIdx, Node, Filters, NVal, ReqId, DownNodes, Service, CHBin, AvailNodeFun) -> RingSize = chashbin:num_partitions(CHBin), Offset = ReqId rem NVal, %% We have our own idea of what nodes are available. The client %% may have a different idea of offline nodes based on network %% partitions, so we take that into account. %% %% The client can't really tell us what nodes it thinks are up (it %% only knows hostnames at best) but the opaque coverage chunks it %% uses have node names embedded in them, and it can tell us which %% chunks do *not* work. UpNodes = AvailNodeFun(Service, CHBin) -- [Node|DownNodes], NeededPartitions = partitions_by_index_or_filter( VnodeIdx, NVal, Filters, RingSize), %% For each partition, create a tuple with that partition (as a %% filter index) mapped to a nested tuple of preflist + (initially %% empty) filter list list. Preflists = lists:map( fun(Partition) -> {convert(Partition, keyspace_filter), safe_hd( partition_to_preflist(Partition, NVal, Offset, UpNodes, CHBin)) } end, NeededPartitions), maybe_create_traditional_replacement(Preflists, lists:keyfind([], 2, Preflists)). %% @private maybe_create_traditional_replacement(Preflists, false) -> %% We do not go to great lengths to minimize the number of %% coverage plan components we'll return, but we do at least sort %% the vnodes we find so that we can consolidate filters later. create_traditional_replacement(lists:sort(Preflists)); maybe_create_traditional_replacement(_Preflists, _) -> {error, primary_partition_unavailable}. %% @private %% Argument to this should be sorted create_traditional_replacement(Preflists) -> dechunk_traditional_replacement( lists:foldl( %% Pattern match the current vnode against the head of the %% accumulator; if the vnodes match, we can consolidate %% partition filters fun({PartIdx, VNode}, [{VNode, Partitions}|Tail]) -> [{VNode, lists:sort([PartIdx|Partitions])}|Tail]; %% Instead if this vnode is new, place it into the %% accumulator as is ({PartIdx, VNode}, Accum) -> [{VNode, [PartIdx]}|Accum] end, [], Preflists)). %% @private %% Take our replacement traditional coverage chunks and consolidate it %% into a traditional coverage plan (that another layer will rechunk, %% but whatchagonnado) dechunk_traditional_replacement(Coverage) -> { lists:map(fun({{Vnode, Node}, _Filters}) -> {Vnode, Node} end, Coverage), lists:filtermap(fun({{_Vnode, _Node}, []}) -> false; ({{Vnode, _Node}, Filters}) -> {true, {Vnode, Filters}} end, Coverage) }. %% @private safe_hd([]) -> []; safe_hd(List) -> hd(List). %% @private replace_subpartition_chunk(_VnodeIdx, Node, {_Mask, _Bits}=SubpID, NVal, ReqId, DownNodes, Service, CHBin, AvailNodeFun) -> Offset = ReqId rem NVal, %% We have our own idea of what nodes are available. The client %% may have a different idea of offline nodes based on network %% partitions, so we take that into account. %% %% The client can't really tell us what nodes it thinks are up (it %% only knows hostnames at best) but the opaque coverage chunks it %% uses have node names embedded in them, and it can tell us which %% chunks do *not* work. UpNodes = AvailNodeFun(Service, CHBin) -- [Node|DownNodes], %% We don't know what partition this subpartition is in, but we %% can request a preflist for it by converting the subpartition to %% a document index. %% %% Unlike traditional coverage filters to document key hash %% mappings which have off-by-one adjustments, subpartition masks %% map directly against the relevant key hashes. DocIdx = convert(SubpID, subpartition_index), PrefList = docidx_to_preflist(DocIdx, NVal, Offset, UpNodes, CHBin), singular_preflist_to_chunk(safe_hd(PrefList), SubpID). %% @private singular_preflist_to_chunk([], _SubpID) -> {error, primary_partition_unavailable}; singular_preflist_to_chunk({VnodeIdx, Node}, SubpID) -> [{VnodeIdx, Node, SubpID}]. %% ==================================================================== %%% Conversion functions %% @private %% We spend a lot of code mapping between data types, mostly %% integer-based. Consolidate that as much as possible here. -spec convert(partition_id() | vnode_id() | subpartition_id(), atom()) -> non_neg_integer(). convert({partition_id, PartitionID, RingSize}, keyspace_filter) -> %% Because data is stored one partition higher than the keyspace %% into which it directly maps, we have to increment a partition %% ID by one to find the relevant keyspace index for traditional %% coverage plan filters id_to_index(add_offset(PartitionID, 1, RingSize), RingSize); convert({partition_id, PartitionID, RingSize}, partition_index) -> id_to_index(PartitionID, RingSize); convert({vnode_id, VNodeID, RingSize}, vnode_index) -> id_to_index(VNodeID, RingSize); convert({SubpID, Bits}, subpartition_index) -> SubpID bsl Bits. %% @private partition_to_preflist(Partition, NVal, Offset, UpNodes, CHBin) -> DocIdx = convert(Partition, partition_index), docidx_to_preflist(DocIdx, NVal, Offset, UpNodes, CHBin). %% @private docidx_to_preflist(DocIdx, NVal, Offset, UpNodes, CHBin) -> %% `chashbin' would be fine with a straight integer, but %% `riak_core_apl' has a -spec that mandates binary(), so we'll %% play its game OrigPreflist = lists:map(fun({{Idx, Node}, primary}) -> {Idx, Node} end, riak_core_apl:get_primary_apl_chbin( <<DocIdx:160/integer>>, NVal, CHBin, UpNodes)), rotate_list(OrigPreflist, length(OrigPreflist), Offset). %% @private rotate_list(List, Len, Offset) when Offset >= Len -> List; rotate_list(List, _Len, Offset) -> {Head, Tail} = lists:split(Offset, List), Tail ++ Head. %% @private index_to_id(Index, RingSize) -> RingIndexInc = chash:ring_increment(RingSize), Index div RingIndexInc. %% @private id_to_index(Id, RingSize) when Id < RingSize-> RingIndexInc = chash:ring_increment(RingSize), Id * RingIndexInc. %% @private find_vnode_partitions(Index, N, RingSize) -> n_keyspaces({vnode_id, index_to_id(Index, RingSize), RingSize}, N). %% @private partitions_by_index_or_filter(Idx, NVal, [], RingSize) -> find_vnode_partitions(Idx, NVal, RingSize); partitions_by_index_or_filter(_Idx, _NVal, Filters, RingSize) -> %% Filters for traditional coverage plans are offset by 1, so when %% converting to partition indexes or IDs, we have to subtract 1 lists:map(fun(P) -> {partition_id, add_offset(index_to_id(P, RingSize), -1, RingSize), RingSize} end, Filters). %% @private %% Must be able to comply with PVC if the target is 'all' check_pvc(List, _PVC, allup) -> List; check_pvc(List, PVC, all) -> check_pvc2(List, length(List), PVC). %% @private check_pvc2(List, Len, PVC) when Len >= PVC -> List; check_pvc2(_List, _Len, _PVC) -> []. %% @private create_subpartition_plan(VNodeTarget, NVal, Count, PVC, ReqId, Service, CHBin, AvailNodeFun) -> MaskBSL = data_bits(Count), %% Calculate an offset based on the request id to offer the %% possibility of different sets of VNodes being used even when %% all nodes are available. Offset = ReqId rem NVal, RingSize = chashbin:num_partitions(CHBin), UpNodes = AvailNodeFun(Service, CHBin), SubpList = lists:map(fun(SubpCounter) -> SubpID = {SubpCounter, MaskBSL}, SubpIndex = convert(SubpID, subpartition_index), PartID = {partition_id, chashbin:responsible_position(SubpIndex, CHBin), RingSize}, %% PVC is much like R; if the number of %% available primary partitions won't reach %% the specified PVC value, don't bother %% including this partition at all. We can %% decide later (based on all vs allup) %% whether to return the successful %% components of the coverage plan {PartID, SubpID, check_pvc( docidx_to_preflist(SubpIndex, NVal, Offset, UpNodes, CHBin), PVC, VNodeTarget)} end, lists:seq(0, Count - 1)), %% Now we have a list of tuples; each subpartition maps to a %% partition ID, subpartition ID, and a list of zero or more %% {vnode_index, node} tuples for that partition maybe_create_subpartition_plan(VNodeTarget, SubpList, PVC). %% @private maybe_create_subpartition_plan(allup, SubpList, PVC) -> map_subplist_to_plan(SubpList, PVC); maybe_create_subpartition_plan(all, SubpList, PVC) -> maybe_create_subpartition_plan(all, lists:keyfind([], 3, SubpList), SubpList, PVC). %% @private maybe_create_subpartition_plan(all, false, SubpList, PVC) -> map_subplist_to_plan(SubpList, PVC); maybe_create_subpartition_plan(all, _, _SubpList, _PVC) -> {error, insufficient_vnodes_available}. %% @private map_subplist_to_plan(SubpList, PVC) -> lists:flatten( lists:filtermap(fun({_PartID, _SubpID, []}) -> false; ({_PartID, SubpID, Vnodes}) -> {true, map_pvc_vnodes(Vnodes, SubpID, PVC)} end, SubpList)). %% @private map_pvc_vnodes(Vnodes, SubpID, PVC) -> map_pvc_vnodes(Vnodes, SubpID, PVC, []). %% @private map_pvc_vnodes(_Vnodes, _SubpID, 0, Accum) -> lists:reverse(Accum); map_pvc_vnodes([{VnodeIdx, Node}|Tail], SubpID, PVC, Accum) -> map_pvc_vnodes(Tail, SubpID, PVC-1, [{ VnodeIdx, Node, SubpID }] ++ Accum). %% @private increment_vnode({vnode_id, Position, RingSize}, Offset) -> {vnode_id, add_offset(Position, Offset, RingSize), RingSize}. %% @private list_all_vnode_ids(RingSize) -> lists:map(fun(Id) -> {vnode_id, Id, RingSize} end, lists:seq(0, RingSize - 1)). %% @private %% Note that these Id values are tagged tuples, not integers merge_coverage_results({VnodeId, PartitionIds}, Acc) -> case proplists:get_value(VnodeId, Acc) of undefined -> [{VnodeId, PartitionIds} | Acc]; MorePartitionIds -> UniqueValues = lists:usort(PartitionIds ++ MorePartitionIds), [{VnodeId, UniqueValues} | proplists:delete(VnodeId, Acc)] end. %% @private %% @doc Generates a minimal set of vnodes and partitions to find the requested data -spec find_minimal_coverage(list(partition_id()), list(vnode_id()), non_neg_integer(), non_neg_integer(), non_neg_integer(), list({vnode_id(), list(partition_id())})) -> {ok, list({vnode_id(), list(partition_id())})} | {error, term()}. find_minimal_coverage(_AllPartitions, _AvailableVnodes, _Offset, _NVal, 0, Results) -> {ok, Results}; find_minimal_coverage(AllPartitions, AvailableVnodes, Offset, NVal, PVC, ResultsAcc) -> %% Calculate the available keyspaces. The list of %% keyspaces for each vnode that have already been %% covered by the plan are subtracted from the complete %% list of keyspaces so that coverage plans that %% want to cover more one preflist vnode work out %% correctly. AvailableKeySpaces = [{increment_vnode(VNode, Offset), VNode, n_keyspaces(VNode, NVal) -- proplists:get_value(VNode, ResultsAcc, [])} || VNode <- (AvailableVnodes)], case find_coverage_vnodes(ordsets:from_list(AllPartitions), AvailableKeySpaces, ResultsAcc) of {ok, CoverageResults} -> UpdatedResults = lists:foldl(fun merge_coverage_results/2, ResultsAcc, CoverageResults), find_minimal_coverage(AllPartitions, AvailableVnodes, Offset, NVal, PVC-1, UpdatedResults); Error -> Error end. %% @private %% @doc Find the N key spaces for a VNode. These are *not* the same as %% the filters for the traditional cover plans; the filters would be %% incremented by 1 -spec n_keyspaces(vnode_id(), pos_integer()) -> list(partition_id()). n_keyspaces({vnode_id, VNode, RingSize}, N) -> ordsets:from_list([{partition_id, X rem RingSize, RingSize} || X <- lists:seq(RingSize + VNode - N, RingSize + VNode - 1)]). %% @private %% @doc Find a minimal set of covering VNodes. %% %% Takes: %% A list of all partition IDs still needed for coverage %% A list of available vnode IDs with partitions they cover %% An accumulator for results %% Returns a list of {vnode_id, [partition_id,...]} tuples. -spec find_coverage_vnodes(list(partition_id()), list(vnode_id()), list({vnode_id(), list(partition_id())})) -> {ok, list({vnode_id(), list(partition_id())})}| {insufficient_vnodes_available, list(vnode_id()), list({vnode_id(), list(partition_id())})}. find_coverage_vnodes([], _, Coverage) -> {ok, lists:sort(Coverage)}; find_coverage_vnodes(Partitions, [], Coverage) -> {insufficient_vnodes_available, Partitions, lists:sort(Coverage)}; find_coverage_vnodes(Partitions, AvailableVnodes, Coverage) -> case find_best_vnode_for_keyspace(Partitions, AvailableVnodes) of {error, no_coverage} -> %% Bail find_coverage_vnodes(Partitions, [], Coverage); VNode -> {value, {_, VNode, Covers}, UpdAvailable} = lists:keytake(VNode, 2, AvailableVnodes), UpdCoverage = [{VNode, ordsets:intersection(Partitions, Covers)} | Coverage], UpdPartitions = ordsets:subtract(Partitions, Covers), find_coverage_vnodes(UpdPartitions, UpdAvailable, UpdCoverage) end. %% @private %% Find the vnode that covers the most of the remaining keyspace. Use %% VNode ID + offset (determined by request ID) as the tiebreaker %% (more precisely, the tagged tuple that contains the offset vnode's %% ID) find_best_vnode_for_keyspace(PartitionIDs, Available) -> CoverCount = [{covers(PartitionIDs, CoversKeys), VNode, TieBreaker} || {TieBreaker, VNode, CoversKeys} <- Available], %% Head of the list is the best result unless all of them have %% zero overlap with the partitions for which we need coverage interpret_best_vnode(hd(lists:sort(fun compare_vnode_keyspaces/2, CoverCount))). %% @private interpret_best_vnode({0, _, _}) -> {error, no_coverage}; interpret_best_vnode({_, VNode, _}) -> VNode. %% @private %% There is a potential optimization here once %% the partition claim logic has been changed %% so that physical nodes claim partitions at %% regular intervals around the ring. %% The optimization is for the case %% when the partition count is not evenly divisible %% by the n_val and when the coverage counts of the %% two arguments are equal and a tiebreaker is %% required to determine the sort order. In this %% case, choosing the lower node for the final %% vnode to complete coverage will result %% in an extra physical node being involved %% in the coverage plan so the optimization is %% to choose the upper node to minimize the number %% of physical nodes. compare_vnode_keyspaces({CA, _VA, TBA}, {CB, _VB, TBB}) -> if CA > CB -> %% Descending sort on coverage true; CA < CB -> false; true -> TBA < TBB %% If equal coverage choose the lower node. end. %% @private %% @doc Count how many of CoversKeys appear in KeySpace covers(KeySpace, CoversKeys) -> ordsets:size(ordsets:intersection(KeySpace, CoversKeys)). %% @private %% Determines the number of non-mask bits in the 2^160 keyspace. %% Note that PartitionCount does not have to be ring size; we could be %% creating a coverage plan for subpartitions data_bits(PartitionCount) -> 160 - round(math:log(PartitionCount) / math:log(2)). %% @private next_power_of_two(X) -> round(math:pow(2, round(log2(X - 1) + 0.5))). log2(X) -> math:log(X) / math:log(2.0). %% =================================================================== %% EUnit tests %% =================================================================== -ifdef(TEST). -define(SET(X), ordsets:from_list(X)). bits_test() -> %% 160 - log2(8) ?assertEqual(157, data_bits(8)), %% 160 - log2(65536) ?assertEqual(144, data_bits(65536)). partid(Partitions) -> lists:map(fun({partition_id, P, _RingSize}) -> P end, Partitions). n_keyspaces_test() -> %% First vnode in a cluster with ring size 64 should (with nval 3) %% cover keyspaces 61-63 ?assertEqual([61, 62, 63], partid(n_keyspaces({vnode_id, 0, 64}, 3))), %% 4th vnode in a cluster with ring size 8 should (with nval 5) %% cover the first 3 and last 2 keyspaces ?assertEqual([0, 1, 2, 6, 7], partid(n_keyspaces({vnode_id, 3, 8}, 5))), %% First vnode in a cluster with a single partition should (with %% any nval) cover the only keyspace ?assertEqual([0], partid(n_keyspaces({vnode_id, 0, 1}, 1))). covers_test() -> %% Count the overlap between the sets ?assertEqual(2, covers(?SET([1, 2]), ?SET([0, 1, 2, 3]))), ?assertEqual(1, covers(?SET([1, 2]), ?SET([0, 1]))), ?assertEqual(0, covers(?SET([1, 2, 3]), ?SET([4, 5, 6, 7]))). %% `find_best_vnode_for_keyspace' actually takes tagged tuples, but it %% works fine with plain ol' integers, and the logic is much easier to %% see this way best_vnode_test() -> %% Given two vnodes 0 and 7, pick 0 because it has more of the %% desired keyspaces ?assertEqual(0, find_best_vnode_for_keyspace( ?SET([0, 1, 2, 3, 4]), [{2, 0, ?SET([6, 7, 0, 1, 2])}, {1, 7, ?SET([5, 6, 7, 0, 1])}])), %% Given two vnodes 0 and 7, pick 7 because they cover the same %% keyspaces and 7 has the lower tiebreaker ?assertEqual(7, find_best_vnode_for_keyspace( ?SET([0, 1, 2, 3, 4]), [{2, 0, ?SET([6, 7, 0, 1, 2])}, {1, 7, ?SET([6, 7, 0, 1, 2])}])), %% Given two vnodes 0 and 7, pick 0 because they cover the same %% keyspaces and 0 has the lower tiebreaker ?assertEqual(0, find_best_vnode_for_keyspace( ?SET([0, 1, 2, 3, 4]), [{2, 0, ?SET([6, 7, 0, 1, 2])}, {3, 7, ?SET([6, 7, 0, 1, 2])}])). create_plan_test_() -> {foreach, fun cpsetup/0, [fun test_create_traditional_plan/1, fun test_create_subpartition_plan/1, fun test_replace_traditional/1, fun test_replace_traditional2/1, fun test_replace_traditional3/1, fun test_replace_subpartition/1] }. chash_init() -> {8, [{0,node1}, {182687704666362864775460604089535377456991567872,node2}, {365375409332725729550921208179070754913983135744,node3}, {548063113999088594326381812268606132370974703616,node1}, {730750818665451459101842416358141509827966271488,node2}, {913438523331814323877303020447676887284957839360,node3}, {1096126227998177188652763624537212264741949407232,node1}, {1278813932664540053428224228626747642198940975104,node2}]}. cpsetup() -> CHash = chash_init(), chashbin:create(CHash). test_replace_traditional3(CHBin) -> %% Unlike test_replace_traditional, this function will iterate %% through through N request IDs. Seems like an obvious place to %% use QuickCheck for more coverage OldFilters = [1278813932664540053428224228626747642198940975104, 0], ?_assertMatch([true, true, true], lists:map(fun(N) -> {NewVnodes, NewFilters} = replace_traditional_chunk( 182687704666362864775460604089535377456991567872, node2, OldFilters, 3, N, [], riak_kv, CHBin, fun(_, _) -> [node1, node3] end), equivalent_coverage( OldFilters, NewFilters) andalso %% Make sure none of the new %% vnodes live on our "down" %% node, node3 [] == lists:filter(fun({_, node2}) -> true; (_) -> false end, NewVnodes) end, lists:seq(0, 2))). test_replace_traditional2(CHBin) -> %% Unlike test_replace_traditional, this function will iterate %% through through N request IDs. Seems like an obvious place to %% use QuickCheck for more coverage ?_assertMatch([true, true, true], lists:map(fun(N) -> {NewVnodes, NewFilters} = replace_traditional_chunk( 913438523331814323877303020447676887284957839360, node3, [], 3, N, [], riak_kv, CHBin, fun(_, _) -> [node1, node2] end), equivalent_coverage( 913438523331814323877303020447676887284957839360, NewFilters, 3, CHBin) andalso %% Make sure none of the new %% vnodes live on our "down" %% node, node3 [] == lists:filter(fun({_, node3}) -> true; (_) -> false end, NewVnodes) end, lists:seq(0, 2))). equivalent_coverage(OldVNode, NewFilters, NVal, CHBin) -> RingSize = chashbin:num_partitions(CHBin), PartitionSize = chash:ring_increment(RingSize), OldFilters = test_vnode_to_filters(OldVNode, NVal, RingSize, PartitionSize), equivalent_coverage(OldFilters, NewFilters). equivalent_coverage(OldFilters, NewFilters) -> %% This logic relies on the fact that all of the new vnodes must %% have an explicit filter list since we're replacing a full vnode %% and there is no other vnode which has exactly the same %% partitions ConsolidatedFilters = lists:foldl( fun({_VNode, FilterList}, Acc) -> Acc ++ FilterList end, [], NewFilters), lists:sort(OldFilters) == lists:sort(ConsolidatedFilters). test_vnode_to_filters(Index, NVal, RingSize, PartitionSize) -> lists:map(fun(N) -> (((Index div PartitionSize) + (RingSize - N)) rem RingSize) * PartitionSize end, lists:seq(0, NVal-1)). test_replace_traditional(CHBin) -> %% We're asking or a replacement for the 4th vnode (id 3), with %% nval of 3. This means it is responsible for partitions 0, 1, 2, %% but given the off-by-one behavior for filter lists the implicit %% filters for this vnode are 1, 2, and 3. %% Because we need to replace all 3 partitions that this vnode is %% responsible for, we're going to need at least 2 vnodes in the %% new chunk. %% We're reporting that node2 is the only node online, so the %% vnodes we have available with relevant partitions are limited %% to these two: 182687704666362864775460604089535377456991567872 (id 1) %% and 730750818665451459101842416358141509827966271488 (id 4) %% vnode 1 has partition 0 (filter 1) %% vnode 4 has partitions 1 and 2 (filters 2 and 3) {ExpectedVnodes, ExpectedFilters} = { [ {182687704666362864775460604089535377456991567872, node2}, {730750818665451459101842416358141509827966271488, node2} ], [ {182687704666362864775460604089535377456991567872, [182687704666362864775460604089535377456991567872] }, {730750818665451459101842416358141509827966271488, [365375409332725729550921208179070754913983135744, 548063113999088594326381812268606132370974703616] } ] }, {NewVnodes, NewFilters} = replace_traditional_chunk( 548063113999088594326381812268606132370974703616, node1, [], 3, 0, [], riak_kv, CHBin, fun(_, _) -> [node2] end), [?_assertEqual(ExpectedVnodes, lists:sort(NewVnodes)), ?_assertEqual(ExpectedFilters, lists:sort(NewFilters))]. test_replace_subpartition(CHBin) -> %% Our riak_core_node_watcher replacement function says only node1 %% is up, so the code will have no choice but to give us back the %% only node1 vnode with the zeroth partition NewChunk = [{548063113999088594326381812268606132370974703616, node1, {0, 156}}], [?_assertEqual(NewChunk, replace_subpartition_chunk(182687704666362864775460604089535377456991567872, node2, {0, 156}, 3, 0, [], riak_kv, CHBin, fun(_, _) -> [node1] end))]. test_create_subpartition_plan(CHBin) -> Plan = [ {182687704666362864775460604089535377456991567872, node2, {0, 156}}, {182687704666362864775460604089535377456991567872, node2, {1, 156}}, {365375409332725729550921208179070754913983135744, node3, {2, 156}}, {365375409332725729550921208179070754913983135744, node3, {3, 156}}, {548063113999088594326381812268606132370974703616, node1, {4, 156}}, {548063113999088594326381812268606132370974703616, node1, {5, 156}}, {730750818665451459101842416358141509827966271488, node2, {6, 156}}, {730750818665451459101842416358141509827966271488, node2, {7, 156}}, {913438523331814323877303020447676887284957839360, node3, {8, 156}}, {913438523331814323877303020447676887284957839360, node3, {9, 156}}, {1096126227998177188652763624537212264741949407232, node1, {10, 156}}, {1096126227998177188652763624537212264741949407232, node1, {11, 156}}, {1278813932664540053428224228626747642198940975104, node2, {12, 156}}, {1278813932664540053428224228626747642198940975104, node2, {13, 156}}, {0,node1, {14, 156}}, {0,node1, {15, 156}} ], [?_assertEqual(Plan, create_subpartition_plan(all, 3, 16, 1, 3, riak_kv, CHBin, fun(_, _) -> [node1, node2, node3] end))]. test_create_traditional_plan(CHBin) -> Plan = {[{1278813932664540053428224228626747642198940975104, node2}, {730750818665451459101842416358141509827966271488, node2}, {365375409332725729550921208179070754913983135744, node3}], [{730750818665451459101842416358141509827966271488, [548063113999088594326381812268606132370974703616, 730750818665451459101842416358141509827966271488]}]}, [?_assertEqual(Plan, create_traditional_plan(all, 3, 1, 1234, riak_kv, CHBin, fun(_, _) -> [node1, node2, node3] end))]. -endif.
src/riak_core_coverage_plan.erl
0.708112
0.478224
riak_core_coverage_plan.erl
starcoder
-module(blockchain_assert_loc_v2_SUITE). -include_lib("common_test/include/ct.hrl"). -include_lib("eunit/include/eunit.hrl"). -include("blockchain_vars.hrl"). -include("blockchain_utils.hrl"). -export([all/0, init_per_testcase/2, end_per_testcase/2]). -export([ basic_test/1, bad_owner_sig_test/1, bad_payer_sig_test/1, zero_staking_fee_test/1, same_loc_diff_gain_test/1, same_loc_diff_elevation_test/1, invalid_gain_test/1, insufficient_assert_version_test/1, insufficient_assert_res_test/1, bad_nonce_test/1, min_antenna_gain_not_set_test/1, max_antenna_gain_not_set_test/1 ]). all() -> [ basic_test, bad_owner_sig_test, bad_payer_sig_test, zero_staking_fee_test, same_loc_diff_gain_test, same_loc_diff_elevation_test, invalid_gain_test, insufficient_assert_version_test, insufficient_assert_res_test, bad_nonce_test, min_antenna_gain_not_set_test, max_antenna_gain_not_set_test ]. %%-------------------------------------------------------------------- %% TEST CASE SETUP %%-------------------------------------------------------------------- init_per_testcase(TestCase, Config) -> Config0 = blockchain_ct_utils:init_base_dir_config(?MODULE, TestCase, Config), Balance = 5000, {ok, Sup, {PrivKey, PubKey}, Opts} = test_utils:init(?config(base_dir, Config0)), ExtraVars = extra_vars(TestCase), {ok, GenesisMembers, _GenesisBlock, ConsensusMembers, Keys} = test_utils:init_chain(Balance, {PrivKey, PubKey}, true, ExtraVars), Chain = blockchain_worker:blockchain(), Swarm = blockchain_swarm:swarm(), N = length(ConsensusMembers), % Check ledger to make sure everyone has the right balance Ledger = blockchain:ledger(Chain), Entries = blockchain_ledger_v1:entries(Ledger), _ = lists:foreach( fun(Entry) -> Balance = blockchain_ledger_entry_v1:balance(Entry), 0 = blockchain_ledger_entry_v1:nonce(Entry) end, maps:values(Entries) ), [ {balance, Balance}, {sup, Sup}, {pubkey, PubKey}, {privkey, PrivKey}, {opts, Opts}, {chain, Chain}, {swarm, Swarm}, {n, N}, {consensus_members, ConsensusMembers}, {genesis_members, GenesisMembers}, {ledger, Ledger}, {keys, Keys} | Config0 ]. %%-------------------------------------------------------------------- %% TEST CASE TEARDOWN %%-------------------------------------------------------------------- end_per_testcase(_, Config) -> Sup = ?config(sup, Config), % Make sure blockchain saved on file = in memory case erlang:is_process_alive(Sup) of true -> true = erlang:exit(Sup, normal), ok = test_utils:wait_until(fun() -> false =:= erlang:is_process_alive(Sup) end); false -> ok end, ok. %%-------------------------------------------------------------------- %% TEST CASES %%-------------------------------------------------------------------- basic_test(Config) -> Chain = ?config(chain, Config), ConsensusMembers = ?config(consensus_members, Config), [_, {Owner, {_OwnerPubkey, _OwnerPrivKey, OwnerSigFun}} | _] = ConsensusMembers, %% NOTE: gateway pubkey bin = owner = payer for the tests GatewayPubkeyBin = Owner, Payer = Owner, %% Construct a simple assert_location_v2 transaction NewLoc = 631252734740306943, {_Txn1, Txn2} = base_assert_loc_v2_txn(GatewayPubkeyBin, Owner, Payer, NewLoc, 1, Chain), STxn0 = blockchain_txn_assert_location_v2:sign(Txn2, OwnerSigFun), STxn1 = blockchain_txn_assert_location_v2:sign_payer(STxn0, OwnerSigFun), %% This transaction should be valid ok = blockchain_txn_assert_location_v2:is_valid(STxn1, Chain), %% Check that the transaction propogates properly {ok, Block} = test_utils:create_block(ConsensusMembers, [STxn1]), _ = blockchain_gossip_handler:add_block(Block, Chain, self(), blockchain_swarm:swarm()), ?assertEqual({ok, blockchain_block:hash_block(Block)}, blockchain:head_hash(Chain)), ?assertEqual({ok, Block}, blockchain:head_block(Chain)), ?assertEqual({ok, 2}, blockchain:height(Chain)), ?assertEqual({ok, Block}, blockchain:get_block(2, Chain)), ok. bad_owner_sig_test(Config) -> Chain = ?config(chain, Config), ConsensusMembers = ?config(consensus_members, Config), [ {_NotOwner, {_NotOwnerPubkey, _NotOwnerPrivKey, NotOwnerSigFun}}, {Owner, {_OwnerPubkey, _OwnerPrivKey, OwnerSigFun}} | _ ] = ConsensusMembers, %% NOTE: gateway pubkey bin = owner = payer for the tests GatewayPubkeyBin = Owner, Payer = Owner, %% Construct a simple assert_location_v2 transaction NewLoc = 631252734740306943, {_Txn1, Txn2} = base_assert_loc_v2_txn(GatewayPubkeyBin, Owner, Payer, NewLoc, 1, Chain), STxn0 = blockchain_txn_assert_location_v2:sign(Txn2, NotOwnerSigFun), STxn1 = blockchain_txn_assert_location_v2:sign_payer(STxn0, OwnerSigFun), %% This transaction should be invalid {error, {bad_owner_signature, _}} = blockchain_txn_assert_location_v2:is_valid(STxn1, Chain), ok. bad_payer_sig_test(Config) -> Chain = ?config(chain, Config), ConsensusMembers = ?config(consensus_members, Config), [ {_NotPayer, {_NotPayerPubkey, _NotPayerPrivKey, NotPayerSigFun}}, {Owner, {_OwnerPubkey, _OwnerPrivKey, OwnerSigFun}} | _ ] = ConsensusMembers, %% NOTE: gateway pubkey bin = owner = payer for the tests GatewayPubkeyBin = Owner, Payer = Owner, %% Construct a simple assert_location_v2 transaction NewLoc = 631252734740306943, {_Txn1, Txn2} = base_assert_loc_v2_txn(GatewayPubkeyBin, Owner, Payer, NewLoc, 1, Chain), STxn0 = blockchain_txn_assert_location_v2:sign(Txn2, OwnerSigFun), STxn1 = blockchain_txn_assert_location_v2:sign_payer(STxn0, NotPayerSigFun), %% This transaction should be invalid {error, {bad_payer_signature, _}} = blockchain_txn_assert_location_v2:is_valid(STxn1, Chain), ok. zero_staking_fee_test(Config) -> Chain = ?config(chain, Config), ConsensusMembers = ?config(consensus_members, Config), [_, {Owner, {_OwnerPubkey, _OwnerPrivKey, OwnerSigFun}} | _] = ConsensusMembers, %% NOTE: gateway pubkey bin = owner = payer for the tests GatewayPubkeyBin = Owner, Payer = Owner, %% Construct a simple assert_location_v2 transaction NewLoc = 631252734740306943, {Txn1, _Txn2} = base_assert_loc_v2_txn(GatewayPubkeyBin, Owner, Payer, NewLoc, 1, Chain), %% Zero-ing out the staking_fee ZeroStakingFee = 0, ZeroStakingFeeTxn = blockchain_txn_assert_location_v2:staking_fee(Txn1, ZeroStakingFee), ZeroStakingFeeSTxn0 = blockchain_txn_assert_location_v2:sign(ZeroStakingFeeTxn, OwnerSigFun), ZeroStakingFeeSTxn1 = blockchain_txn_assert_location_v2:sign_payer( ZeroStakingFeeSTxn0, OwnerSigFun ), %% This transaction should be invalid %% Reason: We have zero-ed the staking_fee but a new location was specified {error, {wrong_staking_fee, {assert_location_v2, _, _}}} = blockchain_txn_assert_location_v2:is_valid( ZeroStakingFeeSTxn1, Chain ), ok. same_loc_diff_gain_test(Config) -> Ledger = ?config(ledger, Config), Chain = ?config(chain, Config), ConsensusMembers = ?config(consensus_members, Config), [_, {Owner, {_OwnerPubkey, _OwnerPrivKey, OwnerSigFun}} | _] = ConsensusMembers, %% NOTE: gateway pubkey bin = owner = payer for the tests GatewayPubkeyBin = Owner, Payer = Owner, %% assert that the default gain is set correctly {ok, GW} = blockchain_ledger_v1:find_gateway_info(GatewayPubkeyBin, Ledger), ?assertEqual(?DEFAULT_GAIN, blockchain_ledger_gateway_v2:gain(GW)), ExistingLocation = blockchain_ledger_gateway_v2:location(GW), %% We will not change the location of the gateway but will update the gain %% and supply 0 staking_fee ZeroStakingFee = 0, NewGain = 23, SameLocDiffGainTxn0 = blockchain_txn_assert_location_v2:new( GatewayPubkeyBin, Owner, Payer, ExistingLocation, 1 ), SameLocDiffGainFee = blockchain_txn_assert_location_v2:calculate_fee( SameLocDiffGainTxn0, Chain ), SameLocDiffGainTxn1 = blockchain_txn_assert_location_v2:fee( SameLocDiffGainTxn0, SameLocDiffGainFee ), SameLocDiffGainTxn2 = blockchain_txn_assert_location_v2:gain(SameLocDiffGainTxn1, NewGain), SameLocDiffGainTxn3 = blockchain_txn_assert_location_v2:staking_fee( SameLocDiffGainTxn2, ZeroStakingFee ), SameLocDiffGainSTxn0 = blockchain_txn_assert_location_v2:sign(SameLocDiffGainTxn3, OwnerSigFun), SameLocDiffGainSTxn1 = blockchain_txn_assert_location_v2:sign_payer( SameLocDiffGainSTxn0, OwnerSigFun ), %% This transaction should be valid ok = blockchain_txn_assert_location_v2:is_valid(SameLocDiffGainSTxn1, Chain), %% Check that the transaction propogates properly {ok, Block} = test_utils:create_block(ConsensusMembers, [SameLocDiffGainSTxn1]), _ = blockchain_gossip_handler:add_block(Block, Chain, self(), blockchain_swarm:swarm()), ?assertEqual({ok, blockchain_block:hash_block(Block)}, blockchain:head_hash(Chain)), ?assertEqual({ok, Block}, blockchain:head_block(Chain)), ?assertEqual({ok, 2}, blockchain:height(Chain)), ?assertEqual({ok, Block}, blockchain:get_block(2, Chain)), %% Check that the new elevation reflects properly Ledger1 = blockchain:ledger(Chain), {ok, GW1} = blockchain_ledger_v1:find_gateway_info(GatewayPubkeyBin, Ledger1), ?assertEqual(NewGain, blockchain_ledger_gateway_v2:gain(GW1)), ok. same_loc_diff_elevation_test(Config) -> Ledger = ?config(ledger, Config), Chain = ?config(chain, Config), ConsensusMembers = ?config(consensus_members, Config), [_, {Owner, {_OwnerPubkey, _OwnerPrivKey, OwnerSigFun}} | _] = ConsensusMembers, %% NOTE: gateway pubkey bin = owner = payer for the tests GatewayPubkeyBin = Owner, Payer = Owner, {ok, GW} = blockchain_ledger_v1:find_gateway_info(GatewayPubkeyBin, Ledger), ?assertEqual(?DEFAULT_ELEVATION, blockchain_ledger_gateway_v2:elevation(GW)), ExistingLocation = blockchain_ledger_gateway_v2:location(GW), %% We will not change the location of the gateway but will update the gain %% and supply 0 staking_fee ZeroStakingFee = 0, NewElevation = 5, SameLocDiffElevationTxn0 = blockchain_txn_assert_location_v2:new( GatewayPubkeyBin, Owner, Payer, ExistingLocation, 1 ), SameLocDiffElevationFee = blockchain_txn_assert_location_v2:calculate_fee( SameLocDiffElevationTxn0, Chain ), SameLocDiffElevationTxn1 = blockchain_txn_assert_location_v2:fee( SameLocDiffElevationTxn0, SameLocDiffElevationFee ), SameLocDiffElevationTxn2 = blockchain_txn_assert_location_v2:elevation( SameLocDiffElevationTxn1, NewElevation ), SameLocDiffElevationTxn3 = blockchain_txn_assert_location_v2:staking_fee( SameLocDiffElevationTxn2, ZeroStakingFee ), SameLocDiffElevationSTxn0 = blockchain_txn_assert_location_v2:sign( SameLocDiffElevationTxn3, OwnerSigFun ), SameLocDiffElevationSTxn1 = blockchain_txn_assert_location_v2:sign_payer( SameLocDiffElevationSTxn0, OwnerSigFun ), %% This transaction should be valid ok = blockchain_txn_assert_location_v2:is_valid(SameLocDiffElevationSTxn1, Chain), %% Check that the transaction propogates properly {ok, Block} = test_utils:create_block(ConsensusMembers, [SameLocDiffElevationSTxn1]), _ = blockchain_gossip_handler:add_block(Block, Chain, self(), blockchain_swarm:swarm()), ?assertEqual({ok, blockchain_block:hash_block(Block)}, blockchain:head_hash(Chain)), ?assertEqual({ok, Block}, blockchain:head_block(Chain)), ?assertEqual({ok, 2}, blockchain:height(Chain)), ?assertEqual({ok, Block}, blockchain:get_block(2, Chain)), %% Check that the new elevation reflects properly Ledger1 = blockchain:ledger(Chain), {ok, GW1} = blockchain_ledger_v1:find_gateway_info(GatewayPubkeyBin, Ledger1), ?assertEqual(NewElevation, blockchain_ledger_gateway_v2:elevation(GW1)), ok. invalid_gain_test(Config) -> Ledger = ?config(ledger, Config), Chain = ?config(chain, Config), ConsensusMembers = ?config(consensus_members, Config), [_, {Owner, {_OwnerPubkey, _OwnerPrivKey, OwnerSigFun}} | _] = ConsensusMembers, %% NOTE: gateway pubkey bin = owner = payer for the tests GatewayPubkeyBin = Owner, Payer = Owner, %% assert that the default gain is set correctly {ok, GW} = blockchain_ledger_v1:find_gateway_info(GatewayPubkeyBin, Ledger), ?assertEqual(?DEFAULT_GAIN, blockchain_ledger_gateway_v2:gain(GW)), ExistingLocation = blockchain_ledger_gateway_v2:location(GW), %% We will not change the location of the gateway and supply 0 staking_fee ZeroStakingFee = 0, %% NOTE: Supply a gain value outside of the allowed range (16 dbi = 160) NewGain = 160, SameLocDiffGainTxn0 = blockchain_txn_assert_location_v2:new( GatewayPubkeyBin, Owner, Payer, ExistingLocation, 1 ), SameLocDiffGainFee = blockchain_txn_assert_location_v2:calculate_fee( SameLocDiffGainTxn0, Chain ), SameLocDiffGainTxn1 = blockchain_txn_assert_location_v2:fee( SameLocDiffGainTxn0, SameLocDiffGainFee ), SameLocDiffGainTxn2 = blockchain_txn_assert_location_v2:gain(SameLocDiffGainTxn1, NewGain), SameLocDiffGainTxn3 = blockchain_txn_assert_location_v2:staking_fee( SameLocDiffGainTxn2, ZeroStakingFee ), SameLocDiffGainSTxn0 = blockchain_txn_assert_location_v2:sign(SameLocDiffGainTxn3, OwnerSigFun), SameLocDiffGainSTxn1 = blockchain_txn_assert_location_v2:sign_payer( SameLocDiffGainSTxn0, OwnerSigFun ), %% This transaction should be invalid {error, {invalid_assert_loc_txn_v2, {invalid_antenna_gain, 160, 10, 150}}} = blockchain_txn_assert_location_v2:is_valid( SameLocDiffGainSTxn1, Chain ), ok. insufficient_assert_version_test(Config) -> Chain = ?config(chain, Config), ConsensusMembers = ?config(consensus_members, Config), [_, {Owner, {_OwnerPubkey, _OwnerPrivKey, OwnerSigFun}} | _] = ConsensusMembers, %% NOTE: gateway pubkey bin = owner = payer for the tests GatewayPubkeyBin = Owner, Payer = Owner, %% Construct a simple assert_location_v2 transaction NewLoc = 631252734740306943, {_Txn1, Txn2} = base_assert_loc_v2_txn(GatewayPubkeyBin, Owner, Payer, NewLoc, 1, Chain), STxn0 = blockchain_txn_assert_location_v2:sign(Txn2, OwnerSigFun), STxn1 = blockchain_txn_assert_location_v2:sign_payer(STxn0, OwnerSigFun), {error, {invalid_assert_loc_txn_v2, insufficient_assert_loc_txn_version}} = blockchain_txn_assert_location_v2:is_valid( STxn1, Chain ), ok. insufficient_assert_res_test(Config) -> Chain = ?config(chain, Config), ConsensusMembers = ?config(consensus_members, Config), [_, {Owner, {_OwnerPubkey, _OwnerPrivKey, OwnerSigFun}} | _] = ConsensusMembers, %% NOTE: gateway pubkey bin = owner = payer for the tests GatewayPubkeyBin = Owner, Payer = Owner, %% Construct a simple assert_location_v2 transaction NewLoc = h3:parent(631252734740306943, 9), {_Txn1, Txn2} = base_assert_loc_v2_txn(GatewayPubkeyBin, Owner, Payer, NewLoc, 1, Chain), STxn0 = blockchain_txn_assert_location_v2:sign(Txn2, OwnerSigFun), STxn1 = blockchain_txn_assert_location_v2:sign_payer(STxn0, OwnerSigFun), %% This transaction should be valid {error, {insufficient_assert_res, _}} = blockchain_txn_assert_location_v2:is_valid( STxn1, Chain ), ok. bad_nonce_test(Config) -> Chain = ?config(chain, Config), ConsensusMembers = ?config(consensus_members, Config), [_, {Owner, {_OwnerPubkey, _OwnerPrivKey, OwnerSigFun}} | _] = ConsensusMembers, %% NOTE: gateway pubkey bin = owner = payer for the tests GatewayPubkeyBin = Owner, Payer = Owner, %% Some random bad nonce BadNonce = 10, %% Construct a simple assert_location_v2 transaction NewLoc = 631252734740306943, {_Txn1, Txn2} = base_assert_loc_v2_txn(GatewayPubkeyBin, Owner, Payer, NewLoc, BadNonce, Chain), STxn0 = blockchain_txn_assert_location_v2:sign(Txn2, OwnerSigFun), STxn1 = blockchain_txn_assert_location_v2:sign_payer(STxn0, OwnerSigFun), %% This transaction should be valid {error, {bad_nonce, {assert_location_v2, BadNonce, 0}}} = blockchain_txn_assert_location_v2:is_valid( STxn1, Chain ), ok. min_antenna_gain_not_set_test(Config) -> Chain = ?config(chain, Config), ConsensusMembers = ?config(consensus_members, Config), [_, {Owner, {_OwnerPubkey, _OwnerPrivKey, OwnerSigFun}} | _] = ConsensusMembers, %% NOTE: gateway pubkey bin = owner = payer for the tests GatewayPubkeyBin = Owner, Payer = Owner, %% Construct a simple assert_location_v2 transaction NewLoc = 631252734740306943, {_Txn1, Txn2} = base_assert_loc_v2_txn(GatewayPubkeyBin, Owner, Payer, NewLoc, 1, Chain), STxn0 = blockchain_txn_assert_location_v2:sign(Txn2, OwnerSigFun), STxn1 = blockchain_txn_assert_location_v2:sign_payer(STxn0, OwnerSigFun), {error, {invalid_assert_loc_txn_v2, min_antenna_gain_not_set}} = blockchain_txn_assert_location_v2:is_valid( STxn1, Chain ), ok. max_antenna_gain_not_set_test(Config) -> Chain = ?config(chain, Config), ConsensusMembers = ?config(consensus_members, Config), [_, {Owner, {_OwnerPubkey, _OwnerPrivKey, OwnerSigFun}} | _] = ConsensusMembers, %% NOTE: gateway pubkey bin = owner = payer for the tests GatewayPubkeyBin = Owner, Payer = Owner, %% Construct a simple assert_location_v2 transaction NewLoc = 631252734740306943, {_Txn1, Txn2} = base_assert_loc_v2_txn(GatewayPubkeyBin, Owner, Payer, NewLoc, 1, Chain), STxn0 = blockchain_txn_assert_location_v2:sign(Txn2, OwnerSigFun), STxn1 = blockchain_txn_assert_location_v2:sign_payer(STxn0, OwnerSigFun), %% This transaction should be invalid {error, {invalid_assert_loc_txn_v2, max_antenna_gain_not_set}} = blockchain_txn_assert_location_v2:is_valid( STxn1, Chain ), ok. %%-------------------------------------------------------------------- %% HELPERS %%-------------------------------------------------------------------- base_assert_loc_v2_txn(GatewayPubkeyBin, Owner, Payer, Loc, Nonce, Chain) -> Txn0 = blockchain_txn_assert_location_v2:new(GatewayPubkeyBin, Owner, Payer, Loc, Nonce), Fee = blockchain_txn_assert_location_v2:calculate_fee(Txn0, Chain), SFee = blockchain_txn_assert_location_v2:calculate_staking_fee(Txn0, Chain), Txn1 = blockchain_txn_assert_location_v2:fee(Txn0, Fee), {Txn1, blockchain_txn_assert_location_v2:staking_fee(Txn1, SFee)}. extra_vars(min_antenna_gain_not_set_test) -> #{ txn_fees => true, staking_fee_txn_assert_location_v1 => 1000000, assert_loc_txn_version => 2, max_antenna_gain => 150 }; extra_vars(max_antenna_gain_not_set_test) -> #{ txn_fees => true, staking_fee_txn_assert_location_v1 => 1000000, assert_loc_txn_version => 2, min_antenna_gain => 10 }; extra_vars(insufficient_assert_version_test) -> #{ txn_fees => true, staking_fee_txn_assert_location_v1 => 1000000, assert_loc_txn_version => 1, min_antenna_gain => 10, max_antenna_gain => 150 }; extra_vars(_) -> #{ txn_fees => true, staking_fee_txn_assert_location_v1 => 1000000, assert_loc_txn_version => 2, min_antenna_gain => 10, max_antenna_gain => 150 }.
test/blockchain_assert_loc_v2_SUITE.erl
0.545286
0.601067
blockchain_assert_loc_v2_SUITE.erl
starcoder
%%------------------------------------------------------------------- %% %% Copyright (c) 2015, <NAME> <<EMAIL>> %% %% This file is provided to you under the Apache License, %% Version 2.0 (the "License"); you may not use this file %% except in compliance with the License. You may obtain %% a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, %% software distributed under the License is distributed on an %% "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY %% KIND, either express or implied. See the License for the %% specific language governing permissions and limitations %% under the License. %% %%------------------------------------------------------------------- %% @doc `backoff_supervisor' is a simple supervisor for a single child. If the %% child is not started the supervisor will make attempts to start it, with %% increasing backoff between attempts. If the child returns an error from its %% start function when an automatic start attempt is made the supervisor will %% shutdown. %% %% If the restart type of the child is `permanent' the supervisor will shutdown %% if the child exits with any reason. %% %% If the restart type is `transient' the supervisor will shutdown if the %% child exits abnormally. If a `transient' child exits with a normal reason %% (`normal', `shutdown', `{shutdown, any()}') the child is removed and a new %% child is started after an initial delay. %% %% If the restart type of the child is `temporary' the child is removed on exit %% and a new child is started after an initial delay. %% %% For the supervisor to continue making start attempts the child start function %% must return `ignore'. However if a start attempt is made using %% `start_child/2' an error will be ignored and the supervisor will not %% shutdown. However the current delay between start attempts will be restarted. %% If a child is alive `start_child/2' returns %% `{error, {already_started, Pid}}', where `Pid' is the `pid()' of the child. %% %% `terminate_child/2', `delete_child/2' and `restart_child/2' behave the same %% as the equivalent functions for a `simple_one_for_one' `supervisor'. %% Therefore a `backoff_supervisor' is equivalent to a `simple_one_for_one' %% `supervisor' with `0' maximum restarts - except only one child can be %% running and the supervisor will try to start a child if no child is running %% for a period of time. If one of theses attempts fails the supervisor will %% shutdown. If one of these attempts returns `ignore' the supervisor will %% increase the time period (up to a maximum) and try again. %% %% The specification for a `backoff_supervisor' uses a similar behaviour to a %% supervisor. A callback module must implement a single function, `init/1'. %% `init/1' should return `{ok, {{BackoffType, Start, Max}, [ChildSpec]}}' or %% `ignore'. `BackoffType' is the `backoff' type to use, either `normal' for %% exponential backoff or `jitter' for backoff with random increments. `Start' %% is the initial delay, in milliseconds, before making a first attempt to start %% the child. `Max' is the maximum delay, in milliseconds, between start %% attempts. `ChildSpec' is the child specification of the child, which is a %% `supervisor:child_spec()'. There must be a single `ChildSpec'. In the case of %% `ignore' the `backoff_supervisor' will not start and return `ignore' from %% `start_link/2,3'. %% %% The specification of a `backoff_supervisor' can be changed in the same way as %% a `supervisor' using a code change. The backoff strategy will be used to set %% the next delay, and all subsequent delays, but will not effect an active %% delay. The child specification is updated for an active child and all %% subsequent children. In the case of `ignore' the specification is unchanged. %% %% An example of a `backoff_supervisor' callback module: %% ``` %% -module(backoff_example). %% %% -export([start_link/0]). %% -export([init/1]). %% %% start_link() -> %% backoff_supervisor:start_link(?MODULE, []). %% %% init([]) -> %% {ok, {{jitter, 100, 5000}, %% {name, {mod, fun, []}, transient, 5000, worker, [mod]}}}. %% ''' %% %% When declaring a child spec of a `backoff_supervisor' in its parent %% supervisor the shutdown should be `infinity' and the child type `supervisor'. %% A `backoff_supervisor' should not be used as the top level supervisor in an %% application because that process is assumed to be a `supervisor'. %% %% @see supervisor %% @see backoff -module(backoff_supervisor). -behaviour(gen_server). -define(TIMER_MAX, 4294967295). %% public api -export([start_child/2]). -export([restart_child/2]). -export([terminate_child/2]). -export([delete_child/2]). -export([which_children/1]). -export([count_children/1]). -export([start_link/2]). -export([start_link/3]). %% gen_server api -export([init/1]). -export([handle_call/3]). -export([handle_cast/2]). -export([handle_info/2]). -export([terminate/2]). -export([code_change/3]). %% types -type backoff_supervisor() :: pid() | atom() | {global, any()} | {via, module(), any()}. -type backoff_spec() :: {normal | jitter, pos_integer(), pos_integer() | infinity}. -export_type([backoff_supervisor/0]). -export_type([backoff_spec/0]). -callback init(Args :: any()) -> {ok, {backoff_spec(), [supervisor:child_spec(), ...]}} | ignore. -record(state, {name :: {pid(), module()} | {local, atom()} | {global, any()} | {via, module(), any()}, module :: module(), args :: any(), id :: any(), mfargs :: {module(), atom(), list()}, restart :: temporary | transient | permanent, shutdown :: brutal_kill | timeout(), child_type :: worker | supervisor, modules :: dynamic | [module()], child :: undefined | pid(), child_mfargs :: undefined | {module(), atom(), list()}, start :: pos_integer(), max :: pos_integer() | infinity, backoff_type :: normal | jitter, backoff :: backoff:backoff(), backoff_ref :: undefined | reference()}). %% public api %% @doc Start a child under the supervisor, if one is not already started. %% %% `ExtraArgs' will be appended to the list of arguments in the child %% specification for this child. `ExtraArgs' will only be appended to this %% child, subsequent children started by the supervisor automatically will not %% have `ExtraArgs'. %% %% @see supervisor:start_child/2 -spec start_child(Supervisor, ExtraArgs) -> Result when Supervisor :: backoff_supervisor(), ExtraArgs :: [any()], Result :: supervisor:startchild_ret(). start_child(Supervisor, ExtraArgs) -> call(Supervisor, {start_child, ExtraArgs}). %% @doc Restart a child under the supervisor, always returns %% `{error, simple_one_for_one}'. %% %% @see supervisor:restart_child/2 -spec restart_child(Supervisor, Child) -> {error, simple_one_for_one} when Supervisor :: backoff_supervisor(), Child :: pid(). restart_child(Supervisor, Child) -> call(Supervisor, {restart_child, Child}). %% @doc Terminate the child, `Child', under the supervisor, if it is the %% supervisor's child. %% %% If `Child' is the supervisor's child it will be terminated. A new child will %% automatically be started by the supervisor after an initial delay. %% %% @see supervisor:terminate_child/2 -spec terminate_child(Supervisor, Child) -> ok | {error, not_found} when Supervisor :: backoff_supervisor(), Child :: pid(). terminate_child(Supervisor, Child) -> call(Supervisor, {terminate_child, Child}). %% @doc Delete a child under the supervisor, always returns %% `{error, simple_one_for_one}'. %% %% @see supervisor:delete_child/2 -spec delete_child(Supervisor, Child) -> {error, simple_one_for_one} when Supervisor :: backoff_supervisor(), Child :: pid(). delete_child(Supervisor, Child) -> call(Supervisor, {delete_child, Child}). %% @doc Return a list of child specifications and child processes. %% %% The list will be empty if there is no children under the supervisor, or %% contain the information for the supervisors single child if is alive. The %% name of the child will always be `undefined'. %% %% @see supervisor:which_children/1 -spec which_children(Supervisor) -> [Child] when Supervisor :: backoff_supervisor(), Child :: {undefined, pid(), supervisor | worker, dynamic | [module()]}. which_children(Supervisor) -> call(Supervisor, which_children). %% @doc Return a property list containing counts relating to the supervisor's %% child specifications and child processes. %% %% @see supervisor:count_children/1 -spec count_children(Supervisor) -> [Count, ...] when Supervisor :: backoff_supervisor(), Count :: {specs, 1} | {active, 0 | 1} | {supervisors, 0 | 1} | {workers, 0 | 1}. count_children(Supervisor) -> call(Supervisor, count_children). %% @doc Starts a supervisor with callback module `Module' and argument `Args'. %% %% @see supervisor:start_link/2 -spec start_link(Module, Args) -> {ok, Pid} | ignore | {error, Reason} when Module :: module(), Args :: any(), Pid :: pid(), Reason :: term(). start_link(Module, Args) -> gen_server:start_link(?MODULE, {self, Module, Args}, []). %% @doc Starts a supervisor with name `Name', callback module `Module' and %% argument `Args'. %% %% @see supervisor:start_link/3 -spec start_link(Name, Module, Args) -> {ok, Pid} | ignore | {error, Reason} when Name :: {local, atom()} | {global, any()} | {via, module(), any()}, Module :: module(), Args :: any(), Pid :: pid(), Reason :: term(). start_link(Name, Module, Args) -> gen_server:start_link(Name, ?MODULE, {Name, Module, Args}, []). %% gen_server api %% @private init({self, Module, Args}) -> init({{self(), Module}, Module, Args}); init({Name, Module, Args}) -> process_flag(trap_exit, true), case catch Module:init(Args) of {ok, {BackoffSpec, StartSpec}} -> init(Name, Module, Args, BackoffSpec, StartSpec); ignore -> ignore; {'EXIT', Reason} -> {stop, Reason}; Other -> {stop, {bad_return, {Module, init, Other}}} end. %% @private handle_call({start_child, ExtraArgs}, _, State) -> handle_start_child(ExtraArgs, State); handle_call({restart_child, _}, _, State) -> {reply, {error, simple_one_for_one}, State}; handle_call({terminate_child, Terminate}, _, State) -> handle_terminate_child(Terminate, State); handle_call({delete_child, _}, _, State) -> {reply, {error, simple_one_for_one}, State}; handle_call(which_children, _, State) -> handle_which_children(State); handle_call(count_children, _, State) -> handle_count_children(State). %% @private handle_cast(Cast, State) -> {stop, {bad_cast, Cast}, State}. %% @private handle_info({timeout, TRef, ?MODULE}, State) -> handle_timeout(TRef, State); handle_info({'EXIT', Pid, Reason}, State) -> handle_exit(Pid, Reason, State); handle_info(Msg, State) -> error_logger:error_msg("Backoff Supervisor received unexpected message: ~p", [Msg]), {noreply, State}. %% @private code_change(_, #state{module=Module, args=Args} = State, _) -> case catch Module:init(Args) of {ok, {BackoffSpec, StartSpec}} -> handle_code_change(BackoffSpec, StartSpec, State); ignore -> {ok, State}; {'EXIT', _} = Exit -> {error, Exit}; Other -> {error, {bad_return, {Module, init, Other}}} end. %% @private terminate(_Reason, State) -> shutdown(State). %% Internal call(Supervisor, Request) -> gen_server:call(Supervisor, Request, infinity). init(Name, Module, Args, {_, _, _} = BackoffSpec, StartSpec) -> case check_backoff_spec(BackoffSpec) of ok -> do_init(Name, Module, Args, BackoffSpec, StartSpec); {error, Reason} -> {stop, {backoff_spec, Reason}} end; init(_, _, _, Other, _) -> {stop, {bad_backoff_spec, Other}}. do_init(Name, Module, Args, {BackoffType, Start, Max}, [ChildSpec]) -> case supervisor:check_childspecs([ChildSpec]) of ok -> {Id, MFA, Restart, Shutdown, ChildType, Modules} = ChildSpec, Backoff = backoff_init(BackoffType, Start, Max), BRef = backoff:fire(Backoff), State = #state{name=Name, module=Module, args=Args, backoff_type=BackoffType, start=Start, max=Max, backoff=Backoff, backoff_ref=BRef, id=Id, mfargs=MFA, restart=Restart, shutdown=Shutdown, child_type=ChildType, modules=Modules}, {ok, State}; {error, Reason} -> {stop, {start_spec, Reason}} end; do_init(_, _, _, _, StartSpec) -> {stop, {bad_start_spec, StartSpec}}. check_backoff_spec({BackoffType, _, _}) when not (BackoffType =:= normal orelse BackoffType =:= jitter) -> {error, {invalid_type, BackoffType}}; check_backoff_spec({_, Start, _}) when not (is_integer(Start) andalso Start > 0 andalso Start =< ?TIMER_MAX) -> {error, {invalid_start, Start}}; check_backoff_spec({_, Start, Max}) when not (Max >= Start andalso (is_integer(Max) orelse Max =:= infinity)) -> {error, {invalid_max, Max}}; check_backoff_spec({_, _, _}) -> ok. backoff_init(BackoffType, Start, Max) -> Backoff = backoff:init(Start, min(Max, ?TIMER_MAX), self(), ?MODULE), backoff:type(Backoff, BackoffType). handle_start_child(ExtraArgs, #state{child=undefined, backoff_ref=BRef} = State) when is_reference(BRef) -> _ = erlang:cancel_timer(BRef), do_start_child(ExtraArgs, State#state{backoff_ref=undefined}); handle_start_child(_, #state{child=Child, backoff_ref=undefined} = State) when is_pid(Child) -> {reply, {error, {already_started, Child}}, State}. do_start_child(ExtraArgs, #state{mfargs={Mod, Fun, Args}} = State) -> NArgs = Args ++ ExtraArgs, case catch apply(Mod, Fun, NArgs) of {ok, Child} = OK when is_pid(Child) -> {reply, OK, started(Child, {Mod, Fun, NArgs}, State)}; {ok, Child, _} = OK when is_pid(Child) -> {reply, OK, started(Child, {Mod, Fun, NArgs}, State)}; ignore -> {reply, {ok, undefined}, fire(State)}; {error, _} = Error -> {reply, Error, fire(State)}; Reason -> {reply, {error, Reason}, fire(State)} end. started(Child, MFA, #state{backoff=Backoff} = State) -> {_, NBackoff} = backoff:succeed(Backoff), State#state{child=Child, child_mfargs=MFA, backoff=NBackoff}. fire(#state{backoff=Backoff} = State) -> State#state{backoff_ref=backoff:fire(Backoff)}. handle_terminate_child(Child, #state{child=Child} = State) when is_pid(Child) -> Reply = shutdown(State), {reply, Reply, fire(State#state{child=undefined, child_mfargs=undefined})}; handle_terminate_child(Pid, State) when is_pid(Pid) -> case erlang:is_process_alive(Pid) of true -> {reply, {error, not_found}, State}; false -> {reply, ok, State} end; handle_terminate_child(_, State) -> {reply, {error, simple_one_for_one}, State}. shutdown(#state{child=Child} = State) when is_pid(Child) -> Monitor = monitor(process, Child), shutdown(Monitor, State); shutdown(#state{child=undefined}) -> ok. shutdown(Monitor, #state{shutdown=brutal_kill, child=Child} = State) -> exit(Child, kill), shutdown_await(Monitor, killed, infinity, State); shutdown(Monitor, #state{shutdown=Timeout, child=Child} = State) -> exit(Child, shutdown), shutdown_await(Monitor, shutdown, Timeout, State). shutdown_await(Monitor, Reason, Timeout, #state{child=Child, restart=Restart} = State) -> case do_shutdown_await(Monitor, Child, Reason, Timeout) of ok -> ok; {error, normal} when Restart =/= permanent -> ok; {error, Reason2} -> supervisor_report(shutdown_error, Reason2, State) end. do_shutdown_await(Monitor, Pid, Reason, Timeout) -> receive {'DOWN', Monitor, _, _, Reason} -> ok; {'DOWN', Monitor, _, _, Other} -> {error, Other} after Timeout -> shutdown_timeout(Monitor, Pid) end. shutdown_timeout(Monitor, Pid) -> exit(Pid, kill), receive {'DOWN', Monitor, _, _, Reason} -> {error, Reason} end. supervisor_report(Error, Reason, #state{name=Name} = State) -> Report = [{supervisor, Name}, {errorContext, Error}, {reason, Reason}, {offender, child(State)}], error_logger:error_report(supervisor_report, Report). child(#state{child_mfargs=undefined, mfargs=MFA} = State) -> child(State#state{child_mfargs=MFA}); child(#state{child=Child, child_mfargs=MFA, id=Id, restart=Restart, shutdown=Shutdown, child_type=ChildType}) -> [{pid, Child}, {name, Id}, {mfargs, MFA}, {restart_type, Restart}, {shutdown, Shutdown}, {child_type, ChildType}]. handle_which_children(#state{child=Child, child_type=ChildType, modules=Modules} = State) when is_pid(Child) -> {reply, [{undefined, Child, ChildType, Modules}], State}; handle_which_children(#state{child=undefined} = State) -> {reply, [], State}. handle_count_children(State) -> {reply, [{specs, 1} | do_count_children(State)], State}. do_count_children(#state{child=Child, child_type=supervisor}) when is_pid(Child) -> [{active, 1}, {supervisors, 1}, {workers, 0}]; do_count_children(#state{child=Child, child_type=worker}) when is_pid(Child) -> [{active, 1}, {supervisors, 0}, {workers, 1}]; do_count_children(#state{child=undefined}) -> [{active, 0}, {supervisors, 0}, {workers, 0}]. handle_timeout(BRef, #state{backoff_ref=BRef, mfargs={Mod, Fun, Args}} = State) when is_reference(BRef) -> NState = State#state{backoff_ref=undefined}, case catch apply(Mod, Fun, Args) of {ok, Child} when is_pid(Child) -> restarted(Child, NState); {ok, Child, _} when is_pid(Child) -> restarted(Child, NState); ignore -> backoff(NState); {error, Reason} -> failed(Reason, NState); Reason -> failed(Reason, NState) end; handle_timeout(_, State) -> {noreply, State}. restarted(Child, #state{mfargs=MFA} = State) -> {noreply, started(Child, MFA, State)}. backoff(#state{backoff=Backoff} = State) -> {_, NBackoff} = backoff:fail(Backoff), {noreply, fire(State#state{backoff=NBackoff})}. failed(Reason, #state{id=Id, mfargs=MFA} = State) -> supervisor_report(start_error, Reason, State#state{child_mfargs=MFA}), NReason = {shutdown, {failed_to_start_child, Id, Reason}}, {stop, NReason, State}. handle_exit(Child, Reason, #state{child=Child} = State) when is_pid(Child) -> terminated(Reason, State); handle_exit(_, _, State) -> {noreply, State}. terminated(Reason, #state{restart=permanent} = State) -> terminated_stop(Reason, State); terminated(normal, State) -> terminated(State); terminated(shutdown, State) -> terminated(State); terminated({shutdown, _}, State) -> terminated(State); terminated(Reason, #state{restart=transient} = State) -> terminated_stop(Reason, State); terminated(Reason, #state{restart=temporary} = State) -> supervisor_report(child_terminated, Reason, State), terminated(State). terminated_stop(Reason, #state{id=Id} = State) -> supervisor_report(child_terminated, Reason, State), supervisor_report(shutdown, reached_max_restart_intensity, State), NReason = {shutdown, {reached_max_restart_intensity, Id, Reason}}, {stop, NReason, State#state{child=undefined, child_mfargs=undefined}}. terminated(State) -> {noreply, fire(State#state{child=undefined, child_mfargs=undefined})}. handle_code_change({_, _, _} = BackoffSpec, StartSpec, State) -> case check_backoff_spec(BackoffSpec) of ok -> NState = change_backoff_spec(BackoffSpec, State), change_start_spec(StartSpec, NState); {error, Reason} -> {error, {backoff_spec, Reason}} end; handle_code_change(Other, _, _) -> {error, {bad_backoff_spec, Other}}. change_backoff_spec({BackoffType, Start, Max}, #state{start=Start, max=Max, backoff=Backoff} = State) -> State#state{backoff=backoff:type(Backoff, BackoffType)}; change_backoff_spec({BackoffType, Start, Max}, State) -> State#state{backoff=backoff_init(BackoffType, Start, Max)}. change_start_spec([ChildSpec], State) -> case supervisor:check_childspecs([ChildSpec]) of ok -> {Id, MFA, Restart, Shutdown, ChildType, Modules} = ChildSpec, NState = State#state{id=Id, mfargs=MFA, restart=Restart, shutdown=Shutdown, child_type=ChildType, modules=Modules}, {ok, NState}; {error, Reason} -> {error, {start_spec, Reason}} end; change_start_spec(Other, _) -> {stop, {bad_start_spec, Other}}.
src/backoff_supervisor.erl
0.716219
0.441492
backoff_supervisor.erl
starcoder
%%%------------------------------------------------------------------------ %% Copyright 2019, OpenTelemetry Authors %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %% %% @doc %% Span behaviour. %% @end %%%------------------------------------------------------------------------- -module(ot_span). -export([start_span/3, end_span/2, get_ctx/2, is_recording_events/2, set_attribute/4, set_attributes/3, add_event/4, add_events/3, set_status/3, update_name/3]). -include("opentelemetry.hrl"). -type start_opts() :: #{parent => undefined | opentelemetry:span() | opentelemetry:span_ctx(), attributes => opentelemetry:attributes(), sampler => ot_sampler:sampler(), sampling_hint => ot_sampler:sampling_decision(), links => opentelemetry:links(), is_recorded => boolean(), start_time => wts:timestamp(), kind => opentelemetry:span_kind()}. -export_type([start_opts/0]). -callback start_span(opentelemetry:span_name(), start_opts()) -> opentelemetry:span_ctx(). -callback end_span(opentelemetry:span_ctx()) -> boolean() | {error, term()}. -callback get_ctx(opentelemetry:span()) -> opentelemetry:span_ctx(). -callback is_recording_events(opentelemetry:span_ctx()) -> boolean(). -callback set_attribute(opentelemetry:span_ctx(), opentelemetry:attribute_key(), opentelemetry:attribute_value()) -> boolean(). -callback set_attributes(opentelemetry:span_ctx(), opentelemetry:attributes()) -> boolean(). -callback add_event(opentelemetry:span_ctx(), unicode:unicode_binary(), opentelemetry:attributes()) -> boolean(). -callback add_events(opentelemetry:span_ctx(), opentelemetry:time_events()) -> boolean(). -callback set_status(opentelemetry:span_ctx(), opentelemetry:status()) -> boolean(). -callback update_name(opentelemetry:span_ctx(), opentelemetry:span_name()) -> boolean(). %% handy macros so we don't have function name typos -define(DO(Tracer, SpanCtx, Args), do_span_function(?FUNCTION_NAME, Tracer, SpanCtx, Args)). -spec start_span(Tracer, SpanName, Opts) -> SpanCtx when Tracer :: opentelemetry:tracer(), SpanName :: opentelemetry:span_name(), Opts :: start_opts(), SpanCtx :: opentelemetry:span_ctx(). start_span(Tracer, SpanName, Opts) -> SpanModule = ot_tracer:span_module(Tracer), SpanModule:start_span(SpanName, Opts). -spec end_span(Tracer, SpanCtx) -> boolean() | {error, term()} when Tracer :: opentelemetry:tracer(), SpanCtx :: opentelemetry:span_ctx(). end_span(Tracer, SpanCtx) -> SpanModule = ot_tracer:span_module(Tracer), SpanModule:end_span(SpanCtx). -spec get_ctx(Tracer, Span) -> SpanCtx when Tracer :: opentelemetry:tracer(), Span :: opentelemetry:span(), SpanCtx :: opentelemetry:span_ctx(). get_ctx(Tracer, Span) -> SpanModule = ot_tracer:span_module(Tracer), SpanModule:get_ctx(Span). -spec is_recording_events(Tracer, SpanCtx) -> boolean() when Tracer :: opentelemetry:tracer(), SpanCtx :: opentelemetry:span_ctx(). is_recording_events(Tracer, SpanCtx) -> ?DO(Tracer, SpanCtx, []). -spec set_attribute(Tracer, SpanCtx, Key, Value) -> boolean() when Tracer :: opentelemetry:tracer(), Key :: opentelemetry:attribute_key(), Value :: opentelemetry:attribute_value(), SpanCtx :: opentelemetry:span_ctx(). set_attribute(Tracer, SpanCtx, Key, Value) -> ?DO(Tracer, SpanCtx, [Key, Value]). -spec set_attributes(Tracer, SpanCtx, Attributes) -> boolean() when Tracer :: opentelemetry:tracer(), Attributes :: opentelemetry:attributes(), SpanCtx :: opentelemetry:span_ctx(). set_attributes(Tracer, SpanCtx, Attributes) -> ?DO(Tracer, SpanCtx, [Attributes]). -spec add_event(Tracer, SpanCtx, Name, Attributes) -> boolean() when Tracer :: opentelemetry:tracer(), Name :: unicode:unicode_binary(), Attributes :: opentelemetry:attributes(), SpanCtx :: opentelemetry:span_ctx(). add_event(Tracer, SpanCtx, Name, Attributes) -> ?DO(Tracer, SpanCtx, [Name, Attributes]). -spec add_events(Tracer, SpanCtx, Events) -> boolean() when Tracer :: opentelemetry:tracer(), Events :: opentelemetry:events(), SpanCtx :: opentelemetry:span_ctx(). add_events(Tracer, SpanCtx, Events) -> ?DO(Tracer, SpanCtx, [Events]). -spec set_status(Tracer, SpanCtx, Status) -> boolean() when Tracer :: opentelemetry:tracer(), Status :: opentelemetry:status(), SpanCtx :: opentelemetry:span_ctx(). set_status(Tracer, SpanCtx, Status) -> ?DO(Tracer, SpanCtx, [Status]). -spec update_name(Tracer, SpanCtx, Name) -> boolean() when Tracer :: opentelemetry:tracer(), Name :: opentelemetry:span_name(), SpanCtx :: opentelemetry:span_ctx(). update_name(Tracer, SpanCtx, SpanName) -> ?DO(Tracer, SpanCtx, [SpanName]). %% internal functions do_span_function(Function, Tracer, SpanCtx, Args) -> SpanModule = ot_tracer:span_module(Tracer), apply_span_function(SpanModule, Function, [SpanCtx | Args]). apply_span_function(ot_span_noop, _Function, _Args) -> ok; apply_span_function(SpanModule, Function, Args) -> erlang:apply(SpanModule, Function, Args).
src/ot_span.erl
0.686055
0.450964
ot_span.erl
starcoder
% Licensed under the Apache License, Version 2.0 (the "License"); you may not % use this file except in compliance with the License. You may obtain a copy of % the License at % % http://www.apache.org/licenses/LICENSE-2.0 % % Unless required by applicable law or agreed to in writing, software % distributed under the License is distributed on an "AS IS" BASIS, WITHOUT % WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the % License for the specific language governing permissions and limitations under % the License. -module(couch_btree). -export([open/2, open/3, query_modify/4, add/2, add_remove/3]). -export([fold/4, full_reduce/1, final_reduce/2, size/1, foldl/3, foldl/4]). -export([fold_reduce/4, lookup/2, get_state/1, set_options/2]). -export([less/3]). -include("couch_db.hrl"). -define(CHUNK_THRESHOLD, 16#4ff). extract(#btree{extract_kv=Extract}, Value) -> Extract(Value). assemble(#btree{assemble_kv=Assemble}, Key, Value) -> Assemble(Key, Value). less(#btree{less=Less}, A, B) -> Less(A, B). % pass in 'nil' for State if a new Btree. open(State, Fd) -> {ok, #btree{root=State, fd=Fd}}. set_options(Bt, []) -> Bt; set_options(Bt, [{split, Extract}|Rest]) -> set_options(Bt#btree{extract_kv=Extract}, Rest); set_options(Bt, [{join, Assemble}|Rest]) -> set_options(Bt#btree{assemble_kv=Assemble}, Rest); set_options(Bt, [{less, Less}|Rest]) -> set_options(Bt#btree{less=Less}, Rest); set_options(Bt, [{reduce, Reduce}|Rest]) -> set_options(Bt#btree{reduce=Reduce}, Rest); set_options(Bt, [{compression, Comp}|Rest]) -> set_options(Bt#btree{compression=Comp}, Rest). open(State, Fd, Options) -> {ok, set_options(#btree{root=State, fd=Fd}, Options)}. get_state(#btree{root=Root}) -> Root. final_reduce(#btree{reduce=Reduce}, Val) -> final_reduce(Reduce, Val); final_reduce(Reduce, {[], []}) -> Reduce(reduce, []); final_reduce(_Bt, {[], [Red]}) -> Red; final_reduce(Reduce, {[], Reductions}) -> Reduce(rereduce, Reductions); final_reduce(Reduce, {KVs, Reductions}) -> Red = Reduce(reduce, KVs), final_reduce(Reduce, {[], [Red | Reductions]}). fold_reduce(#btree{root=Root}=Bt, Fun, Acc, Options) -> Dir = couch_util:get_value(dir, Options, fwd), StartKey = couch_util:get_value(start_key, Options), EndKey = case couch_util:get_value(end_key_gt, Options) of undefined -> couch_util:get_value(end_key, Options); LastKey -> LastKey end, KeyGroupFun = couch_util:get_value(key_group_fun, Options, fun(_,_) -> true end), {StartKey2, EndKey2} = case Dir of rev -> {EndKey, StartKey}; fwd -> {StartKey, EndKey} end, try {ok, Acc2, GroupedRedsAcc2, GroupedKVsAcc2, GroupedKey2} = reduce_stream_node(Bt, Dir, Root, StartKey2, EndKey2, undefined, [], [], KeyGroupFun, Fun, Acc), if GroupedKey2 == undefined -> {ok, Acc2}; true -> case Fun(GroupedKey2, {GroupedKVsAcc2, GroupedRedsAcc2}, Acc2) of {ok, Acc3} -> {ok, Acc3}; {stop, Acc3} -> {ok, Acc3} end end catch throw:{stop, AccDone} -> {ok, AccDone} end. full_reduce(#btree{root=nil,reduce=Reduce}) -> {ok, Reduce(reduce, [])}; full_reduce(#btree{root=Root}) -> {ok, element(2, Root)}. size(#btree{root = nil}) -> 0; size(#btree{root = {_P, _Red}}) -> % pre 1.2 format nil; size(#btree{root = {_P, _Red, Size}}) -> Size. % wraps a 2 arity function with the proper 3 arity function convert_fun_arity(Fun) when is_function(Fun, 2) -> fun (visit, KV, _Reds, AccIn) -> Fun(KV, AccIn); (traverse, _K, _Red, AccIn) -> {ok, AccIn} end; convert_fun_arity(Fun) when is_function(Fun, 3) -> fun (visit, KV, Reds, AccIn) -> Fun(KV, Reds, AccIn); (traverse, _K, _Red, AccIn) -> {ok, AccIn} end; convert_fun_arity(Fun) when is_function(Fun, 4) -> Fun. % Already arity 4 make_key_in_end_range_function(#btree{less=Less}, fwd, Options) -> case couch_util:get_value(end_key_gt, Options) of undefined -> case couch_util:get_value(end_key, Options) of undefined -> fun(_Key) -> true end; LastKey -> fun(Key) -> not Less(LastKey, Key) end end; EndKey -> fun(Key) -> Less(Key, EndKey) end end; make_key_in_end_range_function(#btree{less=Less}, rev, Options) -> case couch_util:get_value(end_key_gt, Options) of undefined -> case couch_util:get_value(end_key, Options) of undefined -> fun(_Key) -> true end; LastKey -> fun(Key) -> not Less(Key, LastKey) end end; EndKey -> fun(Key) -> Less(EndKey, Key) end end. foldl(Bt, Fun, Acc) -> fold(Bt, Fun, Acc, []). foldl(Bt, Fun, Acc, Options) -> fold(Bt, Fun, Acc, Options). fold(#btree{root=nil}, _Fun, Acc, _Options) -> {ok, {[], []}, Acc}; fold(#btree{root=Root}=Bt, Fun, Acc, Options) -> Dir = couch_util:get_value(dir, Options, fwd), InRange = make_key_in_end_range_function(Bt, Dir, Options), Result = case couch_util:get_value(start_key, Options) of undefined -> stream_node(Bt, [], Bt#btree.root, InRange, Dir, convert_fun_arity(Fun), Acc); StartKey -> stream_node(Bt, [], Bt#btree.root, StartKey, InRange, Dir, convert_fun_arity(Fun), Acc) end, case Result of {ok, Acc2}-> FullReduction = element(2, Root), {ok, {[], [FullReduction]}, Acc2}; {stop, LastReduction, Acc2} -> {ok, LastReduction, Acc2} end. add(Bt, InsertKeyValues) -> add_remove(Bt, InsertKeyValues, []). add_remove(Bt, InsertKeyValues, RemoveKeys) -> {ok, [], Bt2} = query_modify(Bt, [], InsertKeyValues, RemoveKeys), {ok, Bt2}. query_modify(Bt, LookupKeys, InsertValues, RemoveKeys) -> #btree{root=Root} = Bt, InsertActions = lists:map( fun(KeyValue) -> {Key, Value} = extract(Bt, KeyValue), {insert, Key, Value} end, InsertValues), RemoveActions = [{remove, Key, nil} || Key <- RemoveKeys], FetchActions = [{fetch, Key, nil} || Key <- LookupKeys], SortFun = fun({OpA, A, _}, {OpB, B, _}) -> case A == B of % A and B are equal, sort by op. true -> op_order(OpA) < op_order(OpB); false -> less(Bt, A, B) end end, Actions = lists:sort(SortFun, lists:append([InsertActions, RemoveActions, FetchActions])), {ok, KeyPointers, QueryResults} = modify_node(Bt, Root, Actions, []), {ok, NewRoot} = complete_root(Bt, KeyPointers), {ok, QueryResults, Bt#btree{root=NewRoot}}. % for ordering different operations with the same key. % fetch < remove < insert op_order(fetch) -> 1; op_order(remove) -> 2; op_order(insert) -> 3. lookup(#btree{root=Root, less=Less}=Bt, Keys) -> SortedKeys = lists:sort(Less, Keys), {ok, SortedResults} = lookup(Bt, Root, SortedKeys), % We want to return the results in the same order as the keys were input % but we may have changed the order when we sorted. So we need to put the % order back into the results. couch_util:reorder_results(Keys, SortedResults). lookup(_Bt, nil, Keys) -> {ok, [{Key, not_found} || Key <- Keys]}; lookup(Bt, Node, Keys) -> Pointer = element(1, Node), {NodeType, NodeList} = get_node(Bt, Pointer), case NodeType of kp_node -> lookup_kpnode(Bt, list_to_tuple(NodeList), 1, Keys, []); kv_node -> lookup_kvnode(Bt, list_to_tuple(NodeList), 1, Keys, []) end. lookup_kpnode(_Bt, _NodeTuple, _LowerBound, [], Output) -> {ok, lists:reverse(Output)}; lookup_kpnode(_Bt, NodeTuple, LowerBound, Keys, Output) when tuple_size(NodeTuple) < LowerBound -> {ok, lists:reverse(Output, [{Key, not_found} || Key <- Keys])}; lookup_kpnode(Bt, NodeTuple, LowerBound, [FirstLookupKey | _] = LookupKeys, Output) -> N = find_first_gteq(Bt, NodeTuple, LowerBound, tuple_size(NodeTuple), FirstLookupKey), {Key, PointerInfo} = element(N, NodeTuple), SplitFun = fun(LookupKey) -> not less(Bt, Key, LookupKey) end, case lists:splitwith(SplitFun, LookupKeys) of {[], GreaterQueries} -> lookup_kpnode(Bt, NodeTuple, N + 1, GreaterQueries, Output); {LessEqQueries, GreaterQueries} -> {ok, Results} = lookup(Bt, PointerInfo, LessEqQueries), lookup_kpnode(Bt, NodeTuple, N + 1, GreaterQueries, lists:reverse(Results, Output)) end. lookup_kvnode(_Bt, _NodeTuple, _LowerBound, [], Output) -> {ok, lists:reverse(Output)}; lookup_kvnode(_Bt, NodeTuple, LowerBound, Keys, Output) when tuple_size(NodeTuple) < LowerBound -> % keys not found {ok, lists:reverse(Output, [{Key, not_found} || Key <- Keys])}; lookup_kvnode(Bt, NodeTuple, LowerBound, [LookupKey | RestLookupKeys], Output) -> N = find_first_gteq(Bt, NodeTuple, LowerBound, tuple_size(NodeTuple), LookupKey), {Key, Value} = element(N, NodeTuple), case less(Bt, LookupKey, Key) of true -> % LookupKey is less than Key lookup_kvnode(Bt, NodeTuple, N, RestLookupKeys, [{LookupKey, not_found} | Output]); false -> case less(Bt, Key, LookupKey) of true -> % LookupKey is greater than Key lookup_kvnode(Bt, NodeTuple, N+1, RestLookupKeys, [{LookupKey, not_found} | Output]); false -> % LookupKey is equal to Key lookup_kvnode(Bt, NodeTuple, N, RestLookupKeys, [{LookupKey, {ok, assemble(Bt, LookupKey, Value)}} | Output]) end end. complete_root(_Bt, []) -> {ok, nil}; complete_root(_Bt, [{_Key, PointerInfo}])-> {ok, PointerInfo}; complete_root(Bt, KPs) -> {ok, ResultKeyPointers} = write_node(Bt, kp_node, KPs), complete_root(Bt, ResultKeyPointers). %%%%%%%%%%%%% The chunkify function sucks! %%%%%%%%%%%%% % It is inaccurate as it does not account for compression when blocks are % written. Plus with the "case byte_size(term_to_binary(InList)) of" code % it's probably really inefficient. chunkify(InList) -> case ?term_size(InList) of Size when Size > ?CHUNK_THRESHOLD -> NumberOfChunksLikely = ((Size div ?CHUNK_THRESHOLD) + 1), ChunkThreshold = Size div NumberOfChunksLikely, chunkify(InList, ChunkThreshold, [], 0, []); _Else -> [InList] end. chunkify([], _ChunkThreshold, [], 0, OutputChunks) -> lists:reverse(OutputChunks); chunkify([], _ChunkThreshold, OutList, _OutListSize, OutputChunks) -> lists:reverse([lists:reverse(OutList) | OutputChunks]); chunkify([InElement | RestInList], ChunkThreshold, OutList, OutListSize, OutputChunks) -> case ?term_size(InElement) of Size when (Size + OutListSize) > ChunkThreshold andalso OutList /= [] -> chunkify(RestInList, ChunkThreshold, [], 0, [lists:reverse([InElement | OutList]) | OutputChunks]); Size -> chunkify(RestInList, ChunkThreshold, [InElement | OutList], OutListSize + Size, OutputChunks) end. modify_node(Bt, RootPointerInfo, Actions, QueryOutput) -> case RootPointerInfo of nil -> NodeType = kv_node, NodeList = []; _Tuple -> Pointer = element(1, RootPointerInfo), {NodeType, NodeList} = get_node(Bt, Pointer) end, NodeTuple = list_to_tuple(NodeList), {ok, NewNodeList, QueryOutput2} = case NodeType of kp_node -> modify_kpnode(Bt, NodeTuple, 1, Actions, [], QueryOutput); kv_node -> modify_kvnode(Bt, NodeTuple, 1, Actions, [], QueryOutput) end, case NewNodeList of [] -> % no nodes remain {ok, [], QueryOutput2}; NodeList -> % nothing changed {LastKey, _LastValue} = element(tuple_size(NodeTuple), NodeTuple), {ok, [{LastKey, RootPointerInfo}], QueryOutput2}; _Else2 -> {ok, ResultList} = write_node(Bt, NodeType, NewNodeList), {ok, ResultList, QueryOutput2} end. reduce_node(#btree{reduce=nil}, _NodeType, _NodeList) -> []; reduce_node(#btree{reduce=R}, kp_node, NodeList) -> R(rereduce, [element(2, Node) || {_K, Node} <- NodeList]); reduce_node(#btree{reduce=R}=Bt, kv_node, NodeList) -> R(reduce, [assemble(Bt, K, V) || {K, V} <- NodeList]). reduce_tree_size(kv_node, NodeSize, _KvList) -> NodeSize; reduce_tree_size(kp_node, NodeSize, []) -> NodeSize; reduce_tree_size(kp_node, _NodeSize, [{_K, {_P, _Red}} | _]) -> % pre 1.2 format nil; reduce_tree_size(kp_node, _NodeSize, [{_K, {_P, _Red, nil}} | _]) -> nil; reduce_tree_size(kp_node, NodeSize, [{_K, {_P, _Red, Sz}} | NodeList]) -> reduce_tree_size(kp_node, NodeSize + Sz, NodeList). get_node(#btree{fd = Fd}, NodePos) -> {ok, {NodeType, NodeList}} = couch_file:pread_term(Fd, NodePos), {NodeType, NodeList}. write_node(#btree{fd = Fd, compression = Comp} = Bt, NodeType, NodeList) -> % split up nodes into smaller sizes NodeListList = chunkify(NodeList), % now write out each chunk and return the KeyPointer pairs for those nodes ResultList = [ begin {ok, Pointer, Size} = couch_file:append_term( Fd, {NodeType, ANodeList}, [{compression, Comp}]), {LastKey, _} = lists:last(ANodeList), SubTreeSize = reduce_tree_size(NodeType, Size, ANodeList), {LastKey, {Pointer, reduce_node(Bt, NodeType, ANodeList), SubTreeSize}} end || ANodeList <- NodeListList ], {ok, ResultList}. modify_kpnode(Bt, {}, _LowerBound, Actions, [], QueryOutput) -> modify_node(Bt, nil, Actions, QueryOutput); modify_kpnode(_Bt, NodeTuple, LowerBound, [], ResultNode, QueryOutput) -> {ok, lists:reverse(ResultNode, bounded_tuple_to_list(NodeTuple, LowerBound, tuple_size(NodeTuple), [])), QueryOutput}; modify_kpnode(Bt, NodeTuple, LowerBound, [{_, FirstActionKey, _}|_]=Actions, ResultNode, QueryOutput) -> Sz = tuple_size(NodeTuple), N = find_first_gteq(Bt, NodeTuple, LowerBound, Sz, FirstActionKey), case N =:= Sz of true -> % perform remaining actions on last node {_, PointerInfo} = element(Sz, NodeTuple), {ok, ChildKPs, QueryOutput2} = modify_node(Bt, PointerInfo, Actions, QueryOutput), NodeList = lists:reverse(ResultNode, bounded_tuple_to_list(NodeTuple, LowerBound, Sz - 1, ChildKPs)), {ok, NodeList, QueryOutput2}; false -> {NodeKey, PointerInfo} = element(N, NodeTuple), SplitFun = fun({_ActionType, ActionKey, _ActionValue}) -> not less(Bt, NodeKey, ActionKey) end, {LessEqQueries, GreaterQueries} = lists:splitwith(SplitFun, Actions), {ok, ChildKPs, QueryOutput2} = modify_node(Bt, PointerInfo, LessEqQueries, QueryOutput), ResultNode2 = lists:reverse(ChildKPs, bounded_tuple_to_revlist(NodeTuple, LowerBound, N - 1, ResultNode)), modify_kpnode(Bt, NodeTuple, N+1, GreaterQueries, ResultNode2, QueryOutput2) end. bounded_tuple_to_revlist(_Tuple, Start, End, Tail) when Start > End -> Tail; bounded_tuple_to_revlist(Tuple, Start, End, Tail) -> bounded_tuple_to_revlist(Tuple, Start+1, End, [element(Start, Tuple)|Tail]). bounded_tuple_to_list(Tuple, Start, End, Tail) -> bounded_tuple_to_list2(Tuple, Start, End, [], Tail). bounded_tuple_to_list2(_Tuple, Start, End, Acc, Tail) when Start > End -> lists:reverse(Acc, Tail); bounded_tuple_to_list2(Tuple, Start, End, Acc, Tail) -> bounded_tuple_to_list2(Tuple, Start + 1, End, [element(Start, Tuple) | Acc], Tail). find_first_gteq(_Bt, _Tuple, Start, End, _Key) when Start == End -> End; find_first_gteq(Bt, Tuple, Start, End, Key) -> Mid = Start + ((End - Start) div 2), {TupleKey, _} = element(Mid, Tuple), case less(Bt, TupleKey, Key) of true -> find_first_gteq(Bt, Tuple, Mid+1, End, Key); false -> find_first_gteq(Bt, Tuple, Start, Mid, Key) end. modify_kvnode(_Bt, NodeTuple, LowerBound, [], ResultNode, QueryOutput) -> {ok, lists:reverse(ResultNode, bounded_tuple_to_list(NodeTuple, LowerBound, tuple_size(NodeTuple), [])), QueryOutput}; modify_kvnode(Bt, NodeTuple, LowerBound, [{ActionType, ActionKey, ActionValue} | RestActions], ResultNode, QueryOutput) when LowerBound > tuple_size(NodeTuple) -> case ActionType of insert -> modify_kvnode(Bt, NodeTuple, LowerBound, RestActions, [{ActionKey, ActionValue} | ResultNode], QueryOutput); remove -> % just drop the action modify_kvnode(Bt, NodeTuple, LowerBound, RestActions, ResultNode, QueryOutput); fetch -> % the key/value must not exist in the tree modify_kvnode(Bt, NodeTuple, LowerBound, RestActions, ResultNode, [{not_found, {ActionKey, nil}} | QueryOutput]) end; modify_kvnode(Bt, NodeTuple, LowerBound, [{ActionType, ActionKey, ActionValue} | RestActions], AccNode, QueryOutput) -> N = find_first_gteq(Bt, NodeTuple, LowerBound, tuple_size(NodeTuple), ActionKey), {Key, Value} = element(N, NodeTuple), ResultNode = bounded_tuple_to_revlist(NodeTuple, LowerBound, N - 1, AccNode), case less(Bt, ActionKey, Key) of true -> case ActionType of insert -> % ActionKey is less than the Key, so insert modify_kvnode(Bt, NodeTuple, N, RestActions, [{ActionKey, ActionValue} | ResultNode], QueryOutput); remove -> % ActionKey is less than the Key, just drop the action modify_kvnode(Bt, NodeTuple, N, RestActions, ResultNode, QueryOutput); fetch -> % ActionKey is less than the Key, the key/value must not exist in the tree modify_kvnode(Bt, NodeTuple, N, RestActions, ResultNode, [{not_found, {ActionKey, nil}} | QueryOutput]) end; false -> % ActionKey and Key are maybe equal. case less(Bt, Key, ActionKey) of false -> case ActionType of insert -> modify_kvnode(Bt, NodeTuple, N+1, RestActions, [{ActionKey, ActionValue} | ResultNode], QueryOutput); remove -> modify_kvnode(Bt, NodeTuple, N+1, RestActions, ResultNode, QueryOutput); fetch -> % ActionKey is equal to the Key, insert into the QueryOuput, but re-process the node % since an identical action key can follow it. modify_kvnode(Bt, NodeTuple, N, RestActions, ResultNode, [{ok, assemble(Bt, Key, Value)} | QueryOutput]) end; true -> modify_kvnode(Bt, NodeTuple, N + 1, [{ActionType, ActionKey, ActionValue} | RestActions], [{Key, Value} | ResultNode], QueryOutput) end end. reduce_stream_node(_Bt, _Dir, nil, _KeyStart, _KeyEnd, GroupedKey, GroupedKVsAcc, GroupedRedsAcc, _KeyGroupFun, _Fun, Acc) -> {ok, Acc, GroupedRedsAcc, GroupedKVsAcc, GroupedKey}; reduce_stream_node(Bt, Dir, Node, KeyStart, KeyEnd, GroupedKey, GroupedKVsAcc, GroupedRedsAcc, KeyGroupFun, Fun, Acc) -> P = element(1, Node), case get_node(Bt, P) of {kp_node, NodeList} -> reduce_stream_kp_node(Bt, Dir, NodeList, KeyStart, KeyEnd, GroupedKey, GroupedKVsAcc, GroupedRedsAcc, KeyGroupFun, Fun, Acc); {kv_node, KVs} -> reduce_stream_kv_node(Bt, Dir, KVs, KeyStart, KeyEnd, GroupedKey, GroupedKVsAcc, GroupedRedsAcc, KeyGroupFun, Fun, Acc) end. reduce_stream_kv_node(Bt, Dir, KVs, KeyStart, KeyEnd, GroupedKey, GroupedKVsAcc, GroupedRedsAcc, KeyGroupFun, Fun, Acc) -> GTEKeyStartKVs = case KeyStart of undefined -> KVs; _ -> lists:dropwhile(fun({Key,_}) -> less(Bt, Key, KeyStart) end, KVs) end, KVs2 = case KeyEnd of undefined -> GTEKeyStartKVs; _ -> lists:takewhile( fun({Key,_}) -> not less(Bt, KeyEnd, Key) end, GTEKeyStartKVs) end, reduce_stream_kv_node2(Bt, adjust_dir(Dir, KVs2), GroupedKey, GroupedKVsAcc, GroupedRedsAcc, KeyGroupFun, Fun, Acc). reduce_stream_kv_node2(_Bt, [], GroupedKey, GroupedKVsAcc, GroupedRedsAcc, _KeyGroupFun, _Fun, Acc) -> {ok, Acc, GroupedRedsAcc, GroupedKVsAcc, GroupedKey}; reduce_stream_kv_node2(Bt, [{Key, Value}| RestKVs], GroupedKey, GroupedKVsAcc, GroupedRedsAcc, KeyGroupFun, Fun, Acc) -> case GroupedKey of undefined -> reduce_stream_kv_node2(Bt, RestKVs, Key, [assemble(Bt,Key,Value)], [], KeyGroupFun, Fun, Acc); _ -> case KeyGroupFun(GroupedKey, Key) of true -> reduce_stream_kv_node2(Bt, RestKVs, GroupedKey, [assemble(Bt,Key,Value)|GroupedKVsAcc], GroupedRedsAcc, KeyGroupFun, Fun, Acc); false -> case Fun(GroupedKey, {GroupedKVsAcc, GroupedRedsAcc}, Acc) of {ok, Acc2} -> reduce_stream_kv_node2(Bt, RestKVs, Key, [assemble(Bt,Key,Value)], [], KeyGroupFun, Fun, Acc2); {stop, Acc2} -> throw({stop, Acc2}) end end end. reduce_stream_kp_node(Bt, Dir, NodeList, KeyStart, KeyEnd, GroupedKey, GroupedKVsAcc, GroupedRedsAcc, KeyGroupFun, Fun, Acc) -> Nodes = case KeyStart of undefined -> NodeList; _ -> lists:dropwhile( fun({Key,_}) -> less(Bt, Key, KeyStart) end, NodeList) end, NodesInRange = case KeyEnd of undefined -> Nodes; _ -> {InRange, MaybeInRange} = lists:splitwith( fun({Key,_}) -> less(Bt, Key, KeyEnd) end, Nodes), InRange ++ case MaybeInRange of [] -> []; [FirstMaybe|_] -> [FirstMaybe] end end, reduce_stream_kp_node2(Bt, Dir, adjust_dir(Dir, NodesInRange), KeyStart, KeyEnd, GroupedKey, GroupedKVsAcc, GroupedRedsAcc, KeyGroupFun, Fun, Acc). reduce_stream_kp_node2(Bt, Dir, [{_Key, NodeInfo} | RestNodeList], KeyStart, KeyEnd, undefined, [], [], KeyGroupFun, Fun, Acc) -> {ok, Acc2, GroupedRedsAcc2, GroupedKVsAcc2, GroupedKey2} = reduce_stream_node(Bt, Dir, NodeInfo, KeyStart, KeyEnd, undefined, [], [], KeyGroupFun, Fun, Acc), reduce_stream_kp_node2(Bt, Dir, RestNodeList, KeyStart, KeyEnd, GroupedKey2, GroupedKVsAcc2, GroupedRedsAcc2, KeyGroupFun, Fun, Acc2); reduce_stream_kp_node2(Bt, Dir, NodeList, KeyStart, KeyEnd, GroupedKey, GroupedKVsAcc, GroupedRedsAcc, KeyGroupFun, Fun, Acc) -> {Grouped0, Ungrouped0} = lists:splitwith(fun({Key,_}) -> KeyGroupFun(GroupedKey, Key) end, NodeList), {GroupedNodes, UngroupedNodes} = case Grouped0 of [] -> {Grouped0, Ungrouped0}; _ -> [FirstGrouped | RestGrouped] = lists:reverse(Grouped0), {RestGrouped, [FirstGrouped | Ungrouped0]} end, GroupedReds = [element(2, Node) || {_, Node} <- GroupedNodes], case UngroupedNodes of [{_Key, NodeInfo}|RestNodes] -> {ok, Acc2, GroupedRedsAcc2, GroupedKVsAcc2, GroupedKey2} = reduce_stream_node(Bt, Dir, NodeInfo, KeyStart, KeyEnd, GroupedKey, GroupedKVsAcc, GroupedReds ++ GroupedRedsAcc, KeyGroupFun, Fun, Acc), reduce_stream_kp_node2(Bt, Dir, RestNodes, KeyStart, KeyEnd, GroupedKey2, GroupedKVsAcc2, GroupedRedsAcc2, KeyGroupFun, Fun, Acc2); [] -> {ok, Acc, GroupedReds ++ GroupedRedsAcc, GroupedKVsAcc, GroupedKey} end. adjust_dir(fwd, List) -> List; adjust_dir(rev, List) -> lists:reverse(List). stream_node(Bt, Reds, Node, StartKey, InRange, Dir, Fun, Acc) -> Pointer = element(1, Node), {NodeType, NodeList} = get_node(Bt, Pointer), case NodeType of kp_node -> stream_kp_node(Bt, Reds, adjust_dir(Dir, NodeList), StartKey, InRange, Dir, Fun, Acc); kv_node -> stream_kv_node(Bt, Reds, adjust_dir(Dir, NodeList), StartKey, InRange, Dir, Fun, Acc) end. stream_node(Bt, Reds, Node, InRange, Dir, Fun, Acc) -> Pointer = element(1, Node), {NodeType, NodeList} = get_node(Bt, Pointer), case NodeType of kp_node -> stream_kp_node(Bt, Reds, adjust_dir(Dir, NodeList), InRange, Dir, Fun, Acc); kv_node -> stream_kv_node2(Bt, Reds, [], adjust_dir(Dir, NodeList), InRange, Dir, Fun, Acc) end. stream_kp_node(_Bt, _Reds, [], _InRange, _Dir, _Fun, Acc) -> {ok, Acc}; stream_kp_node(Bt, Reds, [{Key, Node} | Rest], InRange, Dir, Fun, Acc) -> Red = element(2, Node), case Fun(traverse, Key, Red, Acc) of {ok, Acc2} -> case stream_node(Bt, Reds, Node, InRange, Dir, Fun, Acc2) of {ok, Acc3} -> stream_kp_node(Bt, [Red | Reds], Rest, InRange, Dir, Fun, Acc3); {stop, LastReds, Acc3} -> {stop, LastReds, Acc3} end; {skip, Acc2} -> stream_kp_node(Bt, [Red | Reds], Rest, InRange, Dir, Fun, Acc2) end. drop_nodes(_Bt, Reds, _StartKey, []) -> {Reds, []}; drop_nodes(Bt, Reds, StartKey, [{NodeKey, Node} | RestKPs]) -> case less(Bt, NodeKey, StartKey) of true -> drop_nodes(Bt, [element(2, Node) | Reds], StartKey, RestKPs); false -> {Reds, [{NodeKey, Node} | RestKPs]} end. stream_kp_node(Bt, Reds, KPs, StartKey, InRange, Dir, Fun, Acc) -> {NewReds, NodesToStream} = case Dir of fwd -> % drop all nodes sorting before the key drop_nodes(Bt, Reds, StartKey, KPs); rev -> % keep all nodes sorting before the key, AND the first node to sort after RevKPs = lists:reverse(KPs), case lists:splitwith(fun({Key, _Pointer}) -> less(Bt, Key, StartKey) end, RevKPs) of {_RevsBefore, []} -> % everything sorts before it {Reds, KPs}; {RevBefore, [FirstAfter | Drop]} -> {[element(2, Node) || {_K, Node} <- Drop] ++ Reds, [FirstAfter | lists:reverse(RevBefore)]} end end, case NodesToStream of [] -> {ok, Acc}; [{_Key, Node} | Rest] -> case stream_node(Bt, NewReds, Node, StartKey, InRange, Dir, Fun, Acc) of {ok, Acc2} -> Red = element(2, Node), stream_kp_node(Bt, [Red | NewReds], Rest, InRange, Dir, Fun, Acc2); {stop, LastReds, Acc2} -> {stop, LastReds, Acc2} end end. stream_kv_node(Bt, Reds, KVs, StartKey, InRange, Dir, Fun, Acc) -> DropFun = case Dir of fwd -> fun({Key, _}) -> less(Bt, Key, StartKey) end; rev -> fun({Key, _}) -> less(Bt, StartKey, Key) end end, {LTKVs, GTEKVs} = lists:splitwith(DropFun, KVs), AssembleLTKVs = [assemble(Bt,K,V) || {K,V} <- LTKVs], stream_kv_node2(Bt, Reds, AssembleLTKVs, GTEKVs, InRange, Dir, Fun, Acc). stream_kv_node2(_Bt, _Reds, _PrevKVs, [], _InRange, _Dir, _Fun, Acc) -> {ok, Acc}; stream_kv_node2(Bt, Reds, PrevKVs, [{K,V} | RestKVs], InRange, Dir, Fun, Acc) -> case InRange(K) of false -> {stop, {PrevKVs, Reds}, Acc}; true -> AssembledKV = assemble(Bt, K, V), case Fun(visit, AssembledKV, {PrevKVs, Reds}, Acc) of {ok, Acc2} -> stream_kv_node2(Bt, Reds, [AssembledKV | PrevKVs], RestKVs, InRange, Dir, Fun, Acc2); {stop, Acc2} -> {stop, {PrevKVs, Reds}, Acc2} end end.
src/couchdb/couch_btree.erl
0.551091
0.451447
couch_btree.erl
starcoder
%% ------------------------------------------------------------------- %% %% riak_core: Core Riak Application %% %% Copyright (c) 2007-2015 Basho Technologies, Inc. All Rights Reserved. %% %% This file is provided to you under the Apache License, %% Version 2.0 (the "License"); you may not use this file %% except in compliance with the License. You may obtain %% a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, %% software distributed under the License is distributed on an %% "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY %% KIND, either express or implied. See the License for the %% specific language governing permissions and limitations %% under the License. %% %% ------------------------------------------------------------------- %% @doc The default functions used for claiming partition ownership. Generally, %% a wants_claim function should return either {yes, Integer} or 'no' where %% Integer is the number of additional partitions wanted by this node. A %% choose_claim function should return a riak_core_ring with more %% partitions claimed by this node than in the input ring. %% The usual intention for partition ownership assumes relative heterogeneity of %% capacity and connectivity. Accordingly, the standard claim functions attempt %% to maximize "spread" -- expected distance between partitions claimed by each %% given node. This is in order to produce the expectation that for any %% reasonably short span of consecutive partitions, there will be a minimal %% number of partitions owned by the same node. %% The exact amount that is considered tolerable is determined by the %% application env variable "target_n_val". The functions in riak_core_claim %% will ensure that all sequences up to target_n_val long contain no repeats if %% at all possible. The effect of this is that when the number of nodes in the %% system is smaller than target_n_val, a potentially large number of partitions %% must be moved in order to safely add a new node. After the cluster has grown %% beyond that size, a minimal number of partitions (1/NumNodes) will generally %% be moved. %% If the number of nodes does not divide evenly into the number of partitions, %% it may not be possible to perfectly achieve the maximum spread constraint. %% In that case, Riak will minimize the cases where the constraint is violated %% and they will all exist near the origin point of the ring. -module(riak_core_claim). -type ring() :: riak_core_ring:riak_core_ring(). -export([claim/1, claim/3, claim_until_balanced/2, claim_until_balanced/4]). -export([default_wants_claim/1, default_wants_claim/2, default_choose_claim/1, default_choose_claim/2, default_choose_claim/3, never_wants_claim/1, never_wants_claim/2, random_choose_claim/1, random_choose_claim/2, random_choose_claim/3]). -export([wants_claim_v2/1, wants_claim_v2/2, choose_claim_v2/1, choose_claim_v2/2, choose_claim_v2/3, claim_rebalance_n/2, claim_diversify/3, claim_diagonal/3, wants/1, wants_owns_diff/2, meets_target_n/2, diagonal_stripe/2]). -define(DEF_TARGET_N, 4). %% @doc Run the claim algorithm for the complete ring. %% @param Ring Ring the algorithm is run on. %% @returns The ring after the claim algorithm has been applied. -spec claim(Ring :: ring()) -> ring(). claim(Ring) -> claim(Ring, want, choose). %% @doc Run the claim algorithm for the complete ring. %% @param Ring Ring the algorithm is run on. %% @param Mode1 ignored. %% @param Mode2 ignored. %% @returns The ring after the claim algorithm has been applied. -spec claim(Ring :: ring(), Mode1 :: any(), Mode2 :: any()) -> ring(). claim(Ring, _, _) -> Members = riak_core_ring:claiming_members(Ring), lists:foldl(fun (Node, Ring0) -> claim_until_balanced(Ring0, Node, want, choose) end, Ring, Members). %% @doc Apply the claim algorithm until a given node owns enough partitions. %% @param Ring Ring the algorithm is applied to. %% @param Node Node name of the node to be balanced. %% @returns The balanced ring. -spec claim_until_balanced(Ring :: ring(), Node :: term()) -> ring(). claim_until_balanced(Ring, Node) -> claim_until_balanced(Ring, Node, want, choose). %% @doc Apply the claim algorithm until a given node owns enough partitions. %% @param Ring Ring the algorithm is applied to. %% @param Node Node name of the node to be balanced. %% @param want Fixed guard. %% @param choose Fixed guard %% @returns The balanced ring. -spec claim_until_balanced(Ring :: ring(), Node :: term(), want, choose) -> ring(). claim_until_balanced(Ring, Node, want, choose) -> NeedsIndexes = wants_claim_v2(Ring, Node), case NeedsIndexes of no -> Ring; {yes, _NumToClaim} -> NewRing = choose_claim_v2(Ring, Node), claim_until_balanced(NewRing, Node, want, choose) end. %% =================================================================== %% Claim Function Implementations %% =================================================================== %% @doc Choose a partition at random for the local node. %% @param Ring Ring to claim on. %% @returns Updated ring. -spec default_choose_claim(Ring :: ring()) -> ring(). default_choose_claim(Ring) -> default_choose_claim(Ring, node()). %% @doc Choose a partition for a given node at random. %% @param Ring Ring to claim on. %% @param Node Node to claim for. %% @returns Updated ring. -spec default_choose_claim(Ring :: ring(), Node :: term()) -> ring(). default_choose_claim(Ring, Node) -> choose_claim_v2(Ring, Node). %% @doc Choose a partition for a given node according to the given parameters. %% @param Ring Ring to claim on. %% @param Node Node to claim for. %% @param Params Parameters to consider. %% @returns Updated ring. -spec default_choose_claim(Ring :: ring(), Node :: term(), Params :: [term()]) -> ring(). default_choose_claim(Ring, Node, Params) -> choose_claim_v2(Ring, Node, Params). %% @doc Want a partition if we currently have less than floor(ringsize/nodes). %% @param Ring Ring to claim on. %% @returns `{yes, Difference}' or `no'. -spec default_wants_claim(Ring :: ring()) -> {yes, integer()} | no. default_wants_claim(Ring) -> default_wants_claim(Ring, node()). %% @doc Like {@link default_wants_claim/1} with a given node. %% @param Node Node to decide balance for. -spec default_wants_claim(Ring :: ring(), Node :: term()) -> {yes, integer()} | no. default_wants_claim(Ring, Node) -> wants_claim_v2(Ring, Node). %% @doc Decide if the local node needs more partitions. %% @param Ring Ring to claim on. %% @returns `{yes, Difference}' or `no'. -spec wants_claim_v2(Ring :: ring()) -> {yes, integer()} | no. wants_claim_v2(Ring) -> wants_claim_v2(Ring, node()). %% @doc Like {@link wants_claim_v2/1} for another node. %% @param Node Node to decide balance for. -spec wants_claim_v2(Ring :: ring(), Node :: term()) -> {yes, integer()} | no. wants_claim_v2(Ring, Node) -> Active = riak_core_ring:claiming_members(Ring), Owners = riak_core_ring:all_owners(Ring), Counts = get_counts(Active, Owners), NodeCount = erlang:length(Active), RingSize = riak_core_ring:num_partitions(Ring), Avg = RingSize div NodeCount, Count = proplists:get_value(Node, Counts, 0), case Count < Avg of false -> no; true -> {yes, Avg - Count} end. %% @doc Provide default choose parameters if none given -spec default_choose_params() -> [term()]. default_choose_params() -> default_choose_params([]). %% @doc Provide default NVal if it is not contained in the given parameters. %% @param Params List of claim parameters. %% @returns List of claim parameters containing target NVal. -spec default_choose_params(Params :: term()) -> [term()]. default_choose_params(Params) -> case proplists:get_value(target_n_val, Params) of undefined -> TN = application:get_env(riak_core, target_n_val, ?DEF_TARGET_N), [{target_n_val, TN} | Params]; _ -> Params end. %% @doc Choose a partition the local node should claim. %% @param Ring Ring to claim on. %% @returns Updated ring. -spec choose_claim_v2(Ring :: ring()) -> ring(). choose_claim_v2(Ring) -> choose_claim_v2(Ring, node()). %% @doc Like {@link choose_claim_v2/1} with a specified node. %% @param Node Specified node that claims a partition. -spec choose_claim_v2(Ring :: ring(), Node :: term()) -> ring(). choose_claim_v2(Ring, Node) -> Params = default_choose_params(), choose_claim_v2(Ring, Node, Params). %% @doc Like {@link choose_claim_v2/2} with specified parameters. %% @param Params0 Claim parameter list. -spec choose_claim_v2(Ring :: ring(), Node :: term(), Params0 :: [term()]) -> ring(). choose_claim_v2(Ring, Node, Params0) -> Params = default_choose_params(Params0), %% Active::[node()] Active = riak_core_ring:claiming_members(Ring), %% Owners::[{index(), node()}] Owners = riak_core_ring:all_owners(Ring), %% Counts::[node(), non_neg_integer()] Counts = get_counts(Active, Owners), RingSize = riak_core_ring:num_partitions(Ring), NodeCount = erlang:length(Active), %% Deltas::[node(), integer()] Deltas = get_deltas(RingSize, NodeCount, Owners, Counts), {_, Want} = lists:keyfind(Node, 1, Deltas), TargetN = proplists:get_value(target_n_val, Params), AllIndices = lists:zip(lists:seq(0, length(Owners) - 1), [Idx || {Idx, _} <- Owners]), EnoughNodes = (NodeCount > TargetN) or (NodeCount == TargetN) and (RingSize rem TargetN =:= 0), case EnoughNodes of true -> %% If we have enough nodes to meet target_n, then we prefer to %% claim indices that are currently causing violations, and then %% fallback to indices in linear order. The filtering steps below %% will ensure no new violations are introduced. Violated = lists:flatten(find_violations(Ring, TargetN)), Violated2 = [lists:keyfind(Idx, 2, AllIndices) || Idx <- Violated], Indices = Violated2 ++ AllIndices -- Violated2; false -> %% If we do not have enough nodes to meet target_n, then we prefer %% claiming the same indices that would occur during a %% re-diagonalization of the ring with target_n nodes, falling %% back to linear offsets off these preferred indices when the %% number of indices desired is less than the computed set. Padding = lists:duplicate(TargetN, undefined), Expanded = lists:sublist(Active ++ Padding, TargetN), PreferredClaim = riak_core_claim:diagonal_stripe(Ring, Expanded), PreferredNth = [begin {Nth, Idx} = lists:keyfind(Idx, 2, AllIndices), Nth end || {Idx, Owner} <- PreferredClaim, Owner =:= Node], Offsets = lists:seq(0, RingSize div length(PreferredNth)), AllNth = lists:sublist([(X + Y) rem RingSize || Y <- Offsets, X <- PreferredNth], RingSize), Indices = [lists:keyfind(Nth, 1, AllIndices) || Nth <- AllNth] end, %% Filter out indices that conflict with the node's existing ownership Indices2 = prefilter_violations(Ring, Node, AllIndices, Indices, TargetN, RingSize), %% Claim indices from the remaining candidate set Claim = select_indices(Owners, Deltas, Indices2, TargetN, RingSize), Claim2 = lists:sublist(Claim, Want), NewRing = lists:foldl(fun (Idx, Ring0) -> riak_core_ring:transfer_node(Idx, Node, Ring0) end, Ring, Claim2), RingChanged = [] /= Claim2, RingMeetsTargetN = meets_target_n(NewRing, TargetN), case {RingChanged, EnoughNodes, RingMeetsTargetN} of {false, _, _} -> %% Unable to claim, fallback to re-diagonalization sequential_claim(Ring, Node, TargetN); {_, true, false} -> %% Failed to meet target_n, fallback to re-diagonalization sequential_claim(Ring, Node, TargetN); _ -> NewRing end. %% @private for each node in owners return a tuple of owner and delta %% where delta is an integer that expresses how many nodes the owner %% needs it's ownership to change by. A positive means the owner needs %% that many more partitions, a negative means the owner can lose that %% many paritions. -spec get_deltas(RingSize :: pos_integer(), NodeCount :: pos_integer(), Owners :: [{Index :: non_neg_integer(), node()}], Counts :: [{node(), non_neg_integer()}]) -> Deltas :: [{node(), integer()}]. get_deltas(RingSize, NodeCount, Owners, Counts) -> Avg = RingSize / NodeCount, %% the most any node should own Max = ceiling(RingSize / NodeCount), ActiveDeltas = [{Member, Count, normalise_delta(Avg - Count)} || {Member, Count} <- Counts], BalancedDeltas = rebalance_deltas(ActiveDeltas, Max, RingSize), add_default_deltas(Owners, BalancedDeltas, 0). %% @private a node can only claim whole partitions, but if RingSize %% rem NodeCount /= 0, a delta will be a float. This function decides %% if that float should be floored or ceilinged -spec normalise_delta(float()) -> integer(). normalise_delta(Delta) when Delta < 0 -> %% if the node has too many (a negative delta) give up the most %% you can (will be rebalanced) ceiling(abs(Delta)) * -1; normalise_delta(Delta) -> %% if the node wants partitions, ask for the fewest for least %% movement trunc(Delta). %% @private so that we don't end up with an imbalanced ring where one %% node has more vnodes than it should (e.g. [{n1, 6}, {n2, 6}, {n3, %% 6}, {n4, 8}, {n5,6} we rebalance the deltas so that select_indices %% doesn't leave some node not giving up enough partitions -spec rebalance_deltas([{node(), integer()}], pos_integer(), pos_integer()) -> [{node(), integer()}]. rebalance_deltas(NodeDeltas, Max, RingSize) -> AppliedDeltas = [Own + Delta || {_, Own, Delta} <- NodeDeltas], case lists:sum(AppliedDeltas) - RingSize of 0 -> [{Node, Delta} || {Node, _Cnt, Delta} <- NodeDeltas]; N when N < 0 -> increase_keeps(NodeDeltas, N, Max, []) end. %% @private increases the delta for (some) nodes giving away %% partitions to the max they can keep -spec increase_keeps(Deltas :: [{node(), integer()}], WantsError :: integer(), Max :: pos_integer(), Acc :: [{node(), integer()}]) -> Rebalanced :: [{node(), integer()}]. increase_keeps(Rest, 0, _Max, Acc) -> [{Node, Delta} || {Node, _Own, Delta} <- lists:usort(lists:append(Rest, Acc))]; increase_keeps([], N, Max, Acc) when N < 0 -> increase_takes(lists:reverse(Acc), N, Max, []); increase_keeps([{Node, Own, Delta} | Rest], N, Max, Acc) when Delta < 0 -> WouldOwn = Own + Delta, Additive = case WouldOwn + 1 =< Max of true -> 1; false -> 0 end, increase_keeps(Rest, N + Additive, Max, [{Node, Own + Delta + Additive} | Acc]); increase_keeps([NodeDelta | Rest], N, Max, Acc) -> increase_keeps(Rest, N, Max, [NodeDelta | Acc]). %% @private increases the delta for (some) nodes taking partitions to the max %% they can ask for -spec increase_takes(Deltas :: [{node(), integer()}], WantsError :: integer(), Max :: pos_integer(), Acc :: [{node(), integer()}]) -> Rebalanced :: [{node(), integer()}]. increase_takes(Rest, 0, _Max, Acc) -> [{Node, Delta} || {Node, _Own, Delta} <- lists:usort(lists:append(Rest, Acc))]; increase_takes([], N, _Max, Acc) when N < 0 -> [{Node, Delta} || {Node, _Own, Delta} <- lists:usort(Acc)]; increase_takes([{Node, Own, Delta} | Rest], N, Max, Acc) when Delta > 0 -> WouldOwn = Own + Delta, Additive = case WouldOwn + 1 =< Max of true -> 1; false -> 0 end, increase_takes(Rest, N + Additive, Max, [{Node, Own, Delta + Additive} | Acc]); increase_takes([NodeDelta | Rest], N, Max, Acc) -> increase_takes(Rest, N, Max, [NodeDelta | Acc]). %% @doc Check if the given ring can provide enough owners for each node to meet %% the target NVal. %% @param Ring Ring to check. %% @param TargetN NVal to check. %% @returns Boolean indicating if the ring meets the requirement. -spec meets_target_n(Ring :: ring(), TargetN :: pos_integer()) -> boolean(). meets_target_n(Ring, TargetN) -> Owners = lists:keysort(1, riak_core_ring:all_owners(Ring)), meets_target_n(Owners, TargetN, 0, [], []). %% @private %% @doc Helper function for {@link meets_target_n/2}. -spec meets_target_n(Owners :: [{integer(), term()}], TargetN :: pos_integer(), Index :: non_neg_integer(), First :: [{integer(), term()}], Last :: [{integer(), term()}]) -> boolean(). meets_target_n([{Part, Node} | Rest], TargetN, Index, First, Last) -> case lists:keytake(Node, 1, Last) of {value, {Node, LastIndex, _}, NewLast} -> if Index - LastIndex >= TargetN -> %% node repeat respects TargetN meets_target_n(Rest, TargetN, Index + 1, First, [{Node, Index, Part} | NewLast]); true -> %% violation of TargetN false end; false -> %% haven't seen this node yet meets_target_n(Rest, TargetN, Index + 1, [{Node, Index} | First], [{Node, Index, Part} | Last]) end; meets_target_n([], TargetN, Index, First, Last) -> %% start through end guarantees TargetN %% compute violations at wrap around, but don't fail %% because of them: handle during reclaim Violations = lists:filter(fun ({Node, L, _}) -> {Node, F} = proplists:lookup(Node, First), Index - L + F < TargetN end, Last), {true, [Part || {_, _, Part} <- Violations]}. %% @doc Claim diversify tries to build a perfectly diverse ownership list that %% meets target N. It uses wants to work out which nodes want partitions, %% but does not honor the counts currently. The algorithm incrementally %% builds the ownership list, updating the adjacency matrix needed to %% compute the diversity score as each node is added and uses it to drive %% the selection of the next nodes. %% @param Wants List of Node names and the respective number of partition they %% want to claim. %% @param Owners List of indices and the name of their owning node. %% @param Params Parameters. %% @returns New owner list and a list of attributes, in this case `diversified'. -spec claim_diversify(Wants :: [{term(), integer()}], Owners :: [{integer(), term()}], Params :: [term()]) -> {[{integer(), term()}], [atom()]}. claim_diversify(Wants, Owners, Params) -> TN = proplists:get_value(target_n_val, Params, ?DEF_TARGET_N), Q = length(Owners), Claiming = [N || {N, W} <- Wants, W > 0], {ok, NewOwners, _AM} = riak_core_claim_util:construct(riak_core_claim_util:gen_complete_len(Q), Claiming, TN), {NewOwners, [diversified]}. %% @doc Claim nodes in seq a,b,c,a,b,c trying to handle the wraparound case to %% meet target N %% @param Wants List of Node names and the respective number of partition they %% want to claim. %% @param Owners List of indices and the name of their owning node. %% @param Params Parameters. %% @returns Diagonalized list of owners and a list of attributes, in this case %% `diagonalized'. -spec claim_diagonal(Wants :: [{term(), integer()}], Owners :: [{integer(), term()}], Params :: [term()]) -> {[term()], [atom()]}. claim_diagonal(Wants, Owners, Params) -> TN = proplists:get_value(target_n_val, Params, ?DEF_TARGET_N), Claiming = lists:sort([N || {N, W} <- Wants, W > 0]), S = length(Claiming), Q = length(Owners), Reps = Q div S, %% Handle the ring wrapround case. If possible try to pick nodes %% that are not within the first TN of Claiming, if enough nodes %% are available. Tail = Q - Reps * S, Last = case S >= TN + Tail of true -> % If number wanted can be filled excluding first TN nodes lists:sublist(lists:nthtail(TN - Tail, Claiming), Tail); _ -> lists:sublist(Claiming, Tail) end, {lists:flatten([lists:duplicate(Reps, Claiming), Last]), [diagonalized]}. %% @private fall back to diagonal striping vnodes across nodes in a %% sequential round robin (eg n1 | n2 | n3 | n4 | n5 | n1 | n2 | n3 %% etc) However, different to `claim_rebalance_n', this function %% attempts to eliminate tail violations (for example a ring that %% starts/ends n1 | n2 | ...| n3 | n4 | n1) -spec sequential_claim(riak_core_ring:riak_core_ring(), node(), integer()) -> riak_core_ring:riak_core_ring(). sequential_claim(Ring, Node, TargetN) -> Nodes = lists:usort([Node | riak_core_ring:claiming_members(Ring)]), NodeCount = length(Nodes), RingSize = riak_core_ring:num_partitions(Ring), Overhang = RingSize rem NodeCount, HasTailViolation = Overhang > 0 andalso Overhang < TargetN, Shortfall = TargetN - Overhang, CompleteSequences = RingSize div NodeCount, MaxFetchesPerSeq = NodeCount - TargetN, MinFetchesPerSeq = ceiling(Shortfall / CompleteSequences), CanSolveViolation = CompleteSequences * MaxFetchesPerSeq >= Shortfall, Zipped = case HasTailViolation andalso CanSolveViolation of true -> Partitions = lists:sort([I || {I, _} <- riak_core_ring:all_owners(Ring)]), Nodelist = solve_tail_violations(RingSize, Nodes, Shortfall, MinFetchesPerSeq), lists:zip(Partitions, lists:flatten(Nodelist)); false -> diagonal_stripe(Ring, Nodes) end, lists:foldl(fun ({P, N}, Acc) -> riak_core_ring:transfer_node(P, N, Acc) end, Ring, Zipped). %% @private every module has a ceiling function -spec ceiling(float()) -> integer(). ceiling(F) -> T = trunc(F), case F - T == 0 of true -> T; false -> T + 1 end. %% @private rem_fill increase the tail so that there is no wrap around %% preflist violation, by taking a `Shortfall' number nodes from %% earlier in the preflist -spec solve_tail_violations(integer(), [node()], integer(), integer()) -> [node()]. solve_tail_violations(RingSize, Nodes, Shortfall, MinFetchesPerSeq) -> StartingNode = RingSize rem length(Nodes) + 1, build_nodelist(RingSize, Nodes, Shortfall, StartingNode, MinFetchesPerSeq, []). %% @private build the node list by building tail to satisfy TargetN, then removing %% the added nodes from earlier segments -spec build_nodelist(integer(), [node()], integer(), integer(), integer(), [node()]) -> [node()]. build_nodelist(RingSize, Nodes, _Shortfall = 0, _NodeCounter, _MinFetchesPerSeq, Acc) -> %% Finished shuffling, backfill if required ShuffledRing = lists:flatten(Acc), backfill_ring(RingSize, Nodes, (RingSize - length(ShuffledRing)) div length(Nodes), Acc); build_nodelist(RingSize, Nodes, Shortfall, NodeCounter, MinFetchesPerSeq, _Acc = []) -> %% Build the tail with sufficient nodes to satisfy TargetN NodeCount = length(Nodes), LastSegLength = RingSize rem NodeCount + Shortfall, NewSeq = lists:sublist(Nodes, 1, LastSegLength), build_nodelist(RingSize, Nodes, Shortfall, NodeCounter, MinFetchesPerSeq, NewSeq); build_nodelist(RingSize, Nodes, Shortfall, NodeCounter, MinFetchesPerSeq, Acc) -> %% Build rest of list, subtracting minimum of MinFetchesPerSeq, Shortfall %% or (NodeCount - NodeCounter) each time NodeCount = length(Nodes), NodesToRemove = min(min(MinFetchesPerSeq, Shortfall), NodeCount - NodeCounter), RemovalList = lists:sublist(Nodes, NodeCounter, NodesToRemove), NewSeq = lists:subtract(Nodes, RemovalList), NewNodeCounter = NodeCounter + NodesToRemove, build_nodelist(RingSize, Nodes, Shortfall - NodesToRemove, NewNodeCounter, MinFetchesPerSeq, [NewSeq | Acc]). %% @private Backfill the ring with full sequences -spec backfill_ring(integer(), [node()], integer(), [node()]) -> [node()]. backfill_ring(_RingSize, _Nodes, _Remaining = 0, Acc) -> Acc; backfill_ring(RingSize, Nodes, Remaining, Acc) -> backfill_ring(RingSize, Nodes, Remaining - 1, [Nodes | Acc]). %% @doc Rebalance the expected load on nodes using a diagonal stripe. %% @param Ring :: Ring to rebalance. %% @param Node :: Node to rebalance from. %% @returns Rebalanced ring. %% @see diagonal_stripe/2. -spec claim_rebalance_n(Ring :: ring(), Node :: term()) -> ring(). claim_rebalance_n(Ring, Node) -> Nodes = lists:usort([Node | riak_core_ring:claiming_members(Ring)]), Zipped = diagonal_stripe(Ring, Nodes), lists:foldl(fun ({P, N}, Acc) -> riak_core_ring:transfer_node(P, N, Acc) end, Ring, Zipped). %% @doc Creates a diagonal stripw of the given nodes over the partitions of the %% ring. %% @param Ring Ring on which the stripes are built. %% @param Nodes Nodes that are to be distributed. %% @returns List of indices and assigned nodes. -spec diagonal_stripe(Ring :: ring(), Nodes :: [term()]) -> [{integer(), term()}]. diagonal_stripe(Ring, Nodes) -> %% diagonal stripes guarantee most disperse data Partitions = lists:sort([I || {I, _} <- riak_core_ring:all_owners(Ring)]), Zipped = lists:zip(Partitions, lists:sublist(lists:flatten(lists:duplicate(1 + length(Partitions) div length(Nodes), Nodes)), 1, length(Partitions))), Zipped. %% @doc Choose a random partition for the local node. %% @param Ring Ring to claim on. %% @returns Updated ring. -spec random_choose_claim(Ring :: ring()) -> ring(). random_choose_claim(Ring) -> random_choose_claim(Ring, node()). %% @doc Like {@link random_choose_claim/1} with a specified node. %% @param Node Node to choose a partition for. -spec random_choose_claim(Ring :: ring(), Node :: term()) -> ring(). random_choose_claim(Ring, Node) -> random_choose_claim(Ring, Node, []). %% @doc Like {@link random_choose_claim/2} with specified parameters. %% @param Params List of parameters, currently ignored. -spec random_choose_claim(Ring :: ring(), Node :: term(), Params :: [term()]) -> ring(). random_choose_claim(Ring, Node, _Params) -> riak_core_ring:transfer_node(riak_core_ring:random_other_index(Ring), Node, Ring). %% @doc For use by nodes that should not claim any partitions. -spec never_wants_claim(ring()) -> no. never_wants_claim(_) -> no. %% @doc For use by nodes that should not claim any partitions. -spec never_wants_claim(ring(), term()) -> no. never_wants_claim(_, _) -> no. %% =================================================================== %% Private %% =================================================================== %% @private %% %% @doc Determines indices that violate the given target_n spacing %% property. find_violations(Ring, TargetN) -> Owners = riak_core_ring:all_owners(Ring), Suffix = lists:sublist(Owners, TargetN - 1), Owners2 = Owners ++ Suffix, %% Use a sliding window to determine violations {Bad, _} = lists:foldl(fun (P = {Idx, Owner}, {Out, Window}) -> Window2 = lists:sublist([P | Window], TargetN - 1), case lists:keyfind(Owner, 2, Window) of {PrevIdx, Owner} -> {[[PrevIdx, Idx] | Out], Window2}; false -> {Out, Window2} end end, {[], []}, Owners2), lists:reverse(Bad). %% @private %% @doc Counts up the number of partitions owned by each node. -spec get_counts([node()], [{integer(), _}]) -> [{node(), non_neg_integer()}]. get_counts(Nodes, Ring) -> Empty = [{Node, 0} || Node <- Nodes], Counts = lists:foldl(fun ({_Idx, Node}, Counts) -> case lists:member(Node, Nodes) of true -> dict:update_counter(Node, 1, Counts); false -> Counts end end, dict:from_list(Empty), Ring), dict:to_list(Counts). %% @private %% @doc Add default delta values for all owners to the delta list. -spec add_default_deltas(IdxOwners :: [{integer(), term()}], Deltas :: [{term(), integer()}], Default :: integer()) -> [{term(), integer()}]. add_default_deltas(IdxOwners, Deltas, Default) -> {_, Owners} = lists:unzip(IdxOwners), Owners2 = lists:usort(Owners), Defaults = [{Member, Default} || Member <- Owners2], lists:ukeysort(1, Deltas ++ Defaults). %% @private %% @doc Filter out candidate indices that would violate target_n given a node's %% current partition ownership. -spec prefilter_violations(Ring :: ring(), Node :: term(), AllIndices :: [{term(), integer()}], Indices :: [{term(), integer()}], TargetN :: pos_integer(), RingSize :: non_neg_integer()) -> [{term(), integer()}]. prefilter_violations(Ring, Node, AllIndices, Indices, TargetN, RingSize) -> CurrentIndices = riak_core_ring:indices(Ring, Node), CurrentNth = [lists:keyfind(Idx, 2, AllIndices) || Idx <- CurrentIndices], [{Nth, Idx} || {Nth, Idx} <- Indices, lists:all(fun ({CNth, _}) -> spaced_by_n(CNth, Nth, TargetN, RingSize) end, CurrentNth)]. %% @private %% @doc Select indices from a given candidate set, according to two %% goals. %% %% 1. Ensure greedy/local target_n spacing between indices. Note that this %% goal intentionally does not reject overall target_n violations. %% %% 2. Select indices based on the delta between current ownership and %% expected ownership. In other words, if A owns 5 partitions and %% the desired ownership is 3, then we try to claim at most 2 partitions %% from A. -spec select_indices(Owners :: [], Deltas :: [{term(), integer()}], Indices :: [{term(), integer()}], TargetN :: pos_integer(), RingSize :: pos_integer()) -> [integer()]. select_indices(_Owners, _Deltas, [], _TargetN, _RingSize) -> []; select_indices(Owners, Deltas, Indices, TargetN, RingSize) -> OwnerDT = dict:from_list(Owners), {FirstNth, _} = hd(Indices), %% The `First' symbol indicates whether or not this is the first %% partition to be claimed by this node. This assumes that the %% node doesn't already own any partitions. In that case it is %% _always_ safe to claim the first partition that another owner %% is willing to part with. It's the subsequent partitions %% claimed by this node that must not break the target_n invariant. {Claim, _, _, _} = lists:foldl(fun ({Nth, Idx}, {Out, LastNth, DeltaDT, First}) -> Owner = dict:fetch(Idx, OwnerDT), Delta = dict:fetch(Owner, DeltaDT), MeetsTN = spaced_by_n(LastNth, Nth, TargetN, RingSize), case (Delta < 0) and (First or MeetsTN) of true -> NextDeltaDT = dict:update_counter(Owner, 1, DeltaDT), {[Idx | Out], Nth, NextDeltaDT, false}; false -> {Out, LastNth, DeltaDT, First} end end, {[], FirstNth, dict:from_list(Deltas), true}, Indices), lists:reverse(Claim). %% @private %% @doc Determine if two positions in the ring meet target_n spacing. -spec spaced_by_n(Ntha :: integer(), NthB :: integer(), TargetN :: pos_integer(), RingSize :: pos_integer()) -> boolean(). spaced_by_n(NthA, NthB, TargetN, RingSize) -> case NthA > NthB of true -> NFwd = NthA - NthB, NBack = NthB - NthA + RingSize; false -> NFwd = NthA - NthB + RingSize, NBack = NthB - NthA end, (NFwd >= TargetN) and (NBack >= TargetN). %% @doc For each node in wants, work out how many more partition each node wants %% (positive) or is overloaded by (negative) compared to what it owns. %% @param Wants List of node names and their target number of partitions. %% @param Owns List of node names and their actual number of partitions. -spec wants_owns_diff(Wants :: [{term(), integer()}], Owns :: [{term(), integer()}]) -> [{term(), integer()}]. wants_owns_diff(Wants, Owns) -> [case lists:keyfind(N, 1, Owns) of {N, O} -> {N, W - O}; false -> {N, W} end || {N, W} <- Wants]. %% @doc Given a ring, work out how many partition each wants to be %% considered balanced. %% @param Ring Ring to figure out wants for. %% @returns List of node names and the number of wanted partitions. -spec wants(Ring :: ring()) -> [{term(), integer()}]. wants(Ring) -> Active = lists:sort(riak_core_ring:claiming_members(Ring)), Inactive = riak_core_ring:all_members(Ring) -- Active, Q = riak_core_ring:num_partitions(Ring), ActiveWants = lists:zip(Active, wants_counts(length(Active), Q)), InactiveWants = [{N, 0} || N <- Inactive], lists:sort(ActiveWants ++ InactiveWants). %% @private %% @doc Given a number of nodes and ring size, return a list of %% desired ownership, S long that add up to Q -spec wants_counts(S :: non_neg_integer(), Q :: non_neg_integer()) -> [integer()]. wants_counts(S, Q) -> Max = roundup(Q / S), case S * Max - Q of 0 -> lists:duplicate(S, Max); X -> lists:duplicate(X, Max - 1) ++ lists:duplicate(S - X, Max) end. %% @private %% @doc Round up to next whole integer - ceil -spec roundup(float()) -> integer(). roundup(I) when I >= 0 -> T = erlang:trunc(I), case I - T of Neg when Neg < 0 -> T; Pos when Pos > 0 -> T + 1; _ -> T end. %% =================================================================== %% Unit tests %% =================================================================== -ifdef(TEST). -compile(export_all). -include_lib("eunit/include/eunit.hrl"). wants_claim_test() -> riak_core_ring_manager:setup_ets(test), riak_core_test_util:setup_mockring1(), {ok, Ring} = riak_core_ring_manager:get_my_ring(), ?assertEqual({yes, 1}, (default_wants_claim(Ring))), riak_core_ring_manager:cleanup_ets(test), riak_core_ring_manager:stop(). %% @private console helper function to return node lists for claiming %% partitions -spec gen_diag(pos_integer(), pos_integer()) -> [Node :: atom()]. gen_diag(RingSize, NodeCount) -> Nodes = [list_to_atom(lists:concat(["n_", N])) || N <- lists:seq(1, NodeCount)], {HeadNode, RestNodes} = {hd(Nodes), tl(Nodes)}, R0 = riak_core_ring:fresh(RingSize, HeadNode), RAdded = lists:foldl(fun (Node, Racc) -> riak_core_ring:add_member(HeadNode, Racc, Node) end, R0, RestNodes), Diag = diagonal_stripe(RAdded, Nodes), {_P, N} = lists:unzip(Diag), N. %% @private call with result of gen_diag/1 only, does the list have %% tail violations, returns true if so, false otherwise. -spec has_violations([Node :: atom()]) -> boolean(). has_violations(Diag) -> RS = length(Diag), NC = length(lists:usort(Diag)), Overhang = RS rem NC, Overhang > 0 andalso Overhang < 4. %% hardcoded target n of 4 -endif.
src/riak_core_claim.erl
0.724578
0.469703
riak_core_claim.erl
starcoder
%%%------------------------------------------------------------------- %%% @author <NAME> %%% @copyright (C) 2018 ACK CYFRONET AGH %%% This software is released under the MIT license cited in 'LICENSE.txt'. %%% @end %%%------------------------------------------------------------------- %%% @doc %%% This module contains utilities for verifying DNS configuration. %%% It is used to ensure that DNS contains records necessary %%% for the cluster to work correctly. %%% @end %%%------------------------------------------------------------------- -module(onepanel_dns). -author("<NAME>"). -include("modules/errors.hrl"). -include("modules/onepanel_dns.hrl"). -include_lib("ctool/include/logging.hrl"). -export([check_any/4, check_all/4, make_results_stub/2]). -export([build_bind_record/3]). % allowed query types -type dns_type() :: a | txt. % names used in queries -type dns_name() :: binary() | string(). % IP tuple for A queries, binary for TXT queries -type dns_value() :: inet:ip4_address() | binary(). % Result type -type dns_check() :: #dns_check{}. %% Possible results of dns check: %% unresolvable - no values, expected or not, could be resolved %% bad_recods - none of expected values are found in the resolved ones %% missing_records - some, but not all, expected IPs are present in resolved ones %% ok - all of expected values are found in the resolved ones -type summary() :: ok | unresolvable | bad_records | missing_records. -export_type([summary/0, dns_check/0, dns_type/0, dns_name/0, dns_value/0]). %%%=================================================================== %%% Public API %%%=================================================================== %%-------------------------------------------------------------------- %% @doc %% Queries all given DNS servers and compares result with the expected %% list. Returns the most optimistic result among DNS servers. %% @end %%-------------------------------------------------------------------- -spec check_any(Expected :: [dns_value()], Names :: [dns_name()], Type :: dns_type(), Servers :: [inet:ip4_address()]) -> dns_check(). check_any(Expected, Names, Type, Servers) -> Diffs = check(Expected, Names, Type, Servers), Best = lists:last(lists:sort(fun compare/2, Diffs)), log_result(Names, Type, Servers, Best), Best. %%-------------------------------------------------------------------- %% @doc %% Queries all given DNS servers and compares result with the expected %% list. Returns the most pessimistic result among responsive DNS servers. %% @end %%-------------------------------------------------------------------- -spec check_all(Expected :: [dns_value()], Names :: [dns_name()], Type :: dns_type(), Servers :: [inet:ip4_address()]) -> dns_check(). check_all(Expected, Names, Type, Servers) -> Diffs = check(Expected, Names, Type, Servers), Worst = hd(lists:sort(fun compare/2, Diffs)), log_result(Names, Type, Servers, Worst), Worst. %%-------------------------------------------------------------------- %% @doc %% Creates dns_check record for when no actual results could be %% obtained. %% @end %%-------------------------------------------------------------------- -spec make_results_stub(Summary :: summary(), Expected :: [dns_value()]) -> dns_check(). make_results_stub(Summary, Expected) -> Unique = lists:usort(Expected), #dns_check{summary = Summary, expected = Unique, missing = Unique}. %%-------------------------------------------------------------------- %% @doc %% Build DNS record in the popular BIND server zone format. %% @end %%-------------------------------------------------------------------- -spec build_bind_record(Domain :: binary(), Type :: atom(), Value :: inet:ip4_address () | binary()) -> binary(). build_bind_record(Domain, a, IP) when is_tuple(IP) -> build_bind_record(Domain, a, onepanel_ip:ip4_to_binary(IP)); build_bind_record(Domain, txt, Value) -> QuotedValue = <<$", Value/binary, $">>, onepanel_utils:join([Domain, <<"IN">>, <<"TXT">>, QuotedValue], <<" ">>); build_bind_record(Domain, Type, Value) -> TypeBin = string:uppercase(onepanel_utils:convert(Type, binary)), onepanel_utils:join([Domain, <<"IN">>, TypeBin, Value], <<" ">>). %%%=================================================================== %%% Internal functions %%%=================================================================== %%-------------------------------------------------------------------- %% @private %% @doc %% Performs DNS check on each DNS server. Filters out servers %% which could not be contacted. %% @end %%-------------------------------------------------------------------- -spec check(Expected :: [dns_value()], Names :: [dns_name()], Type :: dns_type(), Servers :: [inet:ip4_address() | default]) -> [dns_check()]. check(Expected, Names, Type, []) -> check(Expected, Names, Type, [default]); check([], [], _Type, _Servers) -> [#dns_check{summary = ok, expected = [], got = []}]; check(Expected, Names, Type, Servers) -> Results = lists_utils:pmap(fun(Server) -> check_on_server(Expected, Names, Type, Server) end, Servers), WithoutErrors = lists:filter(fun (error) -> false; (#dns_check{}) -> true end, Results), case WithoutErrors of [] -> throw(?ERROR_DNS_SERVERS_UNREACHABLE(Servers)); _ -> WithoutErrors end. %%-------------------------------------------------------------------- %% @private %% @doc %% Queries given Server for given Names using DNS query type Type. %% Compares results with the list of Expected values. %% @end %%-------------------------------------------------------------------- -spec check_on_server(Expected :: [dns_value()], Names :: [dns_name()], Type :: dns_type(), Servers :: inet:ip4_address() | default) -> dns_check() | error. check_on_server(Expected, Names, Type, ServerIP) -> case lookup(Names, Type, ServerIP) of error -> error; Resolved -> Correct = lists_utils:intersect(Resolved, Expected), Missing = lists_utils:subtract(Expected, Resolved), Additional = lists_utils:subtract(Resolved, Expected), Summary = if Resolved == [] -> unresolvable; Correct == [] -> bad_records; Missing /= [] -> missing_records; true -> ok end, #dns_check{summary = Summary, expected = lists:usort(Expected), got = lists:usort(Resolved), missing = Missing, excessive = Additional} end. %%-------------------------------------------------------------------- %% @private %% @doc %% Returns aggregated results of DNS queries for all Names. %% If any of the queries causes connection error, all results %% are discarded and 'error' is returned. %% @end %%-------------------------------------------------------------------- -spec lookup(Quries :: [dns_name()], Type :: dns_type(), DnsServerIP :: inet:ip4_address() | default) -> [dns_value()] | error. lookup(Names, Type, DnsServerIP) -> Results = lists:flatten(lists_utils:pmap(fun(Name) -> NameStr = onepanel_utils:convert(Name, list), Opts = case DnsServerIP of default -> []; IP -> [{nameservers, [{IP, 53}]}] end, Resolved = inet_res:resolve(NameStr, in, Type, Opts), case Resolved of {error, {_Reason, _DnsMsg}} -> []; {error, Reason} -> ?warning("Error querying server ~p for DNS check ~p of name ~p: ~p", [DnsServerIP, Type, NameStr, Reason]), error; {ok, Msg} -> Answers = inet_dns:msg(Msg, anlist), % filter_answer will convert TXT query results from string to binary % preventing them from being caught by lists:flatten. filter_answer(Type, Answers) end end, Names)), case lists:any(fun(error) -> true; (_) -> false end, Results) of true -> error; false -> Results end. %%-------------------------------------------------------------------- %% @private %% @doc %% Filters query response by expected class and type. %% This is similar to inet_res:lookup, which was not used %% in order to expose server connection errors. %% @end %%-------------------------------------------------------------------- -spec filter_answer(Type :: dns_type(), Anlist :: [DnsRR :: term()]) -> [dns_value()]. filter_answer(Type, Anlist) -> [normalize_dns_data(Type, inet_dns:rr(RR, data)) || RR <- Anlist, inet_dns:rr(RR, class) == in, inet_dns:rr(RR, type) == Type]. %%-------------------------------------------------------------------- %% @private %% @doc %% Converts dns_data value returned by inet_res to expected format. %% @end %%-------------------------------------------------------------------- -spec normalize_dns_data(Type :: dns_type(), Data :: term()) -> dns_value(). normalize_dns_data(txt, Data) -> list_to_binary(Data); normalize_dns_data(a, IPTuple) -> IPTuple. %%-------------------------------------------------------------------- %% @private %% @doc %% Comparator for the check results. %% When summaries are equal, result with higher number of excessive %% records is considered worse. %% @end %%-------------------------------------------------------------------- -spec compare(dns_check(), dns_check()) -> boolean(). compare(#dns_check{summary = SameSummary, excessive = Exc1}, #dns_check{summary = SameSummary, excessive = Exc2}) -> length(Exc1) >= length(Exc2); compare(#dns_check{summary = Result1}, #dns_check{summary = Result2}) -> summary_to_integer(Result1) < summary_to_integer(Result2). %%-------------------------------------------------------------------- %% @private %% @doc %% Provides results summary ordering with lower numbers indicating %% more problematic results. %% @end %%-------------------------------------------------------------------- -spec summary_to_integer(summary()) -> integer(). summary_to_integer(unresolvable) -> 1; summary_to_integer(bad_records) -> 2; summary_to_integer(missing_records) -> 3; summary_to_integer(ok) -> 4. %%-------------------------------------------------------------------- %% @private %% @doc %% Logs message describing negative results of the check. %% @end %%-------------------------------------------------------------------- -spec log_result(Names :: [dns_name()], Type :: dns_type(), Servers :: [inet:ip4_address()], Result :: dns_check()) -> ok. log_result(_Names, _Type, _Servers, #dns_check{summary = ok}) -> ok; log_result(Names, Type, Servers, #dns_check{summary = Summary, expected = Expected, got = Got}) -> ?warning("DNS check for records ~s named ~p failed with \"~s\":~n" "Servers used: ~p~nExpected values: ~p~nObtained values: ~p", [string:uppercase(atom_to_list(Type)), Names, Summary, Servers, Expected, Got]).
src/modules/onepanel_dns.erl
0.55254
0.436682
onepanel_dns.erl
starcoder
%% ------------------------------------------------------------------- %% %% Copyright <2013-2018> < %% Technische Universität Kaiserslautern, Germany %% Université Pierre et Marie Curie / Sorbonne-Université, France %% Universidade NOVA de Lisboa, Portugal %% Université catholique de Louvain (UCL), Belgique %% INESC TEC, Portugal %% > %% %% This file is provided to you under the Apache License, %% Version 2.0 (the "License"); you may not use this file %% except in compliance with the License. You may obtain %% a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, %% software distributed under the License is distributed on an %% "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY %% KIND, either expressed or implied. See the License for the %% specific language governing permissions and limitations %% under the License. %% %% List of the contributors to the development of Antidote: see AUTHORS file. %% Description and complete License: see LICENSE file. %% ------------------------------------------------------------------- %% @doc Provides an API to interact with DC nodes. Interactions include: %% - Getting descriptors. %% - Connecting Nodes to a DC. %% - Check the DC start status. %% - Check node restart status. %% - Tell nodes to drop ping messages. -module(inter_dc_manager). -include("antidote.hrl"). -include("inter_dc_repl.hrl"). %% =================================================================== %% Public API %% =================================================================== -define(DC_CONNECT_RETRIES, 5). -define(DC_CONNECT_RETY_SLEEP, 1000). -include_lib("kernel/include/logger.hrl"). -export([ get_descriptor/0, start_bg_processes/1, observe_dcs_sync/1, dc_successfully_started/0, check_node_restart/0, forget_dcs/1, drop_ping/1 ]). -spec get_descriptor() -> {ok, descriptor()}. get_descriptor() -> %% Wait until all needed vnodes are spawned, so that the heartbeats are already being sent ok = dc_utilities:ensure_all_vnodes_running_master(inter_dc_log_sender_vnode_master), Nodes = dc_utilities:get_my_dc_nodes(), Publishers = lists:map( fun(Node) -> rpc:call(Node, inter_dc_pub, get_address_list, []) end, Nodes ), LogReaders = lists:map( fun(Node) -> rpc:call(Node, inter_dc_query_router, get_address_list, []) end, Nodes ), {ok, #descriptor{ dcid = dc_utilities:get_my_dc_id(), partition_num = dc_utilities:get_partitions_num(), publishers = Publishers, logreaders = LogReaders }}. %% This will connect the list of local nodes to the DC given by the descriptor %% When a connecting to a new DC, Nodes will be all the nodes in the local DC %% Otherwise this will be called with a single node that is reconnecting (for example after one of the nodes in the DC crashes and restarts) %% Note this is an internal function, to instruct the local DC to connect to a new DC the observe_dcs_sync(Descriptors) function should be used -spec observe_dc(descriptor(), [node()]) -> ok | inter_dc_conn_err(). observe_dc( Desc = #descriptor{ dcid = DCID, partition_num = PartitionsNumRemote, publishers = Publishers, logreaders = LogReaders }, Nodes ) -> PartitionsNumLocal = dc_utilities:get_partitions_num(), case PartitionsNumRemote == PartitionsNumLocal of false -> ?LOG_ERROR("Cannot observe remote DC: partition number mismatch"), {error, {partition_num_mismatch, PartitionsNumRemote, PartitionsNumLocal}}; true -> case DCID == dc_utilities:get_my_dc_id() of true -> ok; false -> ?LOG_INFO("Observing DC ~p", [DCID]), dc_utilities:ensure_all_vnodes_running_master(inter_dc_log_sender_vnode_master), %% Announce the new publisher addresses to all subscribers in this DC. %% Equivalently, we could just pick one node in the DC and delegate all the subscription work to it. %% But we want to balance the work, so all nodes take part in subscribing. connect_nodes(Nodes, DCID, LogReaders, Publishers, Desc, ?DC_CONNECT_RETRIES) end end. -spec connect_nodes( [node()], dcid(), [socket_address()], [socket_address()], descriptor(), non_neg_integer() ) -> ok | {error, connection_error}. connect_nodes([], _DCID, _LogReaders, _Publishers, _Desc, _Retries) -> ok; connect_nodes(_Nodes, _DCID, _LogReaders, _Publishers, Desc, 0) -> ok = forget_dcs([Desc]), {error, connection_error}; connect_nodes([Node | Rest], DCID, LogReaders, Publishers, Desc, Retries) -> case rpc:call(Node, inter_dc_query_dealer, add_dc, [DCID, LogReaders], ?COMM_TIMEOUT) of ok -> case rpc:call(Node, inter_dc_sub, add_dc, [DCID, Publishers], ?COMM_TIMEOUT) of ok -> connect_nodes(Rest, DCID, LogReaders, Publishers, Desc, ?DC_CONNECT_RETRIES); _ -> timer:sleep(?DC_CONNECT_RETY_SLEEP), ?LOG_ERROR("Unable to connect to publisher ~p", [DCID]), connect_nodes([Node | Rest], DCID, LogReaders, Publishers, Desc, Retries - 1) end; _ -> timer:sleep(?DC_CONNECT_RETY_SLEEP), ?LOG_ERROR("Unable to connect to log reader ~p", [DCID]), connect_nodes([Node | Rest], DCID, LogReaders, Publishers, Desc, Retries - 1) end. %% This should not be called until the local dc's ring is merged -spec start_bg_processes(atom()) -> ok. start_bg_processes(MetaDataName) -> %% Start the meta-data senders Nodes = dc_utilities:get_my_dc_nodes(), %% Ensure vnodes are running and meta_data ok = dc_utilities:ensure_all_vnodes_running_master(inter_dc_log_sender_vnode_master), ok = dc_utilities:ensure_all_vnodes_running_master(clocksi_vnode_master), ok = dc_utilities:ensure_all_vnodes_running_master(logging_vnode_master), ok = dc_utilities:ensure_all_vnodes_running_master(materializer_vnode_master), lists:foreach( fun(Node) -> true = wait_init:wait_ready(Node), ok = rpc:call(Node, dc_utilities, check_registered, [meta_data_sender_sup]), ok = rpc:call(Node, dc_utilities, check_registered, [meta_data_manager_sup]), ok = rpc:call(Node, dc_utilities, check_registered_global, [ stable_meta_data_server:generate_server_name(Node) ]), ok = rpc:call(Node, meta_data_sender, start, [MetaDataName]) end, Nodes ), %% Load the internal meta-data ok = dc_meta_data_utilities:store_meta_data_name(MetaDataName), %% Start the timers sending the heartbeats ?LOG_INFO("Starting heartbeat sender timers"), Responses = dc_utilities:bcast_vnode_sync(logging_vnode_master, {start_timer, undefined}), %% Be sure they all started ok, crash otherwise ok = lists:foreach( fun({_, ok}) -> ok end, Responses ), ok. %% This should be called once the DC is up and running successfully %% It sets a flag on disk to true. When this is true on fail and %% restart the DC will load its state from disk -spec dc_successfully_started() -> ok. dc_successfully_started() -> dc_meta_data_utilities:dc_start_success(). %% Checks is the node is restarting when it had already been running %% If it is then all the background processes and connections are restarted -spec check_node_restart() -> boolean(). check_node_restart() -> case dc_meta_data_utilities:is_restart() of true -> ?LOG_INFO("This node was previously configured, will restart from previous config"), MyNode = node(), %% Load any env variables ok = dc_utilities:check_registered_global( stable_meta_data_server:generate_server_name(MyNode) ), ok = dc_meta_data_utilities:load_env_meta_data(), %% Ensure vnodes are running and meta_data ok = dc_utilities:ensure_local_vnodes_running_master(inter_dc_log_sender_vnode_master), ok = dc_utilities:ensure_local_vnodes_running_master(clocksi_vnode_master), ok = dc_utilities:ensure_local_vnodes_running_master(logging_vnode_master), ok = dc_utilities:ensure_local_vnodes_running_master(materializer_vnode_master), wait_init:wait_ready(MyNode), ok = dc_utilities:check_registered(meta_data_sender_sup), ok = dc_utilities:check_registered(meta_data_manager_sup), ok = dc_utilities:check_registered(inter_dc_query_router), ok = dc_utilities:check_registered(inter_dc_sub), ok = dc_utilities:check_registered(inter_dc_pub), ok = dc_utilities:check_registered(inter_dc_query_response_sup), ok = dc_utilities:check_registered(inter_dc_query_dealer), {ok, MetaDataName} = dc_meta_data_utilities:get_meta_data_name(), ok = meta_data_sender:start(MetaDataName), %% Start the timers sending the heartbeats ?LOG_INFO("Starting heartbeat sender timers"), Responses = dc_utilities:bcast_my_vnode_sync( logging_vnode_master, {start_timer, undefined} ), %% Be sure they all started ok, crash otherwise ok = lists:foreach( fun({_, ok}) -> ok end, Responses ), %% Reconnect this node to other DCs OtherDCs = dc_meta_data_utilities:get_dc_descriptors(), Responses3 = reconnect_dcs_after_restart(OtherDCs, MyNode), %% Ensure all connections were successful, crash otherwise Responses3 = [X = ok || X <- Responses3], true; false -> false end. -spec reconnect_dcs_after_restart([descriptor()], node()) -> [ok | inter_dc_conn_err()]. reconnect_dcs_after_restart(Descriptors, MyNode) -> ok = forget_dcs(Descriptors, [MyNode]), observe_dcs_sync(Descriptors, [MyNode]). %% This should be called when connecting the local DC to a new external DC -spec observe_dcs_sync([descriptor()]) -> [ok | inter_dc_conn_err()]. observe_dcs_sync(Descriptors) -> Nodes = dc_utilities:get_my_dc_nodes(), observe_dcs_sync(Descriptors, Nodes). -spec observe_dcs_sync([descriptor()], [node()]) -> [ok | inter_dc_conn_err()]. observe_dcs_sync(Descriptors, Nodes) -> {ok, SS} = dc_utilities:get_stable_snapshot(), DCs = lists:map( fun(DC) -> {observe_dc(DC, Nodes), DC} end, Descriptors ), lists:foreach( fun({Res, Desc = #descriptor{dcid = DCID}}) -> case Res of ok -> Value = vectorclock:get(DCID, SS), wait_for_stable_snapshot(DCID, Value), ok = dc_meta_data_utilities:store_dc_descriptors([Desc]); _ -> ok end end, DCs ), [Result1 || {Result1, _DC1} <- DCs]. -spec forget_dc(descriptor(), [node()]) -> ok. forget_dc(#descriptor{dcid = DCID}, Nodes) -> case DCID == dc_utilities:get_my_dc_id() of true -> ok; false -> ?LOG_NOTICE("Forgetting DC ~p", [DCID]), lists:foreach( fun(Node) -> ok = rpc:call(Node, inter_dc_query_dealer, del_dc, [DCID]) end, Nodes ), lists:foreach(fun(Node) -> ok = rpc:call(Node, inter_dc_sub, del_dc, [DCID]) end, Nodes) end. -spec forget_dcs([descriptor()]) -> ok. forget_dcs(Descriptors) -> Nodes = dc_utilities:get_my_dc_nodes(), forget_dcs(Descriptors, Nodes). -spec forget_dcs([descriptor()], [node()]) -> ok. forget_dcs(Descriptors, Nodes) -> lists:foreach( fun(Descriptor) -> forget_dc(Descriptor, Nodes) end, Descriptors ). %% Tell nodes within the DC to drop heartbeat ping messages from other %% DCs, used for debugging -spec drop_ping(boolean()) -> ok. drop_ping(DropPing) -> Responses = dc_utilities:bcast_vnode_sync(inter_dc_dep_vnode_master, {drop_ping, DropPing}), %% Be sure they all returned ok, crash otherwise ok = lists:foreach( fun({_, ok}) -> ok end, Responses ). %%%%%%%%%%%%% %% Utils wait_for_stable_snapshot(DCID, MinValue) -> case DCID == dc_utilities:get_my_dc_id() of true -> ok; false -> {ok, SS} = dc_utilities:get_stable_snapshot(), Value = vectorclock:get(DCID, SS), case Value > MinValue of true -> ?LOG_INFO("Connected to DC ~p", [DCID]), ok; false -> ?LOG_INFO("Waiting for DC ~p", [DCID]), timer:sleep(1000), wait_for_stable_snapshot(DCID, MinValue) end end.
apps/antidote/src/inter_dc_manager.erl
0.546496
0.401923
inter_dc_manager.erl
starcoder
-module(rstar_search). -export([search_within/2, search_near/3]). -ifdef(TEST). -compile(export_all). -endif. -include("../include/rstar.hrl"). % Searches the tree for any geometries within or intersecting % with the given geometry search_within(Root, Geo) -> search_within(Root, Geo, []). search_within(#geometry{value=Value}, Geo, Results) when is_record(Value, leaf) -> lists:foldl(fun(Child, Res) -> case rstar_geometry:intersect(Child, Geo) of undefined -> Res; % Add the child to the result set _ -> [Child | Res] end end, Results, Value#leaf.entries); search_within(Node, Geo, Results) -> lists:foldl(fun(Child, Res) -> case rstar_geometry:intersect(Child, Geo) of undefined -> Res; % Recurse into the matching node _ -> search_within(Child, Geo, Res) end end, Results, Node#geometry.value#node.children). % Returns the K nearest points to the given Geometry search_near(Node, Geo, K) -> Results = search_near_recursive(Node, Geo, K, gb_sets:new()), [R || {_, R} <- gb_sets:to_list(Results)]. search_near_recursive(#geometry{value=Value}, Geo, K, Results) when is_record(Value, leaf) -> lists:foldl(fun(C, Res) -> % Compute the center the distance to the target geometry Distance = rstar_geometry:min_dist(Geo, C), case gb_sets:size(Res) of % Handle the case where we have less than K matches L when L < K -> gb_sets:add({Distance, C}, Res); % Handle when we have K matches already _ -> {MaxDist, _} = MaxNN = gb_sets:largest(Res), if MaxDist > Distance -> gb_sets:add({Distance, C}, gb_sets:delete(MaxNN, Res)); true -> Res end end end, Results, Value#leaf.entries); search_near_recursive(Node, Geo, K, Results) -> % Create an Active Branch List based on our children sorted by their % minimum distance from the query region Children = Node#geometry.value#node.children, Sorted = lists:sort([{rstar_geometry:min_dist(Geo, C), C} || C <- Children]), ActiveBranchList = [B || {_, B} <- Sorted], % Prune the children and iterate over the branches Pruned = prune_branches(Geo, K, ActiveBranchList, Results), search_near_branches(Pruned, Geo, K, Results). % Helper to iterate through the Active Branch List gathering results % Return the result set when there are no further branches search_near_branches([], _, _, Results) -> Results; % Check each branch and prune after updating the results search_near_branches([Branch | ABL], Geo, K, Results) -> NewResults = search_near_recursive(Branch, Geo, K, Results), Pruned = prune_branches(Geo, K, ABL, NewResults), search_near_branches(Pruned, Geo, K, NewResults). % Prunes the search branches based on the existing results prune_branches(Geo, K, Branches, Results) -> case gb_sets:size(Results) of % Do not prune if we don't have K neighbors yet L when L < K -> Branches; _ -> {MaxDist, _} = gb_sets:largest(Results), lists:filter(fun(B) -> rstar_geometry:min_dist(Geo, B) =< MaxDist end, Branches) end.
src/rstar_search.erl
0.620047
0.659967
rstar_search.erl
starcoder
-module(poly_SUITE). -include_lib("eunit/include/eunit.hrl"). -export([all/0, init_per_testcase/2, end_per_testcase/2]). -export( [ eval_test/1, from_fr_test/1, interpolate_with_fr_test/1, zeroize_test/1, self_subtract_test/1, add_zero_test/1, sub_zero_test/1, mul_poly_test/1, add_different_sizes_poly_test/1, negative_cmp_test/1, f_of_x_test/1, serde_test/1 ] ). all() -> [ eval_test, from_fr_test, interpolate_with_fr_test, zeroize_test, self_subtract_test, add_zero_test, sub_zero_test, mul_poly_test, add_different_sizes_poly_test, negative_cmp_test, f_of_x_test, serde_test ]. init_per_testcase(_, Config) -> Config. end_per_testcase(_, Config) -> Config. eval_test(_Config) -> %% poly = 5x³ + x - 2. Poly = tc_poly:from_coeffs([-2, 1, 0, 5]), Samples = [{-1, -8}, {2, 40}, {3, 136}, {5, 628}], %% check f(a) = b ?assert( lists:all( fun({Point, Answer}) -> AnswerFr = tc_fr:into(Answer), EvalFr = tc_poly:eval(Poly, Point), tc_fr:cmp(AnswerFr, EvalFr) end, Samples ) ), %% poly can be interpolated because num_sample >= degree + 1 ?assert(tc_poly:cmp(Poly, tc_poly:interpolate(Samples))), ?assertEqual(3, tc_poly:degree(Poly)), ok. from_fr_test(_Config) -> %% poly = 5x³ + x - 2. Coeffs = [-2, 1, 0, 5], Frs = [tc_fr:into(I) || I <- Coeffs], Poly = tc_poly:from_frs(Frs), Samples = [{-1, -8}, {2, 40}, {3, 136}, {5, 628}], %% check f(a) = b ?assert( lists:all( fun({Point, Answer}) -> AnswerFr = tc_fr:into(Answer), EvalFr = tc_poly:eval(Poly, Point), tc_fr:cmp(AnswerFr, EvalFr) end, Samples ) ), %% poly can be interpolated because num_sample >= degree + 1 ?assert(tc_poly:cmp(Poly, tc_poly:interpolate(Samples))), ?assertEqual(3, tc_poly:degree(Poly)), ok. interpolate_with_fr_test(_Config) -> %% poly = 5x³ + x - 2. Poly = tc_poly:from_coeffs([-2, 1, 0, 5]), Samples = [{-1, -8}, {2, 40}, {3, 136}, {5, 628}], FrSamples = [{tc_fr:into(A), tc_fr:into(B)} || {A, B} <- Samples], %% check f(a) = b ?assert( lists:all( fun({Point, Answer}) -> AnswerFr = tc_fr:into(Answer), EvalFr = tc_poly:eval(Poly, Point), tc_fr:cmp(AnswerFr, EvalFr) end, Samples ) ), %% poly can be interpolated because num_sample >= degree + 1 ?assert(tc_poly:cmp(Poly, tc_poly:interpolate(Samples))), %% we should also be able to interpolate poly using fr values of samples ?assert(tc_poly:cmp(Poly, tc_poly:interpolate_from_fr(FrSamples))), ?assertEqual(3, tc_poly:degree(Poly)), ok. zeroize_test(_Config) -> %% random_poly -> zeroize -> is_zero ?assert(tc_poly:is_zero(tc_poly:zeroize(tc_poly:random(4)))), BiPoly = tc_bipoly:random(3), BiCommitment = tc_bipoly:commitment(BiPoly), ZeroBiPoly = tc_bipoly:zeroize(BiPoly), ZeroBiCommitment = tc_bipoly:commitment(ZeroBiPoly), ?assertEqual(false, tc_bicommitment:cmp(ZeroBiCommitment, BiCommitment)), ?assert( tc_g1:cmp( tc_g1:zero(), tc_bicommitment:eval( ZeroBiCommitment, rand:uniform(100), rand:uniform(100) ) ) ), ok. self_subtract_test(_Config) -> %% f(x) - f(x) = 0 P = tc_poly:random(2), ?assert(tc_poly:cmp(tc_poly:zero(), tc_poly:sub(P, P))). add_zero_test(_Config) -> %% f(x) + 0 = f(x) P = tc_poly:random(2), ?assert(tc_poly:cmp(P, tc_poly:add_scalar(0, P))). sub_zero_test(_Config) -> %% f(x) - 0 = f(x) P = tc_poly:random(2), ?assert(tc_poly:cmp(P, tc_poly:sub_scalar(0, P))). mul_poly_test(_Config) -> %% p1 = (x² + 1) %% p2 = (x - 1) %% p1 * p2 = p3 = x³ - x² + x - 1 %% p1(p) * p2(p) = p3(p) P1 = tc_poly:from_coeffs([1, 0, 1]), P2 = tc_poly:from_coeffs([-1, 1]), P3 = tc_poly:from_coeffs([-1, 1, -1, 1]), ?assert(tc_poly:cmp(P3, tc_poly:mul(P1, P2))), P1Eval = tc_poly:eval(P1, 5), P2Eval = tc_poly:eval(P2, 5), P3Eval = tc_poly:eval(P3, 5), ?assert(tc_fr:cmp(tc_fr:into(26), P1Eval)), ?assert(tc_fr:cmp(tc_fr:into(4), P2Eval)), ?assert(tc_fr:cmp(tc_fr:into(104), P3Eval)), ok. add_different_sizes_poly_test(_Config) -> P1 = tc_poly:random(5), P2 = tc_poly:random(8), AddedPoly = tc_poly:add(P1, P2), %% result should be of degree 8 ?assertEqual(8, tc_poly:degree(AddedPoly)), %% if we subtract B from the result, we should get back A with degree 5 SubPoly = tc_poly:sub(AddedPoly, P2), ?assertEqual(5, tc_poly:degree(SubPoly)), ?assert(tc_poly:cmp(P1, SubPoly)), ok. negative_cmp_test(_Config) -> P1 = tc_poly:random(5), P2 = tc_poly:add(P1, P1), %% since P1 /= 2*P1 ?assertEqual(false, tc_poly:cmp(P1, P2)), ok. f_of_x_test(_Config) -> %% f(x) = 5x², f(2) = 5 * 2 * 2 P = tc_poly:from_coeffs([0, 0, 5]), Eval = tc_poly:eval(P, 2), ?assert(tc_fr:cmp(tc_fr:into(5 * 2 * 2), Eval)), ok. serde_test(_Config) -> %% f(x) = 5x², f(2) = 5 * 2 * 2 P = tc_poly:from_coeffs([0, 0, 5]), SerializedPoly = tc_poly:serialize(P), DeserializePoly = tc_poly:deserialize(SerializedPoly), ?assert(tc_fr:cmp(tc_fr:into(5 * 2 * 2), tc_poly:eval(DeserializePoly, 2))), ok.
test/poly_SUITE.erl
0.712332
0.612628
poly_SUITE.erl
starcoder
%% @copyright 2011 <NAME> %% @author <NAME> <<EMAIL>> %% @version {@date} {@time} %% @doc Function call timing analysis. Parts lifted from hipe_timer.erl (thanks OTP) %% @end -module(funalysis). -export([advanced/2, par_advanced/3]). -export([par_proc_iterations/2, analyse/2]). t(F) -> NullTime = empty_time(), {Time, _} = timer:tc(F, []), erlang:max(Time - NullTime, 0). % -spec empty_time() -> microseconds(). empty_time() -> {Time, _} = timer:tc(fun () -> ok end, []), Time. advanced(_Fun, I) when I < 2 -> false; advanced(Fun, Iterations) -> Measurements = [t(Fun) || _ <- lists:seq(1, Iterations)], analyse(Measurements, Iterations). par_advanced(Fun, Procs, Iterations) when Iterations > Procs, Procs > 0 -> Master = self(), Ref = make_ref(), Pids = [ spawn( fun () -> timing_loop(Master, Ref, Fun, ProcIters ) end ) || ProcIters <- par_proc_iterations(Procs, Iterations) ], Measurements = lists:flatmap( fun (Pid) -> receive {Pid, Ref, M} -> M end end, Pids), analyse(Measurements, Iterations). par_proc_iterations(Procs, Iterations) -> Rem = (Iterations rem Procs), [ case N =< Rem of true -> (Iterations div Procs) + 1; false -> (Iterations div Procs) end || N <- lists:seq(1, Procs) ]. timing_loop(Master, Ref, Fun, ProcIters) -> Master ! {self(), Ref, [t(Fun) || _ <- lists:seq(1, ProcIters) ]}, exit(normal). analyse(Measurements, Iterations) -> Wallclock = Measurements, WMin = lists:min(Wallclock), WMax = lists:max(Wallclock), WMean = mean(Wallclock), WMedian = median(Wallclock), WVariance = variance(Wallclock), WStddev = stddev(Wallclock), WVarCoff = 100 * WStddev / WMean, WSum = lists:sum(Wallclock), [{wallclock,[{min, WMin}, {max, WMax}, {mean, WMean}, {median, WMedian}, {variance, WVariance}, {stdev, WStddev}, {varcoff, WVarCoff}, {sum, WSum}, {values, Wallclock}]}, {iterations, Iterations}]. split(M) -> split(M, [], []). split([{W,R}|More], AccW, AccR) -> split(More, [W|AccW], [R|AccR]); split([], AccW, AccR) -> {AccW, AccR}. mean(L) -> mean(L, 0, 0). mean([V|Vs], No, Sum) -> mean(Vs, No+1, Sum+V); mean([], No, Sum) when No > 0 -> Sum/No; mean([], _No, _Sum) -> exit(empty_list). median(L) -> S = length(L), SL = lists:sort(L), case even(S) of true -> (lists:nth((S div 2), SL) + lists:nth((S div 2) + 1, SL)) / 2; false -> lists:nth((S div 2), SL) end. even(S) -> (S band 1) =:= 0. %% diffs(L, V) -> %% [X - V || X <- L]. square_diffs(L, V) -> [(X - V) * (X - V) || X <- L]. variance(L) -> Mean = mean(L), N = length(L), if N > 1 -> lists:sum(square_diffs(L,Mean)) / (N-1); true -> exit('too few values') end. stddev(L) -> math:sqrt(variance(L)).
funalysis.erl
0.532425
0.44571
funalysis.erl
starcoder
% Licensed under the Apache License, Version 2.0 (the "License"); you may not % use this file except in compliance with the License. You may obtain a copy of % the License at % % http://www.apache.org/licenses/LICENSE-2.0 % % Unless required by applicable law or agreed to in writing, software % distributed under the License is distributed on an "AS IS" BASIS, WITHOUT % WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the % License for the specific language governing permissions and limitations under % the License. -module(ejson). -export([encode/1, decode/1]). -on_load(init/0). init() -> SoName = case code:priv_dir(ejson) of {error, bad_name} -> case filelib:is_dir(filename:join(["..", priv])) of true -> filename:join(["..", priv, ejson]); false -> filename:join([priv, ejson]) end; Dir -> filename:join(Dir, ejson) end, (catch erlang:load_nif(SoName, 0)), case erlang:system_info(otp_release) of "R13B03" -> true; _ -> ok end. decode(undefined) -> throw({invalid_json, undefined}); decode(IoList) -> try nif_decode(IoList) catch exit:ejson_nif_not_loaded -> erl_decode(IoList) end. encode(EJson) -> try nif_encode(EJson) catch exit:ejson_nif_not_loaded -> erl_encode(EJson) end. nif_decode(IoList) -> case reverse_tokens(IoList) of {ok, ReverseTokens} -> [[EJson]] = make_ejson(ReverseTokens, [[]]), EJson; Error -> throw({invalid_json, {Error, IoList}}) end. erl_decode(IoList) -> try (mochijson2:decoder([{object_hook, fun({struct, L}) -> {L} end}]))(IoList) catch _Type:Error -> throw({invalid_json, {Error, IoList}}) end. nif_encode(EJson) -> RevList = encode_rev(EJson), final_encode(lists:reverse(lists:flatten([RevList]))). erl_encode(EJson) -> Opts = [{handler, fun mochi_encode_handler/1}], iolist_to_binary((mochijson2:encoder(Opts))(EJson)). mochi_encode_handler({L}) when is_list(L) -> {struct, L}; mochi_encode_handler(Bad) -> exit({json_encode, {bad_term, Bad}}). % Encode the json into a reverse list that's almost an iolist % everything in the list is the final output except for tuples with % {0, Strings} and {1, Floats}, which are to be converted to strings % inside the NIF. encode_rev(true) -> <<"true">>; encode_rev(false) -> <<"false">>; encode_rev(null) -> <<"null">>; encode_rev(I) when is_integer(I) -> list_to_binary(integer_to_list(I)); encode_rev(S) when is_binary(S) -> {0, S}; encode_rev(S) when is_atom(S) -> {0, list_to_binary(atom_to_list(S))}; encode_rev(F) when is_float(F) -> {1, F}; encode_rev({Props}) when is_list(Props) -> encode_proplist_rev(Props, [<<"{">>]); encode_rev(Array) when is_list(Array) -> encode_array_rev(Array, [<<"[">>]); encode_rev(Bad) -> throw({json_encode, {bad_term, Bad}}). encode_array_rev([], Acc) -> [<<"]">> | Acc]; encode_array_rev([Val | Rest], [<<"[">>]) -> encode_array_rev(Rest, [encode_rev(Val), <<"[">>]); encode_array_rev([Val | Rest], Acc) -> encode_array_rev(Rest, [encode_rev(Val), <<",">> | Acc]). encode_proplist_rev([], Acc) -> [<<"}">> | Acc]; encode_proplist_rev([{Key,Val} | Rest], [<<"{">>]) -> encode_proplist_rev( Rest, [encode_rev(Val), <<":">>, {0, as_binary(Key)}, <<"{">>]); encode_proplist_rev([{Key,Val} | Rest], Acc) -> encode_proplist_rev( Rest, [encode_rev(Val), <<":">>, {0, as_binary(Key)}, <<",">> | Acc]). as_binary(B) when is_binary(B) -> B; as_binary(A) when is_atom(A) -> list_to_binary(atom_to_list(A)); as_binary(L) when is_list(L) -> list_to_binary(L). make_ejson([], Stack) -> Stack; make_ejson([0 | RevEvs], [ArrayValues, PrevValues | RestStack]) -> % 0 ArrayStart make_ejson(RevEvs, [[ArrayValues | PrevValues] | RestStack]); make_ejson([1 | RevEvs], Stack) -> % 1 ArrayEnd make_ejson(RevEvs, [[] | Stack]); make_ejson([2 | RevEvs], [ObjValues, PrevValues | RestStack]) -> % 2 ObjectStart make_ejson(RevEvs, [[{ObjValues} | PrevValues] | RestStack]); make_ejson([3 | RevEvs], Stack) -> % 3 ObjectEnd make_ejson(RevEvs, [[] | Stack]); make_ejson([{0, Value} | RevEvs], [Vals | RestStack] = _Stack) -> % {0, IntegerString} make_ejson(RevEvs, [[list_to_integer(binary_to_list(Value)) | Vals] | RestStack]); make_ejson([{1, Value} | RevEvs], [Vals | RestStack] = _Stack) -> % {1, FloatString} make_ejson(RevEvs, [[list_to_float(binary_to_list(Value)) | Vals] | RestStack]); make_ejson([{3, String} | RevEvs], [[PrevValue|RestObject] | RestStack] = _Stack) -> % {3 , ObjectKey} make_ejson(RevEvs, [[{String, PrevValue}|RestObject] | RestStack]); make_ejson([Value | RevEvs], [Vals | RestStack] = _Stack) -> make_ejson(RevEvs, [[Value | Vals] | RestStack]). reverse_tokens(_) -> exit(ejson_nif_not_loaded). final_encode(_) -> exit(ejson_nif_not_loaded).
src/ejson/ejson.erl
0.504639
0.44059
ejson.erl
starcoder
-module(nanometer). -type piece() :: atom() | integer() | binary(). -type name() :: [piece()]. -type options() :: proplists:proplist(). -type values() :: [{piece(), number()}]. -type type() :: counter | meter | gauge | histogram | external. -type error() :: {error, any()}. -callback create(name(), options()) -> any(). -callback notify(name(), number()) -> any(). -callback acquire(name()) -> values(). -callback release(name()) -> any(). -callback reset(name()) -> any(). -callback exists(name()) -> boolean(). -callback list() -> [name()]. -callback count() -> non_neg_integer(). %% low-level API -export([create/3, notify/3, acquire/2, release/2, reset/2, list/1, types/0, callback_module/1, stats/0]). %% convenience API -export([count/2, meter/2, gauge/2, histogram/2, histogram/3, time/2, time/3]). -export_type([name/0, error/0, options/0, values/0, type/0]). -type result() :: ok | not_started | {error, any()}. %% don't die! -define(SAVE_OUR_SOULS(X), case catch(begin X end) of {'EXIT', {badarg, [{ets, _, _, _} | _]}} -> not_started; {'EXIT', {Error, _}} -> {error, Error}; _ -> ok end). -spec create(type(), name(), options()) -> result(). create(Type, Name, Options) -> Mod = callback_module(Type), ?SAVE_OUR_SOULS(Mod:create(Name, Options)). -spec notify(type(), name(), number()) -> result(). notify(Type, Name, Number) -> Mod = callback_module(Type), ?SAVE_OUR_SOULS(Mod:notify(Name, Number)). -spec acquire(type(), name()) -> values(). acquire(Type, Name) -> Mod = callback_module(Type), Mod:acquire(Name). -spec release(type(), name()) -> ok. release(Type, Name) -> Mod = callback_module(Type), Mod:release(Name), ok. -spec reset(type(), name()) -> ok. reset(Type, Name) -> Mod = callback_module(Type), Mod:reset(Name), ok. -spec list(type()) -> [name()]. list(Type) -> Mod = callback_module(Type), Mod:list(). -spec types() -> [type()]. types() -> [counter, meter, gauge, histogram, external]. -spec callback_module(type()) -> module(). callback_module(counter) -> nanometer_counter; callback_module(meter) -> nanometer_meter; callback_module(gauge) -> nanometer_gauge; callback_module(histogram) -> nanometer_histogram; callback_module(external) -> nanometer_external. -spec stats() -> values(). stats() -> lists:map( fun(Type) -> Mod = callback_module(Type), {Type, Mod:count()} end, types()). -spec count(name(), integer()) -> result(). count(Name, Delta) when is_integer(Delta) -> ?SAVE_OUR_SOULS( begin nanometer_counter:create(Name, []), nanometer_counter:notify(Name, Delta) end ). -spec meter(name(), integer()) -> result(). meter(Name, Delta) when is_integer(Delta) -> ?SAVE_OUR_SOULS( begin nanometer_meter:create(Name, []), nanometer_meter:notify(Name, Delta) end ). -spec gauge(name(), number()) -> result(). gauge(Name, Value) -> ?SAVE_OUR_SOULS( begin nanometer_gauge:create(Name, []), nanometer_gauge:notify(Name, Value) end ). -spec histogram(name(), number()) -> result(). histogram(Name, Value) -> histogram(Name, Value, []). -spec histogram(name(), number(), options()) -> result(). histogram(Name, Value, Options) -> ?SAVE_OUR_SOULS( begin nanometer_histogram:create(Name, Options), nanometer_histogram:notify(Name, Value) end ). -spec time(name(), fun(() -> X)) -> X when X :: any(). time(Name, Fun) -> time(Name, Fun, []). -spec time(name(), fun(() -> X), options()) -> X when X :: any(). time(Name, Fun, Options) -> Time0 = nanometer_compat:monotonic_us(), Result = Fun(), Time1 = nanometer_compat:monotonic_us(), histogram(Name, Time1 - Time0, Options), Result.
src/nanometer.erl
0.551815
0.431884
nanometer.erl
starcoder
%% ============================================================================= %% bondy_consistent_hashing.erl - %% Copyright (c) 2016-2021 Leapsight. All rights reserved. %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %% ============================================================================= %% ----------------------------------------------------------------------------- %% @doc %% It uses Jump Consistent Hash algorithm described in %% [A Fast, Minimal Memory, Consistent Hash Algorithm](https://arxiv.org/ftp/ %% arxiv/papers/1406/1406.2294.pdf). %% @end %% ----------------------------------------------------------------------------- -module(bondy_consistent_hashing). -define(MAGIC, 16#27BB2EE687B0B0FD). -define(MASK, 16#FFFFFFFFFFFFFFFF). -export([bucket/2]). -export([bucket/3]). %% ============================================================================= %% API %% ============================================================================= %% ----------------------------------------------------------------------------- %% @doc %% @end %% ----------------------------------------------------------------------------- -spec bucket(Key :: term(), Buckets :: pos_integer()) -> Bucket :: integer(). bucket(Key, Buckets) -> bucket(Key, Buckets, jch). %% ----------------------------------------------------------------------------- %% @doc %% @end %% ----------------------------------------------------------------------------- -spec bucket(Key :: term(), Buckets :: pos_integer(), Algo :: atom()) -> Bucket :: integer(). bucket(_, 1, _) -> 0; bucket(Key, Buckets, jch) when is_integer(Key) andalso is_integer(Buckets) andalso Buckets > 1 -> jump_consistent_hash(Key, Buckets); bucket(Key, Buckets, Algo) -> bucket(erlang:phash2(Key), Buckets, Algo). %% ============================================================================= %% PRIVATE %% ============================================================================= jump_consistent_hash(Key, N) -> jump_consistent_hash(Key, N, -1, 0). %% ----------------------------------------------------------------------------- %% @private %% @doc %% The following is the C++ implementation in %% A Fast, Minimal Memory, Consistent Hash Algorithm %% https://arxiv.org/pdf/1406.2294.pdf %% %% @end %% ----------------------------------------------------------------------------- %% static int32_t jump_consistent_hash(uint64_t key, int32_t num_buckets) { %% int64_t b = -1, j = 0; %% while (j < num_buckets) { %% b = j; %% key = key * 2862933555777941757ULL + 1; %% j = (b + 1) * ((double)(1LL << 31) / (double)((key >> 33) + 1)); %% } %% return (int32_t)b; %% } %% jump_consistent_hash(Key, N, _, J0) when J0 < N -> %% B1 = J0, NewKey = (Key * ?MAGIC + 1) band ?MASK, J1 = trunc((J0 + 1) * ((1 bsl 31) / ((NewKey bsr 33) + 1)) ), jump_consistent_hash(NewKey, N, J0, J1); jump_consistent_hash(_, _, B, _) -> B. -ifdef(TEST). -include_lib("eunit/include/eunit.hrl"). jch_test_() -> Cases = %% {Expect, Key, Buckets} [ {0, 0, 1}, {0, 3, 1}, {0, 0, 2}, {1, 4, 2}, {0, 7, 2}, {55, 1, 128}, {120, 129, 128}, {0, 0, 100000000}, {38172097, 128, 100000000}, {1644467860, 128, 2147483648}, {92, 18446744073709551615, 128} ], [?_assertEqual(Expect, bucket(K, B, jch)) || {Expect, K, B} <- Cases]. -endif.
apps/bondy/src/bondy_consistent_hashing.erl
0.555073
0.433442
bondy_consistent_hashing.erl
starcoder
%% Copyright 2016-2017 TensorHub, Inc. %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. -module(guild_collector_protocol). -export([new_input_buffer/0, input/2]). -define(PART_DELIM, <<"\n\n">>). new_input_buffer() -> {[], undefined}. input(Buf, Bin) -> Now = input_timestamp(), handle_parts(split_input(Bin), Now, Buf). input_timestamp() -> erlang:system_time(milli_seconds). split_input(Bin) -> re:split(Bin, ?PART_DELIM, [{return, binary}]). handle_parts([<<>>], _Now, Buf) -> finalize_decoded(Buf); handle_parts([Part], Now, Buf) -> finalize_decoded(buffer_part(Part, Now, Buf)); handle_parts([Part|Rest], Now, Buf) -> NextBuf = finalize_parts(buffer_part(Part, Now, Buf)), handle_parts(Rest, Now, finalize_parts(NextBuf)). buffer_part(Part, Now, {Decoded, undefined}) -> {Decoded, {Now, [Part]}}; buffer_part(Part, _Now, {Decoded, {Time, Parts}}) -> {Decoded, {Time, [Part|Parts]}}. finalize_decoded({Decoded, Working}) -> {lists:reverse(Decoded), {[], Working}}. finalize_parts({Decoded, {Time, Parts}}) -> {[{Time, decode(lists:reverse(Parts))}|Decoded], undefined}; finalize_parts({Decoded, undefined}) -> {Decoded, undefined}. decode([<<>>]) -> eof; decode(Bin) -> case guild_json:try_decode(Bin) of {ok, Decoded} -> format_decoded(Decoded); {error, _Err} -> {invalid, Bin} end. format_decoded({[{<<"kv">>, {KeyVals}}]}) -> {kv, KeyVals}; format_decoded({[{<<"ktsv">>, {KeyTimeStepVals}}]}) -> {ktsv, KeyTimeStepVals}; format_decoded(Other) -> {other, Other}.
src/guild_collector_protocol.erl
0.573201
0.413004
guild_collector_protocol.erl
starcoder
-module(gen_rate_limiter). -behaviour(gen_server). %% API -export([start_link/2, run/2]). %% Gen Server -export([init/1, handle_call/3, handle_cast/2, handle_info/2, terminate/2, code_change/3]). %% This module defines a generic rate limiter (rl) intended to be used with APIs that %% only allow a certain number of requests per timespan. %% =================================================================== %% Rate Limiter Callback Module %% =================================================================== %% The callback module has to define 4 function that define the behaviour of the %% rl: -callback init(Args :: term()) -> State :: term(). %% Called after startup. Args is the same as in %% gen_rate_limiter:start_link(Module, Args) -> [See Gen Server start_link return] -callback before_run(State :: term()) -> State :: term(). %% Called every time before Module:run is called. Runs in the rl process. -callback run(Args :: term(), State :: term()) -> {Return :: term(), State :: term()}. %% Called when gen_rate_limiter:run(Pid, Args) -> Return is called. Runs in %% the caller's process, if this call fails the rl's old State will be kept. -callback to_wait(State :: term()) -> {ToWait :: non_neg_integer(), State :: term()}. %% Called every time after Module:run was called (even if it failed). ToWait %% (measured in milliseconds) determines how long the rl will sleep before %% allowing Module:run to be called again. Runs in the rl process. %% =================================================================== %% Gen Rate Limiter %% =================================================================== start_link(Module, Args) -> gen_server:start_link(?MODULE, {Module, Args}, []). run(Pid, Args) -> {M, RLState} = gen_server:call(Pid, {request, self()}, infinity), {Return, NewRLState} = M:run(Args, RLState), Pid ! {ok, NewRLState}, Return. init({Module, Args}) -> {ok, {Module, Module:init(Args)}}. handle_call({request, Pid}, From, {Module, RLState}) when is_pid(Pid), is_atom(Module) -> IntermediateRLState = Module:before_run(RLState), MRef = monitor(process, Pid), gen_server:reply(From, {Module, IntermediateRLState}), AfterRunRLState = receive {ok, S} -> demonitor(MRef, [flush]), S; {'DOWN', MRef, process, Pid, Reason} -> io:format("Gen Rate Limiter: Run of ~p failed with reason ~p.~n", [Pid, Reason]), IntermediateRLState end, {ToWait, FinalRLState} = Module:to_wait(AfterRunRLState), if ToWait > 0 -> timer:sleep(ToWait); true -> ok end, {noreply, {Module, FinalRLState}}; handle_call(Msg, {Pid, _Tag} = _From, {Module, RLState} = State) -> io:format("Gen Rate Limiter ~p (callback module ~p): Received unexpected call ~p from ~p while in state ~p.", [self(), Module, Msg, Pid, RLState]), {noreply, State}. handle_cast(Msg, {Module, RLState} = State) -> io:format("Gen Rate Limiter ~p (callback module ~p): Received unexpected cast ~p while in state ~p.", [self(), Module, Msg, RLState]), {noreply, State}. handle_info(Msg, {Module, RLState} = State) -> io:format("Gen Rate Limiter ~p (callback module ~p): Received unexpected info ~p while in state ~p.", [self(), Module, Msg, RLState]), {noreply, State}. code_change(_OldVsn, State, _Extra) -> {ok, State}. terminate(_Reason, _State) -> ok.
src/gen_rate_limiter.erl
0.583322
0.410993
gen_rate_limiter.erl
starcoder
%%============================================================================== %% Copyright 2013-2021 <NAME> <<EMAIL>> %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %%============================================================================== %%%------------------------------------------------------------------- %%% @doc %%% A lazy lib. %%% %%% This library provides a simple pull oriented interface towards different %%% data sources using a lazy data abstraction. %%% %%% The primary use case is the decoding of protocols where the sizes of parts %%% are not known until they have been completely decoded and one wishes to %%% decouple the handling of the data source and the protocol code. %%% %%% The user of the library provides a promise of data: a function that %%% given a timeout returns either a data or eol (end of lazy) within %%% the timespan given by the timeout. A promise can also have a state %%% in which case the initial state has to be provided with the function. %%% The stateful promise function must retun a tuple of {Data, NewState}. %%% If the timeout is eol is given the promise should return eol and any %%% resources, such as streams and sockets, should be deallocated. %%% %%% The library provides the functions create/1 and create/2 to create %%% lazy data structures given a promise. The structure can then be %%% applied to a timeout to generate a tuple of data and new lazy data %%% {Data, Lazy} or eol. %%% %%% Several utility functions are provided to construct new LazyData. %%% %%% One function is provided that creates an "empty" LazyData empty/0. %%% %%% Two functions to add a data to existing lazy data are provided: %%% prepend/2 and append/2 that adds the data before or after the lazy data %%% respectively. %%% %%% One function that combines lazy data is provided: concat/2. %%% %%% A number of utility funtions are provided, functioning both as simple %%% examples as well as convenience for basic uses: %%% list_to_data/1, iolist_to_data/1 %%% tcp_to_data/2, tcp_to_data/3,tcp_to_data/4, tcp_socket_to_data/1 %%% tcp_reconnect_to_data/3 %%% file_to_data/2, file_stream_to_data/2 %%% %%% N.B. This library relies heavily on the construction of lambda functions %%% and for the sake of clearity and efficiency should be avoided if a %%% more direct approach, that does not suffer heavily from these very %%% drawback itself, exists. %%% @end %%% %% @author <NAME> <<EMAIL>> %% @copyright (C) 2013-2021, <NAME> <<EMAIL>> %%%------------------------------------------------------------------- -module(lazy). -copyright('<NAME> <<EMAIL>>'). %% Library functions -export([create/1, create/2]). -export([empty/0, prepend/2, append/2, concat/2]). -export([list_to_data/1, iolist_to_data/1, tcp_to_data/2, tcp_to_data/3,tcp_to_data/4, tcp_socket_to_data/1, tcp_reconnect_to_data/3, file_to_data/2, file_stream_to_data/2 ]). %% Types -type data(Type) :: fun(([timeout() | eol]) -> {Type, data(Type)}) | eol. -type promise(Type) :: fun(([timeout() | eol]) -> Type | eol). -type promise(Type, State) :: fun(([timeout() | eol], State) -> {Type, State} | eol). %% Exported Types -export_type([data/1, promise/1, promise/2]). %% =================================================================== %% Library functions. %% =================================================================== %%-------------------------------------------------------------------- %% Function: create(Promise) -> LazyData %% @doc %% Given a promise, lazy data is created. %% @end %%-------------------------------------------------------------------- -spec create(promise(Type)) -> data(Type). %%-------------------------------------------------------------------- create(F) -> fun(Timeout) -> case F(Timeout) of eol -> eol; Data -> {Data, create(F)} end end. %%-------------------------------------------------------------------- %% Function: create(Promise, State) -> LazyData %% @doc %% Given a promise and an initial state, stateful lazy data is created. %% @end %%-------------------------------------------------------------------- -spec create(promise(Type, State), State) -> data(Type). %%-------------------------------------------------------------------- create(F, State) -> fun(Timeout) -> case F(Timeout, State) of {Data, State1} -> {Data, create(F, State1)}; eol -> eol end end. %%-------------------------------------------------------------------- %% Function: empty() -> LazyData. %% @doc %% Creates an empty LazyData that just returns eol. %% @end %%-------------------------------------------------------------------- -spec empty() -> data(_). %%-------------------------------------------------------------------- empty() -> fun(_) -> eol end. %%-------------------------------------------------------------------- %% Function: prepend(Data, LazyData) -> LazyData. %% @doc %% Lazy data is constructed from data and lazy data where when cosumed %% the data comes before any of the lazy data. %% @end %%-------------------------------------------------------------------- -spec prepend(Type, data(Type)) -> data(Type). %%-------------------------------------------------------------------- prepend(Data, Lazy) -> fun(eol) -> Lazy(eol); (_) -> {Data, Lazy} end. %%-------------------------------------------------------------------- %% Function: append(Data, LazyData) -> LazyData. %% @doc %% Lazy data is constructed from data and lazy data where when cosumed %% the data comes after all of the lazy data. %% @end %%-------------------------------------------------------------------- -spec append(Type, data(Type)) -> data(Type). %%-------------------------------------------------------------------- append(Data, Lazy) -> fun(eol) -> Lazy(eol); (Timeout) -> case Lazy(Timeout) of eol -> {Data, empty()}; {Data1, Lazy1} -> {Data1, append(Data, Lazy1)} end end. %%-------------------------------------------------------------------- %% Function: concat(LazyData1, LazyData2) -> LazyData. %% @doc %% Lazy data is constructed from lazy data and lazy data where when cosumed %% the LazyData1 comes before all the data of of the LazyData2. %% @end %%-------------------------------------------------------------------- -spec concat(data(Type), data(Type)) -> data(Type). %%-------------------------------------------------------------------- concat(Lazy1, Lazy2) -> fun(eol) -> Lazy1(eol), Lazy2(eol); (Timeout) -> case Lazy1(Timeout) of eol -> Lazy2(Timeout); {Data1, Lazy11} -> {Data1, concat(Lazy11, Lazy2)} end end. %%-------------------------------------------------------------------- %% Function: list_to_data(DataList) -> LazyData %% @doc %% Lazy data is constructed from a list. %% @end %%-------------------------------------------------------------------- -spec list_to_data([Type]) -> data(Type). %%-------------------------------------------------------------------- list_to_data(List) -> Promise = fun(eol, _) -> eol; (_, []) -> eol; (_, [H | T]) -> {H, T} end, create(Promise, List). %%-------------------------------------------------------------------- %% Function: iolist_to_data(IOList) -> LazyBinary %% @doc %% A Lazy binary is constructed from an iolist. %% @end %%-------------------------------------------------------------------- -spec iolist_to_data(iolist()) -> data(binary()). %%-------------------------------------------------------------------- iolist_to_data(List) -> Promise = fun(eol, _) -> eol; (_, []) -> eol; (_, H) when is_binary(H) -> {H, []}; (_, [H | T]) when is_binary(H) -> {H, T}; (_, [H | T]) -> {iolist_to_binary(H), T} end, create(Promise, List). %%-------------------------------------------------------------------- %% Function: tcp_to_data(Host, Port) -> LazyBinary | Error %% @doc %% A Lazy binary is constructed from the socket that opening a tcp connetion %% to the host in binary mode with packet size 0. If an error occurs during %% connection an error is returned. %% @end %%-------------------------------------------------------------------- -spec tcp_to_data(HostName, Port) -> data(binary()) | {error, inet:posix()} when HostName:: inet:ip_address() | inet:hostname(), Port ::inet:port_number(). %%-------------------------------------------------------------------- tcp_to_data(HostName, Port) -> tcp_to_data(HostName, Port, infinity). %%-------------------------------------------------------------------- %% Function: tcp_to_data(Host, Port, Timeout) -> LazyBinary | Error %% @doc %% A Lazy binary is constructed from the socket that opening a tcp connetion %% to the host in binary mode with packet size 0. If an error or timeout %% occurs during connection an error is returned. %% @end %%-------------------------------------------------------------------- -spec tcp_to_data(HostName, Port, timeout()) -> data(binary()) | {error, inet:posix()} when HostName:: inet:ip_address() | inet:hostname(), Port ::inet:port_number(). %%-------------------------------------------------------------------- tcp_to_data(HostName, Port, Timeout) -> tcp_to_data(HostName, Port, Timeout,[]). %%-------------------------------------------------------------------- %% Function: tcp_to_data(Host, Port, Timeout, TCPOptions) -> LazyBinary | Error %% @doc %% A Lazy binary is constructed from the socket that opening a tcp connetion %% to the host in binary mode with packet size 0. If an error or timeout %% occurs during connection an error is returned. %% If the options provided are inconsistent with: %% {packet, 0}, binary, {active, false} %% unexpected and undefined behaviour will be the result. %% @end %%-------------------------------------------------------------------- -spec tcp_to_data(HostName, Port, timeout(), [gen_tcp:connect_option()]) -> data(binary()) | {error, inet:posix()} when HostName:: inet:ip_address() | inet:hostname(), Port ::inet:port_number(). %%-------------------------------------------------------------------- tcp_to_data(HostName, Port, Timeout, OptionsIn) -> Options = [{packet, 0}, binary, {active, false} | OptionsIn], case gen_tcp:connect(HostName, Port, Options, Timeout) of {ok, Socket} -> tcp_socket_to_data(Socket); Error = {error, _} -> Error end. %%-------------------------------------------------------------------- %% Function: tcp_socket_to_data(Socket) -> LazyBinary %% @doc %% A Lazy binary is constructed from the socket, it is expected to be %% connected to the host in binary mode with packet size 0. %% On errors/closure reading from the socket results in the closure %% and eol is returned. Timeout in reading gives an empty binary. %% @end %%-------------------------------------------------------------------- -spec tcp_socket_to_data(inet:socket()) -> data(binary()). %%-------------------------------------------------------------------- tcp_socket_to_data(Socket) -> Promise = fun(eol) -> gen_tcp:close(Socket), eol; (Timeout) -> case gen_tcp:recv(Socket, 0, Timeout) of {ok, Packet} -> Packet; {error, timeout} -> <<>>; {error, closed} -> eol; {error, _} -> gen_tcp:close(Socket), eol end end, create(Promise). %%-------------------------------------------------------------------- %% Function: tcp_reconnect_to_data(Host, Port, Timeout) -> LazyBinary %% @doc %% A Lazy binary is constructed from the socket that opening a tcp connetion %% to the host in binary mode with packet size 0. If an error occurs during %% connection an error is returned. If the connection is closed it is %% reconnected. %% @end %%-------------------------------------------------------------------- -spec tcp_reconnect_to_data(HostName, Port, timeout()) -> data(binary()) when HostName:: inet:ip_address() | inet:hostname(), Port ::inet:port_number(). %%-------------------------------------------------------------------- tcp_reconnect_to_data(HostName, Port, Timeout) -> case tcp_to_data(HostName, Port, Timeout) of Lazy when is_function(Lazy, 1) -> concat(Lazy, fun(eol) -> eol; (CallTimeout) -> Lazy1 = tcp_reconnect_to_data(HostName, Port, Timeout), Lazy1(CallTimeout) end); _ -> empty() end. %%-------------------------------------------------------------------- %% Function: file_to_data(Mode, FileName) -> LazyBinary %% @doc %% A Lazy binary is constructed from the stream, when opening the file %% in binary raw mode with read_ahead. %% On errors reading from the stream results in the closure %% and eol is returned. %% The mode determines if the data is read linewise or in chunks of %% Mode characters. %% @end %%-------------------------------------------------------------------- -spec file_to_data(line | integer(), file:filename()) -> data(binary()) | {error, file:posix() | badarg | system_limit}. %%-------------------------------------------------------------------- file_to_data(Type, Name) -> case file:open(Name, [read, binary, raw, read_ahead]) of {ok, Device} -> file_stream_to_data(Type, Device); Error = {error, _} -> Error end. %%-------------------------------------------------------------------- %% Function: file_stream_to_data(Mode, Stream) -> LazyBinary %% @doc %% A Lazy binary is constructed from the stream, it is expected to be %% opened in binary raw mode with read_ahead. %% On errors reading from the stream results in the closure %% and eol is returned. %% The mode determines if the data is read linewise or in chunks of %% Mode characters. %% @end %%-------------------------------------------------------------------- -spec file_stream_to_data(line | integer(), file:io_device()) -> data(binary()). %%-------------------------------------------------------------------- file_stream_to_data(line, Stream) -> Promise = fun(eol) -> file:close(Stream), eol; (_) -> case file:read_line(Stream) of {ok, Data} -> Data; {error, _} -> file:close(Stream), eol; eof -> file:close(Stream), eol end end, create(Promise); file_stream_to_data(ChunkSize, Stream) -> Promise = fun(eol) -> file:close(Stream), eol; (_) -> case file:read(Stream, ChunkSize) of {ok, Data} -> Data; {error, _} -> file:close(Stream), eol; eof -> file:close(Stream), eol end end, create(Promise). %% =================================================================== %% Internal functions. %% ===================================================================
src/lazy.erl
0.591251
0.612252
lazy.erl
starcoder
%% @doc Module for converting ip formats and results -module(ipmangle). -export([ip_results_to_json/1, verify_address/1]). -include("defs.hrl"). -spec check_blacklist(byte()) -> ok. check_blacklist(A) -> case A of 169 -> erlang:error(blacklist_net); _ -> ok end. %% @doc Verifies input is a valid network address. Converts binary input to %% list output. For example, binary "127.0.0.0" returns "127.0.0.". %% This is consistent with input required for the scanner. %% The caller should catch exceptions. %% @end -spec verify_address(binary()) -> [byte()]. verify_address(Network) -> % All functions require lists NetworkList = binary_to_list(Network), % Verifies a valid address, and verifies it is a /24 {ok, AddrTuple} = inet:parse_ipv4strict_address(NetworkList), {A, _B, _C, 0} = AddrTuple, check_blacklist(A), NetworkSubnet = lists:droplast(NetworkList), NetworkSubnet. %% @doc Converts the ip tuple for a JSON format suited for React -spec ip_results_to_json([{inet:ip4_address(), scan_result()}]) -> jiffy:json_value(). ip_results_to_json(Results) -> ScanConverted = [ { single_convert(Result) } || Result <- Results], jiffy:encode(ScanConverted). -spec single_convert({inet:ip4_address(), scan_result()}) -> list({ binary(), binary() | scan_result() }). single_convert({Address, Stat}) -> [{<<"address">>, list_to_binary(inet:ntoa(Address))}, {<<"stat">>, Stat}]. -ifdef(TEST). -include_lib("eunit/include/eunit.hrl"). verify_address_valid_test() -> ?assertEqual("127.0.0.", verify_address(<<"127.0.0.0">>)). verify_address_stupid_test() -> ?assertError(_, verify_address(<<"127.324324234.43432.12321">>)). verify_address_fullip_test() -> ?assertError(_, verify_address(<<"127.0.0.1">>)). ip_results_to_json_test() -> ?assertEqual(ip_results_to_json([{{127, 0, 0, 253}, not_vulnerable}, {{127, 0, 0, 252}, not_vulnerable}]), <<"[{\"address\":\"127.0.0.253\",\"stat\":\"not_vulnerable\"},{\"address\":\"127.0.0.252\",\"stat\":\"not_vulnerable\"}]">>). -endif.
src/ipmangle.erl
0.521227
0.447823
ipmangle.erl
starcoder
%% ------------------------------------------------------------------- %% %% riak_ql_quanta.erl - a library for quantising time for Time Series %% %% @doc This module serves to generate time quanta on multi - (day, hour, minute, %% second) boundaries. The quantum are based on an origin time of Jan 1, 1970 %% 00:00:00 (Unix Epoch). %% The function <em>quantum/3</em> takes a time in milliseconds to bucketize, %% a size of the quantum, and the units of said quantum. %% For instance, the following call would create buckets for timestamps on 15 %% minute boundaries: <em>quantum(Time, 15, m)</em>. The quantum time is returned in %% milliseconds since the Unix epoch. %% The function <em>quanta/4</em> takes 2 times in milliseconds and size of the quantum %% and the of units of said quantum and returns a list of quantum boundaries that span the time %% %% Copyright (c) 2015-2016 Basho Technologies, Inc. All Rights Reserved. %% %% This file is provided to you under the Apache License, %% Version 2.0 (the "License"); you may not use this file %% except in compliance with the License. You may obtain %% a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, %% software distributed under the License is distributed on an %% "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY %% KIND, either express or implied. See the License for the %% specific language governing permissions and limitations %% under the License. %% %% ------------------------------------------------------------------- -module(riak_ql_quanta). -export([ quantum/3, quanta/4, timestamp_to_ms/1, ms_to_timestamp/1, unit_to_millis/2 ]). -type time_ms() :: non_neg_integer(). %% A timestamp in millisconds representing number of millisconds from Unix epoch -type time_unit() :: d | h | m | s | ms. %% The units of quantization available to quantum/3 -type err() :: {error, term()}. %% @doc The Number of Days from Jan 1, 0 to Jan 1, 1970 %% We need this to compute years and months properly including leap years and variable length %% months. -define(DAYS_FROM_0_TO_1970, 719528). -ifdef(TEST). -ifdef(EQC). -include_lib("eqc/include/eqc.hrl"). -define(QC_OUT(P), eqc:on_output(fun(Str, Args) -> io:format(user, Str, Args) end, P)). -compile(export_all). -endif. -endif. %% @clear %% @end %% @doc given an upper and lower bound for time, returns a tuple consisting of %% * the number of slices %% * a list of all the quantum boundaries %% - the length of the list is the number of slices - 1 -spec quanta(time_ms(), time_ms(), non_neg_integer(), time_unit()) -> {integer(), [integer()]} | {error, any()}. quanta(StartTime, EndTime, QuantaSize, Unit) when StartTime > EndTime -> %% cheap trick to handle descending timestamps, reverse the arguments quanta(EndTime, StartTime, QuantaSize, Unit); quanta(StartTime, EndTime, QuantaSize, Unit) -> Start = quantum(StartTime, QuantaSize, Unit), case Start of {error, _} = E -> E; _Other -> End = EndTime, Diff = End - Start, Slice = unit_to_ms(Unit) * QuantaSize, NSlices = accommodate(Diff, Slice), Quanta = gen_quanta(NSlices, Start, Slice, []), {NSlices, Quanta} end. %% compute ceil(Length / Unit) accommodate(Length, Unit) -> Length div Unit + if Length rem Unit > 0 -> 1; el/=se -> 0 end. gen_quanta(1, _Start, _Slice, Acc) -> Acc; gen_quanta(N, Start, Slice, Acc) when is_integer(N) andalso N > 1 -> NewA = Start + (N - 1) * Slice, gen_quanta(N - 1, Start, Slice, [NewA | Acc]). %% @doc Given the time in milliseconds since the unix epoch and a time range and unit eg (15, m), %% generate the starting timestamp of the range (quantum) in milliseconds since the epoch where the %% time belongs. Note that Time - Quanta is less than or equal to QuantaSize * Unit (in milliseconds). -spec quantum(time_ms(), non_neg_integer(), time_unit()) -> time_ms() | err(). quantum(Time, QuantaSize, Unit) when is_integer(Time) andalso is_integer(QuantaSize) andalso Unit == d; Unit == h; Unit == m; Unit == s; Unit == ms -> Ms = unit_to_ms(Unit), Diff = Time rem (QuantaSize*Ms), Time - Diff; quantum(_, _, Unit) -> {error, {invalid_unit, Unit}}. %% Convert an integer and a time unit in binary to millis, assumed from the unix %% epoch. -spec unit_to_millis(Value::integer(), Unit::binary() | time_unit()) -> integer() | error. unit_to_millis(V, U) when U == ms; U == <<"ms">> -> V; unit_to_millis(V, U) when U == s; U == <<"s">> -> V*1000; unit_to_millis(V, U) when U == m; U == <<"m">> -> V*1000*60; unit_to_millis(V, U) when U == h; U == <<"h">> -> V*1000*60*60; unit_to_millis(V, U) when U == d; U == <<"d">> -> V*1000*60*60*24; unit_to_millis(_, _) -> error. %% @doc Return the time in milliseconds since 00:00 GMT Jan 1, 1970 (Unix Epoch) -spec timestamp_to_ms(erlang:timestamp()) -> time_ms(). timestamp_to_ms({Mega, Secs, Micro}) -> Mega*1000000000 + Secs*1000 + Micro div 1000. %% @doc Return an erlang:timestamp() given the time in milliseconds since the Unix Epoch -spec ms_to_timestamp(time_ms()) -> erlang:timestamp(). ms_to_timestamp(Time) -> Seconds = Time div 1000, MicroSeconds = (Time rem 1000) * 1000, {0, Seconds, MicroSeconds}. -spec unit_to_ms(time_unit()) -> time_ms(). unit_to_ms(ms) -> 1; unit_to_ms(s) -> 1000; unit_to_ms(m) -> 60 * unit_to_ms(s); unit_to_ms(h) -> 60 * unit_to_ms(m); unit_to_ms(d) -> 24 * unit_to_ms(h). -ifdef(TEST). -include_lib("eunit/include/eunit.hrl"). assert_minutes(Quanta, OkTimes) -> Time = timestamp_to_ms(os:timestamp()), QuantaMs = quantum(Time, Quanta, m), {_, {_, M, _}} = calendar:now_to_universal_time(ms_to_timestamp(QuantaMs)), ?assert(lists:member(M, OkTimes)). quantum_minutes_test() -> assert_minutes(15, [0, 15, 30, 45]), assert_minutes(75, [0, 15, 30, 45]), assert_minutes(5, [0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55]), assert_minutes(6, [0, 6, 12, 18, 24, 30, 36, 42, 48, 54]). assert_hours(Quanta, OkTimes) -> Time = timestamp_to_ms(os:timestamp()), QuantaMs = quantum(Time, Quanta, h), {_, {H, _, _}} = calendar:now_to_universal_time(ms_to_timestamp(QuantaMs)), ?assert(lists:member(H, OkTimes)). quantum_hours_test() -> assert_hours(12, [0, 12]), assert_hours(24, [0]). assert_days(Days) -> Now = os:timestamp(), Time = timestamp_to_ms(Now), QuantaMs = quantum(Time, Days, d), {NowDate, _} = calendar:now_to_universal_time(Now), {QuantaDate, _} = calendar:now_to_universal_time(ms_to_timestamp(QuantaMs)), NowDays = calendar:date_to_gregorian_days(NowDate), QuantaDays = calendar:date_to_gregorian_days(QuantaDate), ?assert((NowDays - QuantaDays) < Days), ?assert((NowDays - QuantaDays) >= 0). quantum_days_test() -> assert_days(1), assert_days(10), assert_days(15), assert_days(28), assert_days(29), assert_days(30), assert_days(31). %% %% test Quanta %% -define(MIN, 60*1000). % minute in miliseconds single_quanta_test() -> Start = 1 * ?MIN, End = 5 * ?MIN, {N, Quanta} = quanta(Start, End, 15, m), Length = length(Quanta), ?assertEqual(1, N), ?assertEqual(N - 1, Length), ?assertEqual([], Quanta). two_quanta_test() -> Start = 1 * ?MIN, End = 16 * ?MIN, {N, Quanta} = quanta(Start, End, 15, m), Length = length(Quanta), ?assertEqual(2, N), ?assertEqual(N -1, Length), ?assertEqual([15 * ?MIN], Quanta). split_quanta_test() -> Start = 14 * ?MIN, End = 16 * ?MIN, {N, Quanta} = quanta(Start, End, 15, m), Length = length(Quanta), ?assertEqual(2, N), ?assertEqual(N - 1, Length), ?assertEqual([15 * ?MIN], Quanta). -ifdef(EQC). prop_quantum_bounded_test() -> ?assertEqual( true, eqc:quickcheck( eqc:numtests(1000, prop_quantum_bounded())) ). %% Ensure that Quantas are always bounded, meaning that any time is no more %% than one quantum ahead of the quantum start. prop_quantum_bounded() -> ?FORALL( {Date, Time, {Quanta, Unit}}, {date_gen(), time_gen(), quantum_gen()}, begin DateTime = {Date, Time}, SecondsFrom0To1970 = ?DAYS_FROM_0_TO_1970 * (unit_to_ms(d) div 1000), DateMs = (calendar:datetime_to_gregorian_seconds(DateTime) - SecondsFrom0To1970)*1000, QuantaMs = quantum(DateMs, Quanta, Unit), QuantaSize = quantum_in_ms(Quanta, Unit), (DateMs - QuantaMs) =< QuantaSize end). quantum_now_from_datetime(DateTime, Quanta, Unit) -> SecondsFrom0To1970 = ?DAYS_FROM_0_TO_1970 * (unit_to_ms(d) div 1000), DateMs = (calendar:datetime_to_gregorian_seconds(DateTime) - SecondsFrom0To1970)*1000, QuantaMs = quantum(DateMs, Quanta, Unit), ms_to_timestamp(QuantaMs). quantum_in_ms(Quanta, Unit) -> Quanta*unit_to_ms(Unit). %% EQC Generators date_gen() -> ?SUCHTHAT(Date, {choose(1970, 2015), choose(1, 12), choose(1, 31)}, calendar:valid_date(Date)). time_gen() -> {choose(0, 23), choose(0, 59), choose(0, 59)}. %% We expect quanta to be bigger than their cardinality %% A quantum of 100 minutes is perfectly reasonable quantum_gen() -> oneof([ {choose(1, 1000), d}, {choose(1, 1000), h}, {choose(1, 1000), m}, {choose(1, 1000), s} ]). -endif. -endif.
src/riak_ql_quanta.erl
0.631253
0.726571
riak_ql_quanta.erl
starcoder
%% @doc This is the behaviour definition for a credential provider %% module and it iterates over a list of providers. You may set the %% `credential_providers` Erlang environment variable if you want to %% restrict checking only a certain subset of the default list. %% %% Default order of checking for credentials is: %% <ol> %% <li>Erlang application environment</li> %% <li>OS environment</li> %% <li>Credentials from AWS file</li> %% <li>ECS Task credentials</li> %% <li>EC2 credentials</li> %% </ol> %% %% Providers are expected to implement a function called `fetch/1' which %% takes as its argument a proplist of options which may influence the %% operation of the provider. The fetch/1 function should return either %% `{ok, Credentials, Expiration}' or `{error, Reason}'. %% %% If a provider returns {ok, ...} then evaluation stops at that provider. %% If it returns {error, ...} then the next provider is executed in order %% until either a set of credentials are returned or the tuple %% `{error, no_credentials}' is returned. %% %% If a new provider is desired, the behaviour interface should be %% implemented and its module name added to the default list. %% @end -module(aws_credentials_provider). -export([fetch/0, fetch/1]). -type options() :: #{provider() => map()}. -type expiration() :: binary() | pos_integer() | infinity. -type provider() :: aws_credentials_env | aws_credentials_file | aws_credentials_ecs | aws_credentials_ec2. -export_type([ options/0, expiration/0 ]). -callback fetch(options()) -> {ok, aws_credentials:credentials(), expiration()} | {error, any()}. -include_lib("kernel/include/logger.hrl"). -define(DEFAULT_PROVIDERS, [aws_credentials_env, aws_credentials_file, aws_credentials_ecs, aws_credentials_ec2]). -spec fetch() -> {ok, aws_credentials:credentials(), expiration()} | {'error', 'no_credentials'}. fetch() -> fetch(#{}). -spec fetch(options()) -> {ok, aws_credentials:credentials(), expiration()} | {'error', 'no_credentials'}. fetch(Options) -> Providers = get_env(credential_providers, ?DEFAULT_PROVIDERS), evaluate_providers(Providers, Options). -spec evaluate_providers([provider() | {provider(), options()}], options()) -> {ok, aws_credentials:credentials(), expiration()} | {'error', no_credentials}. evaluate_providers([], _Options) -> {error, no_credentials}; evaluate_providers([ Provider | Providers ], Options) -> case Provider:fetch(Options) of {error, _} = Error -> ?LOG_ERROR("Provider ~p reports ~p", [Provider, Error], #{domain => [aws_credentials]}), evaluate_providers(Providers, Options); {ok, Credentials, Expiration} -> {ok, Credentials, Expiration} end. -spec get_env(atom(), [provider()]) -> any(). get_env(Key, Default) -> case application:get_env(aws_credentials, Key) of undefined -> Default; {ok, Value} -> Value end.
src/aws_credentials_provider.erl
0.540439
0.438665
aws_credentials_provider.erl
starcoder
%% Puzzle: %% %% transparent paper folding %% https://adventofcode.com/2021/day/17 %% %% explanation: %% https://blog.beerriot.com/2021/12/17/advent-of-code-day-17/ -module(puzzle17). -export([ solveA/0, solveB/0, parse_input/1, find_ys/2, find_xs/2, find_solutions/4 ]). %% - highest Y velocity for target below start is abs(lower_end)-1, %% because it will trace same steps up and down, and then makes one %% more step, which needs to hit bottom end %% %% - highest Y reached during that point is (Y * (Y+1))/2 solveA() -> [_MinX, _MaxX, MinY, _MaxY] = load_file(), Velocity = abs(MinY) - 1, (Velocity * (Velocity + 1)) div 2. solveB() -> [MinX, MaxX, MinY, MaxY] = load_file(), length(puzzle17:find_solutions(MinX, MaxX, MinY, MaxY)). load_file() -> {ok, Data} = file:read_file("puzzles/puzzle17-input.txt"), parse_input(Data). -define(RE, "target area: x=([0-9]+)\\.\\.([0-9]+), y=(-[0-9]+)\\.\\.(-[0-9]+)"). parse_input(Input) -> {match, Params} = re:run(Input, ?RE, [{capture, [1,2,3,4], binary}]), [binary_to_integer(V) || V <- Params ]. %% Notes for Part 2: %% %% - Solve X and Y independently: keep a list of which steps for each %% velocity end up in the target space, then pair up the two, by %% matching the steps. Match X to Y, because of the special case %% noted below. %% %% - Solving X: %% - Skip everything that would have a zero velocity before the target %% - Stop searching when the first step of a velocity is past the target %% - Special case: some velocities might stop in the target, so mark %% that they match every step greater %% %% - Solving Y: %% - Start with highest positive velocity that hits target %% - Work downward until the first velocity has a first step below target find_ys(Min, Max) -> Highest = abs(Min)-1, %% {velocity, step} InitPoints = [{Highest, Highest * 2 + 2}], find_ys(Min, Max, Highest-1, InitPoints). find_ys(Min, _Max, Min, Points) -> [{Min, 1}|Points]; find_ys(Min, Max, Vel, Points) -> case Vel > 0 of true -> FirstStepBelow0 = Vel * 2 + 2, InitVelocityBelow0 = -(Vel+1); false -> FirstStepBelow0 = 1, InitVelocityBelow0 = Vel end, find_ys(Min, Max, Vel-1, [{Vel, Step} || Step <- steps_in_target_y(Min, Max, FirstStepBelow0, InitVelocityBelow0, InitVelocityBelow0-1)] ++ Points). steps_in_target_y(Min, _Max, _Step, Y, _Vel) when Y < Min -> []; steps_in_target_y(Min, Max, Step, Y, Vel) -> case Y =< Max of true -> [Step | steps_in_target_y(Min, Max, Step+1, Y+Vel, Vel-1)]; false -> steps_in_target_y(Min, Max, Step+1, Y+Vel, Vel-1) end. find_xs(Min, Max) -> Slowest = find_first_xs(Min, Max, 1), [{StartVel,_}|_] = Slowest, find_xs(Min, Max, StartVel, Slowest). find_first_xs(Min, Max, Vel) -> case (Vel * (Vel + 1)) div 2 of In when Min =< In, In =< Max -> [{Vel, Step} || Step <- steps_in_target_x(Min, Max, 1, Vel, Vel-1)]; In when In > Max -> {error, unexpected_slowest_too_fast}; _ -> find_first_xs(Min, Max, Vel+1) end. steps_in_target_x(_Min, Max, _Step, X, _Vel) when X > Max -> []; steps_in_target_x(Min, Max, Step, X, 0) -> case (Min =< X) and (X =< Max) of true -> [{all_gte, Step}]; false -> [] end; steps_in_target_x(Min, Max, Step, X, Vel) -> case X >= Min of true -> [Step | steps_in_target_x(Min, Max, Step+1, X+Vel, Vel-1)]; false -> steps_in_target_x(Min, Max, Step+1, X+Vel, Vel-1) end. find_xs(_Min, Max, Vel, Points) when Vel > Max -> Points; find_xs(Min, Max, Vel, Points) -> find_xs(Min, Max, Vel+1, [{Vel, Step} || Step <- steps_in_target_x(Min, Max, 1, Vel, Vel-1)] ++ Points). find_solutions(MinX, MaxX, MinY, MaxY) -> Xs = find_xs(MinX, MaxX), Ys = find_ys(MinY, MaxY), lists:usort( lists:foldl(fun({XVel, Step}, Acc) -> [{XVel, YVel} || YVel <- ys_at_step(Step, Ys)] ++ Acc end, [], Xs)). ys_at_step({all_gte, Step}, Ys) -> [ Y || {Y, S} <- Ys, S >= Step ]; ys_at_step(Step, Ys) -> [ Y || {Y, S} <- Ys, S == Step ].
src/puzzle17.erl
0.562537
0.688724
puzzle17.erl
starcoder
%%-------------------------------------------------------------------- %% Copyright (c) 2020-2021 EMQ Technologies Co., Ltd. All Rights Reserved. %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %%-------------------------------------------------------------------- -module(emqx_limiter_schema). -include_lib("typerefl/include/types.hrl"). -export([ roots/0, fields/1, to_rate/1 , to_bucket_rate/1, minimum_period/0]). -define(KILOBYTE, 1024). -type limiter_type() :: bytes_in | message_in | connection | message_routing. -type bucket_name() :: atom(). -type zone_name() :: atom(). -type rate() :: infinity | float(). -type bucket_rate() :: list(infinity | number()). -typerefl_from_string({rate/0, ?MODULE, to_rate}). -typerefl_from_string({bucket_rate/0, ?MODULE, to_bucket_rate}). -reflect_type([ rate/0 , bucket_rate/0 ]). -export_type([limiter_type/0, bucket_name/0, zone_name/0]). -import(emqx_schema, [sc/2, map/2]). roots() -> [emqx_limiter]. fields(emqx_limiter) -> [ {bytes_in, sc(ref(limiter), #{})} , {message_in, sc(ref(limiter), #{})} , {connection, sc(ref(limiter), #{})} , {message_routing, sc(ref(limiter), #{})} ]; fields(limiter) -> [ {global, sc(rate(), #{})} , {zone, sc(map("zone name", rate()), #{})} , {bucket, sc(map("bucket id", ref(bucket)), #{desc => "Token Buckets"})} ]; fields(bucket) -> [ {zone, sc(atom(), #{desc => "the zone which the bucket in"})} , {aggregated, sc(bucket_rate(), #{})} , {per_client, sc(bucket_rate(), #{})} ]. %% minimum period is 100ms minimum_period() -> 100. %%-------------------------------------------------------------------- %% Internal functions %%-------------------------------------------------------------------- ref(Field) -> hoconsc:ref(?MODULE, Field). to_rate(Str) -> Tokens = [string:trim(T) || T <- string:tokens(Str, "/")], case Tokens of ["infinity"] -> {ok, infinity}; [Quota, Interval] -> {ok, Val} = to_quota(Quota), case emqx_schema:to_duration_ms(Interval) of {ok, Ms} when Ms > 0 -> {ok, Val * minimum_period() / Ms}; _ -> {error, Str} end; _ -> {error, Str} end. to_bucket_rate(Str) -> Tokens = [string:trim(T) || T <- string:tokens(Str, "/,")], case Tokens of [Rate, Capa] -> {ok, infinity} = to_quota(Rate), {ok, CapaVal} = to_quota(Capa), if CapaVal =/= infinity -> {ok, [infinity, CapaVal]}; true -> {error, Str} end; [Quota, Interval, Capacity] -> {ok, Val} = to_quota(Quota), case emqx_schema:to_duration_ms(Interval) of {ok, Ms} when Ms > 0 -> {ok, CapaVal} = to_quota(Capacity), {ok, [Val * minimum_period() / Ms, CapaVal]}; _ -> {error, Str} end; _ -> {error, Str} end. to_quota(Str) -> {ok, MP} = re:compile("^\s*(?:(?:([1-9][0-9]*)([a-zA-z]*))|infinity)\s*$"), Result = re:run(Str, MP, [{capture, all_but_first, list}]), case Result of {match, [Quota, Unit]} -> Val = erlang:list_to_integer(Quota), Unit2 = string:to_lower(Unit), {ok, apply_unit(Unit2, Val)}; {match, [Quota]} -> {ok, erlang:list_to_integer(Quota)}; {match, []} -> {ok, infinity}; _ -> {error, Str} end. apply_unit("", Val) -> Val; apply_unit("kb", Val) -> Val * ?KILOBYTE; apply_unit("mb", Val) -> Val * ?KILOBYTE * ?KILOBYTE; apply_unit("gb", Val) -> Val * ?KILOBYTE * ?KILOBYTE * ?KILOBYTE; apply_unit(Unit, _) -> throw("invalid unit:" ++ Unit).
apps/emqx_limiter/src/emqx_limiter_schema.erl
0.656548
0.432003
emqx_limiter_schema.erl
starcoder
%% Copyright (c) 2018 EMQ Technologies Co., Ltd. All Rights Reserved. %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %% @doc This module manages an opaque collection of statistics data used to %% force garbage collection on `self()' process when hitting thresholds. %% Namely: %% (1) Total number of messages passed through %% (2) Total data volume passed through %% @end -module(emqx_gc). -export([init/1, inc/2, reset/0]). -type st() :: #{ cnt => {integer(), integer()} , oct => {integer(), integer()} }. -define(disabled, disabled). -define(ENABLED(X), (is_integer(X) andalso X > 0)). %% @doc Initialize force GC parameters. -spec init(false | map()) -> ok. init(#{count := Count, bytes := Bytes}) -> Cnt = [{cnt, {Count, Count}} || ?ENABLED(Count)], Oct = [{oct, {Bytes, Bytes}} || ?ENABLED(Bytes)], erlang:put(?MODULE, maps:from_list(Cnt ++ Oct)), ok; init(_) -> erlang:put(?MODULE, #{}), ok. %% @doc Increase count and bytes stats in one call, %% ensure gc is triggered at most once, even if both thresholds are hit. -spec inc(pos_integer(), pos_integer()) -> ok. inc(Cnt, Oct) -> mutate_pd_with(fun(St) -> inc(St, Cnt, Oct) end). %% @doc Reset counters to zero. -spec reset() -> ok. reset() -> mutate_pd_with(fun(St) -> reset(St) end). %% ======== Internals ======== %% mutate gc stats numbers in process dict with the given function mutate_pd_with(F) -> St = F(erlang:get(?MODULE)), erlang:put(?MODULE, St), ok. %% Increase count and bytes stats in one call, %% ensure gc is triggered at most once, even if both thresholds are hit. -spec inc(st(), pos_integer(), pos_integer()) -> st(). inc(St0, Cnt, Oct) -> case do_inc(St0, cnt, Cnt) of {true, St} -> St; {false, St1} -> {_, St} = do_inc(St1, oct, Oct), St end. %% Reset counters to zero. reset(St) -> reset(cnt, reset(oct, St)). -spec do_inc(st(), cnt | oct, pos_integer()) -> {boolean(), st()}. do_inc(St, Key, Num) -> case maps:get(Key, St, ?disabled) of ?disabled -> {false, St}; {Init, Remain} when Remain > Num -> {false, maps:put(Key, {Init, Remain - Num}, St)}; _ -> {true, do_gc(St)} end. do_gc(St) -> erlang:garbage_collect(), reset(St). reset(Key, St) -> case maps:get(Key, St, ?disabled) of ?disabled -> St; {Init, _} -> maps:put(Key, {Init, Init}, St) end.
src/emqx_gc.erl
0.71721
0.440289
emqx_gc.erl
starcoder
%%-------------------------------------------------------------------- %% Copyright (c) 2021 EMQ Technologies Co., Ltd. All Rights Reserved. %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %%-------------------------------------------------------------------- %% Tests that use error injection should go here, to avoid polluting %% the logs and scaring people -module(ekka_mnesia_error_injection_SUITE). -compile(export_all). -compile(nowarn_export_all). -include_lib("eunit/include/eunit.hrl"). -include_lib("snabbkaffe/include/snabbkaffe.hrl"). all() -> ekka_ct:all(?MODULE). init_per_suite(Config) -> Config. end_per_suite(_Config) -> ok. t_agent_restart(_) -> Cluster = ekka_ct:cluster([core, core, replicant], ekka_mnesia_test_util:common_env()), CounterKey = counter, ?check_trace( try Nodes = [N1, N2, N3] = ekka_ct:start_cluster(ekka, Cluster), ekka_mnesia_test_util:wait_shards(Nodes), ekka_mnesia_test_util:stabilize(1000), %% Everything in ekka agent will crash ?inject_crash( #{?snk_meta := #{domain := [ekka, rlog, agent|_]}} , snabbkaffe_nemesis:random_crash(0.4) ), ok = rpc:call(N1, ekka_transaction_gen, counter, [CounterKey, 100, 100]), ekka_mnesia_test_util:stabilize(5100), ekka_mnesia_test_util:compare_table_contents(test_tab, Nodes), N3 after ekka_ct:teardown_cluster(Cluster) end, fun(N3, Trace) -> ?assert(ekka_rlog_props:replicant_bootstrap_stages(N3, Trace)), %ekka_rlog_props:counter_import_check(CounterKey, N3, Trace), ?assert(length(?of_kind(snabbkaffe_crash, Trace)) > 1) end). t_rand_error_injection(_) -> Cluster = ekka_ct:cluster([core, core, replicant], ekka_mnesia_test_util:common_env()), CounterKey = counter, ?check_trace( try Nodes = [N1, N2, N3] = ekka_ct:start_cluster(ekka, Cluster), ekka_mnesia_test_util:wait_shards(Nodes), ekka_mnesia_test_util:stabilize(1000), %% Everything in ekka RLOG will crash ?inject_crash( #{?snk_meta := #{domain := [ekka, rlog|_]}} , snabbkaffe_nemesis:random_crash(0.1) ), ok = rpc:call(N1, ekka_transaction_gen, counter, [CounterKey, 300, 100]), ekka_mnesia_test_util:stabilize(5000), ekka_mnesia_test_util:compare_table_contents(test_tab, Nodes), N3 after ekka_ct:teardown_cluster(Cluster) end, fun(N3, Trace) -> ?assert(ekka_rlog_props:replicant_bootstrap_stages(N3, Trace)), ?assert(ekka_rlog_props:counter_import_check(CounterKey, N3, Trace) > 0), ?assert(length(?of_kind(snabbkaffe_crash, Trace)) > 1) end).
test/ekka_mnesia_error_injection_SUITE.erl
0.523664
0.540863
ekka_mnesia_error_injection_SUITE.erl
starcoder
-module(borda). %% API exports -export([ rankings/2 ]). -include("include/common.hrl"). -include("include/elections.hrl"). %%==================================================================== %% Guard Macros %%==================================================================== -define(IS_VALID_LABEL(Label), Label =:= base0 orelse Label =:= base1 orelse Label =:= dowdell orelse Label =:= nauru). %%==================================================================== %% API functions %%==================================================================== % https://en.wikipedia.org/wiki/Borda_count#Example -spec rankings(label(), [ballot(), ...]) -> [{candidate_name(), number()}, ...] | binary(). rankings(Label, Ballots) when ?IS_VALID_LABEL(Label) -> Candidates = ballot:candidates(hd(Ballots)), Map = rankings(Label, Ballots, length(Candidates), #{}), L = maps:to_list(Map), lists:sort(fun by_votes/2, L); rankings(_, _) -> <<"Please restrict the first argument to base0, base1, dowdell, or nauru.">>. %%==================================================================== %% Internal functions %%==================================================================== -spec add_votes(label(), [candidate(), ...], map()) -> map(). add_votes(Label, [C | Cs], Acc) when is_map(Acc) -> add_votes(Label, [C | Cs], Acc, 1). add_votes(_, [], Acc, _) -> Acc; add_votes(Label, [C | Cs], Acc, Position) -> CN = candidate:name(C), Acc2 = add_votes_at_position(CN, Position, Acc), add_votes(Label, Cs, Acc2, Position + 1). -spec add_votes_at_position(candidate_name(), pos_integer(), map()) -> map(). add_votes_at_position(CN, Position, Acc) -> VotesByPosition = maps:get(CN, Acc, #{}), VotesAtPosition = maps:get(Position, VotesByPosition, 0), NewVotes = maps:put(Position, VotesAtPosition + 1, VotesByPosition), maps:put(CN, NewVotes, Acc). -spec borda_values(label(), pos_integer(), map()) -> map(). borda_values(Label, CandCount, Acc) when is_atom(Label) andalso is_map(Acc) -> borda_values(Label, CandCount, maps:keys(Acc), Acc, #{}). borda_values(_Label, _CandCount, [], #{}, Final) -> Final; borda_values(Label, CandCount, [CandName | CNs], From, Acc) -> Votes = maps:get(CandName, From), From2 = maps:without([CandName], From), Acc2 = maps:put(CandName, value_of_votes(Label, CandCount, Votes), Acc), borda_values(Label, CandCount, CNs, From2, Acc2). by_votes({_, V1}, {_, V2}) -> V1 > V2. rankings(Label, [], CandCount, Acc) -> borda_values(Label, CandCount, Acc); rankings(Label, [B | Bs], CandCount, Acc) when is_map(Acc) -> Acc2 = add_votes(Label, ballot:candidates(B), Acc), rankings(Label, Bs, CandCount, Acc2). -spec value(label(), pos_integer(), pos_integer()) -> non_neg_integer(). value(base0, Position, CandCount) -> CandCount - Position; value(base1, Position, CandCount) -> value(base0, Position, CandCount) + 1; % Cf. https://en.wikipedia.org/wiki/Borda_count#Dowdall_system_(Nauru) value(dowdell, Position, CandCount) -> value(nauru, Position, CandCount); value(nauru, 1, _CandCount) -> 1; value(nauru, Position, _CandCount) -> 1.0 / Position. -spec value_of_votes(label(), pos_integer(), map()) -> non_neg_integer(). value_of_votes(Label, CandCount, Votes) when is_map(Votes) -> Values = lists:map(fun(Idx) -> EachValue = value(Label, Idx, CandCount), VoteCount = maps:get(Idx, Votes), EachValue * VoteCount end, maps:keys(Votes)), lists:sum(Values).
src/borda.erl
0.554953
0.520253
borda.erl
starcoder
%% @doc Model for fetching and updating the Git clone of Zotonic. %% The Git checkout is located in "priv/data/zotonic-git". %% %% This file is a model. You will see functions like 'm_get' that %% implement the model behaviour. Models are always located in the %% directory 'models' and their filename always start with 'm_'. %% %% @author <NAME> <<EMAIL>> %% @copyright 2020 <NAME> %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %% Note that the filename starts with m_, followed by the module name. %% This ensures that there are no unexpected name clashes, which could %% have easily occured if the module was named something like 'm_git'. -module(m_zotonicwww2_git). %% Model behaviour, from zotonic_core/src/behaviours/ %% Only the m_get/3 function is required. -behaviour(zotonic_model). -export([ m_get/3, m_post/3, task_rebuild/1, build_edoc/1, build_doc/1, clone/1, hash/1, pull/1, git_dir/1, edoc_dir/1, doc_dir/1 ]). %% Include the central definitions of Zotonic. Useful for macro %% definitions like ?DEBUG. -include_lib("zotonic_core/include/zotonic.hrl"). %% @doc Handle GET requests for this model. Can be called from the %% templates (m.zotonicwww2_git), the API (/api/model/zotonicwww2_git/get/...) %% or via MQTT (topic model/zotonicwww2_git/get). %% %% The first argument is the split path of the request (after 'get'). %% %% The second argument is the MQTT messages, if any. For template calls %% this could be 'undefined'. API calls also construct a MQTT message, as %% the API routes calls via the MQTT tree and the zotonic_model.erl (in %% zotonic_core/src/support). %% %% The m_get function consumes as much of the Path as is needed, it must %% return its result together with the unconsumed part of the path. The %% zotonic_model functions will do a further lookup of the Path remainder %% in the return value. %% %% On an error an error tuple should be returned. For the template routines %% this maps to 'undefined' and is ignored. The API and MQTT will return %% a payload with the error to the caller. -spec m_get( Path :: list(), zotonic_model:opt_msg(), z:context() ) -> zotonic_model:return(). m_get( [ <<"hash">> | Rest ], _Payload, Context) -> % Only for authenticated users, as we are running a command line % program and don't want to do that to anonymous bots and users. case z_auth:is_auth(Context) of true -> % Return the hash of the current checkout case hash(Context) of {ok, Hash} -> % Note the 'Rest', this is the non-consumed part of the % path. It is often used by templates for further lookups % in returned maps or other structured values. {ok, {Hash, Rest}}; {error, _} = Error -> Error end; false -> {error, eacces} end; m_get( [ <<"rebuild">>, <<"hash">> | Rest ], _Payload, Context) -> % The task_rebuild stores the hash of the last rebuild in the % config key site.rebuild_hash. Hash = m_config:get_value(site, rebuild_hash, Context), {ok, {Hash, Rest}}. %% This is a post handler for HTTP posts to `/api/model/zototonicwww2_git/post` %% and MQTT publish to "model/zotonicwww2_git/post" %% Note that a post handler always consumes the whole path. -spec m_post( Path :: list( binary() ), zotonic_model:opt_msg(), z:context() ) -> {ok, term()} | ok | {error, term()}. m_post( [ <<"rebuild">>, Secret ], _Payload, Context) -> % Compare the value in the config tables with the passed secret. % Config values can be set with m_config:set_value/4 or in the % admin on "/admin/config" case m_config:get_value(site, rebuild_secret, Context) of Secret -> % As a build takes a long time we schedule a build task. % The task is slightly delayed so that repetitive pushes % are started after a small period of inactivity. z_pivot_rsc:insert_task_after( 10, % Seconds or a date ?MODULE, % This module task_rebuild, % Prepend task functions with 'task_' <<>>, % Give a key for multiple tasks with same mod:fun [], % Arguments, non needed for this task Context), {ok, <<"queued">>}; _ -> % Log a message to the lager logs lager:info("Docs rebuild request with wrong secret from ~p", [ m_req:get(peer, Context) ]), % Try to use posix error codes {error, eacces} end. %% @doc Documentation rebuild task scheduled by the 'm_post' handler for %% 'docs-rebuild' above. This task is executed by z_pivot_rsc. Only a single %% task is executed in parallel. -spec task_rebuild( z:context() ) -> ok | {error, term()}. task_rebuild(Context) -> z_utils:pipeline([ fun pull/1, fun build_doc/1, fun zotonicwww2_parse_docs:import/1, fun build_edoc/1, fun() -> {ok, Hash} = hash(Context), lager:info("Rebuild of docs success for '~s'", [ Hash ]), m_config:set_value(site, rebuild_hash, Hash, Context), ok end ], [ Context ]). %% @doc Build Zotonic, the html docs and move them to the doc_dir. Returns %% an error or the current hash. This command takes a long time to run. -spec build_doc( z:context() ) -> {ok, binary()} | {error, term()}. build_doc(Context) -> ok = z_filelib:ensure_dir(doc_dir(Context)), run_gitcmds([ "rm -rf doc/_build", "make docs", "rm -rf ../doc/html", "mv doc/_build/html ../doc/." ], Context). %% @doc Clone a fresh checkout of the zotonic repository. Only running if %% the priv/data/zotonic-git directory is not present. -spec clone( z:context() ) -> {ok, binary()} | {error, term()}. clone(Context) -> Dir = git_dir(Context), case filelib:is_file(Dir) of true -> {error, eexist}; false -> DataDir = unicode:characters_to_list( filename:dirname(Dir) ), ok = z_filelib:ensure_dir(Dir), Cmd = "git clone https://github.com/zotonic/zotonic.git zotonic-git", Options = [ sync, stdout, {cd, DataDir} ], lager:info("Command: \"~s\"", [ Cmd ]), case exec:run(Cmd, Options) of {ok, [ {stdout, Output} ]} -> Output1 = filter_output(Output), lager:info("Command output: ~s", [ Output1 ]), {ok, iolist_to_binary(Output)}; {ok, []} -> lager:info("Command output: "), {ok, <<>>}; {error, _} = Error -> lager:error("Command \"~s\" error: ~p", [ Cmd, Error ]), Error end end. %% Remove escape sequences, especially from the rebar3 output which %% adds color escapes. filter_output(Output) -> O1 = iolist_to_binary(Output), filter_output_1(O1, <<>>). filter_output_1(<<>>, Acc) -> Acc; filter_output_1(<<C, R/binary>>, Acc) when C =:= 9; C =:= 10; C =:= 13 -> filter_output_1(R, <<Acc/binary, C>>); filter_output_1(<<C, R/binary>>, Acc) when C < 32 -> filter_output_1(R, Acc); filter_output_1(<<C, R/binary>>, Acc) -> filter_output_1(R, <<Acc/binary, C>>). %% @doc Build Zotonic, the edoc and move them to the edoc_dir. Returns %% an error or the current hash. This command takes a long time to run. -spec build_edoc( z:context() ) -> {ok, binary()} | {error, term()}. build_edoc(Context) -> ok = z_filelib:ensure_dir(edoc_dir(Context)), run_gitcmds([ "rm -rf doc/_build", "make", "make edocs", "rm -rf ../doc/edoc", "mv doc/_build/edoc ../doc/." ], Context). %% @doc Return the current checkout hash of the git repo. -spec hash( z:context() ) -> {ok, binary()} | {error, term()}. hash(Context) -> run_gitcmd("git log -1 --pretty=format:%H", Context). %% @doc Pull changes in the 'priv/data/zotonic-git' directory. Returns the %% new current hash. -spec pull( z:context() ) -> {ok, binary()} | {error, term()}. pull(Context) -> case run_gitcmd("git clean -f", Context) of {ok, _} -> case run_gitcmd("git pull", Context) of {ok, _} -> hash(Context); {error, _} = Error -> Error end; {error, _} = Error -> Error end. %% @doc Run a series of commands in the 'priv/data/zotonic-git' directory. -spec run_gitcmds( list( string() ), z:context() ) -> {ok, binary()} | {error, term()}. run_gitcmds(Cmds, Context) -> lists:foldl( fun (_Cmd, {error, _} = Error) -> Error; (Cmd, {ok, _}) -> run_gitcmd(Cmd, Context) end, {ok, <<>>}, Cmds). %% @doc Run a supervised command in the 'priv/data/zotonic-git' directory. -spec run_gitcmd( string(), z:context() ) -> {ok, binary()} | {error, term()}. run_gitcmd(Cmd, Context) -> Dir = git_dir(Context), Options = [ sync, stdout, {cd, unicode:characters_to_list(Dir)} ], lager:info("Command: \"~s\"", [ Cmd ]), case exec:run(Cmd, Options) of {ok, [ {stdout, Output} ]} -> lager:info("Command output: ~s", [ Output ]), {ok, iolist_to_binary(Output)}; {ok, []} -> lager:info("Command output: "), {ok, <<>>}; {error, _} = Error -> lager:error("Command \"~s\" error: ~p", [ Cmd, Error ]), Error end. %% @doc Return the directory of the git checkout. -spec git_dir( z:context() ) -> file:filename_all(). git_dir(Context) -> filename:join([ base_dir(Context), <<"zotonic-git">> ]). %% @doc Return the directory for the generated edoc documentation (from the erlang sources) -spec edoc_dir( z:context() ) -> file:filename_all(). edoc_dir(Context) -> filename:join([ base_dir(Context), <<"doc">>, <<"edoc">> ]). %% @doc Return the directory for the generated html documentation (from ReStructuredText) -spec doc_dir( z:context() ) -> file:filename_all(). doc_dir(Context) -> filename:join([ base_dir(Context), <<"doc">>, <<"html">> ]). %% @doc Return the base directory for all data. -spec base_dir( z:context() ) -> file:filename_all(). base_dir(Context) -> z_path:files_subdir_ensure(<<"data">>, Context).
src/models/m_zotonicwww2_git.erl
0.574872
0.464841
m_zotonicwww2_git.erl
starcoder
%% @doc: If you wish to implement your own backend for storing %% registers, your module needs to implement these interfaces. The %% backend modules have quite a lot of responsibility (detailed below) %% to allow for backend-specific optimizations. -module(hyper_register). %% @doc: Creates a new instance of the backend. The return value of %% this function will be passed to all functions in this module. -callback new(P :: hyper:precision()) -> hyper:registers(). %% @doc: Set the register to the given value, *only* if the value %% already stored is lower than the new value. The backend needs to %% ensure the register value is only allowed to increase. -callback set(Index :: integer(), Value :: integer(), hyper:registers()) -> hyper:registers(). %% @doc: Compact is always called before any attempt at reading (sum, %% zero count, etc) or merging. It is intended to give backends that %% buffer the writes a chance to flush the buffer before the registers %% are needed. -callback compact(hyper:registers()) -> hyper:registers(). %% @doc: Merge any number of registers, used to calculate the %% union. For two register values at the same index, the max value %% must be in the resulting register. -callback max_merge([hyper:registers()]) -> hyper:registers(). %% @doc: Same as max_merge/1 but used when we know only two filters %% are merged. -callback max_merge(hyper:registers(), hyper:registers()) -> hyper:registers(). %% @doc: Reduce the precision of the registers. Used for mixed-precision %% union by first reducing the precision to the lowest of all filters. -callback reduce_precision(hyper:precision(), hyper:registers()) -> hyper:registers(). %% @doc: Sum of 2^-R where R is the value in each register. -callback register_sum(hyper:registers()) -> float(). %% @doc: Count of registers set to 0. -callback zero_count(hyper:registers()) -> integer(). %% @doc: Encode and decode are called to convert the in-memory %% representation of the backend to the serialized format. Must return %% one binary where each register is encoded as an 8-bit integer. -callback encode_registers(hyper:registers()) -> binary(). -callback decode_registers(binary(), hyper:precision()) -> hyper:registers(). %% @doc: Size in bytes used to represent the registers in memory. -callback bytes(hyper:registers()) -> integer().
src/hyper_register.erl
0.668556
0.622947
hyper_register.erl
starcoder
%%%------------------------------------------------------------------------ %% Copyright 2019, OpenTelemetry Authors %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %% %% @doc The Accumulator receives metric updates and when `collect' is called %% it sweeps through the instruments calling `checkpoint' with the define %% aggregator for each and submitting to the Integrator. %% @end %%%------------------------------------------------------------------------- -module(ot_metric_accumulator). -behaviour(gen_server). -export([start_link/1, active_table/0, record/2, record/3, observe/3, collect/0, lookup_active/2]). -export([init/1, handle_call/3, handle_cast/2]). -include_lib("stdlib/include/ms_transform.hrl"). -include_lib("opentelemetry_api/include/opentelemetry.hrl"). -include("ot_meter.hrl"). -define(ACTIVE_TAB, active_instrument_updates). -define(active_ms(Name, LabelSet), ets:fun2ms(fun(#active_instrument{key=Key, instrument=#instrument{number_kind=InputType}, aggregator=Aggregator}) when Key =:= {Name, LabelSet} -> {InputType, Aggregator} end)). -record(state, {}). start_link(Opts) -> gen_server:start_link({local, ?MODULE}, ?MODULE, Opts, []). active_table() -> ?ACTIVE_TAB. -spec record(ot_meter:bound_instrument(), number()) -> boolean(). record({Key, {InputType, Aggregator}}, Number) -> Aggregator:update(?ACTIVE_TAB, Key, InputType, Number); record(#active_instrument{key=Key, instrument=#instrument{number_kind=InputType}, aggregator=Aggregator}, Number) -> Aggregator:update(?ACTIVE_TAB, Key, InputType, Number). -spec record(ot_meter:name(), ot_meter:label_set(), number()) -> boolean() | unknown_instrument. record(Name, LabelSet, Number) -> case lookup_active(Name, LabelSet) of unknown_instrument -> %% an Instrument must exist to create an Active Instrument unknown_instrument; {InputType, Aggregator} -> Aggregator:update(?ACTIVE_TAB, {Name, LabelSet}, InputType, Number) end. observe(#instrument{name=Name}, Number, LabelSet) -> _ = lookup_active(Name, LabelSet), ot_metric_aggregator_last_value:update(?ACTIVE_TAB, {Name, LabelSet}, observer, Number), ok. -spec lookup_active(instrument() | ot_meter:name(), ot_meter:label_set()) -> {ot_meter:number_kind(), module()} | unknown_instrument. lookup_active(Instrument=#instrument{name=Name}, LabelSet) -> MatchSpec = ?active_ms(Name, LabelSet), case ets:select(?ACTIVE_TAB, MatchSpec) of [{InputType, Aggregator}] -> {InputType, Aggregator}; [] -> add_active_instrument(Instrument, Name, LabelSet) end; lookup_active(Name, LabelSet) -> MatchSpec = ?active_ms(Name, LabelSet), case ets:select(?ACTIVE_TAB, MatchSpec) of [{InputType, Aggregator}] -> {InputType, Aggregator}; [] -> case ot_meter_default:lookup_instrument(Name) of unknown_instrument -> unknown_instrument; Instrument -> add_active_instrument(Instrument, Name, LabelSet) end end. collect()-> gen_server:call(?MODULE, collect). init(_Opts) -> %% This ETS table is required for other parts to not crash so we create %% it in init and not in a handle_continue or whatever else. %% No heir is worried about since active metrics are created dynamicly _ = ets:new(?ACTIVE_TAB, [named_table, public, {keypos, #active_instrument.key}, {write_concurrency, true}, ordered_set]), {ok, #state{}}. handle_call(collect, _From, State) -> %% TODO: should observers just checkpoint in the first place? %% TODO: should have a timeout on observer callbacks run_observers(ets:tab2list(ot_meter_default:observer_tab())), MS = ets:fun2ms(fun(#active_instrument{key=Key, aggregator=Aggregator}) -> {Key, Aggregator} end), run_checkpoints(ets:select(?ACTIVE_TAB, MS, 20)), {reply, ok, State}; handle_call(_Msg, _From, State) -> {noreply, State}. handle_cast(_Msg, State) -> {noreply, State}. %% run_checkpoints('$end_of_table') -> ok; run_checkpoints({Matches, Continuation}) -> [Aggregator:checkpoint(?ACTIVE_TAB, Key) || {Key, Aggregator} <- Matches], ets:select(Continuation). run_observers([]) -> ok; run_observers([Observer | Rest]) -> run_observer(Observer), run_observers(Rest). run_observer(#observer{instrument=ObserverInstrument, callback=Callback}) -> try Callback(ObserverInstrument) catch _:_ -> %% TODO: log an error ok end. aggregator(#instrument{kind=ot_counter}) -> ot_metric_aggregator_sum; aggregator(#instrument{kind=ot_sum_observer}) -> ot_metric_aggregator_last_value; aggregator(#instrument{kind=ot_value_recorder}) -> ot_metric_aggregator_mmsc. add_active_instrument(Instrument=#instrument{number_kind=InputType}, Name, LabelSet) -> Aggregator = aggregator(Instrument), InitialValue = Aggregator:initial_value(InputType), ActiveInstrument = #active_instrument{key={Name, LabelSet}, instrument=Instrument, aggregator=Aggregator, current=InitialValue}, _ = ets:insert_new(?ACTIVE_TAB, ActiveInstrument), {InputType, Aggregator}.
src/ot_metric_accumulator.erl
0.574992
0.418875
ot_metric_accumulator.erl
starcoder
%% %% Copyright (c) dushin.net %% All rights reserved. %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %% -module(rational). -export([add/2, subtract/2, multiply/2, divide/2, normalize/1, simplify/1, reduce/1, to_decimal/2, round/1]). -type numerator() :: integer(). -type denominator() :: integer(). -type fraction() :: {numerator(), denominator()}. -type composite() :: {integer(), fraction()}. -type rational() :: integer() | fraction() | composite(). %% %% @param A rational %% @param B rational %% @return A + B as a (possibly top-heavy) fraction. %% -spec add(A::rational(), B::rational()) -> fraction(). add(A, B) -> add_rational(normalize(A), normalize(B)). %% @private add_rational({N1, D1}, {N2, D2}) -> {N1 * D2 + N2 * D1, D1 * D2}. %% %% @param A rational %% @param B rational %% @return A - B as a (possibly top-heavy) fraction. %% subtract(A, B) -> subtract_rational(normalize(A), normalize(B)). %% @private subtract_rational({N1, D1}, {N2, D2}) -> {N1 * D2 - N2 * D1, D1 * D2}. %% %% @param A rational %% @param B rational %% @return A * B as a (possibly top-heavy) fraction. %% multiply(A, B) -> multiply_rational(normalize(A), normalize(B)). %% @private multiply_rational({N1, D1}, {N2, D2}) -> {N1 * N2, D1 * D2}. %% %% @param A rational %% @param B rational %% @return A / B as a (possibly top-heavy) fraction. %% divide(A, B) -> divide_rational(normalize(A), normalize(B)). %% @private divide_rational({0, _}, _) -> 0; divide_rational(_, {0, _}) -> undefined; divide_rational({N1, D1}, {N2, D2}) -> {N1 * D2, D1 * N2}. %% %% @param R rational %% @return (possibly top-heavy) fraction. %% -spec normalize(R::rational()) -> fraction(). normalize(X) when is_integer(X) -> {X, 1}; normalize({I, {N, D}}) when is_integer(I) andalso is_integer(N) andalso is_integer(D) -> {I * D + N, D}; normalize({N, D} = R) when is_integer(N) andalso is_integer(D) -> R. %% %% @param X (possibly top-heavy) fraction %% @return simplified rational, a fraction, integer or composite. %% -spec simplify(F::fraction()) -> rational(). % simplify({N, 1}) -> % N; % simplify({N, N}) -> % 1; % simplify({0, D}) -> % 0; % simplify({_N, 0}) -> % undefined; simplify({N, D}) when is_integer(D) andalso N > D -> case N rem D of 0 -> {N div D, {0, D}}; R -> {N div D, reduce({R, D})} end; simplify(X) -> X. %% %% @param X (possibly top-heavy) fraction %% @return reduced fraction (possibly divided by gcd). %% -spec reduce(F::fraction()) -> fraction(). reduce({N, D}) when is_integer(N) andalso is_integer(D) -> case gcd(N, D) of 1 -> {N, D}; G -> {N div G, D div G} end. %% @private gcd(A, B) when B > A -> gcd(B, A); gcd(A, B) -> case A rem B of 0 -> B; R -> gcd(B, R) end. %% %% @param F (possibly top-heavy) fraction %% @param P desired precision %% @return equivalent fraction whose denominator is 10^{Precision}. %% @doc Note that no rounding is performed on the last digit, %% and that in general the returned fraction is an estimate. %% -spec to_decimal(F::fraction(), P::non_neg_integer()) -> rational(). to_decimal({N, D}, Precision) when is_integer(N) andalso is_integer(D) -> case simplify({N, D}) of {I, {N1, D}} -> {I, to_decimal({N1, D}, Precision)}; {N1, D} -> Digits = long_division(D, N1, Precision, []), {to_number(Digits), pow(10, Precision)}; undefined -> undefined; N -> {N, {0, 10}} end. %% @private long_division(_D, _N1, 0, Accum) -> Accum; long_division(D, N1, Precision, Accum) -> N2 = N1 * 10, X = N2 div D, long_division(D, N2 - X * D, Precision - 1, [X | Accum]). %% %% @param R rational %% @return Rounded off integer %% -spec round(R::rational()) -> integer(). round({I, {N, D} = F}) when is_integer(I) andalso is_integer(N) andalso is_integer(D) -> I + ?MODULE:round(F); round({N, D}) when is_integer(N) andalso is_integer(D) -> case N < (D bsr 1) of true -> 0; _ -> 1 end; round(I) when is_integer(I) -> I. %% @private to_number(Digits) -> to_number(Digits, 1, 0). %% @private to_number([], _I, Accum) -> Accum; to_number([D|T], I, Accum) -> to_number(T, I*10, D * I + Accum). %% @private pow(_B, 0) -> 1; pow(B, N) -> B * pow(B, N - 1).
src/rational.erl
0.802981
0.57824
rational.erl
starcoder
%%% -*- mode: Erlang; fill-column: 80; comment-column: 75; -*- %%% vi:ts=4 sw=4 et %%% The MIT License %%% %%% Copyright (c) 2007 <NAME> %%% %%% Permission is hereby granted, free of charge, to any person obtaining a copy %%% of this software and associated documentation files (the "Software"), to deal %%% in the Software without restriction, including without limitation the rights %%% to use, copy, modify, merge, publish, distribute, sublicense, and/or sell %%% copies of the Software, and to permit persons to whom the Software is %%% furnished to do so, subject to the following conditions: %%% %%% The above copyright notice and this permission notice shall be included in %%% all copies or substantial portions of the Software. %%% %%% THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR %%% IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, %%% FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE %%% AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER %%% LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, %%% OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN %%% THE SOFTWARE. %%%--------------------------------------------------------------------------- %%% @author <NAME> %%% @copyright 2007 <NAME> freeyourmind ++ [$@|gmail.<EMAIL>] %%% @doc %%% plists is a drop-in replacement for module <a %%% href="http://www.erlang.org/doc/man/lists.html">lists</a>, making %%% most list operations parallel. It can operate on each element in %%% parallel, for IO-bound operations, on sublists in parallel, for %%% taking advantage of multi-core machines with CPU-bound operations, %%% and across erlang nodes, for parallizing inside a cluster. It %%% handles errors and node failures. It can be configured, tuned, and %%% tweaked to get optimal performance while minimizing overhead. %%% %%% Almost all the functions are identical to equivalent functions in %%% lists, returning exactly the same result, and having both a form %%% with an identical syntax that operates on each element in parallel %%% and a form which takes an optional "malt", a specification for how %%% to parallize the operation. %%% %%% fold is the one exception, parallel fold is different from linear %%% fold. This module also include a simple mapreduce implementation, %%% and the function runmany. All the other functions are implemented %%% with runmany, which is as a generalization of parallel list %%% operations. %%% %%% Malts %%% ===== %%% %%% A malt specifies how to break a list into sublists, and can optionally %%% specify a timeout, which nodes to run on, and how many processes to start %%% per node. %%% %%% Malt = MaltComponent | [MaltComponent] %%% MaltComponent = SubListSize::integer() | {processes, integer()} | %%% {processes, schedulers} | %%% {timeout, Milliseconds::integer()} | {nodes, [NodeSpec]}<br/> %%% %%% NodeSpec = Node::atom() | {Node::atom(), NumProcesses::integer()} | %%% {Node::atom(), schedulers} %%% %%% An integer can be given to specify the exact size for sublists. 1 %%% is a good choice for IO-bound operations and when the operation on %%% each list element is expensive. Larger numbers minimize overhead %%% and are faster for cheap operations. %%% %%% If the integer is omitted, and you have specified a `{processes, %%% X}`, the list is split into X sublists. This is only useful when %%% the time to process each element is close to identical and you %%% know exactly how many lines of execution are available to you. %%% %%% If neither of the above applies, the sublist size defaults to 1. %%% %%% You can use `{processes, X}` to have the list processed by `X` %%% processes on the local machine. A good choice for `X` is the %%% number of lines of execution (cores) the machine provides. This %%% can be done automatically with {processes, schedulers}, which sets %%% the number of processes to the number of schedulers in the erlang %%% virtual machine (probably equal to the number of cores). %%% %%% `{timeout, Milliseconds}` specifies a timeout. This is a timeout %%% for the entire operation, both operating on the sublists and %%% combining the results. exit(timeout) is evaluated if the timeout %%% is exceeded. %%% %%% `{nodes, NodeList}` specifies that the operation should be done %%% across nodes. Every element of NodeList is of the form %%% `{NodeName, NumProcesses}` or NodeName, which means the same as %%% `{NodeName, 1}`. plists runs NumProcesses processes on NodeName %%% concurrently. A good choice for NumProcesses is the number of %%% lines of execution (cores) a node provides plus one. This ensures %%% the node is completely busy even when fetching a new sublist. This %%% can be done automatically with `{NodeName, schedulers}`, in which %%% case plists uses a cached value if it has one, and otherwise finds %%% the number of schedulers in the remote node and adds one. This %%% will ensure at least one busy process per core (assuming the node %%% has a scheduler for each core). %%% %%% plists is able to recover if a node goes down. If all nodes go %%% down, exit(allnodescrashed) is evaluated. %%% %%% Any of the above may be used as a malt, or may be combined into a %%% list. `{nodes, NodeList}` and {processes, X} may not be combined. %%% %%% Examples %%% ======== %%% %%% %%start a process for each element (1-element sublists)< %%% 1 %%% %%% %% start a process for each ten elements (10-element sublists) %%% 10 %%% %%% %% split the list into two sublists and process in two processes %%% {processes, 2} %%% %%% %% split the list into X sublists and process in X processes, %%% %% where X is the number of cores in the machine %%% {processes, schedulers} %%% %%% %% split the list into 10-element sublists and process in two processes %%% [10, {processes, 2}] %%% %%% %% timeout after one second. Assumes that a process should be started %%% %% for each element.<br/> %%% {timeout, 1000} %%% %%% %% Runs 3 processes at a time on apple@desktop, and 2 on orange@laptop %%% %% This is the best way to utilize all the CPU-power of a dual-core<br/> %%% %% desktop and a single-core laptop. Assumes that the list should be<br/> %%% %% split into 1-element sublists.<br/> %%% {nodes, [{apple@desktop, 3}, {orange@laptop, 2}]} %%% %%% %% Like above, but makes plists figure out how many processes to use. %%% {nodes, [{apple@desktop, schedulers}, {orange@laptop, schedulers}]} %%% %%% %% Gives apple and orange three seconds to process the list as<br/> %%% %% 100-element sublists.<br/> %%% [100, {timeout, 3000}, {nodes, [{apple@desktop, 3}, {orange@laptop, 2}]}] %%% %%% Aside: Why Malt? %%% ================ %%% %%% I needed a word for this concept, so maybe my subconsciousness %%% gave me one by making me misspell multiply. Maybe it is an acronym %%% for Malt is A List Tearing Specification. Maybe it is a beer %%% metaphor, suggesting that code only runs in parallel if bribed %%% with spirits. It's jargon, learn it or you can't be part of the %%% in-group. %%% %%% Messages and Errors %%% =================== %%% %%% plists assures that no extraneous messages are left in or will %%% later enter the message queue. This is guaranteed even in the %%% event of an error. %%% %%% Errors in spawned processes are caught and propagated to the %%% calling process. If you invoke %%% %%% plists:map(fun (X) -> 1/X end, [1, 2, 3, 0]). %%% %%% you get a badarith error, exactly like when you use lists:map. %%% %%% plists uses monitors to watch the processes it spawns. It is not a %%% good idea to invoke plists when you are already monitoring %%% processes. If one of them does a non-normal exit, plists receives %%% the 'DOWN' message believing it to be from one of its own %%% processes. The error propagation system goes into effect, which %%% results in the error occuring in the calling process. %%% -module(ec_plists). -export([all/2, all/3, any/2, any/3, filter/2, filter/3, fold/3, fold/4, fold/5, foreach/2, foreach/3, map/2, map/3, ftmap/2, ftmap/3, partition/2, partition/3, sort/1, sort/2, sort/3, usort/1, usort/2, usort/3, mapreduce/2, mapreduce/3, mapreduce/5, runmany/3, runmany/4]). -export_type([malt/0, malt_component/0, node_spec/0, fuse/0, fuse_fun/0]). %%============================================================================ %% types %%============================================================================ -type malt() :: malt_component() | [malt_component()]. -type malt_component() :: SubListSize :: integer() | {processes, integer()} | {processes, schedulers} | {timeout, Milliseconds :: integer()} | {nodes, [node_spec()]}. -type node_spec() :: Node :: atom() | {Node :: atom(), NumProcesses :: integer()} | {Node :: atom(), schedulers}. -type fuse_fun() :: fun((term(), term()) -> term()). -type fuse() :: fuse_fun() | {recursive, fuse_fun()} | {reverse, fuse_fun()}. -type el_fun() :: fun((term()) -> term()). %%============================================================================ %% API %%============================================================================ %% Everything here is defined in terms of runmany. %% The following methods are convient interfaces to runmany. %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec all(el_fun(), list()) -> boolean(). all(Fun, List) -> all(Fun, List, 1). %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec all(el_fun(), list(), malt()) -> boolean(). all(Fun, List, Malt) -> try runmany(fun(L) -> B = lists:all(Fun, L), if B -> nil; true -> erlang:throw(notall) end end, fun(_A1, _A2) -> nil end, List, Malt), true catch throw:notall -> false end. %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec any(fun(), list()) -> boolean(). any(Fun, List) -> any(Fun, List, 1). %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec any(fun(), list(), malt()) -> boolean(). any(Fun, List, Malt) -> try runmany(fun(L) -> B = lists:any(Fun, L), if B -> erlang:throw(any); true -> nil end end, fun(_A1, _A2) -> nil end, List, Malt) of _ -> false catch throw:any -> true end. %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec filter(fun(), list()) -> list(). filter(Fun, List) -> filter(Fun, List, 1). %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec filter(fun(), list(), malt()) -> list(). filter(Fun, List, Malt) -> runmany(fun(L) -> lists:filter(Fun, L) end, {reverse, fun(A1, A2) -> A1 ++ A2 end}, List, Malt). %% Note that with parallel fold there is not foldl and foldr, %% instead just one fold that can fuse Accumlators. %% @doc Like below, but assumes 1 as the Malt. This function is almost useless, %% and is intended only to aid converting code from using lists to plists. -spec fold(fun(), InitAcc :: term(), list()) -> term(). fold(Fun, InitAcc, List) -> fold(Fun, Fun, InitAcc, List, 1). %% @doc Like below, but uses the Fun as the Fuse by default. -spec fold(fun(), InitAcc :: term(), list(), malt()) -> term(). fold(Fun, InitAcc, List, Malt) -> fold(Fun, Fun, InitAcc, List, Malt). %% @doc fold is more complex when made parallel. There is no foldl and %% foldr, accumulators aren't passed in any defined order. The list %% is split into sublists which are folded together. Fun is identical %% to the function passed to lists:fold[lr], it takes (an element, and %% the accumulator) and returns -> a new accumulator. It is used for %% the initial stage of folding sublists. Fuse fuses together the %% results, it takes (Results1, Result2) and returns -> a new result. %% By default sublists are fused left to right, each result of a fuse %% being fed into the first element of the next fuse. The result of %% the last fuse is the result. %% %% Fusing may also run in parallel using a recursive algorithm, %% by specifying the fuse as {recursive, Fuse}. See %% the discussion in {@link runmany/4}. %% %% Malt is the malt for the initial folding of sublists, and for the %% possible recursive fuse. -spec fold(fun(), fuse(), InitAcc :: term(), list(), malt()) -> term(). fold(Fun, Fuse, InitAcc, List, Malt) -> Fun2 = fun(L) -> lists:foldl(Fun, InitAcc, L) end, runmany(Fun2, Fuse, List, Malt). %% @doc Similiar to foreach in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a> %% except it makes no guarantee about the order it processes list elements. -spec foreach(fun(), list()) -> ok. foreach(Fun, List) -> foreach(Fun, List, 1). %% @doc Similiar to foreach in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a> %% except it makes no guarantee about the order it processes list elements. -spec foreach(fun(), list(), malt()) -> ok. foreach(Fun, List, Malt) -> runmany(fun(L) -> lists:foreach(Fun, L) end, fun(_A1, _A2) -> ok end, List, Malt). %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec map(fun(), list()) -> list(). map(Fun, List) -> map(Fun, List, 1). %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec map(fun(), list(), malt()) -> list(). map(Fun, List, Malt) -> runmany(fun(L) -> lists:map(Fun, L) end, {reverse, fun(A1, A2) -> A1 ++ A2 end}, List, Malt). %% @doc values are returned as {value, term()}. -spec ftmap(fun(), list()) -> list(). ftmap(Fun, List) -> map(fun(L) -> try {value, Fun(L)} catch Class:Type -> {error, {Class, Type}} end end, List). %% @doc values are returned as {value, term()}. -spec ftmap(fun(), list(), malt()) -> list(). ftmap(Fun, List, Malt) -> map(fun(L) -> try {value, Fun(L)} catch Class:Type -> {error, {Class, Type}} end end, List, Malt). %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec partition(fun(), list()) -> {list(), list()}. partition(Fun, List) -> partition(Fun, List, 1). %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec partition(fun(), list(), malt()) -> {list(), list()}. partition(Fun, List, Malt) -> runmany(fun(L) -> lists:partition(Fun, L) end, {reverse, fun({True1, False1}, {True2, False2}) -> {True1 ++ True2, False1 ++ False2} end}, List, Malt). %% SORTMALT needs to be tuned -define(SORTMALT, 100). %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec sort(list()) -> list(). sort(List) -> sort(fun(A, B) -> A =< B end, List). %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec sort(fun(), list()) -> list(). sort(Fun, List) -> sort(Fun, List, ?SORTMALT). %% @doc This version lets you specify your own malt for sort. %% %% sort splits the list into sublists and sorts them, and it merges the %% sorted lists together. These are done in parallel. Each sublist is %% sorted in a seperate process, and each merging of results is done in a %% seperate process. Malt defaults to 100, causing the list to be split into %% 100-element sublists. -spec sort(fun(), list(), malt()) -> list(). sort(Fun, List, Malt) -> Fun2 = fun(L) -> lists:sort(Fun, L) end, Fuse = fun(A1, A2) -> lists:merge(Fun, A1, A2) end, runmany(Fun2, {recursive, Fuse}, List, Malt). %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec usort(list()) -> list(). usort(List) -> usort(fun(A, B) -> A =< B end, List). %% @doc Same semantics as in module %% <a href="http://www.erlang.org/doc/man/lists.html">lists</a>. -spec usort(fun(), list()) -> list(). usort(Fun, List) -> usort(Fun, List, ?SORTMALT). %% @doc This version lets you specify your own malt for usort. %% %% usort splits the list into sublists and sorts them, and it merges the %% sorted lists together. These are done in parallel. Each sublist is %% sorted in a seperate process, and each merging of results is done in a %% seperate process. Malt defaults to 100, causing the list to be split into %% 100-element sublists. %% %% usort removes duplicate elments while it sorts. -spec usort(fun(), list(), malt()) -> list(). usort(Fun, List, Malt) -> Fun2 = fun(L) -> lists:usort(Fun, L) end, Fuse = fun(A1, A2) -> lists:umerge(Fun, A1, A2) end, runmany(Fun2, {recursive, Fuse}, List, Malt). %% @doc Like below, assumes default MapMalt of 1. -ifdef(namespaced_types). -spec mapreduce(MapFunc, list()) -> dict:dict() when MapFunc :: fun((term()) -> DeepListOfKeyValuePairs), DeepListOfKeyValuePairs :: [DeepListOfKeyValuePairs] | {Key :: term(), Value :: term()}. -else. -spec mapreduce(MapFunc, list()) -> dict() when MapFunc :: fun((term()) -> DeepListOfKeyValuePairs), DeepListOfKeyValuePairs :: [DeepListOfKeyValuePairs] | {Key :: term(), Value :: term()}. -endif. mapreduce(MapFunc, List) -> mapreduce(MapFunc, List, 1). %% Like below, but uses a default reducer that collects all %% {Key, Value} pairs into a %% <a href="http://www.erlang.org/doc/man/dict.html">dict</a>, %% with values {Key, [Value1, Value2...]}. %% This dict is returned as the result. mapreduce(MapFunc, List, MapMalt) -> mapreduce(MapFunc, List, dict:new(), fun add_key/3, MapMalt). %% @doc This is a very basic mapreduce. You won't write a %% Google-rivaling search engine with it. It has no equivalent in %% lists. Each element in the list is run through the MapFunc, which %% produces either a {Key, Value} pair, or a lists of key value pairs, %% or a list of lists of key value pairs...etc. A reducer process runs %% in parallel with the mapping processes, collecting the key value %% pairs. It starts with a state given by InitState, and for each %% {Key, Value} pair that it receives it invokes ReduceFunc(OldState, %% Key, Value) to compute its new state. mapreduce returns the %% reducer's final state. %% %% MapMalt is the malt for the mapping operation, with a default value of 1, %% meaning each element of the list is mapped by a seperate process. %% %% mapreduce requires OTP R11B, or it may leave monitoring messages in the %% message queue. -ifdef(namespaced_types). -spec mapreduce(MapFunc, list(), InitState :: term(), ReduceFunc, malt()) -> dict:dict() when MapFunc :: fun((term()) -> DeepListOfKeyValuePairs), DeepListOfKeyValuePairs :: [DeepListOfKeyValuePairs] | {Key :: term(), Value :: term()}, ReduceFunc :: fun((OldState :: term(), Key :: term(), Value :: term()) -> NewState :: term()). -else. -spec mapreduce(MapFunc, list(), InitState :: term(), ReduceFunc, malt()) -> dict() when MapFunc :: fun((term()) -> DeepListOfKeyValuePairs), DeepListOfKeyValuePairs :: [DeepListOfKeyValuePairs] | {Key :: term(), Value :: term()}, ReduceFunc :: fun((OldState :: term(), Key :: term(), Value :: term()) -> NewState :: term()). -endif. mapreduce(MapFunc, List, InitState, ReduceFunc, MapMalt) -> Parent = self(), {Reducer, ReducerRef} = erlang:spawn_monitor(fun() -> reducer(Parent, 0, InitState, ReduceFunc) end), MapFunc2 = fun(L) -> Reducer ! lists:map(MapFunc, L), 1 end, SentMessages = try runmany(MapFunc2, fun(A, B) -> A + B end, List, MapMalt) catch exit:Reason -> erlang:demonitor(ReducerRef, [flush]), Reducer ! die, exit(Reason) end, Reducer ! {mappers, done, SentMessages}, Results = receive {Reducer, Results2} -> Results2; {'DOWN', _, _, Reducer, Reason2} -> exit(Reason2) end, receive {'DOWN', _, _, Reducer, normal} -> nil end, Results. reducer(Parent, NumReceived, State, Func) -> receive die -> nil; {mappers, done, NumReceived} -> Parent ! {self(), State}; Keys -> reducer(Parent, NumReceived + 1, each_key(State, Func, Keys), Func) end. each_key(State, Func, {Key, Value}) -> Func(State, Key, Value); each_key(State, Func, [List | Keys]) -> each_key(each_key(State, Func, List), Func, Keys); each_key(State, _, []) -> State. add_key(Dict, Key, Value) -> case dict:is_key(Key, Dict) of true -> dict:append(Key, Value, Dict); false -> dict:store(Key, [Value], Dict) end. %% @doc Like below, but assumes a Malt of 1, %% meaning each element of the list is processed by a seperate process. -spec runmany(fun(), fuse(), list()) -> term(). runmany(Fun, Fuse, List) -> runmany(Fun, Fuse, List, 1). %% Begin internal stuff (though runmany/4 is exported). %% @doc All of the other functions are implemented with runmany. runmany %% takes a List, splits it into sublists, and starts processes to operate on %% each sublist, all done according to Malt. Each process passes its sublist %% into Fun and sends the result back. %% %% The results are then fused together to get the final result. There are two %% ways this can operate, lineraly and recursively. If Fuse is a function, %% a fuse is done linearly left-to-right on the sublists, the results %% of processing the first and second sublists being passed to Fuse, then %% the result of the first fuse and processing the third sublits, and so on. If %% Fuse is {reverse, FuseFunc}, then a fuse is done right-to-left, the results %% of processing the second-to-last and last sublists being passed to FuseFunc, %% then the results of processing the third-to-last sublist and %% the results of the first fuse, and and so forth. %% Both methods preserve the original order of the lists elements. %% %% To do a recursive fuse, pass Fuse as {recursive, FuseFunc}. %% The recursive fuse makes no guarantee about the order the results of %% sublists, or the results of fuses are passed to FuseFunc. It %% continues fusing pairs of results until it is down to one. %% %% Recursive fuse is down in parallel with processing the sublists, and a %% process is spawned to fuse each pair of results. It is a parallized %% algorithm. Linear fuse is done after all results of processing sublists %% have been collected, and can only run in a single process. %% %% Even if you pass {recursive, FuseFunc}, a recursive fuse is only done if %% the malt contains {nodes, NodeList} or {processes, X}. If this is not the %% case, a linear fuse is done. -spec runmany(fun(([term()]) -> term()), fuse(), list(), malt()) -> term(). runmany(Fun, Fuse, List, Malt) when erlang:is_list(Malt) -> runmany(Fun, Fuse, List, local, no_split, Malt); runmany(Fun, Fuse, List, Malt) -> runmany(Fun, Fuse, List, [Malt]). runmany(Fun, Fuse, List, Nodes, no_split, [MaltTerm | Malt]) when erlang:is_integer(MaltTerm) -> runmany(Fun, Fuse, List, Nodes, MaltTerm, Malt); runmany(Fun, Fuse, List, local, Split, [{processes, schedulers} | Malt]) -> %% run a process for each scheduler S = erlang:system_info(schedulers), runmany(Fun, Fuse, List, local, Split, [{processes, S} | Malt]); runmany(Fun, Fuse, List, local, no_split, [{processes, X} | _] = Malt) -> %% Split the list into X sublists, where X is the number of processes L = erlang:length(List), case (L rem X) of 0 -> runmany(Fun, Fuse, List, local, (L / X), Malt); _ -> runmany(Fun, Fuse, List, local, (L / X) + 1, Malt) end; runmany(Fun, Fuse, List, local, Split, [{processes, X} | Malt]) -> %% run X process on local machine Nodes = lists:duplicate(X, node()), runmany(Fun, Fuse, List, Nodes, Split, Malt); runmany(Fun, Fuse, List, Nodes, Split, [{timeout, X} | Malt]) -> Parent = erlang:self(), Timer = proc_lib:spawn(fun() -> receive stoptimer -> Parent ! {timerstopped, erlang:self()} after X -> Parent ! {timerrang, erlang:self()}, receive stoptimer -> Parent ! {timerstopped, erlang:self()} end end end), Ans = try runmany(Fun, Fuse, List, Nodes, Split, Malt) catch %% we really just want the after block, the syntax %% makes this catch necessary. willneverhappen -> nil after Timer ! stoptimer, cleanup_timer(Timer) end, Ans; runmany(Fun, Fuse, List, local, Split, [{nodes, NodeList} | Malt]) -> Nodes = lists:foldl(fun({Node, schedulers}, A) -> X = schedulers_on_node(Node) + 1, lists:reverse(lists:duplicate(X, Node), A); ({Node, X}, A) -> lists:reverse(lists:duplicate(X, Node), A); (Node, A) -> [Node | A] end, [], NodeList), runmany(Fun, Fuse, List, Nodes, Split, Malt); runmany(Fun, {recursive, Fuse}, List, local, Split, []) -> %% local recursive fuse, for when we weren't invoked with {processes, X} %% or {nodes, NodeList}. Degenerates recursive fuse into linear fuse. runmany(Fun, Fuse, List, local, Split, []); runmany(Fun, Fuse, List, Nodes, no_split, []) -> %% by default, operate on each element seperately runmany(Fun, Fuse, List, Nodes, 1, []); runmany(Fun, Fuse, List, local, Split, []) -> List2 = splitmany(List, Split), local_runmany(Fun, Fuse, List2); runmany(Fun, Fuse, List, Nodes, Split, []) -> List2 = splitmany(List, Split), cluster_runmany(Fun, Fuse, List2, Nodes). cleanup_timer(Timer) -> receive {timerrang, Timer} -> cleanup_timer(Timer); {timerstopped, Timer} -> nil end. schedulers_on_node(Node) -> case erlang:get(ec_plists_schedulers_on_nodes) of undefined -> X = determine_schedulers(Node), erlang:put(ec_plists_schedulers_on_nodes, dict:store(Node, X, dict:new())), X; Dict -> case dict:is_key(Node, Dict) of true -> dict:fetch(Node, Dict); false -> X = determine_schedulers(Node), erlang:put(ec_plists_schedulers_on_nodes, dict:store(Node, X, Dict)), X end end. determine_schedulers(Node) -> Parent = erlang:self(), Child = proc_lib:spawn(Node, fun() -> Parent ! {self(), erlang:system_info(schedulers)} end), erlang:monitor(process, Child), receive {Child, X} -> receive {'DOWN', _, _, Child, _Reason} -> nil end, X; {'DOWN', _, _, Child, Reason} when Reason =/= normal -> 0 end. %% @doc local runmany, for when we weren't invoked with {processes, X} %% or {nodes, NodeList}. Every sublist is processed in parallel. local_runmany(Fun, Fuse, List) -> Parent = self(), Pids = lists:map(fun(L) -> F = fun() -> Parent ! {self(), Fun(L)} end, {Pid, _} = erlang:spawn_monitor(F), Pid end, List), Answers = try lists:map(fun receivefrom/1, Pids) catch throw:Message -> {BadPid, Reason} = Message, handle_error(BadPid, Reason, Pids) end, lists:foreach(fun(Pid) -> normal_cleanup(Pid) end, Pids), fuse(Fuse, Answers). receivefrom(Pid) -> receive {Pid, R} -> R; {'DOWN', _, _, Pid, Reason} when Reason =/= normal -> erlang:throw({Pid, Reason}); {timerrang, _} -> erlang:throw({nil, timeout}) end. %% Convert List into [{Number, Sublist}] cluster_runmany(Fun, Fuse, List, Nodes) -> {List2, _} = lists:foldl(fun(X, {L, Count}) -> {[{Count, X} | L], Count + 1} end, {[], 0}, List), cluster_runmany(Fun, Fuse, List2, Nodes, [], []). %% @doc Add a pair of results into the TaskList as a fusing task cluster_runmany(Fun, {recursive, Fuse}, [], Nodes, Running, [{_, R1}, {_, R2} | Results]) -> cluster_runmany(Fun, {recursive, Fuse}, [{fuse, R1, R2}], Nodes, Running, Results); cluster_runmany(_, {recursive, _Fuse}, [], _Nodes, [], [{_, Result}]) -> %% recursive fuse done, return result Result; cluster_runmany(_, {recursive, _Fuse}, [], _Nodes, [], []) -> %% edge case where we are asked to do nothing []; cluster_runmany(_, Fuse, [], _Nodes, [], Results) -> %% We're done, now we just have to [linear] fuse the results fuse(Fuse, lists:map(fun({_, R}) -> R end, lists:sort(fun({A, _}, {B, _}) -> A =< B end, lists:reverse(Results)))); cluster_runmany(Fun, Fuse, [Task | TaskList], [N | Nodes], Running, Results) -> %% We have a ready node and a sublist or fuse to be processed, so we start %% a new process Parent = erlang:self(), case Task of {Num, L2} -> Fun2 = fun() -> Parent ! {erlang:self(), Num, Fun(L2)} end; {fuse, R1, R2} -> {recursive, FuseFunc} = Fuse, Fun2 = fun() -> Parent ! {erlang:self(), fuse, FuseFunc(R1, R2)} end end, Fun3 = fun() -> runmany_wrap(Fun2, Parent) end, Pid = proc_lib:spawn(N, Fun3), erlang:monitor(process, Pid), cluster_runmany(Fun, Fuse, TaskList, Nodes, [{Pid, N, Task} | Running], Results); cluster_runmany(Fun, Fuse, TaskList, Nodes, Running, Results) when length(Running) > 0 -> %% We can't start a new process, but can watch over already running ones receive {_Pid, error, Reason} -> RunningPids = lists:map(fun({Pid, _, _}) -> Pid end, Running), handle_error(junkvalue, Reason, RunningPids); {Pid, Num, Result} -> %% throw out the exit message, Reason should be %% normal, noproc, or noconnection receive {'DOWN', _, _, Pid, _Reason} -> nil end, {Running2, FinishedNode, _} = delete_running(Pid, Running, []), cluster_runmany(Fun, Fuse, TaskList, [FinishedNode | Nodes], Running2, [{Num, Result} | Results]); {timerrang, _} -> RunningPids = lists:map(fun({Pid, _, _}) -> Pid end, Running), handle_error(nil, timeout, RunningPids); %% node failure {'DOWN', _, _, Pid, noconnection} -> {Running2, _DeadNode, Task} = delete_running(Pid, Running, []), cluster_runmany(Fun, Fuse, [Task | TaskList], Nodes, Running2, Results); %% could a noproc exit message come before the message from %% the process? we are assuming it can't. %% this clause is unlikely to get invoked due to cluster_runmany's %% spawned processes. It will still catch errors in mapreduce's %% reduce process, however. {'DOWN', _, _, BadPid, Reason} when Reason =/= normal -> RunningPids = lists:map(fun({Pid, _, _}) -> Pid end, Running), handle_error(BadPid, Reason, RunningPids) end; cluster_runmany(_, _, [_Non | _Empty], [] = _Nodes, [] = _Running, _) -> %% We have data, but no nodes either available or occupied erlang:exit(allnodescrashed). -ifdef(fun_stacktrace). runmany_wrap(Fun, Parent) -> try Fun catch exit:siblingdied -> ok; exit:Reason -> Parent ! {erlang:self(), error, Reason}; error:R -> Parent ! {erlang:self(), error, {R, erlang:get_stacktrace()}}; throw:R -> Parent ! {erlang:self(), error, {{nocatch, R}, erlang:get_stacktrace()}} end. -else. runmany_wrap(Fun, Parent) -> try Fun catch exit:siblingdied -> ok; exit:Reason -> Parent ! {erlang:self(), error, Reason}; error:R:Stacktrace -> Parent ! {erlang:self(), error, {R, Stacktrace}}; throw:R:Stacktrace -> Parent ! {erlang:self(), error, {{nocatch, R}, Stacktrace}} end. -endif. delete_running(Pid, [{Pid, Node, List} | Running], Acc) -> {Running ++ Acc, Node, List}; delete_running(Pid, [R | Running], Acc) -> delete_running(Pid, Running, [R | Acc]). handle_error(BadPid, Reason, Pids) -> lists:foreach(fun(Pid) -> erlang:exit(Pid, siblingdied) end, Pids), lists:foreach(fun(Pid) -> error_cleanup(Pid, BadPid) end, Pids), erlang:exit(Reason). error_cleanup(BadPid, BadPid) -> ok; error_cleanup(Pid, BadPid) -> receive {Pid, _} -> error_cleanup(Pid, BadPid); {Pid, _, _} -> error_cleanup(Pid, BadPid); {'DOWN', _, _, Pid, _Reason} -> ok end. normal_cleanup(Pid) -> receive {'DOWN', _, _, Pid, _Reason} -> ok end. %% edge case fuse(_, []) -> []; fuse({reverse, _} = Fuse, Results) -> [RL | ResultsR] = lists:reverse(Results), fuse(Fuse, ResultsR, RL); fuse(Fuse, [R1 | Results]) -> fuse(Fuse, Results, R1). fuse({reverse, FuseFunc} = Fuse, [R2 | Results], R1) -> fuse(Fuse, Results, FuseFunc(R2, R1)); fuse(Fuse, [R2 | Results], R1) -> fuse(Fuse, Results, Fuse(R1, R2)); fuse(_, [], R) -> R. %% @doc Splits a list into a list of sublists, each of size Size, %% except for the last element which is less if the original list %% could not be evenly divided into Size-sized lists. splitmany(List, Size) -> splitmany(List, [], Size). splitmany([], Acc, _) -> lists:reverse(Acc); splitmany(List, Acc, Size) -> {Top, NList} = split(Size, List), splitmany(NList, [Top | Acc], Size). %% @doc Like lists:split, except it splits a list smaller than its first %% parameter split(Size, List) -> split(Size, List, []). split(0, List, Acc) -> {lists:reverse(Acc), List}; split(Size, [H | List], Acc) -> split(Size - 1, List, [H | Acc]); split(_, [], Acc) -> {lists:reverse(Acc), []}.
src/ec_plists.erl
0.503662
0.471892
ec_plists.erl
starcoder
%% @author <NAME> %% @copyright Copyright (c) 2008-2009 <NAME>, Copyright (c) 2009 <NAME> %% %% @doc Conversion functions for all kinds of data types. Changes to %% Rusty's version: added date conversion, undefined handling and more %% to_bool cases. %% Copyright 2009-2012 <NAME> %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. -module(z_convert). -author("<NAME>"). -author("<NAME> <<EMAIL>>"). -author("<NAME> <<EMAIL>>"). -define(ST_JUTTEMIS, {{9999,8,17}, {12,0,0}}). -export ([ clean_lower/1, to_list/1, to_flatlist/1, to_atom/1, to_binary/1, to_binary/2, to_integer/1, to_float/1, to_bool_strict/1, to_bool/1, to_utc/1, to_localtime/1, to_datetime/1, to_date/1, to_time/1, to_isotime/1, to_json/1, unicode_to_utf8/1, convert_json/1, ip_to_list/1, ip_to_long/1, long_to_ip/1 ]). %%% CONVERSION %%% %% @doc Convert to lower case, strip surrounding whitespace. -spec clean_lower(binary()|list()|atom()) -> binary(). clean_lower(L) -> z_string:trim(z_string:to_lower(L)). %% @doc Convert (almost) any value to a list. -spec to_list(term()) -> string(). to_list(undefined) -> []; to_list(<<>>) -> []; to_list({rsc_list, L}) -> L; to_list(L) when is_list(L) -> L; to_list(A) when is_atom(A) -> atom_to_list(A); to_list(B) when is_binary(B) -> binary_to_list(B); to_list(I) when is_integer(I) -> integer_to_list(I); to_list(F) when is_float(F) -> mochinum:digits(F). %% @doc Flatten list and convert to string. -spec to_flatlist(term()) -> string(). to_flatlist(L) when is_list(L) -> case z_string:is_string(L) of true -> L; false -> lists:flatten(to_list(iolist_to_binary(L))) end; to_flatlist(L) -> lists:flatten(to_list(L)). %% @doc Convert (almost) any value to an atom. -spec to_atom(term()) -> atom() | undefined. to_atom(<<>>) -> undefined; to_atom([]) -> undefined; to_atom(A) when is_atom(A) -> A; to_atom(B) when is_binary(B) -> to_atom(binary_to_list(B)); to_atom(I) when is_integer(I) -> to_atom(integer_to_list(I)); to_atom(L) when is_list(L) -> list_to_atom(binary_to_list(iolist_to_binary(L))). %% @doc Convert (almost) any value to an atom. -spec to_binary(term()) -> binary(). to_binary(undefined) -> <<>>; to_binary(A) when is_atom(A) -> to_binary(atom_to_list(A)); to_binary(B) when is_binary(B) -> B; to_binary(I) when is_integer(I) -> to_binary(integer_to_list(I)); to_binary(F) when is_float(F) -> to_binary(to_list(F)); to_binary(L) when is_list(L) -> iolist_to_binary(L). %% Specific Zotonic callback, please keep here. to_binary({trans, _} = Tr, Context) -> to_binary(z_trans:lookup_fallback(Tr, Context)); to_binary(A, _Context) -> to_binary(A). %% @doc Convert (almost) any value to an integer. -spec to_integer(term()) -> integer() | undefined. to_integer(undefined) -> undefined; to_integer([]) -> undefined; to_integer(A) when is_atom(A) -> to_integer(atom_to_list(A)); to_integer(B) when is_binary(B) -> to_integer(binary_to_list(B)); to_integer(I) when is_integer(I) -> I; to_integer(F) when is_float(F) -> erlang:round(F); to_integer([C]) when is_integer(C) andalso (C > $9 orelse C < $0) -> C; to_integer(L) when is_list(L) -> list_to_integer(L). %% @doc Convert (almost) any value to a float. -spec to_float(term()) -> float() | undefined. to_float(undefined) -> undefined; to_float([]) -> undefined; to_float(A) when is_atom(A) -> to_float(atom_to_list(A)); to_float(B) when is_binary(B) -> to_float(binary_to_list(B)); to_float(I) when is_integer(I) -> I + 0.0; to_float(F) when is_float(F) -> F; to_float(L) when is_list(L) -> case lists:member($., L) of true -> list_to_float(L); false -> list_to_float(L++".0") %% list_to_float("1") gives a badarg end. %% @doc Quite loose conversion of values to boolean -spec to_bool(term()) -> true | false. to_bool("false") -> false; to_bool("FALSE") -> false; to_bool("n") -> false; to_bool("N") -> false; to_bool("no") -> false; to_bool("NO") -> false; to_bool(<<"false">>) -> false; to_bool(<<"FALSE">>) -> false; to_bool(<<"n">>) -> false; to_bool(<<"N">>) -> false; to_bool(<<"no">>) -> false; to_bool(<<"NO">>) -> false; to_bool("disabled") -> false; to_bool(<<"disabled">>) -> false; to_bool("DISABLED") -> false; to_bool(<<"DISABLED">>) -> false; to_bool([0]) -> false; to_bool(V) -> to_bool_strict(V). % @doc Convert values to boolean values according to the Django rules -spec to_bool_strict(term()) -> true | false. to_bool_strict(undefined) -> false; to_bool_strict(false) -> false; to_bool_strict(0) -> false; to_bool_strict(0.0) -> false; to_bool_strict(<<>>) -> false; to_bool_strict(<<0>>) -> false; to_bool_strict([]) -> false; to_bool_strict({rsc_list, []}) -> false; to_bool_strict({trans, []}) -> false; to_bool_strict("0") -> false; to_bool_strict(<<"0">>) -> false; to_bool_strict({{9999,M,D},{H,I,S}}) % ?ST_JUTTEMIS when is_integer(M), is_integer(D), is_integer(H), is_integer(I), is_integer(S), M >= 1, M =< 12, D >= 1, D =< 31, H >= 0, H =< 23, I >= 0, I =< 59, S >= 0, S =< 60 -> false; to_bool_strict(_) -> true. %% @doc Convert a local date time to utc to_utc(undefined) -> undefined; to_utc({{9999,_,_}, _}) -> ?ST_JUTTEMIS; to_utc(D) -> case catch calendar:local_time_to_universal_time_dst(D) of [] -> D; % This time never existed in the local time, just take it as-is [UTC] -> UTC; [DstUTC, _UTC] -> DstUTC; {'EXIT', _} -> D end. %% @doc Convert a utc date time to local to_localtime(undefined) -> undefined; to_localtime({{9999,_,_},_}) -> ?ST_JUTTEMIS; to_localtime(D) -> case catch calendar:universal_time_to_local_time(D) of {'EXIT', _} -> D; LocalD -> LocalD end. %% @doc Convert an input to a (universal) datetime, using to_date/1 and %% to_time/1. If the input is a string, it is expected to be in iso %% 8601 format, although it can also handle timestamps without time %% zones. The time component of the datatime is optional. to_datetime({{_,_,_},{_,_,_}} = DT) -> DT; to_datetime({_,_,_} = D) -> {D, {0,0,0}}; to_datetime(B) when is_binary(B) -> to_datetime(binary_to_list(B)); to_datetime(L) when is_list(L) -> try case string:tokens(L, " T") of [Date,Time] -> WithTZ = fun(Tm, Tz, Mul) -> TZTime = to_time(Tz), Add = calendar:datetime_to_gregorian_seconds({{0,1,1},TZTime}), Secs = calendar:datetime_to_gregorian_seconds({to_date(Date), to_time(Tm)}), calendar:gregorian_seconds_to_datetime(Secs+(Mul*Add)) end, case string:tokens(Time, "+") of [Time1, TZ] -> %% Timestamp with positive time zone WithTZ(Time1, TZ, -1); _ -> case string:tokens(Time, "-") of [Time1, TZ] -> %% Timestamp with negative time zone WithTZ(Time1, TZ, 1); _ -> case lists:reverse(Time) of [$Z|Rest] -> %% Timestamp ending on Z (= UTC) {to_date(Date), to_time(lists:reverse(Rest))}; _ -> %% Timestamp without time zone {to_date(Date), to_time(Time)} end end end; [Date] -> {to_date(Date), {0,0,0}} end catch _:_ -> undefined end; to_datetime(undefined) -> undefined. %% @doc Convert an input to a date. Input is expected to be YYYY-MM-DD %% or YYYY/MM/DD. to_date({_,_,_} = D) -> D; to_date(B) when is_binary(B) -> to_date(binary_to_list(B)); to_date([]) -> undefined; to_date(L) when is_list(L) -> case string:tokens(L, "-/") of [D,M,Y] when length(Y) =:= 4 -> {to_integer(Y),to_integer(M),to_integer(D)}; [Y,M,D] -> {to_integer(Y),to_integer(M),to_integer(D)}; _ -> undefined end. %% @doc Convert an input to a time. INput is expected to be HH:MM:SS %% or HH.MM.SS. to_time({_,_,_} = D) -> D; to_time(B) when is_binary(B) -> to_time(binary_to_list(B)); to_time([]) -> undefined; to_time([H1,H2,M1,M2]) -> to_time([H1,H2,$:,M1,M2]); to_time(L) when is_list(L) -> [H,I,S|_] = lists:flatten([[to_integer(X) ||X <- string:tokens(L, ":.")], 0, 0]), {H,I,S}. %% @doc Convert a datetime (in universal time) to an ISO time string. -spec to_isotime(calendar:datetime()) -> string(). to_isotime(DateTime) -> to_list(z_dateformat:format(DateTime, "x-m-d\\TH:i:s\\Z", en)). %% %% @doc Convert an Erlang structure to a format that can be serialized by mochijson. %% %% Simple values to_json(undefined) -> null; to_json(X) when is_atom(X) -> X; to_json(X) when is_integer(X) -> X; to_json(X) when is_float(X) -> X; to_json(X) when is_binary(X) -> X; %% Tuple values to_json({{Y,M,D},{H,I,S}} = DateTime) when is_integer(Y), is_integer(M), is_integer(D), is_integer(H), is_integer(I), is_integer(S) -> z_dateformat:format(DateTime, "c", [{utc,DateTime}]); to_json({array, X}) -> %% Explicit request for array (to prevent string conversion for some lists) {array, [to_json(V) || V <- X]}; to_json({struct, X}) -> {struct, X}; to_json({X, Y}) -> {struct, to_json_struct([{X, Y}])}; to_json(X) when is_tuple(X) -> {array, [to_json(V) || V <- tuple_to_list(X)]}; %% List values to_json([{X, Y}]) when is_atom(X) -> {struct, to_json_struct([{X, Y}])}; to_json([{X, Y} | Z]) when is_atom(X) -> {struct, to_json_struct([{X, Y} | Z])}; to_json(X) when is_list(X) -> case z_string:is_string(X) of true -> X; false -> {array, [to_json(V) || V <- X]} end. %% Handle structs specially to_json_struct([]) -> []; to_json_struct([{X,Y}|T]) -> [{to_json_struct_key(X), to_json(Y)} | to_json_struct(T)]. to_json_struct_key(X) when is_atom(X) orelse is_integer(X) orelse is_binary(X) -> X; to_json_struct_key(X) when is_list(X) -> case z_string:is_string(X) of true -> X; false -> invalid_key end; to_json_struct_key(_) -> invalid_key. ip_to_list({IP,Port}) when is_tuple(IP), is_integer(Port) -> ip_to_list(IP); ip_to_list({N1,N2,N3,N4} ) -> lists:flatten([integer_to_list(N1), $., integer_to_list(N2), $., integer_to_list(N3), $., integer_to_list(N4)]); ip_to_list({_K1,_K2,_K3,_K4,_K5,_K6,_K7,_K8} = IPv6) -> L = lists:map(fun(0) -> ""; (N) -> io_lib:format("~.16b", [N]) end, tuple_to_list(IPv6)), lists:flatten(string:join(L, ":")). %% Taken from egeoip (http://code.google.com/p/egeoip/source/browse/trunk/egeoip/src/egeoip.erl?r=19) %% @spec ip_to_long(Address) -> {ok, integer()} | {error, badmatch} %% @doc Convert an IP address from a string, IPv4 tuple or IPv6 tuple to the %% big endian integer representation. ip_to_long({B3, B2, B1, B0}) -> {ok, (B3 bsl 24) bor (B2 bsl 16) bor (B1 bsl 8) bor B0}; ip_to_long({W7, W6, W5, W4, W3, W2, W1, W0}) -> {ok, (W7 bsl 112) bor (W6 bsl 96) bor (W5 bsl 80) bor (W4 bsl 64) bor (W3 bsl 48) bor (W2 bsl 32) bor (W1 bsl 16) bor W0}; ip_to_long(_) -> {error, badmatch}. %% @doc Convert long int to IP address tuple. FIXME: ipv6 long_to_ip(L) -> {ok, {(L band (255 bsl 24)) bsr 24, (L band (255 bsl 16)) bsr 16, (L band (255 bsl 8)) bsr 8, L band 255}}. %% @doc Convert json from facebook favour to an easy to use format for zotonic templates. convert_json({K, V}) when is_binary(K) -> {z_convert:to_atom(K), convert_json(V)}; convert_json({struct, PropList}) when is_list(PropList) -> convert_json(PropList); convert_json(L) when is_list(L) -> [convert_json(V) || V <- L]; convert_json(V) -> V. unicode_to_utf8(List) when is_list(List) -> lists:flatmap(fun unicode_to_utf8/1, List); unicode_to_utf8(Ch) -> char_to_utf8(Ch). char_to_utf8(Ch) when is_integer(Ch), Ch >= 0 -> if Ch < 128 -> %% 0yyyyyyy [Ch]; Ch < 16#800 -> %% 110xxxxy 10yyyyyy [16#C0 + (Ch bsr 6), 128+(Ch band 16#3F)]; Ch < 16#10000 -> %% 1110xxxx 10xyyyyy 10yyyyyy if Ch < 16#D800; Ch > 16#DFFF, Ch < 16#FFFE -> [16#E0 + (Ch bsr 12), 128+((Ch bsr 6) band 16#3F), 128+(Ch band 16#3F)]; true -> [$?] end; Ch < 16#200000 -> %% 11110xxx 10xxyyyy 10yyyyyy 10yyyyyy [16#F0+(Ch bsr 18), 128+((Ch bsr 12) band 16#3F), 128+((Ch bsr 6) band 16#3F), 128+(Ch band 16#3F)]; Ch < 16#4000000 -> %% 111110xx 10xxxyyy 10yyyyyy 10yyyyyy 10yyyyyy [16#F8+(Ch bsr 24), 128+((Ch bsr 18) band 16#3F), 128+((Ch bsr 12) band 16#3F), 128+((Ch bsr 6) band 16#3F), 128+(Ch band 16#3F)]; Ch < 16#80000000 -> %% 1111110x 10xxxxyy 10yyyyyy 10yyyyyy 10yyyyyy 10yyyyyy [16#FC+(Ch bsr 30), 128+((Ch bsr 24) band 16#3F), 128+((Ch bsr 18) band 16#3F), 128+((Ch bsr 12) band 16#3F), 128+((Ch bsr 6) band 16#3F), 128+(Ch band 16#3F)]; true -> [$?] end.
src/z_convert.erl
0.513425
0.468669
z_convert.erl
starcoder
%% Copyright 2016-2017 TensorHub, Inc. %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. -module(port_io). -behavior(gen_server). -export([start_link/1]). -export([init/1, handle_info/2, handle_cast/2, handle_call/3, terminate/2, code_change/3]). -record(state, {port, buf, buflen, readers}). %% =================================================================== %% Start / init %% =================================================================== start_link(Exe) -> gen_server:start_link(?MODULE, [Exe], []). init([Exe]) -> Port = erlang:open_port({spawn_executable, Exe}, [binary, exit_status]), {ok, #state{port=Port, buf=[], buflen=0, readers=[]}}. handle_info({io_request, From, Ref, Req}, State) -> handle_io_request(Req, From, Ref, State); handle_info({Port, {data, Data}}, #state{port=Port}=State) -> handle_data(Data, State); handle_info({Port, {exit_status, 0}}, #state{port=Port}=State) -> {stop, normal, State}; handle_info({Port, {exit_status, N}}, #state{port=Port}=State) -> {stop, {exit_status, N}, State}. %% =================================================================== %% IO request %% =================================================================== handle_io_request({put_chars, latin1, Chars}, From, Ref, State) -> send_to_port(Chars, State), send_io_reply(From, Ref, ok), {noreply, State}; handle_io_request({get_chars, _Prompt, Len}, From, Ref, State) -> Next = apply_readers(add_reader(Len, From, Ref, State)), {noreply, Next}. send_to_port(Data, #state{port=Port}) -> erlang:port_command(Port, Data). send_io_reply(From, Ref, Reply) -> erlang:send(From, {io_reply, Ref, Reply}). add_reader(Len, From, Ref, #state{readers=Readers}=S) -> Reader = {Len, From, Ref}, S#state{readers=Readers++[Reader]}. apply_readers(State) -> case pop_available_reader(State) of {Reader, NextState} -> apply_readers(apply_reader(Reader, NextState)); false -> State end. pop_available_reader(#state{readers=[]}) -> false; pop_available_reader(#state{readers=[Reader|Rest]}=State) -> case data_available(Reader, State) of true -> {Reader, State#state{readers=Rest}}; false -> false end. data_available({Len, _, _}, #state{buflen=Available}) -> Len =< Available. apply_reader({Len, From, Ref}, #state{buf=Buf}=State) -> {Chars, NextBuf} = get_chars(Len, Buf), send_io_reply(From, Ref, {ok, Chars}), set_buffer(NextBuf, State). get_chars(Len, [B1|RestBuf]) when Len =< byte_size(B1) -> {Chars, RestChars} = split_binary(B1, Len), {Chars, get_chars_finalize_buf(RestChars, RestBuf)}; get_chars(Len, [B1, B2|RestBuf]) -> get_chars(Len, [<<B1/binary, B2/binary>>|RestBuf]). get_chars_finalize_buf(<<>>, Buf) -> Buf; get_chars_finalize_buf(Chars, Buf) -> [Chars|Buf]. set_buffer(Buf, S) -> S#state{buf=Buf, buflen=iolist_size(Buf)}. %% =================================================================== %% Handle data %% =================================================================== handle_data(Data, State) -> Next = apply_readers(buffer(Data, State)), {noreply, Next}. buffer(Data, #state{buf=Buf, buflen=Len}=S) -> S#state{buf=Buf++[Data], buflen=Len+byte_size(Data)}. %% =================================================================== %% gen_server boilerplate %% =================================================================== handle_cast(_Msg, State) -> {noreply, State}. handle_call(_Msg, _From, State) -> {noreply, State}. terminate(_Reason, _State) -> ok. code_change(_OldVsn, State, _Extra) -> {ok, State}.
src/port_io.erl
0.532911
0.449997
port_io.erl
starcoder
% @doc % <a href="https://reference.digilentinc.com/reference/pmod/pmodgyro/reference-manual"> % PmodGYRO % </a> % module that gets the gyroscopes data via SPI. % % Start the driver with % ``` % 1> grisp:add_device(spi1, pmod_gyro). % ''' % @end -module(pmod_gyro). -behaviour(gen_server). -include("grisp.hrl"). -include("pmod_gyro.hrl"). % API -export([start_link/2]). -export([read/0]). % Callbacks -export([init/1]). -export([handle_call/3]). -export([handle_cast/2]). -export([handle_info/2]). -export([code_change/3]). -export([terminate/2]). -define(SPI_MODE, #{clock => {low, leading}}). %--- API ----------------------------------------------------------------------- % @private start_link(Slot, Opts) -> gen_server:start_link(?MODULE, [Slot, Opts], []). % @doc Read the gyroscopes X, Y and Z values in degrees per second. % % === Example === % ``` % 2> pmod_gyro:read(). % {249.28279313922965,-26.078862235243843,12.764756149667337} % ''' -spec read() -> {X::float(), Y::float(), Z::float()}. read() -> Dev = grisp_devices:default(?MODULE), case gen_server:call(Dev#device.pid, read) of {error, Reason} -> error(Reason); Result -> Result end. %--- Callbacks ----------------------------------------------------------------- % @private init([Slot, Opts]) -> Bus = grisp_spi:open(Slot), verify_device(Bus), Res = maps:get(resolution, Opts, 250), ResOpt = case Res of 250 -> 2#00000000; 500 -> 2#00010000; 2000 -> 2#00100000; _ -> error({invalid_option, Res}) end, % Set the resolution <<>> = write(Bus, ?CTRL_REG4, ResOpt), % Enable the device and axis sensors <<>> = write(Bus, ?CTRL_REG1, 2#00001111), grisp_devices:register(Slot, ?MODULE), {ok, #{bus => Bus, unit_degree => (32766 / Res)}}. % @private handle_call(read, _From, #{bus := Bus, unit_degree := UnitDeg} = State) -> <<X:16/signed-little, Y:16/signed-little, Z:16/signed-little>> = read(Bus, ?OUT_X_L, 6), {reply, {X / UnitDeg, Y / UnitDeg, Z / UnitDeg}, State}; handle_call(Request, From, _State) -> error({unknown_request, Request, From}). % @private handle_cast(Request, _State) -> error({unknown_cast, Request}). % @private handle_info(Info, _State) -> error({unknown_info, Info}). % @private code_change(_OldVsn, State, _Extra) -> {ok, State}. % @private terminate(_Reason, _State) -> ok. %--- Internal ------------------------------------------------------------------ verify_device(Bus) -> case read(Bus, ?WHO_AM_I, 1) of <<?DEVID>> -> ok; Other -> error({device_mismatch, {who_am_i, Other}}) end. read(Bus, Reg, Pad) -> transfer(Bus, {?SPI_MODE, <<?RW_READ:1, ?MS_INCR:1, Reg:6>>, 1, Pad}). write(Bus, Reg, Value) -> transfer(Bus, {?SPI_MODE, <<?RW_WRITE:1, ?MS_SAME:1, Reg:6, Value:8>>, 2, 0}). transfer(Bus, Message) -> [Response] = grisp_spi:transfer(Bus, [Message]), Response.
src/pmod_gyro.erl
0.530723
0.705747
pmod_gyro.erl
starcoder
%%This file is licensed under the terms of the Modified BSD License. %%This is a counter to keep track of simulated time. The timed %%semantics work without it but having now() in ABS is useful. Global %%time is also necessary to calculate the next deployment component %%update barrier (distance_to_next_boundary/0). -module(clock). -behaviour(gen_server). -export([start_link/2,stop/0,advance/1,advance_limit/1,is_at_limit/0,now/0,time_since_model_start/0,next_boundary/0,distance_to_next_boundary/0]). -export([code_change/3,handle_call/3,handle_cast/2,handle_info/2,init/1,terminate/2]). -record(state,{now, limit, start_real_time}). %% Interface start_link(Clocklimit, StartTime) -> gen_server:start_link({global, clock}, ?MODULE, [Clocklimit, StartTime], []). stop() -> gen_server:stop({global, clock}). advance(Amount) -> gen_server:call({global, clock}, {advance, Amount}, infinity). advance_limit(Amount) -> %% Advancing the limit here does not cause waiting processes to %% start again; use cog_monitor:increase_clock_limit instead. gen_server:call({global, clock}, {advance_limit, Amount}, infinity). is_at_limit() -> gen_server:call({global, clock}, is_at_limit, infinity). now() -> gen_server:call({global, clock}, now, infinity). time_since_model_start() -> gen_server:call({global, clock}, time_since_model_start, infinity). distance_to_next_boundary() -> %% Returns relative time until the next resource refresh timepoint. gen_server:call({global, clock}, next_int, infinity). next_boundary() -> %% Returns absolute time of the next resource refresh timepoint. gen_server:call({global, clock}, next_boundary, infinity). %% gen_server functions init([Clocklimit, StartTime]) -> {ok, #state{now=rationals:to_r(0), limit=Clocklimit, start_real_time=StartTime}}. handle_call({advance, Amount},_From,State=#state{now=Time,limit=Limit}) -> Newtime = rationals:add(Time, Amount), Reply=case Limit of none -> {reply, {ok, Newtime}, State#state{now=Newtime}}; _ -> case rationals:is_greater(Limit, Newtime) % Limit > Newtime of true -> {reply, {ok, Newtime}, State#state{now=Newtime}}; false -> {reply, {limit_reached, Limit}, State#state{now=Limit}} end end, Reply; handle_call({advance_limit, Amount},_From,State=#state{limit=Limit}) -> case is_integer(Amount) and (Amount > 0) of true -> case Limit of none -> {reply, {ok, none}, State}; _ -> Newlimit=rationals:add(Limit, Amount), {reply, {ok, Newlimit}, State#state{limit=Newlimit}} end; false -> {reply, {error, <<"Need positive integer increment">>}, State} end; handle_call(is_at_limit,_From,State=#state{now=Now,limit=Limit}) -> {reply, cmp:eq(Now, Limit), State}; handle_call(now, _From, State=#state{now=Time}) -> {reply, Time, State}; handle_call(time_since_model_start, _From, State=#state{start_real_time=StartTime}) -> {reply, erlang:system_time(millisecond) - StartTime, State}; handle_call(next_int, _From, State=#state{now=Time}) -> Distance = rationals:sub(Time, rationals:trunc(Time)), case rationals:is_zero(Distance) of true -> {reply, {1,1}, State}; false -> {reply, Distance, State} end; handle_call(next_boundary, _From, State=#state{now=Time}) -> %% Refill boundary is the next integer -- truncate current time and add 1 %% since we don’t have a round-up function, and in case we are at an %% integer we want the next one anyway. {reply,rationals:add(rationals:trunc(Time), {1, 1}) , State}. handle_cast(_Msg,State) -> %% unused {noreply, State}. handle_info(_Info,State) -> %% unused {noreply, State}. terminate(_Reason,_State) -> ok. code_change(_OldVsn,State,_Extra) -> %% not supported {error, State}.
frontend/src/main/resources/erlang/absmodel/src/clock.erl
0.556641
0.412589
clock.erl
starcoder
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Copyright 2018 Pentland Edge Ltd. %% %% Licensed under the Apache License, Version 2.0 (the "License"); you may not %% use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, WITHOUT %% WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the %% License for the specific language governing permissions and limitations %% under the License. %% -module(coords). -export([ lla_to_ecef/1, signed_lon/1, ecef_distance/2, enu_distance/2, deg_to_rad/1, rad_to_deg/1, ecef_to_enu/2, calc_angle/3, dot_product/2, vec_mag/1, dec_to_dms/1, dms_to_dec/1]). %% WGS84 constants. -define(WGS84_A, 6378137). -define(WGS84_B, 6356752.31424518). %% Function to convert from Lat, Lon, Alt to ECEF format lla_to_ecef({Lat,Lon,Alt}) -> % Convert the Lat, Lon into radians. LatRad = deg_to_rad(Lat), LonRad = deg_to_rad(Lon), % Need to check that this height is from the correct reference. H = Alt, Asquared = ?WGS84_A * ?WGS84_A, Bsquared = ?WGS84_B * ?WGS84_B, Esquared = (Asquared - Bsquared) / Asquared, SinLat = math:sin(LatRad), CosLat = math:cos(LatRad), SinLon = math:sin(LonRad), CosLon = math:cos(LonRad), % Calculate the radius of curvature N = ?WGS84_A / math:sqrt(1 - Esquared * SinLat * SinLat), % Calculate the coordinate points. X = (N + H) * CosLat * CosLon, Y = (N + H) * CosLat * SinLon, Z = ((Bsquared/Asquared) * N + H) * SinLat, {X, Y, Z}. %% Convert Longitude to signed degrees format signed_lon(Lon) -> case Lon > 180.0 of true -> Lon - 360.0; false -> Lon end. %% @doc Calculate the magnitude of the difference between two points %% specified in ECEF format. ecef_distance({X1, Y1, Z1}, {X2, Y2, Z2}) -> T1 = math:pow(X2 - X1, 2), T2 = math:pow(Y2 - Y1, 2), T3 = math:pow(Z2 - Z1, 2), math:sqrt(T1 + T2 + T3). %% @doc Calculate the magnitude of the difference between two points %% specified in ENU format. enu_distance(Pt1, Pt2) -> % Can reuse ECEF distance since both are vector magnitude calculations. ecef_distance(Pt1, Pt2). deg_to_rad(Deg) -> Deg * math:pi() / 180. rad_to_deg(Rad) -> Rad * 180 / math:pi(). %% Convert ECEF coordinates to ENU (East, North, Up) local plane. %% Based on the formulae at: %% http://wiki.gis.com/wiki/index.php/Geodetic_system#From_WGS-84_to_ENU:_sample_code ecef_to_enu({RefLat,RefLon,RefH}, {X, Y, Z}) -> % Find reference location in ECEF. {Xr, Yr, Zr} = lla_to_ecef({RefLat,RefLon,RefH}), % Convert the Lat, Lon into radians. LatRad = deg_to_rad(RefLat), LonRad = deg_to_rad(RefLon), SinRefLat = math:sin(LatRad), SinRefLon = math:sin(LonRad), CosRefLat = math:cos(LatRad), CosRefLon = math:cos(LonRad), Xdiff = X - Xr, Ydiff = Y - Yr, Zdiff = Z - Zr, E = -SinRefLon*Xdiff + CosRefLon*Ydiff, N = -SinRefLat*CosRefLon*Xdiff - SinRefLat*SinRefLon*Ydiff + CosRefLat*Zdiff, U = CosRefLat*CosRefLon*Xdiff + CosRefLat*SinRefLon*Ydiff + SinRefLat*Zdiff, {E, N, U}. %% Calculate the angle between two points from a common origin. Return value %% is in radians. Arguments must be supplied in ECEF format. calc_angle({X1,Y1,Z1} = _Orig, {X2,Y2,Z2} = _A, {X3,Y3,Z3} = _B) -> % Create vectors OA and OB. OA = {X2-X1, Y2-Y1, Z2-Z1}, OB = {X3-X1, Y3-Y1, Z3-Z1}, % Calculate the dot product of the two vectors. DotProd = dot_product(OA, OB), % Calculate vector magnitudes. MagOA = vec_mag(OA), MagOB = vec_mag(OB), % Compute the angle. math:acos(DotProd / (MagOA * MagOB)). %% Calculate the dot product of two vectors. dot_product({X1, Y1, Z1}, {X2, Y2, Z2}) -> X1*X2 + Y1*Y2 + Z1*Z2. %% Calculate the magnitude of a vector. vec_mag({X, Y, Z}) -> math:sqrt(X*X + Y*Y + Z*Z). %% Convert a decimal coordinate to {degrees,minutes,seconds}. dec_to_dms(Dec) -> Deg = trunc(Dec), MinF = abs(Dec - Deg), Min = trunc(60.0 * MinF), SecF = MinF - (Min / 60.0), Sec = 3600.0 * SecF, {Deg,Min,Sec}. dms_to_dec({H,M,S}) -> Mag = abs(H) + (M / 60.0) + (S / 3600.0), case H >= 0 of true -> Mag; false -> -Mag end.
src/coords.erl
0.770896
0.553083
coords.erl
starcoder
-module(tql_either). %% API exports -export([ fold/2 , sequence/1 , traverse/2 , map/2 , is_ok/1 , is_error/1 , from_bool/3 , with_default/2 , and_/1 , oks/1 ]). -type either(Result, Reason) :: {ok, Result} | {error, Reason}. -export_type([either/2]). %%%----------------------------------------------------------------------------- %%% API %%%----------------------------------------------------------------------------- %% @doc Fold over a term with a list of functions. %% %% The first function in the list is called with the initial value, and expected %% to produce either an {@type either(Result, Reason)} or a bare `Result'. If %% the produced value is an `{ok, Result}' tuple, the next function is called %% with `Result' as its input. If the produced value is a bare `Result', the %% next function is called with that value as its input. If the produced value %% is an error, processing stops and returns that `{error, Reason}' tuple. %% %% Note that this function will always produce a tuple, so if the final function %% produces a bare value, this will be wrapped in a tuple, too. -spec fold(term(), [fun((Result) -> Return)]) -> either(Result, Reason) when Result :: term(), Reason :: term(), Return :: either(Result, Reason) | Result. fold(Init, Fs) when is_list(Fs) -> Result = lists:foldl(fun fold_handle/2, Init, Fs), fold_create(Result). %% @doc Combine a list of result tuples. %% %% This will result in either an `{error, Reason}' if any of the supplied tuples %% is an error, or `{ok, [Result]}' with all the ok-values sequenced into a %% list. -spec sequence([either(Result, Reason)]) -> either([Result], Reason) when Result :: term(), Reason :: term(). sequence(Eithers) -> lists:foldr(fun ({ok, Success}, {ok, Successes}) -> {ok, [Success | Successes]}; ({ok, _Success}, {error, Failure}) -> {error, Failure}; ({error, Failure}, _) -> {error, Failure} end, {ok, []}, Eithers). %% @doc Collect results of applying either-returning function on inputs. %% %% We'll bail out with the error on the first item for which the function errors %% out. If the function returns an ok value for all inputs, the result will %% contain those values collected in a single result. -spec traverse(F, [A]) -> either([Result], Reason) when F :: fun ((A) -> either(Result, Reason)), A :: term(), Result :: term(), Reason :: term(). traverse(F, Xs) -> traverse_help(F, Xs, []). %% @doc Map a function over an either -spec map(F :: fun ((A) -> B), either(A, Reason)) -> either(B, Reason). map(F, {ok, V}) -> {ok, F(V)}; map(_, {error, _} = R) -> R. %% @doc Check whether the given result tuple is of the form `{ok, Result}'. -spec is_ok(either(Result, Reason)) -> boolean() when Result :: term(), Reason :: term(). is_ok({ok, _}) -> true; is_ok({error, _}) -> false. %% @doc Check whether the given result tuple is of the form `{error, Reason}'. -spec is_error(either(Result, Reason)) -> boolean() when Result :: term(), Reason :: term(). is_error({ok, _}) -> false; is_error({error, _}) -> true. %% @doc Convert a boolean to the form `{ok, Result}' or `{error, Reason}'. -spec from_bool(Result, Reason, boolean()) -> either(Result, Reason) when Result :: term(), Reason :: term(). from_bool(Result, _, true) -> {ok, Result}; from_bool(_, Reason, false) -> {error, Reason}. %% @doc Fold over a term that depend on a previous result -spec and_(fun((X) -> either(Y, Err))) -> fun((X) -> either({X, Y}, Err)). and_(F) -> fun (X) -> case F(X) of {ok, Y} -> {ok, {X,Y}}; {error, E} -> {error, E}; Y -> {ok, {X, Y}} end end. -spec with_default(either(Result, Reason), Default) -> Result | Default when Result :: term(), Reason :: term(), Default :: term(). with_default({ok, Value}, _) -> Value; with_default({error, _}, Default) -> Default. %% @doc Collect the `ok' values from a list of result tuples. -spec oks([either(Result, Reason)]) -> [Result] when Result :: term(), Reason :: term(). oks(Eithers) -> lists:filtermap( fun({ok, Result}) -> {true, Result}; ({error, _}) -> false end , Eithers). %%%----------------------------------------------------------------------------- %%% Internal functions %%%----------------------------------------------------------------------------- -spec fold_handle(fun((Result) -> Return), Return) -> Return when Result :: term(), Reason :: term(), Return :: either(Result, Reason) | Result. fold_handle(F, {ok, Value}) -> F(Value); fold_handle(_, {error, Reason}) -> {error, Reason}; fold_handle(F, Value) -> F(Value). -spec fold_create(either(Result, Reason) | Result) -> either(Result, Reason) when Result :: term(), Reason :: term(). fold_create({error, Reason}) -> {error, Reason}; fold_create({ok, Value}) -> {ok, Value}; fold_create(Value) -> {ok, Value}. -spec traverse_help(F, Xs, Acc) -> either(Result, Reason) when F :: fun ((A) -> either(Result, Reason)), Acc :: [Result], Xs :: [A], A :: term(), Result :: term(), Reason :: term(). traverse_help(_, [], Acc) -> {ok, lists:reverse(Acc)}; traverse_help(F, [X | Xs], Acc) -> case fold_create(F(X)) of {ok, V} -> traverse_help(F, Xs, [V | Acc]); {error, Reason} -> {error, Reason} end. %% Local variables: %% mode: erlang %% erlang-indent-level: 2 %% indent-tabs-mode: nil %% fill-column: 80 %% coding: latin-1 %% End:
src/tql_either.erl
0.677047
0.693337
tql_either.erl
starcoder
% Licensed under the Apache License, Version 2.0 (the "License"); you may not % use this file except in compliance with the License. You may obtain a copy of % the License at % % http://www.apache.org/licenses/LICENSE-2.0 % % Unless required by applicable law or agreed to in writing, software % distributed under the License is distributed on an "AS IS" BASIS, WITHOUT % WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the % License for the specific language governing permissions and limitations under % the License. -module(couch_btree_copy). -export([copy/3]). -include("couch_db.hrl"). -record(acc, { btree, fd, before_kv_write = {fun(Item, Acc) -> {Item, Acc} end, []}, filter = fun(_) -> true end, compression = ?DEFAULT_COMPRESSION, chunk_threshold, nodes = dict:from_list([{1, []}]), cur_level = 1, max_level = 1 }). copy(Btree, Fd, Options) -> case lists:member(override, Options) of true -> ok = couch_file:truncate(Fd, 0); false -> ok end, Acc0 = #acc{ btree = Btree, fd = Fd, chunk_threshold = Btree#btree.chunk_threshold }, Acc = apply_options(Options, Acc0), {ok, _, #acc{cur_level = 1} = FinalAcc0} = couch_btree:fold( Btree, fun fold_copy/3, Acc, []), {ok, CopyRootState, FinalAcc} = finish_copy(FinalAcc0), {_, LastUserAcc} = FinalAcc#acc.before_kv_write, {ok, CopyRootState, LastUserAcc}. apply_options([], Acc) -> Acc; apply_options([{before_kv_write, {Fun, UserAcc}} | Rest], Acc) -> apply_options(Rest, Acc#acc{before_kv_write = {Fun, UserAcc}}); apply_options([{filter, Fun} | Rest], Acc) -> apply_options(Rest, Acc#acc{filter = Fun}); apply_options([override | Rest], Acc) -> apply_options(Rest, Acc); apply_options([{compression, Comp} | Rest], Acc) -> apply_options(Rest, Acc#acc{compression = Comp}); apply_options([{chunk_threshold, Threshold} | Rest], Acc) -> apply_options(Rest, Acc#acc{chunk_threshold = Threshold}). extract(#acc{btree = #btree{extract_kv = Extract}}, Value) -> Extract(Value). assemble(#acc{btree = #btree{assemble_kv = Assemble}}, Key, Value) -> Assemble(Key, Value). before_leaf_write(#acc{before_kv_write = {Fun, UserAcc0}} = Acc, KVs) -> {NewKVs, NewUserAcc} = lists:mapfoldl( fun({K, V}, UAcc) -> Item = assemble(Acc, K, V), {NewItem, UAcc2} = Fun(Item, UAcc), {K, _NewValue} = NewKV = extract(Acc, NewItem), {NewKV, UAcc2} end, UserAcc0, KVs), {NewKVs, Acc#acc{before_kv_write = {Fun, NewUserAcc}}}. write_leaf(#acc{fd = Fd, compression = Comp}, Node, Red) -> {ok, Pos, Size} = couch_file:append_term(Fd, Node, [{compression, Comp}]), {ok, {Pos, Red, Size}}. write_kp_node(#acc{fd = Fd, btree = Bt, compression = Comp}, NodeList) -> {ChildrenReds, ChildrenSize} = lists:foldr( fun({_Key, {_P, Red, Sz}}, {AccR, AccSz}) -> {[Red | AccR], Sz + AccSz} end, {[], 0}, NodeList), Red = case Bt#btree.reduce of nil -> []; _ -> couch_btree:final_reduce(Bt, {[], ChildrenReds}) end, {ok, Pos, Size} = couch_file:append_term( Fd, {kp_node, NodeList}, [{compression, Comp}]), {ok, {Pos, Red, ChildrenSize + Size}}. fold_copy(Item, _Reds, #acc{nodes = Nodes, cur_level = 1, filter = Filter} = Acc) -> case Filter(Item) of false -> {ok, Acc}; true -> {K, V} = extract(Acc, Item), LevelNode = dict:fetch(1, Nodes), LevelNodes2 = [{K, V} | LevelNode], NextAcc = case ?term_size(LevelNodes2) >= Acc#acc.chunk_threshold of true -> {LeafState, Acc2} = flush_leaf(LevelNodes2, Acc), bubble_up({K, LeafState}, Acc2); false -> Acc#acc{nodes = dict:store(1, LevelNodes2, Nodes)} end, {ok, NextAcc} end. bubble_up({Key, NodeState}, #acc{cur_level = Level} = Acc) -> bubble_up({Key, NodeState}, Level, Acc). bubble_up({Key, NodeState}, Level, Acc) -> #acc{max_level = MaxLevel, nodes = Nodes} = Acc, Acc2 = Acc#acc{nodes = dict:store(Level, [], Nodes)}, case Level of MaxLevel -> Acc2#acc{ nodes = dict:store(Level + 1, [{Key, NodeState}], Acc2#acc.nodes), max_level = Level + 1 }; _ when Level < MaxLevel -> NextLevelNodes = dict:fetch(Level + 1, Acc2#acc.nodes), NextLevelNodes2 = [{Key, NodeState} | NextLevelNodes], case ?term_size(NextLevelNodes2) >= Acc#acc.chunk_threshold of true -> {ok, NewNodeState} = write_kp_node( Acc2, lists:reverse(NextLevelNodes2)), bubble_up({Key, NewNodeState}, Level + 1, Acc2); false -> Acc2#acc{ nodes = dict:store(Level + 1, NextLevelNodes2, Acc2#acc.nodes) } end end. finish_copy(#acc{cur_level = 1, max_level = 1, nodes = Nodes} = Acc) -> case dict:fetch(1, Nodes) of [] -> {ok, nil, Acc}; [{_Key, _Value} | _] = KvList -> {RootState, Acc2} = flush_leaf(KvList, Acc), {ok, RootState, Acc2} end; finish_copy(#acc{cur_level = Level, max_level = Level, nodes = Nodes} = Acc) -> case dict:fetch(Level, Nodes) of [{_Key, {Pos, Red, Size}}] -> {ok, {Pos, Red, Size}, Acc}; NodeList -> {ok, RootState} = write_kp_node(Acc, lists:reverse(NodeList)), {ok, RootState, Acc} end; finish_copy(#acc{cur_level = Level, nodes = Nodes} = Acc) -> case dict:fetch(Level, Nodes) of [] -> Acc2 = Acc#acc{cur_level = Level + 1}, finish_copy(Acc2); [{LastKey, _} | _] = NodeList -> {UpperNodeState, Acc2} = case Level of 1 -> flush_leaf(NodeList, Acc); _ when Level > 1 -> {ok, KpNodeState} = write_kp_node(Acc, lists:reverse(NodeList)), {KpNodeState, Acc} end, ParentNode = dict:fetch(Level + 1, Nodes), Acc3 = Acc2#acc{ nodes = dict:store(Level + 1, [{LastKey, UpperNodeState} | ParentNode], Nodes), cur_level = Level + 1 }, finish_copy(Acc3) end. flush_leaf(KVs, #acc{btree = Btree} = Acc) -> {NewKVs, Acc2} = before_leaf_write(Acc, lists:reverse(KVs)), Red = case Btree#btree.reduce of nil -> []; _ -> Items = lists:map( fun({K, V}) -> assemble(Acc2, K, V) end, NewKVs), couch_btree:final_reduce(Btree, {Items, []}) end, {ok, LeafState} = write_leaf(Acc2, {kv_node, NewKVs}, Red), {LeafState, Acc2}.
src/couchdb/couch_btree_copy.erl
0.537284
0.481515
couch_btree_copy.erl
starcoder
%% ------------------------------------------------------------------ %% @doc LoraWAN Adaptive DataRate engine. %% %% This module exposes a simple adaptive DataRate algorithm loosely %% based on Semtech's recommended algorithm [1,2]. %% %% === Usage === %% %% This module has a minimal API and almost zero coupling to external %% code. Because of this, you have to bring your own IO (BYOIO). It %% will not perform any packet reception, transmission, or %% construction. Instead, you must call this module for every uplink %% packet received and the returned terms tell you want actions you %% need to take in order to properly regulate the end-device's %% DataRate and transmit-power. %% %% Example: %% %% ``` %% %% Create a new handle with the device's region. See new/2 for more control. %% State01 = lorawan_adr:new('US915'), %% %% %% First track all packet offers. %% OfferN = #adr_offer{ %% packet_hash = RxOfferHash %% }, %% %% State09 = lorawan_adr:track_offer(State08, OfferN), %% %% %% Assume these RxPkt[X] values all come from uplink packet and %% %% that we have already received several similar packets since State01. %% Packet10 = #adr_packet{ %% rssi = RxPktRssi, %% snr = RxPktSnr, %% wants_adr = RxPktAdrBit == 1, % true %% wants_adr_ack = RxPktAdrAckReqBit == 1, % false %% datarate_config = {RxPktSpreading, RxPktBandwidth}, %% packet_hash = RxPktHash %% }, %% %% {State10, {NewDataRate, NewPower}} = lorawan_adr:track_packet( %% State09, %% Packet10 %% ), %% %% %% %% %% ... here you handle sending a downlink packet with the %% %% `LinkADRReq' MAC command. %% %% %% %% %% If you receive an uplink response with the `LinkADRAns' MAC %% %% command, inform the ADR engine so it can update its state: %% Answer0 = #adr_answer{ %% channel_mask_ack = true, %% datarate_ack = true, %% power_ack = true %% }, %% State11 = lorawan_adr:track_adr_answer(State10, Answer0), %% ''' %% %% ==== A rant about transmit power ==== %% %% LoRaWAN requires us to set both DataRate and transmit power when %% sending a `LinkADRReq' MAC command. Unfortunately, LoRaWAN's %% designers forgot to give devices a mechanism for reporting their %% current TX power. Furthermore, DataRate and transmit power changes %% are communicated to end devices with values, not deltas (e.g., set %% power to 1, not reduce power by 2 steps). Unlike DataRate, which is %% absolute and reported by the forwarding gateway, power degrades %% over distance and is not derivable by the receiving party. Because %% of this design oversight, we track a device's TX power by: %% %% <dl> %% <dt> Devices for which we have not yet received a `LinkADRAns': </dt> %% <dd> Assume the device is transmitting at TXPower index 0 as %% defined its region's 'TX Power Table' defined in the %% regional parameters document. </dd> %% <dt> Otherwise: </dt> %% <dd> Lookup values sent in last `LinkADRReq' </dd> %% </dl> %% %% ==== References ==== %% %% <ol> %% <li>LoRaWAN – simple rate adaptation recommended algorithm (Revision 1.0)</li> %% <li>[https://www.thethingsnetwork.org/docs/lorawan/adaptive-data-rate.html]</li> %% <li>[https://github.com/TheThingsNetwork/ttn/issues/265#issuecomment-329449935]</li> %% </ol> %% @end %% ------------------------------------------------------------------ -module(lorawan_adr). -include("lorawan_adr.hrl"). %% ================================================================== %% Public API Exports %% ================================================================== -export([ datarate_entry/2, max_datarate/1, min_datarate/1, new/1, new/2, track_adr_answer/2, track_offer/2, track_packet/2 ]). -export_type([ adjustment/0, datarate_config/0, handle/0 ]). -type datarate_config() :: {lorawan_utils:spreading(), lorawan_utils:bandwidth()} %% A tuple of `{SpreadingFactor, Bandwidth}'. . -type datarate_entry() :: {pos_integer(), datarate_config()} %% A tuple of `{DataRate, {SpreadingFactor, Bandwidth}}' returned by {@link %% datarate_entry/2}. . -type adjustment() :: hold | {DataRate :: pos_integer(), TxPower :: pos_integer()} %% An adjustment to send to an end-device, or not when `hold', %% returned by {@link track_packet/2}. . %% ================================================================== %% Internal Constants %% ================================================================== %% The default number of seconds we keep historic offers around. %% %% This value can be overridden by calling {@link new/2} with option %% `{offer_lifetime_secs, NumberOfSeconds}'. %% %% We could get more precise about evicting old offers, but %% compared to using time, it doesn't buy us anything and probably makes %% things more complicated than it needs to be. -define(DEFAULT_OFFER_LIFETIME_S, 10). %% The default minimum number of packets we need in history before %% eviction. %% %% This value can be overridden by calling {@link new/2} with option %% `{adr_history_len, Length}'. -define(DEFAULT_ADR_HISTORY_LEN, 20). %% The closest we allow historical packet SNRs to approach the %% theoretical minimum when determining whether or not change %% spreading factor. %% %% This value can be overridden by calling {@link new/2} with option %% `{snr_headroom, Decibels}'. -define(DEFAULT_SNR_HEADROOM_dB, 10.0). %% Default amount to adjust uplink TXPower for every available %% adjustment step in dBm. Note, unlike DataRate, we adjust by dBm and %% then find the closest TXPower index value to match. -define(DEFAULT_TXPOWER_ADJUSTMENT_DBM_STEP_SIZE, 3). %% A map of spreading factor to minimum signal-to-noise ratio needed %% for reliable reception. -define(SNR_THRESHOLD_dB, #{ 7 => -7.5, 8 => -10.0, 9 => -12.5, 10 => -15.0, 11 => -17.5, 12 => -20.0 }). %% ================================================================== %% Internal Types %% ================================================================== -record(packet, { %% Hash is used as a unique identifier. hash :: binary(), %% Number of gateways which heard this packet. gateway_diversity :: pos_integer(), %% Best signal-to-noise ratio for all the gateways that reported SNR. best_snr :: float(), %% Best signal strength for all the gateways that reported RSSI. best_rssi :: float() }). %% {monotonic time (seconds), hash} -type offer() :: {integer(), binary()}. %% Region-specific Data Rate encoding table. %% %% TODO: bandwidth is not currently used because we're artificially %% filtering the regional parameters for only 125kHz channels. Does %% this make sense? %% %% [{DataRate, {Spreading, Bandwidth}}] -type regional_datarates() :: list( {DataRate :: pos_integer(), datarate_config()} ). -record(device, { %% The region who's rules this device is operating under. Not %% necessarily the region the device is physically in. region :: atom(), %% History of previously received packets. packet_history :: list(#packet{}), %% History of previously received offers. offer_history :: list(offer()), %% ADR awaiting an acknowledgment from the device via a %% `LinkADRAns' MAC command. pending_adjustments :: list(adjustment()), %% ADR adjustments acknowledged by the device via a `LinkADRAns' %% MAC command. accepted_adjustments :: list(adjustment()), %% Table of region-specific DataRate parameters. datarates :: regional_datarates(), %% Table of region-specific uplink power. txpowers :: lorawan_mac_region:tx_power_table(), %% Slowest DataRate available in this device's `datarates'. %% %% The word DataRate, as used by the regional parameters document, %% is essentially an index into a table. It is probably safe to %% remove this and assume 0, but let's track it for completeness %% sake for the time being. min_datarate :: pos_integer(), %% Spreading factor for corresponding `min_datarate'. %% %% Q: Why would a `max_spreading' correspond to `min_datarate'? %% A: Because DataRate, aka throughput, decreases as spreading %% increases. %% %% TODO: maybe it's fine to always lookup this value in %% `datarates' when needed. max_spreading :: pos_integer(), %% Fastest DataRate available in this device's `datarates'. max_datarate :: pos_integer(), %% Spreading factor for corresponding `max_datarate'. %% %% Q: Why would a `min_spreading' correspond to `max_datarate'? %% A: Because DataRate, aka throughput, increases as spreading %% factor decreases. %% %% TODO: maybe it's fine to always lookup this value in %% `datarates' when needed. min_spreading :: pos_integer(), %% Index of strongest transmit power in regional parameters table. %% %% TODO: power _decreases_ as the index increases, so %% `max_txpower_idx' is almost certainly index 0. max_txpower_idx :: pos_integer(), %% Value, in dBm, of transmit power at index `max_txpower_idx' in %% the regional parameters document. max_txpower_dbm :: number(), %% Index of weakest transmit power in the regional parameters %% document. min_txpower_idx :: pos_integer(), %% Value, in dBm, of transmit power at index `min_txpower_idx' in %% the regional parameters document. min_txpower_dbm :: number(), %% -------------------------------------------------------------- %% Options %% -------------------------------------------------------------- option_adr_history_len = ?DEFAULT_ADR_HISTORY_LEN :: pos_integer(), option_offer_lifetime_sec = ?DEFAULT_OFFER_LIFETIME_S :: number(), option_snr_headroom = ?DEFAULT_SNR_HEADROOM_dB :: float(), option_txpower_adjustment_dbm_step_size = ?DEFAULT_TXPOWER_ADJUSTMENT_DBM_STEP_SIZE :: number() }). -opaque handle() :: #device{} %% Obtain a `handle()' by calling {@link new/1} or {@link new/2}. %% %% `handle()' are the state object used for all ADR API calls. . %% ================================================================== %% Public API %% ================================================================== %% ------------------------------------------------------------------ %% @doc Returns a new ADR handle with sane defaults suitable for use %% in the specified region. %% %% Use {@link new/2} instead for more granular control of ADR %% behavior. %% @end %% ------------------------------------------------------------------ -spec new(Region :: atom()) -> handle(). new(Region) -> %% Filter gotthardp's table down to only 125kHz uplink DataRates. FilterMapFn = fun ({_, _, down}) -> false; ({DataRate, {Spreading, 125 = Bandwidth}, _Direction}) -> {true, {DataRate, {Spreading, Bandwidth}}}; (_) -> false end, Datarates = lists:filtermap(FilterMapFn, lorawan_mac_region:datars(Region)), %% min-max refer to #device.min_datarate docs [{MinDataRate, {MaxSpreading, _}} | _] = Datarates, {MaxDataRate, {MinSpreading, _}} = lists:last(Datarates), TxPowers = lorawan_mac_region:uplink_power_table(Region), [{MaxTxPowerIdx, MaxTxPowerDBm} | _] = TxPowers, {MinTxPowerIdx, MinTxPowerDBm} = lists:last(TxPowers), #device{ region = Region, offer_history = [], packet_history = [], pending_adjustments = [], accepted_adjustments = [], datarates = Datarates, txpowers = TxPowers, min_datarate = MinDataRate, max_spreading = MaxSpreading, max_datarate = MaxDataRate, min_spreading = MinSpreading, max_txpower_idx = MaxTxPowerIdx, max_txpower_dbm = MaxTxPowerDBm, min_txpower_idx = MinTxPowerIdx, min_txpower_dbm = MinTxPowerDBm }. %% ------------------------------------------------------------------ %% @doc Returns a new ADR handle for the specified region. %% %% <h4>Options:</h4> %% <dl> %% <dt>{@type {offer_lifetime_secs, pos_integer()@}}</dt> %% <dd>The maximum allowable age an offer can be for consideration %% in gateway diversity calculations.</dd> %% %% <dt>{@type {adr_history_len, pos_integer()@}}</dt> %% <dd>The minimum number of packets we need in history before %% {@link track_packet/2} will return DataRate adjustments.</dd> %% %% <dt>{@type {snr_headroom, number()@}}</dt> %% <dd>The closest we allow historical packet SNRs to approach the %% theoretical minimum when determining whether or not change %% DataRate.</dd> %% %% <dt>{@type {txpower_adjustment_dbm_step_size, number()@}}</dt> %% <dd>After maxing out DataRate, adjust uplink transmit power by %% these many dBm for each remaining step. The adjustment is %% added to TxPower when remaining steps is negative, otherwise %% it's subtracted. Read ADR reference algorithm %% for definition of "step".</dd> %% </dl> %% @end %% ------------------------------------------------------------------ -spec new(Region :: term(), Options :: proplists:proplist()) -> handle(). new(Region, Options) -> OptionParserFn = fun ({offer_lifetime_secs, Value}, Device) when Value > 0 -> Device#device{option_offer_lifetime_sec = Value}; ({adr_history_len, Value}, Device) when Value > 0 -> Device#device{option_adr_history_len = Value}; ({snr_headroom, Value}, Device) when Value > 0.0 -> Device#device{option_snr_headroom = Value}; ({txpower_adjustment_dbm_step_size, Value}, Device) when is_number(Value) -> Device#device{option_txpower_adjustment_dbm_step_size = Value} end, lists:foldl(OptionParserFn, new(Region), Options). %% ------------------------------------------------------------------ %% @doc Consider this new offer in future ADR calculations. %% %% Because we have not bought the packet yet, and can't see its %% header, we do not yet know if this packet has its ADR bit %% set. So we will hold onto it until: %% %% <ol> %% <li>We purchase the packet. </li> %% <li>The offer is older than `offer_lifetime_secs' seconds.</li> %% </ol> %% @end %% ------------------------------------------------------------------ -spec track_offer( Device :: handle(), Offer :: #adr_offer{} ) -> handle(). track_offer( Device, Offer ) -> #device{offer_history = Offers} = Device, %% See note for OFFER_LIFETIME_S for an explanation of why we're %% using time and not length. Now = erlang:monotonic_time(seconds), TrimFn = fun({TimeStamp, _}) -> TimeStamp + Device#device.option_offer_lifetime_sec > Now end, Trimmed = lists:takewhile(TrimFn, Offers), Device#device{offer_history = [{Now, Offer#adr_offer.packet_hash} | Trimmed]}. %% ------------------------------------------------------------------ %% @doc Remember this packet and (possibly) return an ADR DataRate %% adjustment. %% %% <h4> Returns </h4> %% <dl> %% <dt>`hold'</dt> %% <dd> Returned when either: %% <ul> %% <li>Have not collected enough historic packets to make an ADR %% adjustment.</li> %% <li>DataRate can not be improved (likely too low historic %% SNR).</li> %% </ul> %% </dd> %% %% <dt>{@type {NewDataRate, NewTxPower@}}</dt> %% <dd>Returned when either: %% <ul> %% <li>Returned when current DataRate is not optimal based on %% SNRs we've seen in historic packets.</li> %% <li>`ADRAckreqbitset' is set, regardless if we've collected %% enough packets yet.</li> %% </ul> %% </dd> %% </dl> %% @end %% ------------------------------------------------------------------ -spec track_packet( handle(), #adr_packet{} ) -> {handle(), adjustment()}. track_packet( Device0, Pkt ) when Pkt#adr_packet.wants_adr == true -> #device{ offer_history = Offers0, packet_history = History0, pending_adjustments = PendingAdjustments0, accepted_adjustments = AcceptedAdjustments0, txpowers = TxPowerTable, max_txpower_idx = MaxTxPowerIdx, min_datarate = MinDataRate, max_datarate = MaxDataRate, option_adr_history_len = ConfiguredAdrHistoryLen, option_snr_headroom = ConfiguredSNRHeadroom, option_txpower_adjustment_dbm_step_size = TxPowerAdjustmentDBmStepSize } = Device0, #adr_packet{ wants_adr_ack = WantsADRAck, rssi = PktRssi, snr = PktSnr, datarate_config = DataRateConfig, packet_hash = PktHash } = Pkt, {OfferGatewayDiversity, Offers1} = count_and_prune_offers_for_hash(PktHash, Offers0), TrimmedHistory = lists:sublist(History0, ConfiguredAdrHistoryLen - 1), History1 = case TrimmedHistory of [H | T] when H#packet.hash == PktHash -> %% Previous diversity count + this packet + offer count %% %% Offer count is unlikely to be > 0 considering we've %% already accepted a different gateway's packet with %% this hash CumulativeDiversity = 1 + OfferGatewayDiversity + H#packet.gateway_diversity, BestSnr = erlang:max(H#packet.best_snr, PktSnr), [H#packet{gateway_diversity = CumulativeDiversity, best_snr = BestSnr} | T]; _ -> NewHistoryHead = #packet{ hash = PktHash, gateway_diversity = OfferGatewayDiversity + 1, best_snr = PktSnr, best_rssi = PktRssi }, [NewHistoryHead | TrimmedHistory] end, {CurrentDataRate, {CurrentSpreading, _CurrentBandwidth}} = case datarate_entry(Device0, DataRateConfig) of %% TODO: what is the best strategy when a device sends a %% packet at a `{Spreading, Bandwidth}' not in our %% regional parameters table? Sort term is to just %% lie to the DataRate adjuster that the device is %% at the minimum DataRate. undefined -> datarate_entry(Device0, MinDataRate); DataRateEntry -> DataRateEntry end, %% We don't (and can't!) know with great confidence what transmit %% power the device is using. We can however guess based on the %% most recently acknowledged ADR adjustment if we have any. If we %% haven't received any acknowledgments yet, we assume the device %% is using its region's maximum transmit power. {LastAcceptedDataRate, LastAcceptedTxPowerIdx} = case AcceptedAdjustments0 of [] -> {CurrentDataRate, MaxTxPowerIdx}; [Accepted | _] -> Accepted end, AcceptedAdjustments1 = case CurrentDataRate == LastAcceptedDataRate of true -> AcceptedAdjustments0; false -> lager:warning("expected data rate ~p, got ~p", [ LastAcceptedDataRate, CurrentDataRate ]), [] end, Adjustment = adjust_uplink_params( CurrentDataRate, CurrentSpreading, MaxDataRate, History1, WantsADRAck, ConfiguredAdrHistoryLen, ConfiguredSNRHeadroom, TxPowerTable, TxPowerAdjustmentDBmStepSize, LastAcceptedTxPowerIdx ), %% We only want to put actual adjustments in %% `pending_adjustments', `hold' is not really an %% adjustment. Perhaps the `adjustment()' type should be %% refactored to indicate that. PendingAdjustments1 = case Adjustment of hold -> PendingAdjustments0; _ -> [Adjustment | PendingAdjustments0] end, Device1 = Device0#device{ offer_history = Offers1, packet_history = History1, pending_adjustments = PendingAdjustments1, accepted_adjustments = AcceptedAdjustments1 }, lager:info( "~p, ADR ~p, ADRAck ~p, SNR ~p, RSSI ~p, diversity ~p," ++ " pend adj len ~p, history len ~p, adjustment ~p", [ datarate_entry(Device0, DataRateConfig), true, WantsADRAck, PktSnr, PktRssi, OfferGatewayDiversity, length(PendingAdjustments1), length(History1), Adjustment ] ), {Device1, Adjustment}; %% We clear all state when device doesn't want ADR control. track_packet( Device, _Pkt ) -> case Device#device.packet_history of [] -> ok; _ -> lager:info("device turned off ADR, clearing history") end, { Device#device{ packet_history = [], offer_history = [], pending_adjustments = [], accepted_adjustments = [] }, hold }. %% ------------------------------------------------------------------ %% @doc Returns this device's minimum (slowest) DataRate index[1]. %% %% 1: what regional parameters spec calls DataRate. %% @end %% ------------------------------------------------------------------ -spec min_datarate(handle()) -> pos_integer(). min_datarate(Device) -> Device#device.min_datarate. %% ------------------------------------------------------------------ %% @doc Returns this device's maximum (fastest) DataRate index[1]. %% %% 1: what regional parameters spec calls DataRate. %% @end %% ------------------------------------------------------------------ -spec max_datarate(handle()) -> pos_integer(). max_datarate(Device) -> Device#device.max_datarate. %% ------------------------------------------------------------------ %% @doc Returns the {@link datarate_entry()} for the given %% DataRate-index[1] or {@link datarate_config()}. %% %% Returns `undefined' when DataRate is not in the device's %% table. This function always returns the full DataRate table %% entry. One of the values in the outer tuple is redundant and will %% be the same as the Index or Config you pass in. %% %% 1: what regional parameters spec calls DataRate. %% @end %% ------------------------------------------------------------------ -spec datarate_entry( Device :: handle(), DataRate :: pos_integer() | datarate_config() ) -> datarate_entry() | undefined. datarate_entry(Device, Config) when is_tuple(Config) -> case lists:keyfind(Config, 2, Device#device.datarates) of false -> undefined; Entry -> Entry end; datarate_entry( Device, DataRate ) when is_integer(DataRate) -> case lists:keyfind(DataRate, 1, Device#device.datarates) of false -> undefined; Entry -> Entry end. %% ------------------------------------------------------------------ %% @doc Must be called when a device uplinks a `LinkADRAns' MAC %% command. %% %% Failure to call this function for every `LinkADRAns' will result in %% growing memory consumption and non-standards-compliant ADR %% behavior. %% @end %% ------------------------------------------------------------------ -spec track_adr_answer(Device :: handle(), Answer :: #adr_answer{}) -> handle(). track_adr_answer(#device{pending_adjustments = []} = Device, _Answer) -> lager:warning( "Device firmware bug or DoS attempt: received an ADR answer" ++ " with no outstanding ADR requests" ), Device; track_adr_answer(Device0, Answer) -> #device{ %% TODO: how to handle the case where `PendingTail =/= []'? pending_adjustments = [{AcceptedDataRate, AcceptedTXPower} = PendingHead | PendingTail], accepted_adjustments = AcceptedAdjustments0 } = Device0, #adr_answer{ channel_mask_ack = ChannelMaskAck, datarate_ack = DataRateAck, power_ack = PowerAck } = Answer, Acceptance = fun(Ack) -> case Ack of true -> accepted; false -> rejected end end, %% NOTE: channel mask is handled outside of this module. All we %% know is if it was accepted. lager:info( "device ~s DataRate ~p, ~s TxPower ~p, and ~s ChannelMask" ++ " with ~p previously unanswered adjustments", [ Acceptance(DataRateAck), AcceptedDataRate, Acceptance(PowerAck), AcceptedTXPower, Acceptance(ChannelMaskAck), length(PendingTail) ] ), case {ChannelMaskAck, DataRateAck, PowerAck} of {_, true, true} -> Device0#device{ pending_adjustments = [], accepted_adjustments = [PendingHead | AcceptedAdjustments0] }; _ -> Device0#device{pending_adjustments = []} end. %% ================================================================== %% Private API %% ================================================================== -spec snr_threshold(lorawan_utils:spreading()) -> number(). snr_threshold(Spreading) -> maps:get(Spreading, ?SNR_THRESHOLD_dB). %% Returns highest SNR from uplink history. -spec best_snr(nonempty_list(#packet{})) -> number(). best_snr([H | T]) -> lists:foldl( fun(Packet, Winner) -> erlang:max(Packet#packet.best_snr, Winner) end, H#packet.best_snr, T ). -spec adjust_uplink_params( AdjSteps :: integer(), DataRate :: pos_integer(), MaxDataRate :: pos_integer(), TxPowerAdjustmentDBmStepSize :: number(), TxPowerDBm :: number(), MinTxPowerDBm :: number(), MaxTxPowerDBm :: number() ) -> {DataRate :: pos_integer(), TxPowerDBm :: number()}. adjust_uplink_params( AdjSteps, DataRate, MaxDataRate, TxPowerAdjustmentDBmStepSize, TxPowerDBm, MinTxPowerDBm, MaxTxPowerDBm ) when (AdjSteps > 0), (DataRate < MaxDataRate) -> %% ------------------------------------------- %% Begin increasing DataRate. %% ------------------------------------------- Take = erlang:min(MaxDataRate - DataRate, AdjSteps), adjust_uplink_params( AdjSteps - Take, DataRate + Take, MaxDataRate, TxPowerAdjustmentDBmStepSize, TxPowerDBm, MinTxPowerDBm, MaxTxPowerDBm ); adjust_uplink_params( AdjSteps, DataRate, MaxDataRate, TxPowerAdjustmentDBmStepSize, TxPowerDBm, MinTxPowerDBm, MaxTxPowerDBm ) when (AdjSteps > 0), (TxPowerDBm > MinTxPowerDBm) -> %% ------------------------------------------- %% DataRate is maxed out; reduce uplink power. %% ------------------------------------------- adjust_uplink_params( AdjSteps - 1, DataRate, MaxDataRate, TxPowerAdjustmentDBmStepSize, max(TxPowerDBm - TxPowerAdjustmentDBmStepSize, MinTxPowerDBm), MinTxPowerDBm, MaxTxPowerDBm ); adjust_uplink_params( AdjSteps, DataRate, MaxDataRate, TxPowerAdjustmentDBmStepSize, TxPowerDBm, MinTxPowerDBm, MaxTxPowerDBm ) when (AdjSteps < 0), (TxPowerDBm < MaxTxPowerDBm) -> %% ------------------------------------------- %% Negative SNR headroom, so we increase power. %% %% NOTE: Decreasing DataRate is not allowed. %% ------------------------------------------- adjust_uplink_params( AdjSteps + 1, DataRate, MaxDataRate, TxPowerAdjustmentDBmStepSize, min(TxPowerDBm + TxPowerAdjustmentDBmStepSize, MaxTxPowerDBm), MinTxPowerDBm, MaxTxPowerDBm ); adjust_uplink_params(_, DataRate, _, _, TxPowerDBm, _, _) -> %% ------------------------------------------- %% We have either: %% - Run out of adjustment steps. %% - Maxed data rate and/or pegged TxPowerDBm %% to either [Max,Min]TxPowerDBm. %% ------------------------------------------- {DataRate, TxPowerDBm}. -spec adjust_uplink_params( CurrentDataRate :: pos_integer(), CurrentSpreading :: lorawan_utils:spreading(), MaxDataRate :: pos_integer(), History :: nonempty_list(#packet{}), WantsADRAck :: boolean(), ConfiguredAdrHistoryLen :: pos_integer(), ConfiguredSNRHeadroom :: number(), TxPowerTable :: lorawan_mac_region:tx_power_table(), TxPowerAdjustmentDBmStepSize :: number(), CurrentTxPowerIdx :: number() ) -> adjustment(). adjust_uplink_params( CurrentDataRate, CurrentSpreading, MaxDataRate, History, WantsADRAck, ConfiguredAdrHistoryLen, ConfiguredSNRHeadroom, TxPowerTable, TxPowerAdjustmentDBmStepSize, CurrentTxPowerIdx ) when (length(History) >= ConfiguredAdrHistoryLen) or (WantsADRAck == true) -> SNR_threshold = snr_threshold(CurrentSpreading), SNR_max = best_snr(History), %% TODO: add a bunch of extra headroom (e.g., perform more %% conservative adjustments, or maybe no adjustments) when %% we do not have enough history but WantsADRAck is true SNR_margin = SNR_max - SNR_threshold - ConfiguredSNRHeadroom, AdjSteps = trunc(SNR_margin / 3), {CurrentTxPowerIdx, CurrentTxPowerDBm} = lists:keyfind( CurrentTxPowerIdx, 1, TxPowerTable ), [{_MaxTxPowerIdx, MaxTxPowerDBm} | _] = TxPowerTable, {_MinTxPowerIdx, MinTxPowerDBm} = lists:last(TxPowerTable), {AdjustedDataRate, AdjustedTxPowerDBm} = adjust_uplink_params( AdjSteps, CurrentDataRate, MaxDataRate, TxPowerAdjustmentDBmStepSize, CurrentTxPowerDBm, MinTxPowerDBm, MaxTxPowerDBm ), %% This part is tricky. Most places deal with %% indices that are absolute, but for TX power, the %% adjustment algorithm deals in dBm. So we need to %% search through the transmit power table and find %% the TxPowerIdx with a dBm closest to the value %% calculated by the adjustment algorithm. TxPowerPredFn = fun({_, DBm}) -> DBm >= AdjustedTxPowerDBm end, {AdjustedTxPowerIdx, _ActualTxPowerDBm} = lists:last( lists:takewhile(TxPowerPredFn, TxPowerTable) ), case WantsADRAck or (AdjustedDataRate /= CurrentDataRate) or (AdjustedTxPowerIdx /= CurrentTxPowerIdx) of true -> {AdjustedDataRate, AdjustedTxPowerIdx}; false -> hold end; adjust_uplink_params( _CurrentDataRate, _CurrentSpreading, _MaxDataRate, _History, _WantsADRAck, _ConfiguredAdrHistoryLen, _ConfiguredSNRHeadroom, _TxPowerTable, _TxPowerAdjustmentDBmStepSize, _CurrentTxPowerIdx ) -> hold. %% Removes all offers from the offer list with hashes matching %% `PktHash'. %% %% Returns the the leftover offers and count of how many offers were %% pruned. This count is needed to calculate gateway diversity (total %% number of gateways who a saw packet). -spec count_and_prune_offers_for_hash( PktHash :: binary(), Offers0 :: list(offer()) ) -> {integer(), list(offer())}. count_and_prune_offers_for_hash( PktHash, Offers0 ) -> %% Remove offers for this hash so they are not counted again. OfferFilterFn = fun({_Timestamp, OfferHash}) -> OfferHash == PktHash end, {Removed, Offers1} = lists:partition(OfferFilterFn, Offers0), {length(Removed), Offers1}. %% ================================================================== %% Tests %% ================================================================== -ifdef(EUNIT). -include_lib("eunit/include/eunit.hrl"). device_history(#device{packet_history = History}) -> History. device_history_len(#device{packet_history = History}) -> length(History). device_offers_len(Device) -> length(Device#device.offer_history). spread_and_bandwidth(State, DataRate) -> {DataRate, {Spread, Bandwidth}} = lorawan_adr:datarate_entry( State, DataRate ), {Spread, Bandwidth}. gen_adr_packet(DRConfig, Snr, Rssi) -> RandNum = rand:uniform(4294967296), Packet0 = #adr_packet{ packet_hash = <<RandNum>>, wants_adr = true, wants_adr_ack = false, datarate_config = DRConfig, snr = Snr, rssi = Rssi }, Packet0. gen_adr_ack(DRConfig, Snr, Rssi) -> RandNum = rand:uniform(4294967296), Packet0 = #adr_packet{ packet_hash = <<RandNum>>, wants_adr = true, wants_adr_ack = true, datarate_config = DRConfig, snr = Snr, rssi = Rssi }, Packet0. post_packet_track(State, DRIdx, Snr, Rssi) -> DRConfig = spread_and_bandwidth(State, DRIdx), Packet0 = gen_adr_packet(DRConfig, Snr, Rssi), {State1, _} = track_packet(State, Packet0), State1. get_packet_adr(State, DRIdx, Snr, Rssi) -> DRConfig = spread_and_bandwidth(State, DRIdx), Packet0 = gen_adr_ack(DRConfig, Snr, Rssi), {_State1, {AdjDataRate, AdjPower}} = track_packet(State, Packet0), {AdjDataRate, AdjPower}. adr_jitter(Range) -> Range * rand:uniform(). exercise_packet_track(State, _DRIdx, Count, _Snr, _Rssi) when Count == 0 -> State; exercise_packet_track(State, DRIdx, Count, Snr0, Rssi0) -> Snr = Snr0 - adr_jitter(0.1), Rssi = Rssi0 - adr_jitter(0.1), State1 = post_packet_track(State, DRIdx, Snr, Rssi), exercise_packet_track(State1, DRIdx, Count - 1, Snr, Rssi). % valid_exercise(DR, AdjustedDR, AdjustedPower) -> % ?assert(AdjustedDR >= DR), % ?assert(AdjustedPower >= 0), % fin. exercise_adr_state(State0, DRIdx, Count, Snr, Rssi) -> % io:format("exercise_adr_state count=~w snr=~w, rssi=~w~n", [Count, Snr, Rssi]), State1 = exercise_packet_track(State0, DRIdx, Count, Snr, Rssi), %% ?assertEqual(Count, device_history_len(State1)), ?assert(device_history_len(State1) < 21), {AdjDR, AdjPower} = get_packet_adr(State1, DRIdx, Snr, Rssi), ?assert(AdjDR >= 0), ?assert(AdjPower >= 0), % io:format("DR = ~w Power = ~w~n", [AdjDR, AdjPower]), {AdjDR, AdjPower}. valid_exercise(State0, DRIdx, Count, Snr, Rssi, ExpectedDR, ExpectedPower) -> {DR, Power} = exercise_adr_state(State0, DRIdx, Count, Snr, Rssi), ?assertEqual(ExpectedDR, DR), ?assertEqual(ExpectedPower, Power). % gen_range(Start, Step, Length) when is_number(Length) -> % [Start + (Step * X) || X <- lists:seq(0, Length)]. gen_startend_range(Start, Step, End) -> Length = round((End - Start) / Step), [Start + (Step * X) || X <- lists:seq(0, Length)]. adr_harness_test_() -> DataRate0 = 0, State0 = new('US915'), [ ?_test(begin valid_exercise(State0, DataRate0, 22, 7.0, X, 3, 1) end) || X <- gen_startend_range(-120.0, 0.1, 0.0) ]. adr_exercise_test_() -> %% DataRate 0 in US915 regional parameters. DataRate0 = 0, % Spreading0 = 10, % Bandwidth0 = 125, %% Snr ranges from 10 to -20 Snr = 10.0, %% Rssi ranges from 0 to -120 Rssi = 0.0, % DRConfig0 = {Spreading0, Bandwidth0}, State0 = new('US915'), ?assertEqual(0, device_offers_len(State0)), ?assertEqual(0, device_history_len(State0)), State1 = post_packet_track(State0, DataRate0, Snr, Rssi), State2 = post_packet_track(State1, DataRate0, Snr, Rssi), ?assertEqual(2, device_history_len(State2)), {_AdjDataRate2, _AdjPower2} = get_packet_adr(State2, DataRate0, Snr, Rssi), %% io:format("NewSpreading2 ~8.16.0B~n", [NewSpreading2]), % io:format("AdjDataRate2 ~w~n", [AdjDataRate2]), % io:format("AdjPower2 ~w~n", [AdjPower2]), PacketLimit = 19, State3 = exercise_packet_track(State0, DataRate0, PacketLimit, Snr, Rssi), % ?assertEqual(19, device_history_len(State3)), ?assert(device_history_len(State3) < 21), {_AdjDataRate3, _AdjPower3} = get_packet_adr(State3, DataRate0, Snr, Rssi), % io:format("AdjDataRate3 ~w~n", [AdjDataRate3]), % io:format("AdjPower3 ~w~n", [AdjPower3]), TestList = [ ?_test(begin valid_exercise(State0, DataRate0, 1, Snr, Rssi, 3, 3) end), ?_test(begin valid_exercise(State0, DataRate0, 3, Snr, Rssi, 3, 3) end), ?_test(begin valid_exercise(State0, DataRate0, 7, Snr, Rssi, 3, 3) end), ?_test(begin valid_exercise(State0, DataRate0, 17, Snr, Rssi, 3, 3) end), ?_test(begin valid_exercise(State0, DataRate0, 19, Snr, Rssi, 3, 3) end), ?_test(begin valid_exercise(State0, DataRate0, 20, Snr, Rssi, 3, 3) end), ?_test(begin valid_exercise(State0, DataRate0, 21, Snr, Rssi, 3, 3) end), ?_test(begin valid_exercise(State0, DataRate0, 22, Snr, Rssi, 3, 3) end), ?_test(begin valid_exercise(State0, DataRate0, 100, Snr, Rssi, 3, 3) end), ?_test(begin valid_exercise(State0, DataRate0, 200, Snr, Rssi, 3, 3) end), ?_test(begin valid_exercise(State0, 0, 22, -20.0, -120.0, 0, 0) end), ?_test(begin valid_exercise(State0, 1, 22, -20.0, -120.0, 1, 0) end), ?_test(begin valid_exercise(State0, 2, 22, -20.0, -120.0, 2, 0) end), ?_test(begin valid_exercise(State0, 3, 22, -20.0, -120.0, 3, 0) end), [ ?_test(begin valid_exercise(StateX, 0, 22, -20.0, -120.0, 0, 0) end) || StateX <- [new('US915'), new('EU868'), new('CN470'), new('AS923'), new('AU915')] ], [ ?_test(begin valid_exercise(State0, 0, X, -20.0, -120.0, 0, 0) end) || X <- lists:seq(1, 200) ], [ ?_test(begin valid_exercise(State0, DataRate0, 22, X, -120.0, 0, 0) end) || X <- gen_startend_range(-20.0, 0.1, -2.5) ], [ ?_test(begin valid_exercise(State0, DataRate0, 22, X, -120.0, 1, 0) end) || X <- gen_startend_range(-1.0, 0.1, 0.9) ], [ ?_test(begin valid_exercise(State0, DataRate0, 22, X, -120.0, 2, 0) end) || X <- gen_startend_range(1.0, 0.1, 3.9) ], [ ?_test(begin valid_exercise(State0, DataRate0, 22, X, -120.0, 3, 0) end) || X <- gen_startend_range(4.0, 0.1, 6.9) ], [ ?_test(begin valid_exercise(State0, DataRate0, 22, X, -120.0, 3, 1) end) || X <- gen_startend_range(7.0, 0.1, 9.9) ], [ ?_test(begin valid_exercise(State0, DataRate0, 22, X, -120.0, 3, 3) end) || X <- gen_startend_range(10.0, 0.1, 12.9) ], [ ?_test(begin valid_exercise(State0, DataRate0, 22, X, -120.0, 3, 4) end) || X <- gen_startend_range(13.0, 0.1, 15.9) ], [ ?_test(begin valid_exercise(State0, DataRate0, 22, X, -120.0, 3, 6) end) || X <- gen_startend_range(16.0, 0.1, 18.9) ], [ ?_test(begin valid_exercise(State0, DataRate0, 22, X, -120.0, 3, 7) end) || X <- gen_startend_range(19.0, 0.1, 21.9) ], [ ?_test(begin valid_exercise(State0, DataRate0, 22, X, -120.0, 3, 9) end) || X <- gen_startend_range(22.0, 0.1, 24.9) ], [ ?_test(begin valid_exercise(State0, DataRate0, 22, 6.9, X, 3, 0) end) || X <- gen_startend_range(-120.0, 0.1, 0.0) ], [ ?_test(begin valid_exercise(State0, DataRate0, 22, 7.0, X, 3, 1) end) || X <- gen_startend_range(-120.0, 0.1, 0.0) ] ], TestList. adr_history_test() -> State0 = new('US915'), ?assertEqual(0, device_offers_len(State0)), ?assertEqual(0, device_history_len(State0)), DataRateConfig = {10, 125}, Offer0 = #adr_offer{ packet_hash = <<0>> }, Packet0WeakSNR = #adr_packet{ packet_hash = <<0>>, wants_adr = true, wants_adr_ack = false, datarate_config = DataRateConfig, snr = 0.0, rssi = 0.0 }, Packet0StrongSNR = Packet0WeakSNR#adr_packet{snr = 10}, Offer1 = #adr_offer{ packet_hash = <<1>> }, Packet1 = Packet0StrongSNR#adr_packet{packet_hash = <<1>>}, %% Make sure offers are stored State1 = track_offer(State0, Offer0), ?assertEqual(1, device_offers_len(State1)), State2 = track_offer(State1, Offer0), ?assertEqual(2, device_offers_len(State2)), State3 = track_offer(State2, Offer1), ?assertEqual(3, device_offers_len(State3)), %% Tracking a packet should clear offers with matching hashes from %% the offer cache. %% %% NOTE: the third offer above has a different hash and will %% remain in the offer cache. {State4, hold} = track_packet(State3, Packet0WeakSNR), ?assertEqual(1, device_offers_len(State4)), ?assertEqual(1, device_history_len(State4)), [HistoryHead0 | _] = device_history(State4), ?assertEqual(Packet0WeakSNR#adr_packet.snr, HistoryHead0#packet.best_snr), ?assertEqual(3, HistoryHead0#packet.gateway_diversity), %% Tracking the same packet but with a stronger SNR should be %% reflected in history and increase gateway diversity. {State5, hold} = track_packet(State4, Packet0StrongSNR), ?assertEqual(1, device_history_len(State5)), [HistoryHead1 | _] = State5#device.packet_history, ?assertEqual(Packet0StrongSNR#adr_packet.snr, HistoryHead1#packet.best_snr), ?assertEqual(4, HistoryHead1#packet.gateway_diversity), %% An interleaved offer to packet sequence, however %% unlikely, should work. {State6, hold} = track_packet(State5, Packet1), ?assertEqual(2, device_history_len(State6)), [HistoryHead2 | _] = State6#device.packet_history, ?assertEqual(Packet1#adr_packet.snr, HistoryHead2#packet.best_snr), ?assertEqual(2, HistoryHead2#packet.gateway_diversity), ?assertEqual(0, device_offers_len(State6)), %% We clear history when the device clears the ADR bit. {State7, hold} = track_packet(State6, Packet1#adr_packet{wants_adr = false}), ?assertEqual(0, device_offers_len(State7)), ?assertEqual(0, device_history_len(State7)), fin. adr_happy_path(State0, DRConfig) -> Packet0 = #adr_packet{ rssi = 0, snr = 10, datarate_config = DRConfig, wants_adr = true, wants_adr_ack = false, packet_hash = <<0>> }, %% Up to MIN_HISTORY_LEN - 1 packets, track_packet should still %% return 'hold'. {State1, {AdjustedDataRate, AdjustedPowerIdx}} = lists:foldl( fun (N, {ADRn, _Action}) -> %% ?assertEqual(hold, Action), lorawan_adr:track_packet(ADRn, Packet0#adr_packet{ packet_hash = <<N>> }); (N, State2) -> io:format("State0 ~w~n", [N]), lorawan_adr:track_packet(State2, Packet0#adr_packet{ packet_hash = <<N>> }) end, State0, lists:seq(1, ?DEFAULT_ADR_HISTORY_LEN) ), {State1, {AdjustedDataRate, AdjustedPowerIdx}}. valid_happy_path(State0, DRConfig) -> {Spreading0, Bandwidth0} = DRConfig, {State1, {AdjustedDataRate, AdjustedPowerIdx}} = adr_happy_path(State0, DRConfig), {AdjustedSpread, AdjustedBandwidth} = spread_and_bandwidth(State1, AdjustedDataRate), ?assert(AdjustedDataRate >= 0), ?assert(AdjustedSpread =< Spreading0), ?assertEqual(AdjustedBandwidth, Bandwidth0), ?assertEqual(0, State1#device.max_txpower_idx), ?assert(AdjustedPowerIdx >= State1#device.max_txpower_idx), ?assert(AdjustedPowerIdx =< State1#device.min_txpower_idx). % io:format("AdjustedDataRate ~w~n", [AdjustedDataRate]), % io:format("AdjustedSpreading ~w~n", [AdjustedSpread]), % io:format("AdjustedBandwidth ~w~n", [AdjustedBandwidth]), % io:format("AdjustedPowerIdx ~w~n", [AdjustedPowerIdx]), % io:format("Min PowerIdx ~w~n", [State1#device.min_txpower_idx]), adr_happy_path_test_() -> [ ?_test(begin valid_happy_path(lorawan_adr:new('US915'), {10, 125}) end), ?_test(begin valid_happy_path(lorawan_adr:new('US915'), {9, 125}) end), ?_test(begin valid_happy_path(lorawan_adr:new('US915'), {8, 125}) end), ?_test(begin valid_happy_path(lorawan_adr:new('US915'), {7, 125}) end), ?_test(begin valid_happy_path(lorawan_adr:new('EU868'), {12, 125}) end), ?_test(begin valid_happy_path(lorawan_adr:new('EU868'), {11, 125}) end), ?_test(begin valid_happy_path(lorawan_adr:new('EU868'), {10, 125}) end), ?_test(begin valid_happy_path(lorawan_adr:new('EU868'), {9, 125}) end), ?_test(begin valid_happy_path(lorawan_adr:new('EU868'), {8, 125}) end), ?_test(begin valid_happy_path(lorawan_adr:new('EU868'), {7, 125}) end) ]. adr_ack_req_test() -> Packet0 = #adr_packet{ rssi = 0, snr = 0, wants_adr = true, wants_adr_ack = true, datarate_config = {7, 125}, packet_hash = <<0>> }, State0 = lorawan_adr:new('US915'), %% We must always respond with new uplink parameters when a device %% requests ADR acknowledgement, even on first packet. {State1, {_AdjustedDataRate, _AdjustedPowerIdx}} = lorawan_adr:track_packet( State0, Packet0 ), ?assertEqual(1, device_history_len(State1)), fin. %% TODO: parameterize and test over a range packet SNRs. adr_does_adr_test() -> %% DataRate 0 in US915 regional parameters. DataRate0 = 0, Spreading0 = 10, Bandwidth0 = 125, DataRateConfig0 = {Spreading0, Bandwidth0}, Packet0 = #adr_packet{ rssi = 0, snr = 6, datarate_config = DataRateConfig0, wants_adr = true, wants_adr_ack = false, packet_hash = <<0>> }, %% Up to MIN_HISTORY_LEN - 1 packets, track_packet should still %% return 'hold'. {State1, {AdjustedDataRate, AdjustedPowerIdx}} = lists:foldl( fun (N, {ADRn, Action}) -> ?assertEqual(hold, Action), lorawan_adr:track_packet(ADRn, Packet0#adr_packet{ packet_hash = <<N>> }); (N, State0) -> lorawan_adr:track_packet(State0, Packet0#adr_packet{ packet_hash = <<N>> }) end, lorawan_adr:new('US915'), lists:seq(1, ?DEFAULT_ADR_HISTORY_LEN) ), {AdjustedDataRate, {AdjustedSpreading, AdjustedBandwidth}} = lorawan_adr:datarate_entry( State1, AdjustedDataRate ), ?assertEqual(?DEFAULT_ADR_HISTORY_LEN, device_history_len(State1)), ?assert(AdjustedDataRate > DataRate0), ?assert(AdjustedSpreading < Spreading0), ?assertEqual(AdjustedBandwidth, Bandwidth0), ?assert(AdjustedPowerIdx >= State1#device.max_txpower_idx), ?assert(AdjustedPowerIdx =< State1#device.min_txpower_idx), %% Check that tracking a fake ADR answer from the device doesn't %% cause a crash. Answer0 = #adr_answer{ channel_mask_ack = true, datarate_ack = true, power_ack = true }, State2 = lorawan_adr:track_adr_answer(State1, Answer0), ?assertEqual([], State2#device.pending_adjustments), Answer1 = #adr_answer{ channel_mask_ack = false, datarate_ack = true, power_ack = true }, State3 = lorawan_adr:track_adr_answer(State1, Answer1), ?assertEqual([], State3#device.pending_adjustments), State4 = lorawan_adr:track_adr_answer(State2, Answer1), ?assertEqual([], State4#device.pending_adjustments), fin. adr_new_test() -> State0 = lorawan_adr:new('US915'), %% See regional parameters document for the reference values %% asserted against below. MinDataRate = lorawan_adr:min_datarate(State0), MinDataRateEntryFromIndex = lorawan_adr:datarate_entry(State0, MinDataRate), ?assertEqual({0, {10, 125}}, MinDataRateEntryFromIndex), MinDataRateEntryFromConfig = lorawan_adr:datarate_entry(State0, {10, 125}), ?assertEqual({0, {10, 125}}, MinDataRateEntryFromConfig), MaxDataRate = lorawan_adr:max_datarate(State0), MaxDataRateEntryFromIndex = lorawan_adr:datarate_entry(State0, MaxDataRate), ?assertEqual({3, {7, 125}}, MaxDataRateEntryFromIndex), MaxDataRateEntryFromConfig = lorawan_adr:datarate_entry(State0, {7, 125}), ?assertEqual({3, {7, 125}}, MaxDataRateEntryFromConfig), InvalidDataRateConfig = lorawan_adr:datarate_entry(State0, MaxDataRate + 1), ?assertEqual(undefined, InvalidDataRateConfig), fin. adr_resists_denial_of_service_test() -> State0 = lorawan_adr:new('US915'), Answer0 = #adr_answer{ channel_mask_ack = true, datarate_ack = true, power_ack = true }, %% This brand new state doesn't have any pending ADR %% requests. It should gracefully handle receiving a bogus ADR %% answer without crashing. State1 = lorawan_adr:track_adr_answer(State0, Answer0), ?assertEqual([], State1#device.accepted_adjustments), fin. -endif.
src/lora/lorawan_adr.erl
0.631708
0.619327
lorawan_adr.erl
starcoder
%%============================================================================== %% Copyright 2021 <NAME> <<EMAIL>> %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %%============================================================================== %%%------------------------------------------------------------------- %%% @doc %% Implements Prefix trees that allows you to find a value associated %% with the longest prefix of the key used. All keys are utf8 binaries. %%% @end %%% %% @author <NAME> <<EMAIL>> %% @copyright (C) 2021, <NAME> <<EMAIL>> %%%------------------------------------------------------------------- -module(pb_tree). -copyright('<NAME> <<EMAIL>>'). %% Library functions -export([new/0, is_pb_tree/1, is_empty/1, add/3, add/4, adds/2, adds/3, delete/2, delete/3, deletes/2, deletes/3, member/2, find/2, find/3, keys/1, values/1, replace/3, replace/4, to_list/1, from_list/1 ]). %% Records -record(bucket, {pivot :: value(), rest = [] :: {value(), value()}, values = [] :: [value()]}). -record(pb_node, {pivot :: value(), left = pb_nil :: pb_tree(), next = pb_nil :: pb_tree(), right = pb_nil :: pb_tree(), value :: value()}). %% Types -opaque pb_tree() :: #pb_node{} | pb_nil. -type key() :: [_]. -type value() :: _. -type default() :: _. -type flag() :: check | nocheck. %% Exported Types -export_type([pb_tree/0]). %% Defines %% =================================================================== %% Library functions. %% =================================================================== %%-------------------------------------------------------------------- %% Function: new() -> Tree. %% @doc %% Creates an empty P-tree. %% @end %%-------------------------------------------------------------------- -spec new() -> pb_tree(). %%-------------------------------------------------------------------- new() -> pb_nil. %%-------------------------------------------------------------------- %% Function: is_pb_tree(X) -> Boolean(). %% @doc %% Returns true if X is a P-tree, false otherwise. %% @end %%-------------------------------------------------------------------- -spec is_pb_tree(_) -> boolean(). %%-------------------------------------------------------------------- is_pb_tree(pb_nil) -> true; is_pb_tree(#pb_node{}) -> true; is_pb_tree(_) -> false. %%-------------------------------------------------------------------- %% Function: is_empty(Tree) -> Boolean. %% @doc %% Returns true if the Tree is empty, false otherwise. %% @end %%-------------------------------------------------------------------- -spec is_empty(_) -> boolean(). %%-------------------------------------------------------------------- is_empty(pb_nil) -> true; is_empty(_) -> false. %%-------------------------------------------------------------------- %% Function: add(Key, Value, Tree) -> Tree. %% @doc %% The Value is saved in Tree under the Key, if that key is present the %% value associated with the key is replaced by Value. %% add(Key, Value, Tree) is equivalent to add(Key, Value, Tree,nocheck). %% @end %%-------------------------------------------------------------------- -spec add(key(), value(), pb_tree()) -> pb_tree(). %%-------------------------------------------------------------------- add(Key, Value, Tree) -> add(Key, Value, Tree, nocheck). %%-------------------------------------------------------------------- %% Function: add(Prefix, Value, Tree, Flag) -> Tree. %% @doc %% The Value is saved in Tree under the Key, if that key is present the %% value associated with the key is replaced by Value. If the flag is %% check an exception is generated and if nocheck the value is replaced. %% @end %%-------------------------------------------------------------------- -spec add(key(), value(), pb_tree(), flag()) -> pb_tree(). %%-------------------------------------------------------------------- add(<<H/utf8>>, Value, pb_nil, _) -> #pb_node{pivot = H, value = Value}; add(<<H/utf8, T/binary>>, Value, pb_nil, _) -> #pb_node{pivot = H, next = add(T, Value, pb_nil, nocheck)}; add(<<H/utf8>>, Value, Tree = #pb_node{pivot = H}, nocheck) -> Tree#pb_node{value = Value}; add(<<H/utf8>>, Value, Tree = #pb_node{pivot = H, value = undefined}, _) -> Tree#pb_node{value = Value}; add(<<H/utf8, T/binary>>, Value, Tree = #pb_node{pivot = H, next =Next},Flag) -> Tree#pb_node{next = add(T, Value, Next, Flag)}; add(I = <<H/utf8,_/binary>>,V,Tree=#pb_node{pivot=P,left=Left},Flag) when H<P -> Tree#pb_node{left = add(I, V, Left, Flag)}; add(I, Value, Tree = #pb_node{right = Right}, Flag) -> Tree#pb_node{right = add(I, Value, Right, Flag)}. %%-------------------------------------------------------------------- %% Function: adds(Pairs, Tree) -> Tree. %% @doc< %% For each {Key, Value} pair in Pairs the value is stored under the %% key in the Tree. The adds(Pairs, Tree) call is equivalent to %% adds(Pairs, Tree, nocheck). %% @end %%-------------------------------------------------------------------- -spec adds([{key(), value()}], pb_tree()) -> pb_tree(). %%-------------------------------------------------------------------- adds(Pairs, Tree) -> adds(Pairs, Tree, nocheck). %%-------------------------------------------------------------------- %% Function: adds(Pairs, Tree, Flag) -> Tree. %% @doc %% For each {Key, Value} pair in Pairs the value is stored under the %% key in the Tree. If an key already has a value associated with %% in the Tree the flag determines what happens. If the flag is check %% an exception is generated if a key has a value associated with it, %% if the flag is nocheck the values will be replaced. %% @end %%-------------------------------------------------------------------- -spec adds([{key(), value()}], pb_tree(), flag()) -> pb_tree(). %%-------------------------------------------------------------------- adds(Pairs, Tree, nocheck) -> Pairs1 = lists:keymerge(1, lists:keysort(1, Pairs), to_list(Tree)), build(bucket_sort(Pairs1, [])); adds(Pairs, Tree, check) -> case lists:any(fun({I, _}) -> member(I, Tree) end, Pairs) of false -> adds(Pairs, Tree, nocheck); true -> erlang:error(badarg) end. %%-------------------------------------------------------------------- %% Function: delete(Key, Tree) -> Tree. %% @doc %% If a value is associated the Key the Tree returned has that %% association removed. The call delete(Key, Tree) is equivalent %% to delete(Key, Tree, nocheck). %% @end %%-------------------------------------------------------------------- -spec delete(key(), pb_tree()) -> pb_tree(). %%-------------------------------------------------------------------- delete(Key, Tree) -> delete(Key, Tree, nocheck). %%-------------------------------------------------------------------- %% Function: delete(Key, Tree, Flag) -> Tree. %% @doc %% If a value is associated the Key the Tree returned has that %% association removed. If there is no value associated with the Key %% in the Tree the flag determines what happens. If the flag is check %% an exception is generated if no association exists, if the flag %% is nocheck the unchanged tree is returned. %% @end %%-------------------------------------------------------------------- -spec delete(key(), pb_tree(), flag()) -> pb_tree(). %%-------------------------------------------------------------------- delete(_, pb_nil, check) -> erlang:error(badarg); delete(_, pb_nil, nocheck) -> pb_nil; delete(<<H/utf8>>, #pb_node{pivot = H, value = undefined}, check) -> erlang:error(badarg); delete(<<H/utf8>>, #pb_node{pivot =H,left=pb_nil,next=pb_nil,right=pb_nil},_) -> pb_nil; delete(<<H/utf8>>, #pb_node{pivot = H, left=Left,next=pb_nil,right=pb_nil},_) -> Left; delete(<<H/utf8>>, #pb_node{pivot = H,left=pb_nil,next=pb_nil,right=Right},_) -> Right; delete(<<H/utf8>>, Tree = #pb_node{pivot = H}, _) -> Tree#pb_node{value = undefined}; delete(<<H/utf8, T/binary>>, #pb_node{pivot = H, next = Next}, Flag) -> delete(T, Next, Flag); delete(I = <<H/utf8, _/binary>>, #pb_node{pivot=P,left=Left},Flag) when H < P -> delete(I, Left, Flag); delete(I, #pb_node{right = Right}, Flag) -> delete(I, Right, Flag). %%-------------------------------------------------------------------- %% Function: deletes(Keys, Tree) -> Tree. %% @doc %% A tree that has all the associations for the keys removed. %% The call deletes(Indces, Tree) is equivalent to %% deletes(Keys, Tree, nocheck). %% @end %%-------------------------------------------------------------------- -spec deletes([key()], pb_tree()) -> pb_tree(). %%-------------------------------------------------------------------- deletes(Keys, Tree) -> deletes(Keys, Tree, nocheck). %%-------------------------------------------------------------------- %% Function: deletes(Key, Tree, Flag) -> Tree. %% @doc %% A tree that has all the associations for the keys removed. %% If there is no value associated with any of the Keys %% in the Tree, the flag determines what happens. If the flag is check %% an exception is generated if, if the flag is nocheck a tree %% is returned with the other associations removed. %% @end %%-------------------------------------------------------------------- -spec deletes([key()], pb_tree(), flag()) -> pb_tree(). %%-------------------------------------------------------------------- deletes(Keys, Tree, nocheck) -> build(bucket_sort(deletes1(lists:sort(Keys), to_list(Tree)), [])); deletes(Keys, Tree, check) -> case lists:all(fun(I) -> member(I, Tree) end, Keys) of true -> deletes(Keys, Tree, nocheck); false -> erlang:error(badarg) end. deletes1(_, []) -> []; deletes1([], L) -> L; deletes1([H | T1], [{H, _} | T2]) -> deletes1(T1, T2); deletes1([H1 | T1], L = [{H2, _} | _]) when H1 > H2 -> deletes1(T1, L); deletes1(I, [_ | T]) -> deletes1(I, T). %%-------------------------------------------------------------------- %% Function: member(Key, Tree) -> Boolean. %% @doc %% Returns true if there is a value associated with Key in the tree, %% otherwise false. %% @end %%-------------------------------------------------------------------- -spec member(key(), pb_tree()) -> boolean(). %%-------------------------------------------------------------------- member(_, pb_nil) -> false; member(<<H/utf8>>, #pb_node{pivot = H, value = Value}) -> Value /= undefined; member(<<H/utf8, T/binary>>, #pb_node{pivot = H, next=Next}) -> member(T, Next); member(I = <<H/utf8, _/binary>>, #pb_node{pivot = P, left = L}) when H < P -> member(I, L); member(I, #pb_node{right = R}) -> member(I, R). %%-------------------------------------------------------------------- %% Function: find(Key, Tree) -> Value. %% @doc %% Returns the value associated with the longest prefix of the Key in the %% Tree or undefined if no such association exists. %% The call find(Key, Tree) is equivalent to find(Key, Tree, undefined). %% @end %%-------------------------------------------------------------------- -spec find(key(), pb_tree()) -> value() | undefined. %%-------------------------------------------------------------------- find(Key, Tree) -> find(Key, Tree, undefined). %%-------------------------------------------------------------------- %% Function: find(Key, Tree, Default) -> Value. %% @doc %% Returns the value associated with the longest prefix of the Key in the %% Tree or Default if no such association exists. %% @end %%-------------------------------------------------------------------- -spec find(key(), pb_tree(), default()) -> value() | default(). %%-------------------------------------------------------------------- find(_, pb_nil, Prev) -> Prev; find(<<H/utf8>>, #pb_node{pivot = H, value = V}, Prev) -> update(V, Prev); find(<<H/utf8, T/binary>>, #pb_node{pivot = H, next = Next, value = V},Prev) -> find(T, Next, update(V, Prev)); find(I = <<H/utf8, _/binary>>, #pb_node{pivot = P,left=Left},Prev) when H < P -> find(I, Left, Prev); find(I, #pb_node{right = Right}, Prev) -> find(I, Right, Prev). update(undefined, Prev) -> Prev; update(New, _) -> New. %%-------------------------------------------------------------------- %% Function: keys(Tree) -> Keys. %% @doc %% Returns all the keys in ascending order. %% @end %%-------------------------------------------------------------------- -spec keys(pb_tree()) -> [key()]. %%-------------------------------------------------------------------- keys(Tree) -> keys(Tree, [], []). keys(pb_nil, _, Acc) -> Acc; keys(#pb_node{pivot = P, left=L,next=N,right=R,value=undefined},Xiferp,Acc) -> Xiferp1 = [P | Xiferp], keys(L, Xiferp, keys(N, Xiferp1, keys(R, Xiferp, Acc))); keys(#pb_node{pivot = P, left=L, next=N, right=R}, Xiferp, Acc) -> Xiferp1 = [P | Xiferp], keys(L, Xiferp, [<< <<X/utf8>> || X <- lists:reverse(Xiferp1)>> | keys(N, Xiferp1, keys(R, Xiferp, Acc))]). %%-------------------------------------------------------------------- %% Function: values(Tree) -> Values. %% @doc %% Returns all the values in ascending order of their keys. %% @end %%-------------------------------------------------------------------- -spec values(pb_tree()) -> [key()]. %%-------------------------------------------------------------------- values(Tree) -> values(Tree, []). values(pb_nil, Acc) -> Acc; values(#pb_node{left = L, next = N, right = R, value = undefined}, Acc) -> values(L, values(N, values(R, Acc))); values(#pb_node{left = L, next = N, right = R, value = V}, Acc) -> values(L, [V | values(N, values(R, Acc))]). %%-------------------------------------------------------------------- %% Function: replace(Key, Value, Tree) -> Tree. %% @doc %% Replaces any existing value associated with Key in the tree, %% otherwise adds a association for the value with the Key. %% The call replace(Key, Value, Tree) is equivalent to %% replace(Key, Value, Tree, nocheck). %% @end %%-------------------------------------------------------------------- -spec replace(key(), value(), pb_tree()) -> pb_tree(). %%-------------------------------------------------------------------- replace(Key, Value, Tree) -> replace(Key, Value, Tree, nocheck). %%-------------------------------------------------------------------- %% Function: replace(Key, Values, Tree, Flag) -> Tree. %% @doc %% Replaces any existing value associated with Key in the tree, %% otherwise the flag determines what happens. If the flag is check %% an exception is generated, otherwise the value is added. %% @end %%-------------------------------------------------------------------- -spec replace(key(), value(), pb_tree(), flag()) -> pb_tree(). %%-------------------------------------------------------------------- replace(Key, Value, Tree, nocheck) -> add(Key, Value, Tree, nocheck); replace(Key, Value, Tree, check) -> replace_check(Key, Value, Tree). replace_check(_, _, pb_nil) -> erlang:error(badarg); replace_check(<<H/utf8>>, _, #pb_node{pivot = H, value = undefined}) -> erlang:error(badarg); replace_check(<<H/utf8>>, Value, Tree = #pb_node{pivot = H}) -> Tree#pb_node{value = Value}; replace_check(<<H/utf8, T/binary>>, Value, Tree =#pb_node{pivot=H,next=Next}) -> Tree#pb_node{next = replace_check(T, Value, Next)}; replace_check(I = <<H/utf8, _/binary>>, Value, Tree=#pb_node{pivot=P,left=Left}) when H < P -> Tree#pb_node{left = replace_check(I, Value, Left)}; replace_check(I, Value, Tree = #pb_node{right = Right}) -> Tree#pb_node{right = replace_check(I, Value, Right)}. %%-------------------------------------------------------------------- %% Function: to_list(Tree) -> Pairs. %% @doc %% From a P-tree a list of {Key, Value} pairs. %% @end %%-------------------------------------------------------------------- -spec to_list(pb_tree()) -> [{key(), value()}]. %%-------------------------------------------------------------------- to_list(Tree) -> to_list(Tree, [], []). to_list(pb_nil, _, Acc) -> Acc; to_list(#pb_node{pivot = P,left=L,next=N,right=R,value=undefined},Xiferp,Acc) -> Xiferp1 = [P | Xiferp], to_list(L, Xiferp, to_list(N, Xiferp1, to_list(R, Xiferp, Acc))); to_list(#pb_node{pivot = P, left=L, next=N, right=R, value=V}, Xiferp, Acc) -> Xiferp1 = [P | Xiferp], to_list(L, Xiferp, [{<< <<X/utf8>> || X <- lists:reverse(Xiferp1)>>, V} | to_list(N, Xiferp1, to_list(R, Xiferp, Acc))]). %%-------------------------------------------------------------------- %% Function: from_list(Pairs) -> Tree. %% @doc %% For each {Key, Value} pair in Pairs the value is stored under the %% key in the Tree. %% @end %%-------------------------------------------------------------------- -spec from_list([{key(), value()}]) -> pb_tree(). %%-------------------------------------------------------------------- from_list(Plist) -> build(bucket_sort(Plist)). %% =================================================================== %% Internal functions. %% =================================================================== bucket_sort(L) -> bucket_sort(lists:sort(L), []). bucket_sort([], []) -> []; bucket_sort([], [B = #bucket{rest = R} | Acc]) -> lists:reverse([B#bucket{rest = bucket_sort(lists:reverse(R))} | Acc]); bucket_sort([{<<H/utf8>>, V} | T], []) -> bucket_sort(T, [#bucket{pivot = H, values = [V]}]); bucket_sort([{<<H/utf8, P/binary>>, V} | T], []) -> bucket_sort(T, [#bucket{pivot = H, rest = [{P, V}]}]); bucket_sort([{<<H/utf8>>, V} | T], [B = #bucket{pivot = H, values = Vs}|Acc]) -> bucket_sort(T, [B#bucket{values = [V | Vs]} | Acc]); bucket_sort([{<<H/utf8, P/binary>>, V} |T],[B=#bucket{pivot=H,rest=T1} |Acc]) -> bucket_sort(T, [B#bucket{rest = [{P, V} | T1]} | Acc]); bucket_sort([{<<H/utf8>>, V} | T], [B = #bucket{rest = R} | Acc]) -> bucket_sort(T, [#bucket{pivot = H, values = [V]}, B#bucket{rest = bucket_sort(lists:reverse(R))} | Acc]); bucket_sort([{<<H/utf8, P/binary>>, V} | T], [B = #bucket{rest = R} | Acc]) -> bucket_sort(T, [#bucket{pivot = H, rest = [{P, V}]}, B#bucket{rest = bucket_sort(lists:reverse(R))} | Acc]). build([]) -> pb_nil; build([#bucket{pivot = P, rest = R, values = [V]}]) -> #pb_node{pivot = P, next = build(R), value = V}; build([#bucket{pivot = P1, rest = R1, values = [V1]}, #bucket{pivot = P2, rest = R2, values = [V2]}]) -> #pb_node{pivot = P1, next = build(R1), value = V1, right = #pb_node{pivot = P2, next = build(R2), value = V2}}; build(Bs) -> Size = length(Bs), case split(Size - 1 - ((Size - 1) div 2), Bs, []) of {Left, #bucket{pivot = P, rest = Rest, values = []}, Right} -> #pb_node{pivot = P, next = build(Rest), left = build(Left), right = build(Right)}; {Left, #bucket{pivot = P, rest = Rest, values = [V]}, Right} -> #pb_node{pivot = P, next = build(Rest), left = build(Left), right = build(Right), value = V} end. split(0, [H | T], Acc) -> {lists:reverse(Acc), H, T}; split(N, [H | T], Acc) -> split(N - 1, T, [H | Acc]).
src/pb_tree.erl
0.63443
0.433622
pb_tree.erl
starcoder
%%%------------------------------------------------------------------- %%% @author <NAME> %%% @copyright (C) 2021 ACK CYFRONET AGH %%% This software is released under the MIT license %%% cited in 'LICENSE.txt'. %%% @end %%%------------------------------------------------------------------- %%% @doc %%% Helper module to ts_persistence operating on time series %%% metric data node. Each time series metric data node is connected with %%% singe metric. Values that exceed time series hub capacity for particular %%% metric (ie. metric's head capacity) are stored in list of time series metric %%% data nodes (capacity of single time series metric data node is also limited so more %%% than one time series metric data node may be needed - see ts_persistence module). %%% @end %%%------------------------------------------------------------------- -module(ts_metric_data_node). -author("<NAME>"). %% API -export([set/1, get/1]). %% datastore_model callbacks -export([get_ctx/0, get_record_struct/1]). -record(ts_metric_data_node, { value :: ts_metric:data_node() }). -type record() :: #ts_metric_data_node{}. -type key() :: datastore:key(). -export_type([key/0]). % Context used only by datastore to initialize internal structures. % Context provided via time_series_collection module functions % overrides it in other cases. -define(CTX, #{ model => ?MODULE, memory_driver => undefined, disc_driver => undefined }). %%%=================================================================== %%% API %%%=================================================================== -spec set(ts_metric:data_node()) -> record(). set(Data) -> #ts_metric_data_node{value = Data}. -spec get(record()) -> ts_metric:data_node(). get(#ts_metric_data_node{value = Data}) -> Data. %%%=================================================================== %%% datastore_model callbacks %%%=================================================================== %%-------------------------------------------------------------------- %% @doc %% Returns model's context needed to initialize internal structure's %% (it is not used to get or save document). %% @end %%-------------------------------------------------------------------- -spec get_ctx() -> datastore:ctx(). get_ctx() -> ?CTX. -spec get_record_struct(datastore_model:record_version()) -> datastore_model:record_struct(). get_record_struct(1) -> {record, [ {value, {record, [ {windows, {custom, json, {ts_windows, encode, decode}}}, {older_node_key, string}, {older_node_timestamp, integer} ]}} ]}.
src/modules/datastore/time_series/persistence/ts_metric_data_node.erl
0.621656
0.437523
ts_metric_data_node.erl
starcoder
% @copyright 2010-2014 Zuse Institute Berlin % Licensed under the Apache License, Version 2.0 (the "License"); % you may not use this file except in compliance with the License. % You may obtain a copy of the License at % % http://www.apache.org/licenses/LICENSE-2.0 % % Unless required by applicable law or agreed to in writing, software % distributed under the License is distributed on an "AS IS" BASIS, % WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. % See the License for the specific language governing permissions and % limitations under the License. %% @author <NAME> <<EMAIL>> %% @doc Integrationstests for the gossip and gossip_load modules. %% @end %% @version $Id$ -module(gossip_SUITE). -author('<EMAIL>'). -vsn('$Id$'). -compile(export_all). -include("unittest.hrl"). -include("scalaris.hrl"). -dialyzer({no_fail_call, test_request_histogram1/1}). -define(NO_OF_NODES, 5). all() -> [ test_no_load, test_load, test_request_histogram1, test_request_histogram2 ]. suite() -> [ {timetrap, {seconds, 80}} ]. init_per_suite(Config) -> Config. end_per_suite(_Config) -> ok. init_per_group(Group, Config) -> unittest_helper:init_per_group(Group, Config). end_per_group(Group, Config) -> unittest_helper:end_per_group(Group, Config). init_per_testcase(_TestCase, Config) -> unittest_helper:make_ring_with_ids( [?MINUS_INFINITY], [{config, [ {monitor_perf_interval, 0}, % deactivate monitor_perf {gossip_load_interval, 100}, % truncated to 0, i.e. immediate delivery {gossip_load_convergence_count_new_round, 5}, {gossip_load_convergence_count_best_values, 1}, {gossip_log_level_warn, warn}, {gossip_log_level_error, error} ] }]), {_, []} = api_vm:add_nodes(?NO_OF_NODES - 1), unittest_helper:wait_for_stable_ring_deep(), [{stop_ring, true} | Config]. end_per_testcase(_TestCase, _Config) -> ok. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Testcases %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% test_no_load(_Config) -> % get values from gossiping (after round finishes) wait_n_rounds(1), send2gossip({cb_msg, {gossip_load, default}, {gossip_get_values_all, self()}}, 1), % {load_info, avg, stddev, size_ldr, size_kr, minLoad, maxLoad, merged} LoadInfoExpected ={load_info, 0.0, 0.0, 5.0, 5.0, 0, 0, whatever, []}, receive {gossip_get_values_all_response, {PrevLoadInfo, _, _}} -> ?compare(fun compare/2, PrevLoadInfo, LoadInfoExpected) end. test_load(_Config) -> write(100), % request load from all nodes and calc expected values manually Loads = lists:map(fun (State) -> dht_node_state:get(State, load) end, get_node_states()), Avg = lists:sum(Loads)/?NO_OF_NODES, Stddev = calc_stddev(Loads), Size = ?NO_OF_NODES, Min = lists:min(Loads), Max = lists:max(Loads), % {load_info, avg, stddev, size_ldr, size_kr, minLoad, maxLoad, merged, other} LoadInfoExpected = {load_info, Avg, Stddev, Size, Size, Min, Max, dc, []}, % get values from gossiping (after round finishes) % first round might be interrupted by node joins, thus wait two rounds wait_n_rounds(1), send2gossip({cb_msg, {gossip_load, default}, {gossip_get_values_all, self()}}, 1), receive {gossip_get_values_all_response, {PrevLoadInfo, _, _}} -> ?compare(fun compare/2, PrevLoadInfo, LoadInfoExpected) end. test_request_histogram1(_Config) -> ?expect_exception(gossip_load:request_histogram(0, comm:this()), error, function_clause). test_request_histogram2(_Config) -> write(100), NoOfBuckets = 10, % get the states from all nodes and calc a histogram manually DHTStates = get_node_states(), Histos = lists:map(fun(State) -> init_histo(State, NoOfBuckets) end, DHTStates), MergeFun = fun(Histo, Acc) -> Combine = fun({Load1, N1}, {Load2, N2}) -> {Load1+Load2, N1+N2}; (unknown, {Load, N}) -> {Load, N}; ({Load, N}, unknown) -> {Load, N} end, lists:zipwith(Combine, Histo, Acc) end, InitialAcc = [ {0,0} || _X <- lists:seq(1, NoOfBuckets)], MergedHisto = lists:foldl(MergeFun, InitialAcc, Histos), % calculate the histogram from Sums and Numbers of Nodes Histo = lists:map(fun({Sum, N}) -> Sum/N end, MergedHisto), %% log:log("Histo: ~w", [Histo]), gossip_load:request_histogram(NoOfBuckets, comm:this()), %% GossipedHisto = receive {histogram, Histo} -> Histo end, % doesn't work, don't know why! GossipedHisto = receive Msg -> element(2, Msg) end, % remove the intervals from gossiped histo: GossipedHisto1 = lists:map(fun({_Interval, Value}) -> Value end, GossipedHisto), % calc the estimates GossipedHisto2 = lists:map(fun({Value, Weight}) -> Value/Weight end, GossipedHisto1), %% log:log("GossipedHisto: ~w", [GossipedHisto2]), ?compare(fun compare/2, GossipedHisto2, Histo). %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% Helper %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% @doc Sends a Msg to the gossip process -spec send2gossip(Msg::comm:message(), Delay::non_neg_integer()) -> ok. send2gossip(Msg, Delay) -> Group = pid_groups:group_with(gossip), Pid = pid_groups:pid_of(Group, gossip), msg_delay:send_local(Delay, Pid, Msg). %% @doc Waits n rounds, pulling the web_debug_info every second. -spec wait_n_rounds(NoOfRounds::pos_integer()) -> ok. wait_n_rounds(NoOfRounds) -> Round = get_current_round(), wait_for_round(Round+NoOfRounds). %% @doc Helper for wait_n_round/0 -spec wait_for_round(TargetRound::pos_integer()) -> ok. wait_for_round(TargetRound) -> Round = get_current_round(), %% log:log("CurrentRound: ~w, TargetRound: ~w", [Round, TargetRound]), if Round >= TargetRound -> ok; Round =/= TargetRound -> wait_for_round(TargetRound) end. %% @doc Get the current round (with one second delay). -spec get_current_round() -> non_neg_integer(). get_current_round() -> send2gossip({web_debug_info, self()}, 1), receive {web_debug_info_reply, KeyValueList} -> case lists:keyfind("cur_round", 1, KeyValueList) of {"cur_round", Round} -> Round; false -> get_current_round() % happens if cb module not yet initiated end end. %% @doc Writes the given number of dummy entries to scalaris. -spec write(NoOfEntries::pos_integer()) -> ok. write(NoOfEntries) -> _CommitLog = [ begin K = io_lib:format("k~w", [N]), V = io_lib:format("v~w", [N]), api_tx:write(K,V) end || N <- lists:seq(1,NoOfEntries) ], %% log:log("CommitLog: ~w", [_CommitLog]), ok. %% @doc Get the states from all dht nodes. -spec get_node_states() -> [dht_node_state:state(), ...]. get_node_states() -> % request load from all nodes and calc expected values manually [gen_component:get_state(Pid) || Pid <- pid_groups:find_all(dht_node)]. %% @doc Builds a initial histogram, i.e. a histogram for a single node. -spec init_histo(DHTNodeState::dht_node_state:state(), NoOfBuckets::pos_integer()) -> [{non_neg_integer(), 1}, ...]. init_histo(DHTNodeState, NoOfBuckets) -> DB = dht_node_state:get(DHTNodeState, db), MyRange = dht_node_state:get(DHTNodeState, my_range), Buckets = intervals:split(intervals:all(), NoOfBuckets), _Histo = [ get_load_for_interval(BucketInterval, MyRange, DB) || BucketInterval <- Buckets ], %% log:log("Histo ~w", [_Histo]), _Histo. %% @doc Gets the load for a given interval. -spec get_load_for_interval(BucketInterval::intervals:interval(), MyRange::intervals:interval(), DB::db_dht:db()) -> {non_neg_integer(), 1}. get_load_for_interval(BucketInterval, MyRange, DB) -> case intervals:is_empty(intervals:intersection(BucketInterval, MyRange)) of true -> unknown; false -> Load = db_dht:get_load(DB, BucketInterval), {Load, 1} end. %% @doc Compares LoadInfo records and Histo. Returns true if the difference is less than 5 %. -spec compare(LoadInfo1::gossip_load:load_info(), LoadInfo1::gossip_load:load_info()) -> boolean(); (Histo1::[float(),...], Histo2::[float(),...]) -> boolean(). compare(LoadInfo1, LoadInfo2) when is_tuple(LoadInfo1) andalso is_tuple(LoadInfo2) -> %% log:log("LoadInfo1: ~w~n", [LoadInfo1]), %% log:log("LoadInfo2: ~w~n", [LoadInfo2]), Fun = fun(Key, {Acc, LI1, LI2}) -> Value1 = gossip_load:load_info_get(Key, LI1), Value2 = gossip_load:load_info_get(Key, LI2), {Acc andalso calc_diff(Value1, Value2) < 5.0, LI1, LI2} end, % merged counter is excluded from comparison element(1, lists:foldl(Fun, {true, LoadInfo1, LoadInfo2}, [avgLoad, stddev, size_ldr, size_kr, minLoad, maxLoad])); compare(Histo1, Histo2) when is_list(Histo1) andalso is_list(Histo2) -> Fun = fun(Val1, Val2) -> calc_diff(Val1, Val2) < 5.0 end, ComparisonResult = lists:zipwith(Fun, Histo1, Histo2), lists:all(fun(X) -> X end, ComparisonResult). %% @doc Calculates the difference in percent from one value to another value. -spec calc_diff(Value1::float()|unknown, Value2::float()|unknown) -> float(). calc_diff(Value1, Value2) -> if (Value1 =/= unknown) andalso (Value1 =:= Value2) -> 0.0; (Value1 =:= unknown) orelse (Value2 =:= unknown) orelse (Value1 == 0) -> 100.0; true -> ((Value1 + abs(Value2 - Value1)) * 100.0 / Value1) - 100 end. %% @doc Calculate the standard deviation for a given list of numbers. -spec calc_stddev(Loads::[number()]) -> float(). calc_stddev(Loads) -> Avg = lists:sum(Loads)/?NO_OF_NODES, LoadsSquared = lists:map(fun (Load) -> math:pow(Avg-Load, 2) end, Loads), _Stddev = math:sqrt(lists:sum(LoadsSquared)/?NO_OF_NODES).
test/gossip_SUITE.erl
0.506836
0.406626
gossip_SUITE.erl
starcoder
%%% @doc %%% A utility to create small memory buffers for an application without needing to allocate %%% one big memory buffer. For example, maybe you need to reference data by bytes at random %%% locations but you don't know which locations use. Or, you're processing %%% chunks that may live at different byte positions. Rather than trying to create 1 large %%% buffer to handle this, you can use memory_page to allocate only the buffers needed. %%% %%% Say you have data at bit positions: 10, 1048, and will have others at higher ranges. You %%% could use this: %%% ``` %%% 1 big buffer %%% [ | | | | | | ... ] %%% 0 4k %%% ''' %%% Even though you may not need it all. Or you could do this: Use Smaller buffers at %%% differnt offsets on demand: %%% %%% ``` %%% [0] -> [ | | ...] %%% 0 1023 %%% ''' %%% ``` %%% [1] -> [ | | ...] %%% 1024 2048 %%% ''' %%% %%% Memory Pager does the latter. %%% @end -module(memory_pager). -export([ new/0, new/1, get/2, set/3, num_of_pages/1, pagesize_in_bytes/1, pagenum_for_byte_index/2, collect/1, truncate_buffer/3 ]). %% Default page size -define(PAGE_SIZE, 1024). %% Initial capacity of the page list -define(NUM_PAGES, 32768). -type state() :: {Pages :: [page()], PageSize :: pos_integer()}. -type page() :: {Offset :: pos_integer(), Buffer :: binary()}. %% @doc Create a new memory pager with a default page size of 1024 bytes -spec new() -> state(). new() -> new(?PAGE_SIZE). %% @doc Create a new memory pager with the given page size. Note: %% 'PageSize' must be a power of 2 in 'bytes' or an error occurs. -spec new(PageSize :: pos_integer()) -> state() | erlang:throw({badarg, not_power_of_two}). new(PageSize) -> case power_of_two(PageSize) of true -> { array:new([{size, ?NUM_PAGES}, {fixed, false}, {default, nil}]), PageSize }; _ -> erlang:error({badarg, not_power_of_two}) end. %% @doc Get a page by page num. -spec get(PageNum :: pos_integer(), State :: state()) -> {none, state()} | {ok, Page :: page(), state()}. get(PageNum, {Pages, _} = State) -> case array:get(PageNum, Pages) of nil -> {none, State}; Page -> {ok, Page, State} end. %% @doc Set a buffer at the given page number. If the incoming buffer %% is not the same as the configured Page size, it will be truncated to fit the Page size. -spec set( PageNum :: pos_integer(), Buffer :: binary(), State :: state() ) -> {ok, state()}. set(PageNum, Buffer, {Pages, PageSize}) -> BufSize = byte_size(Buffer), Buf = truncate_buffer(Buffer, BufSize, PageSize), %% Calculate the offset of the buffer in relation to the pagenum Offset = PageNum * PageSize, UpdatedPages = array:set( PageNum, {Offset, Buf}, Pages ), {ok, {UpdatedPages, PageSize}}. %% @doc Return the number of pages -spec num_of_pages(State :: state()) -> pos_integer(). num_of_pages({Pages, _}) -> array:sparse_size(Pages). %% @doc return the page number for the given byte index -spec pagenum_for_byte_index( Index :: pos_integer(), State :: state() ) -> pos_integer(). pagenum_for_byte_index(Index, {_, PageSize}) -> Index div PageSize. %% @doc Return the page size. This is set on creation -spec pagesize_in_bytes(State :: state()) -> pos_integer(). pagesize_in_bytes({_, PageSize}) -> PageSize. %% @doc Return a list of all valid pages. Each entry in the list %% contains {PageNum, Page}. -spec collect( State :: state() ) -> [{PageNum :: pos_integer(), Page :: page()}]. collect({Pages, _}) -> array:sparse_to_orddict(Pages). %% @doc Determine if 'Value' is a power of 2 -spec power_of_two(Value :: pos_integer()) -> boolean(). power_of_two(Value) -> case (Value band (Value - 1)) of 0 -> true; _ -> false end. %% @private %% Deal with incoming buffers that may not have the comfigured page size. %% - If the incoming buffer is the page size, do nothing and return the buffer %% - If the incoming buffer is larger than the page size, Copy page size bytes %% to a new buffer, return it, and disgard the rest. %% - If the incoming buffer is smaller than the page size, append '0s' to the %% buffer to make it the correct page size. truncate_buffer(Buffer, BufferSize, PageSize) when BufferSize =:= PageSize -> %% Same size, ok. Buffer; truncate_buffer(Buffer, BufferSize, PageSize) when BufferSize > PageSize -> %% Buffer is bigger <<A:PageSize/binary, _/binary>> = Buffer, A; truncate_buffer(Buffer, _, PageSize) -> %% Buffer is smaller append '0s'. Diff = PageSize - byte_size(Buffer), Zeros = <<0:Diff/unit:8>>, <<Buffer/binary, Zeros/binary>>.
src/memory_pager.erl
0.564939
0.732197
memory_pager.erl
starcoder
%%%------------------------------------------------------------------- %%% @doc %%% A set of optics specific to tuples. %%% @end %%%------------------------------------------------------------------- -module(optic_tuples). %% API -export([all/0, all/1, element/1, element/2, field/3, field/4]). %%%=================================================================== %%% API %%%=================================================================== %% @see all/1 -spec all() -> optic:optic(). all() -> all(#{}). %% @doc %% Focus on all elements of a tuple. %% %% Example: %% %% ``` %% > optic:get([optic_tuples:all()], {1,2,3}). %% {ok,[1,2,3]} %% ''' %% @end %% @param Options Common optic options. %% @returns An opaque optic record. -spec all(Options) -> optic:optic() when Options :: optic:variations(). all(Options) -> Fold = fun (Fun, Acc, Tuple) when is_tuple(Tuple) -> {ok, lists:foldl(Fun, Acc, tuple_to_list(Tuple))}; (_, _, _) -> {error, undefined} end, MapFold = fun (Fun, Acc, Tuple) when is_tuple(Tuple) -> {NewList, NewAcc} = lists:mapfoldl(Fun, Acc, tuple_to_list(Tuple)), {ok, {list_to_tuple(NewList), NewAcc}}; (_, _, _) -> {error, undefined} end, New = fun (_Data, _Template) -> {} end, Optic = optic:new(MapFold, Fold), optic:variations(Optic, Options, New). %% @see element/2 -spec element(N) -> optic:optic() when N :: pos_integer(). element(N) -> optic_tuples:element(N, #{}). %% @doc %% Focus on the nth element of a tuple. As with `erlang:element/2', %% indexing begins at 1. %% %% Example: %% %% ``` %% > optic:get([optic_tuples:element(1)], {1,2,3}). %% {ok,[1]} %% ''' %% @end %% @param N The index of the tuple element to focus on. %% @param Options Common optic options. %% @returns An opaque optic record. -spec element(N, Options) -> optic:optic() when N :: pos_integer(), Options :: optic:variations(). element(N, Options) -> Fold = fun (Fun, Acc, Tuple) when N =< tuple_size(Tuple) -> Nth = erlang:element(N, Tuple), {ok, Fun(Nth, Acc)}; (_, _, _) -> {error, undefined} end, MapFold = fun (Fun, Acc, Tuple) when N =< tuple_size(Tuple) -> Nth = erlang:element(N, Tuple), {NewNth, NewAcc} = Fun(Nth, Acc), {ok, {erlang:setelement(N, Tuple, NewNth), NewAcc}}; (_, _, _) -> {error, undefined} end, New = fun (Tuple, Template) when is_tuple(Tuple) -> list_to_tuple(tuple_to_list(Tuple) ++ lists:duplicate(N - tuple_size(Tuple), Template)); (_Data, Template) -> list_to_tuple(lists:duplicate(N, Template)) end, Optic = optic:new(MapFold, Fold), optic:variations(Optic, Options, New). %% @see field/4 -spec field(Tag, Size, N) -> optic:optic() when Tag :: atom(), Size :: pos_integer(), N :: pos_integer(). field(Tag, Size, N) -> field(Tag, Size, N, #{}). %% @doc %% Focus on a record field. As records are a compiler construct, this %% depends on the `?OPTIC_FIELD' macro in `include/optic_tuples.hrl' %% to construct the required arguments from the record definition. %% %% Given the record definition: %% %% ``` %% -include_lib("optic/include/optic_tuples.hrl"). %% -record(example, {first}). %% ''' %% %% Example: %% %% ``` %% > optic:get([optic_tuples:field(?OPTIC_FIELD(example, first))], %% #example{first=1}). %% {ok,[1]} %% ''' %% @end %% @param Tag The expected record tag. %% @param Size The expected record size. %% @param N The index of the field in the record tuple. %% @param Options Common optic options. %% @returns An opaque optic record. -spec field(Tag, Size, N, Options) -> optic:optic() when Tag :: atom(), Size :: pos_integer(), N :: pos_integer(), Options :: optic:variations(). field(Tag, Size, N, Options) -> Fold = fun (Fun, Acc, Tuple) when erlang:element(1, Tuple) == Tag, Size == tuple_size(Tuple), N > 1, N =< tuple_size(Tuple) -> Nth = erlang:element(N, Tuple), {ok, Fun(Nth, Acc)}; (_, _, _) -> {error, undefined} end, MapFold = fun (Fun, Acc, Tuple) when erlang:element(1, Tuple) == Tag, Size == tuple_size(Tuple), N > 1, N =< tuple_size(Tuple) -> Nth = erlang:element(N, Tuple), {NewNth, NewAcc} = Fun(Nth, Acc), {ok, {erlang:setelement(N, Tuple, NewNth), NewAcc}}; (_, _, _) -> {error, undefined} end, New = fun (_Data, Template) -> list_to_tuple([Tag] ++ lists:duplicate(Size - 1, Template)) end, Optic = optic:new(MapFold, Fold), optic:variations(Optic, Options, New).
src/optic_tuples.erl
0.635336
0.533094
optic_tuples.erl
starcoder
%% Copyright (c) 2021 <NAME> %% %% Permission is hereby granted, free of charge, to any person obtaining a %% copy of this software and associated documentation files (the "Software"), %% to deal in the Software without restriction, including without limitation %% the rights to use, copy, modify, merge, publish, distribute, sublicense, %% and/or sell copies of the Software, and to permit persons to whom the %% Software is furnished to do so, subject to the following conditions: %% %% The above copyright notice and this permission notice shall be included in %% all copies or substantial portions of the Software. %% %% THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR %% IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, %% FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE %% AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER %% LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING %% FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER %% DEALINGS IN THE SOFTWARE. %% %% locus is an independent project and has not been authorized, sponsored, %% or otherwise approved by MaxMind. %% @private -module(locus_shared_bitarray). -include_lib("stdlib/include/assert.hrl"). %% ------------------------------------------------------------------ %% API Function Exports %% ------------------------------------------------------------------ -export([new/1, set/2, is_set/2, get_positions_set_at_cell/2]). %% ------------------------------------------------------------------ %% Macro Definitions %% ------------------------------------------------------------------ -define(BITSHIFT, 6). %% ------------------------------------------------------------------ %% Type Definitions %% ------------------------------------------------------------------ -opaque t() :: atomics:atomics_ref(). -export_type([t/0]). %% ------------------------------------------------------------------ %% API Function Definitions %% ------------------------------------------------------------------ -spec new(pos_integer()) -> t(). new(Length) -> Size = ceil(Length / (1 bsl ?BITSHIFT)), atomics:new(Size, [{signed, false}]). -spec set(t(), non_neg_integer()) -> ok. set(Array, Position) -> OneBasedIndex = (Position bsr ?BITSHIFT) + 1, Offset = Position band ((1 bsl ?BITSHIFT) - 1), UpdateMask = 1 bsl Offset, Cell = atomics:get(Array, OneBasedIndex), set_recur(Array, Cell, OneBasedIndex, UpdateMask). -spec is_set(t(), non_neg_integer()) -> boolean(). is_set(Array, Position) -> OneBasedIndex = (Position bsr ?BITSHIFT) + 1, Offset = Position band ((1 bsl ?BITSHIFT) - 1), ReadMask = 1 bsl Offset, try atomics:get(Array, OneBasedIndex) of Cell -> (Cell band ReadMask) =/= 0 catch error:badarg when is_reference(Array), is_integer(Position), Position >= 0 -> throw({position_out_of_bounds, Position}) end. -spec get_positions_set_at_cell(t(), non_neg_integer()) -> [non_neg_integer(), ...]. get_positions_set_at_cell(Array, Index) -> try atomics:get(Array, _OneBasedIndex = Index + 1) of Cell -> BasePosition = Index bsl ?BITSHIFT, positions_set_at_cell(BasePosition, Cell) catch error:badarg -> #{size := Size} = atomics:info(Array), ?assertMatch({true, _, _}, {Index >= Size, Index, Size}), throw({index_out_of_bounds, Index}) end. %% ------------------------------------------------------------------ %% Internal Function Definitions %% ------------------------------------------------------------------ set_recur(Array, Cell, OneBasedIndex, UpdateMask) -> case Cell bor UpdateMask of Cell -> ok; NewCell -> case atomics:exchange(Array, OneBasedIndex, NewCell) of Cell -> ok; UpdatedCell -> set_recur(Array, UpdatedCell, OneBasedIndex, UpdateMask) end end. positions_set_at_cell(BasePosition, Cell) when Cell =/= 0 -> case Cell band 1 of 1 -> [BasePosition | positions_set_at_cell(BasePosition + 1, Cell bsr 1)]; _ -> positions_set_at_cell(BasePosition + 1, Cell bsr 1) end; positions_set_at_cell(_BasePosition, _Cell) -> [].
src/locus_shared_bitarray.erl
0.552057
0.476641
locus_shared_bitarray.erl
starcoder
-module(gh3). -export([to_polyfills/2, from_polyfill/1]). %% @doc Transforms a pre-parsed GeoJSON map into a list of polyfills. %% %% Example: %% %% ``` %% %% First download a map from https://geojson-maps.ash.ms %% {ok, JSONb} = file:read_file("custom.geo.json"), %% JSON = jsx:decode(JSONb, []), %% Poly = h3:to_polyfills(JSON, 8), %% ''' -spec to_polyfills(JSON :: map(), Resolution :: h3:resolution()) -> [[h3:h3index(), ...], ...]. to_polyfills(#{<<"features">> := Features}, Resolution) -> lists:map( fun (Feature) -> to_polyfills(Feature, Resolution) end, Features ); to_polyfills(#{<<"geometry">> := Geometry}, Resolution) -> to_polyfills(Geometry, Resolution); to_polyfills( #{<<"type">> := <<"MultiPolygon">>, <<"coordinates">> := Coordinates}, Resolution ) -> geojson_parse_polygons(Coordinates, Resolution); to_polyfills( #{<<"type">> := <<"Polygon">>, <<"coordinates">> := Coordinates}, Resolution ) -> h3:polyfill(geojson_parse_polygon(Coordinates), Resolution). geojson_parse_polygons(Polygons, Resolution) -> lists:map( fun (P) -> h3:polyfill(geojson_parse_polygon(P), Resolution) end, Polygons ). geojson_parse_polygon(OutlineAndHoles) -> lists:map(fun (OH) -> geojson_transform_coordinates(OH) end, OutlineAndHoles). geojson_transform_coordinates(CoordinateList) -> lists:map(fun ([Lat, Lon]) -> {float(Lon), float(Lat)} end, CoordinateList). %% @doc Converts a polyfill into a GeoJSON map. %% %% Round-trip example: %% %% ``` %% %% First download a map from https://geojson-maps.ash.ms %% {ok, JSONb} = file:read_file("custom.geo.json"), %% JSON = jsx:decode(JSONb, []), %% Polys = h3:to_polyfills(JSON, 9), %% Unsorted = lists:flatten(Polys), %% Sorted = lists:sort(fun(A, B) -> h3:get_resolution(A) < h3:get_resolution(B) end, Unsorted), %% Compacted = h3:compact(Sorted), %% NewJSON = h3:from_polyfill(Compacted), %% NewJSONb = jsx:encode(NewJSON), %% ok = file:write_file("custom.reconstructed.json", NewJSONb). %% ## Try plotting at http://geojson.tools %% ''' from_polyfill(Polyfill) -> #{ <<"type">> => <<"MultiPolygon">>, <<"coordinates">> => multi_polygon_to_geojson(h3:set_to_multi_polygon(Polyfill)) }. multi_polygon_to_geojson({Lat, Lon}) -> [Lon, Lat]; multi_polygon_to_geojson(Polygons) -> lists:map(fun (P) -> multi_polygon_to_geojson(P) end, Polygons).
src/gh3.erl
0.687105
0.73307
gh3.erl
starcoder
%%--------------------------------------------------------------------------- %% | %% Module : String %% Copyright : (c) 2020-2021 EMQ Technologies Co., Ltd. %% License : BSD-style (see the LICENSE file) %% %% Maintainer : <NAME>, <EMAIL> %% <NAME>, <EMAIL> %% Stability : experimental %% Portability : portable %% %% The UTF-8 String FFI module. %% %%--------------------------------------------------------------------------- -module('String'). -include("../Foreign/Maybe.hrl"). -export([ concat/2 , reverse/1 , replicate/2 , strlen/1 , equalIgnoreCase/2 , hasPrefix/2 , hasSuffix/2 , indexOf/2 , lastIndexOf/2 , find/2 , findLast/2 , replace/3 , replaceFirst/3 , replaceLast/3 , split/2 , lines/1 , words/1 , sliceTo/3 , pad/2 , padLeft/2 , padBoth/2 , trimChars/2 , trimLeft/1 , trimLeftChars/2 , trimRight/1 , trimRightChars/2 ]). -define(Whitespace, "\x{0009}\x{000B}\x{000C}\x{0020}\x{00A0}"). -define(LineFeed, "\x{000A}\x{000D}\x{2028}\x{2029}"). -type(prefix() :: string()). -type(suffix() :: string()). -type(pattern() :: string()). -type(replacement() :: string()). -spec(concat(string(), string()) -> string()). concat(S1, S2) -> string:concat(S1, S2). -spec(reverse(string()) -> string()). reverse(String) -> lists:reverse(String). -spec(replicate(pos_integer(), string()) -> string()). replicate(N, S) -> lists:flatten(lists:duplicate(N, S)). -spec(strlen(string()) -> integer()). strlen(String) -> string:length(String). -spec(equalIgnoreCase(string(), string()) -> boolean()). equalIgnoreCase(S1, S2) -> string:equal(S1, S2, true). -spec(hasPrefix(string(), prefix()) -> boolean()). hasPrefix(String, Prefix) -> string:prefix(String, Prefix) /= nomatch. -spec(hasSuffix(string(), suffix()) -> boolean()). hasSuffix(String, Suffix) -> Pos = (strlen(String) - strlen(Suffix)), if Pos >= 0 -> string:equal(string:slice(String, Pos), Suffix); true -> false end. -spec(indexOf(char(), string()) -> integer()). indexOf(Char, String) -> string:chr(String, Char). -spec(lastIndexOf(char(), string()) -> integer()). lastIndexOf(Char, String) -> string:rchr(String, Char). -spec(find(string(), pattern()) -> maybe(string())). find(String, Pattern) -> doFind(String, Pattern, leading). -spec(findLast(string(), pattern()) -> maybe(string())). findLast(String, Pattern) -> doFind(String, Pattern, trailing). doFind(String, Pattern, Dir) -> case string:find(String, Pattern, Dir) of nomatch -> ?Nothing; SubStr -> ?Just(SubStr) end. -spec(replace(string(), pattern(), replacement()) -> string()). replace(String, Pattern, Replacement) -> string:replace(String, Pattern, Replacement, all). -spec(replaceFirst(string(), pattern(), replacement()) -> string()). replaceFirst(String, Pattern, Replacement) -> string:replace(String, Pattern, Replacement, leading). -spec(replaceLast(string(), pattern(), replacement()) -> string()). replaceLast(String, Pattern, Replacement) -> string:replace(String, Pattern, Replacement, trailing). -spec(split(string(), Sep :: string()) -> [string()]). split(String, Sep) -> string:split(String, Sep, all). -spec(lines(string()) -> [string()]). lines(String) -> string:split(String, "\n", all). -spec(words(string()) -> [string()]). words(String) -> string:tokens(String, ?Whitespace ++ ?LineFeed). -spec(sliceTo(string(), pos_integer(), pos_integer()) -> string()). sliceTo(String, Start, End) -> string:slice(String, Start, (End - Start)). flat(X) -> lists:flatten(X). -spec(pad(string(), pos_integer()) -> string()). pad(String, Len) -> flat(string:pad(String, Len)). -spec(padLeft(string(), pos_integer()) -> string()). padLeft(String, Len) -> flat(string:pad(String, Len, leading)). -spec(padBoth(string(), pos_integer()) -> string()). padBoth(String, Len) -> flat(string:pad(String, Len, both)). -spec(trimChars(string(), string()) -> string()). trimChars(String, Chars) -> string:trim(String, both, Chars). -spec(trimLeft(string()) -> string()). trimLeft(String) -> string:trim(String, leading). -spec(trimLeftChars(string(), string()) -> string()). trimLeftChars(String, Chars) -> string:trim(String, leading, Chars). -spec(trimRight(string()) -> string()). trimRight(String) -> string:trim(String, trailing). -spec(trimRightChars(string(), string()) -> string()). trimRightChars(String, Chars) -> string:trim(String, trailing, Chars).
lib/Data/String.erl
0.511473
0.47591
String.erl
starcoder
%% Copyright 2011 <NAME> <<EMAIL>> % % Licensed under the Apache License, Version 2.0 (the "License"); % you may not use this file except in compliance with the License. % You may obtain a copy of the License at % % http://www.apache.org/licenses/LICENSE-2.0 % % Unless required by applicable law or agreed to in writing, software % distributed under the License is distributed on an "AS IS" BASIS, % WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. % See the License for the specific language governing permissions and % limitations under the License. -module(qrcode_mask). -include("qrcode_params.hrl"). -export([generate/2, select/1]). -define(PENALTY_RULE_1, 3). -define(PENALTY_RULE_2, 3). -define(PENALTY_RULE_3, 40). -define(PENALTY_RULE_4, 10). %% Generates all eight masked versions of the bit matrix generate(#qr_params{dimension = Dim}, Matrix) -> Sequence = lists:seq(0, 7), Functions = [mask(X) || X <- Sequence], Masks = [generate_mask(Dim, MF) || MF <- Functions], [apply_mask(Matrix, Mask, []) || Mask <- Masks]. %% Selects the lowest penalty candidate from a list of bit matrices select([H|T]) -> Score = score_candidate(H), select_candidate(T, 0, 0, Score, H). %% Internal % generate_mask(Max, MF) -> Sequence = lists:seq(0, Max - 1), [generate_mask(Sequence, Y, MF) || Y <- Sequence]. generate_mask(Sequence, Y, MF) -> [case MF(X, Y) of true -> 1; false -> 0 end || X <- Sequence]. apply_mask([H|T], [H0|T0], Acc) -> Row = apply_mask0(H, H0, []), apply_mask(T, T0, [Row|Acc]); apply_mask([], [], Acc) -> lists:reverse(Acc). apply_mask0([H|T], [H0|T0], Acc) when is_integer(H) -> apply_mask0(T, T0, [H bxor H0|Acc]); apply_mask0([H|T], [_|T0], Acc) -> apply_mask0(T, T0, [H|Acc]); apply_mask0([], [], Acc) -> lists:reverse(Acc). % (i + j) mod 2 = 0 mask(0) -> fun(X, Y) -> (X + Y) rem 2 =:= 0 end; % i mod 2 = 0 mask(1) -> fun(_X, Y) -> Y rem 2 =:= 0 end; % j mod 3 = 0 mask(2) -> fun(X, _Y) -> X rem 3 =:= 0 end; % (i + j) mod 3 = 0 mask(3) -> fun(X, Y) -> (X + Y) rem 3 =:= 0 end; % ((i div 2) + (j div 3)) mod 2 = 0 mask(4) -> fun(X, Y) -> (X div 3 + Y div 2) rem 2 =:= 0 end; %101 (i * j) mod 2 + (i *j) mod 3 = 0 mask(5) -> fun(X, Y) -> Sum = X * Y, Sum rem 2 + Sum rem 3 =:= 0 end; % ((i * j) mod 2 + (i* j) mod 3) mod 2 = 0 mask(6) -> fun(X, Y) -> Sum = X * Y, (Sum rem 2 + Sum rem 3) rem 2 =:= 0 end; %((i * j) mod 3 + (i + j) mod 2) mod 2 = 0 mask(7) -> fun(X, Y) -> ((X * Y rem 3) + ((X + Y) rem 2)) rem 2 =:= 0 end. select_candidate([H|T], Count, Mask, Score, C) -> case score_candidate(H) of X when X < Score -> select_candidate(T, Count + 1, Count + 1, X, H); _ -> select_candidate(T, Count + 1, Mask, Score, C) end; select_candidate([], _, Mask, _Score, C) -> %?TTY({selected, Mask, {score, Score}}), {Mask, C}. score_candidate(C) -> Rule1 = apply_penalty_rule_1(C), Rule2 = apply_penalty_rule_2(C), Rule3 = apply_penalty_rule_3(C), Rule4 = apply_penalty_rule_4(C), Total = Rule1 + Rule2 + Rule3 + Rule4, %?TTY({score, Total, [Rule1, Rule2, Rule3, Rule4]}), Total. %% Section 8.2.2 apply_penalty_rule_1(Candidate) -> ScoreRows = rule1(Candidate, 0), ScoreCols = rule1(rows_to_columns(Candidate), 0), ScoreRows + ScoreCols. % rule1([Row|T], Score) -> Score0 = rule1_row(Row, Score), rule1(T, Score0); rule1([], Score) -> Score. % rule1_row(L = [H|_], Score) -> F = fun (1) when H =:= 1 -> true; (1) -> false; (_) when H =:= 0 orelse is_integer(H) =:= false -> true; (_) -> false end, {H0,T0} = lists:splitwith(F, L), case length(H0) of Repeats when Repeats >= 5 -> Penalty = ?PENALTY_RULE_1 + Repeats - 5, rule1_row(T0, Score + Penalty); _ -> rule1_row(T0, Score) end; rule1_row([], Score) -> Score. %% apply_penalty_rule_2(_M = [H, H0|T]) -> % ?TTY(M), Blocks = rule2(1, 1, H, H0, [H0|T], []), Blocks0 = composite_blocks(Blocks, []), Blocks1 = composite_blocks(Blocks0, []), % ?TTY(Blocks1), score_blocks(Blocks1, 0). score_blocks([{_, {M, N}, _}|T], Acc) -> Score = ?PENALTY_RULE_2 * (M - 1) * (N - 1), score_blocks(T, Acc + Score); score_blocks([], Acc) -> Acc. rule2(X, Y, [H, H|T], [H, H|T0], Rows, Acc) -> rule2(X + 1, Y, [H|T], [H|T0], Rows, [{{X, Y}, {2, 2}, H}|Acc]); rule2(X, Y, [_|T], [_|T0], Rows, Acc) -> rule2(X + 1, Y, T, T0, Rows, Acc); rule2(_, Y, [], [], [H, H0|T], Acc) -> rule2(1, Y + 1, H, H0, [H0|T], Acc); rule2(_, _, [], [], [_], Acc) -> lists:reverse(Acc). composite_blocks([H|T], Acc) -> {H0, T0} = composite_block(H, T, []), composite_blocks(T0, [H0|Acc]); composite_blocks([], Acc) -> lists:reverse(Acc). composite_block(B, [H|T], Acc) -> case combine_block(B, H) of false -> composite_block(B, T, [H|Acc]); B0 -> composite_block(B0, T, Acc) end; composite_block(B, [], Acc) -> {B, lists:reverse(Acc)}. % Does Block 0 contain the Block 1 coordinate? combine_block(B = {{X, Y}, {SX, SY}, _}, B0 = {{X0, Y0}, _, _}) when X0 < X + SX orelse Y0 < Y + SY -> combine_block0(B, B0); combine_block(_, _) -> false. % are they same valued? combine_block0(B = {_, _, V}, B0 = {_, _, V0}) when V =:= V0 orelse (V =/= 1 andalso V0 =/= 1) -> combine_block1(B, B0); combine_block0(_, _) -> false. % is B extended by B0 horizontally? combine_block1({{X, Y}, {SX, SY}, V}, {{X0, Y}, {SX0, SY}, _}) when X0 =:= X + SX - 1 -> {{X, Y}, {SX + SX0 - 1, SY}, V}; % is B extended by B0 vertically? combine_block1({{X, Y}, {SX, SY}, V}, {{X, Y0}, {SX, SY0}, _}) when Y0 =:= Y + SY - 1 -> {{X, Y}, {SX, SY + SY0 - 1}, V}; combine_block1(_, _) -> false. %% apply_penalty_rule_3(Candidate) -> RowScores = [rule3(Row, 0) || Row <- Candidate], ColumnScores = [rule3(Col, 0) || Col <- rows_to_columns(Candidate)], lists:sum(RowScores) + lists:sum(ColumnScores). % rule3(Row = [1|T], Score) -> Ones = lists:takewhile(fun(X) -> X =:= 1 end, Row), Scale = length(Ones), case Scale * 7 of Length when Length > length(Row) -> rule3(T, Score); Length -> case is_11311_pattern(lists:sublist(Row, Length), Scale) of true -> rule3(T, Score + ?PENALTY_RULE_3); false -> rule3(T, Score) end end; rule3([_|T], Score) -> rule3(T, Score); rule3([], Acc) -> Acc. % is_11311_pattern(List, Scale) -> List0 = lists:map(fun(X) when X =:= 1 -> 1; (_) -> 0 end, List), Result = condense(List0, Scale, []), Result =:= [1,0,1,1,1,0,1]. % condense([], _, Acc) -> lists:reverse(Acc); condense(L, Scale, Acc) -> {H, T} = lists:split(Scale, L), case lists:sum(H) of Scale -> condense(T, Scale, [1|Acc]); 0 -> condense(T, Scale, [0|Acc]); _ -> undefined end. %% apply_penalty_rule_4(Candidate) -> Proportion = rule4(Candidate, 0, 0), %?TTY({proportion, Proportion}), ?PENALTY_RULE_4 * (trunc(abs(Proportion * 100 - 50)) div 5). % rule4([H|T], Dark, All) -> All0 = All + length(H), Dark0 = Dark + length([X || X <- H, X =:= 1]), rule4(T, Dark0, All0); rule4([], Dark, All) -> Dark / All. % rows_to_columns(L) -> rows_to_columns(L, []). rows_to_columns([[]|_], Acc) -> lists:reverse(Acc); rows_to_columns(L, Acc) -> Heads = [H || [H|_] <- L], Tails = [T || [_|T] <- L], rows_to_columns(Tails, [Heads|Acc]).
src/qrcode_mask.erl
0.513181
0.484685
qrcode_mask.erl
starcoder
%%%------------------------------------------------------------------- %%% @author <NAME> %%% @copyright (C) 2017: <NAME> %%% This software is released under the MIT license cited in 'LICENSE.md'. %%% @end %%%------------------------------------------------------------------- %%% @doc %%% This module provides search functionality for leaf nodes of B+ tree. %%% @end %%%------------------------------------------------------------------- -module(bp_tree_leaf). -author("<NAME>"). -include("bp_tree.hrl"). %% API exports -export([find_leftmost/1, find_key/2]). %%==================================================================== %% API functions %%==================================================================== %%-------------------------------------------------------------------- %% @doc %% Returns leftmost leaf in a B+ tree. %% @end %%-------------------------------------------------------------------- -spec find_leftmost(bp_tree:tree()) -> {{ok, bp_tree_node:id(), bp_tree:tree_node()} | {error, term()}, bp_tree:tree()}. find_leftmost(Tree) -> case bp_tree_store:get_root_id(Tree) of {{ok, RootId}, Tree2} -> find_leftmost(RootId, Tree2); {{error, Reason}, Tree2} -> {{error, Reason}, Tree2} end. %%-------------------------------------------------------------------- %% @doc %% Returns key at given offset starting from leftmost node. Along with the key %% a node holding it and its position within this node is returned. %% @end %%-------------------------------------------------------------------- -spec find_key(non_neg_integer(), bp_tree:tree()) -> { {ok, bp_tree:key(), pos_integer(), bp_tree_node:id(), bp_tree:tree_node()} | {error, term() }, bp_tree:tree()}. find_key(Offset, Tree) -> case find_leftmost(Tree) of {{ok, NodeId, Node}, Tree2} -> find_key(Offset, NodeId, Node, Tree2); {{error, Reason}, Tree2} -> {{error, Reason}, Tree2} end. %%==================================================================== %% Internal functions %%==================================================================== %%-------------------------------------------------------------------- %% @private %% @doc %% Returns leftmost leaf in a B+ tree rooted in a node. %% @end %%-------------------------------------------------------------------- -spec find_leftmost(bp_tree_node:id(), bp_tree:tree()) -> {{ok, bp_tree_node:id(), bp_tree:tree_node()} | {error, term()}, bp_tree:tree()}. find_leftmost(NodeId, Tree) -> {{ok, Node}, Tree2} = bp_tree_store:get_node(NodeId, Tree), case bp_tree_node:leftmost_child(Node) of {ok, NodeId2} -> find_leftmost(NodeId2, Tree2); {error, not_found} -> {{ok, NodeId, Node}, Tree2} end. %%-------------------------------------------------------------------- %% @private %% @doc %% Returns key at given offset starting from a node. Along with the key %% a node holding it and its position within this node is returned. %% @end %%-------------------------------------------------------------------- -spec find_key(non_neg_integer(), bp_tree_node:id(), bp_tree:tree_node(), bp_tree:tree()) -> {{ok, bp_tree:key(), pos_integer(), bp_tree_node:id(), bp_tree:tree_node()} | {error, term()}, bp_tree:tree()}. find_key(Offset, NodeId, Node, Tree) -> Size = bp_tree_node:size(Node), case Size =< Offset of true -> case bp_tree_node:right_sibling(Node) of {ok, NodeId2} -> {{ok, Node2}, Tree2} = bp_tree_store:get_node(NodeId2, Tree), find_key(Offset - Size, NodeId2, Node2, Tree2); {error, Reason} -> {{error, Reason}, Tree} end; false -> {ok, Key} = bp_tree_node:key(Offset + 1, Node), {{ok, Key, Offset + 1, NodeId, Node}, Tree} end.
src/bp_tree_leaf.erl
0.624523
0.40251
bp_tree_leaf.erl
starcoder
%% @doc This is the behaviour definition for a credential provider module %% and it iterates over a list of providers. You may set the `credential_providers` %% Erlang environment variable if you want to restrict checking only a certain %% subset of the default list. %% %% Default order of checking for credentials is: %% <ol> %% <li>Erlang application environment</li> %% <li>OS environment</li> %% <li>Credentials from AWS file</li> %% <li>ECS Task credentials</li> %% <li>EC2 credentials</li> %% </ol> %% %% Providers are expected to implement a function called `fetch/1' which %% takes as its argument a proplist of options which may influence the %% operation of the provider. The fetch/1 function should return either %% `{ok, Credentials, Expiration}' or `{error, Reason}'. %% %% If a provider returns {ok, ...} then evaluation stops at that provider. %% If it returns {error, ...} then the next provider is executed in order %% until either a set of credentials are returns or the tuple %% `{error, no_credentials}' is returned. %% %% If a new provider is desired, the behaviour interface should be %% implemented and its module name added to the default list. %% @end -module(aws_credentials_provider). -export([fetch/0, fetch/1]). -callback fetch ( Options :: proplists:proplist() ) -> {ok, Credentials :: map(), Expiration :: binary() | pos_integer() | infinity} | {error, Reason :: term()}. -define(DEFAULT_PROVIDERS, [aws_credentials_env, aws_credentials_file, aws_credentials_ecs, aws_credentials_ec2]). fetch() -> fetch([]). fetch(Options) -> Providers = get_env(credential_providers, ?DEFAULT_PROVIDERS), evaluate_providers(Providers, Options). evaluate_providers([], _Options) -> {error, no_credentials}; evaluate_providers([ H | T ], Options) -> case H:fetch(Options) of {error, _} = Error -> error_logger:error_msg("Provider ~p reports ~p", [H, Error]), evaluate_providers(T, Options); Credentials -> Credentials end. get_env(Key, Default) -> case application:get_env(aws_credentials, Key) of undefined -> Default; {ok, Value} -> Value end.
src/aws_credentials_provider.erl
0.543833
0.485539
aws_credentials_provider.erl
starcoder
% Licensed under the Apache License, Version 2.0 (the "License"); you may not % use this file except in compliance with the License. You may obtain a copy of % the License at % % http://www.apache.org/licenses/LICENSE-2.0 % % Unless required by applicable law or agreed to in writing, software % distributed under the License is distributed on an "AS IS" BASIS, WITHOUT % WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the % License for the specific language governing permissions and limitations under % the License. -module(vtree_bulk). -include_lib("couch/include/couch_db.hrl"). -ifndef(makecheck). -define(MAX_FILLED, 40). -else. -define(MAX_FILLED, 4). -compile(export_all). -endif. -export([omt_load/2, omt_write_tree/2, bulk_load/4]). -export([log_n_ceil/2]). % XXX vmx: check if tree has correct meta information set for every node % (type=inner/type=leaf) % The seedtree is kept in memory, therefore it makes sense to restrict the % maximum height of it. -define(MAX_SEEDTREE_HEIGHT, 3). % same as in vtree -record(node, { % type = inner | leaf type=leaf}). -record(seedtree_root, { tree = [] :: list(), outliers = [] :: list(), height = 0 :: integer() }). -record(seedtree_leaf, { orig = [] :: list(), new = [] :: list(), % position in file (of this node) pos = -1 :: integer() }). -type mbr() :: {number(), number(), number(), number()}. -type vtree_node() :: {mbr(), tuple(), list()}. -type seedtree_root() :: tuple(). -type seedtree_node() :: tuple(). -type omt_node() :: tuple(). -type omt_tree() :: [omt_node()] | [omt_tree()]. % @doc Bulk load a tree. Returns the position of the root node and the height % of the tree -spec bulk_load(Fd::file:io_device(), RootPos::integer(), TargetTreeHeight::integer(), Nodes::[vtree_node()]) -> {ok, integer(), integer()}. % No nodes to load bulk_load(_Fd, RootPos, TargetTreeHeight, []) -> {ok, RootPos, TargetTreeHeight}; % Tree is empty bulk_load(Fd, _RootPos, TargetTreeHeight, Nodes) when TargetTreeHeight==0 -> % Tree is empty => bulk load it {Omt, TreeHeight} = omt_load(Nodes, ?MAX_FILLED), {ok, MbrAndPosList} = omt_write_tree(Fd, Omt), {_Mbrs, PosList} = lists:unzip(MbrAndPosList), NewNodes = load_nodes(Fd, PosList), {ok, NewPos} = case length(NewNodes) of % single node as root 1 -> Result = couch_file:append_term(Fd, hd(NewNodes)), ok = couch_file:sync(Fd), Result; % multiple nodes _ -> write_parent(Fd, NewNodes) end, {ok, NewPos, TreeHeight}; bulk_load(Fd, RootPos, TargetTreeHeight, Nodes) -> SeedtreeHeight = floor(TargetTreeHeight/2), Seedtree = seedtree_init(Fd, RootPos, SeedtreeHeight), Seedtree2 = seedtree_insert_list(Seedtree, Nodes), % There might be a huge about of outliers, insert them in a different % way. Create an OMT tree and insert the whole treet into the existing % one at the approriate height. Outliers = Seedtree2#seedtree_root.outliers, OutliersNum = length(Outliers), Seedtree3 = if OutliersNum > 100 -> Seedtree2#seedtree_root{outliers=[]}; true -> Seedtree2 end, {ok, Result, NewHeight, _HeightDiff} = seedtree_write( Fd, Seedtree3, TargetTreeHeight - Seedtree3#seedtree_root.height), {NewPos, NewHeight2, NewMbr} = case length(Result) of % single node as root 1 -> {ok, ResultPos} = couch_file:append_term(Fd, hd(Result)), ok = couch_file:sync(Fd), {ResultPos, NewHeight, element(1, hd(Result))}; % multiple nodes _ -> {ok, ResultPos} = write_parent(Fd, Result), {ResultPos, NewHeight+1, vtree:calc_nodes_mbr(Result)} end, % Insert outliers as whole subtree {NewPos2, NewHeight3} = if OutliersNum > 100 -> insert_outliers(Fd, NewPos, NewMbr, NewHeight2, Outliers); true -> {NewPos, NewHeight2} end, %?debugVal(NewPos2), {ok, NewPos2, NewHeight3}. % @doc If there is a huge number of outliers, we bulk load them into a new % tree and insert that tree directly to the original target tree. Returns % the position of the root node in file and the height of the tree. -spec insert_outliers(Fd::file:io_device(), TargetPos::integer(), TargetMbr::mbr(), TargetHeight::integer(), Nodes::[vtree_node()]) -> {integer(), integer()}. insert_outliers(Fd, TargetPos, TargetMbr, TargetHeight, Nodes) -> {Omt, OmtHeight} = omt_load(Nodes, ?MAX_FILLED), {ok, MbrAndPosList} = omt_write_tree(Fd, Omt), {Mbrs, PosList} = lists:unzip(MbrAndPosList), MergedMbr = vtree:calc_mbr([TargetMbr|Mbrs]), % OmtHeight - 1 because this node might contain several children case TargetHeight - (OmtHeight - 1) of % Both, the new bulk loaded tree and targt tree have the same height % => create new common root Diff when Diff == 0 -> %?debugMsg("same height"), if length(MbrAndPosList) + 1 =< ?MAX_FILLED -> % XXX vmx: is using type=inner always valid? NewRootNode = {MergedMbr, #node{type=inner}, [TargetPos|PosList]}, {ok, NewOmtPos} = couch_file:append_term(Fd, NewRootNode), ok = couch_file:sync(Fd), {NewOmtPos, TargetHeight+1}; % split the node and create new root node true -> % NOTE vmx: These are not a normal nodes. The last element in the % tuple is not the position of the children, but the position % of the node itself. % XXX vmx: is type=inner always right? NewChildren = [{Mbr, #node{type=inner}, Pos} || {Mbr, Pos} <- [{TargetMbr, TargetPos}|MbrAndPosList]], NodeToSplit = {MergedMbr, #node{type=inner}, NewChildren}, {_SplittedMbr, Node1, Node2} = vtree:split_node(NodeToSplit), {ok, NewOmtPos} = write_parent(Fd, [Node1|[Node2]]), {NewOmtPos, TargetHeight+2} end; % insert new tree into target tree Diff when Diff > 0 -> %?debugMsg("target tree higher"), {ok, _, SubPos, Inc} = insert_subtree( Fd, TargetPos, MbrAndPosList, Diff-1), {SubPos, TargetHeight+Inc}; % insert target tree into new tree Diff when Diff < 0 -> %?debugMsg("target tree smaller"), {OmtRootMbrs, OmtRootPosList} = lists:unzip(MbrAndPosList), OmtRootNode = {vtree:calc_mbr(OmtRootMbrs), #node{type=inner}, OmtRootPosList}, {ok, NewOmtPos} = couch_file:append_term(Fd, OmtRootNode), ok = couch_file:sync(Fd), {ok, _, SubPos, Inc} = insert_subtree( Fd, NewOmtPos, [{TargetMbr, TargetPos}], abs(Diff)), {SubPos, OmtHeight+Inc} end. % @doc OMT bulk loading. MaxNodes is the number of maximum children per node. % Returns the OMT tree and the height of the tree % Based on (but modified): % OMT: Overlap Minimizing Top-down Bulk Loading Algorithm for R-tree -spec omt_load(Nodes::[omt_node()], MaxNodes::integer()) -> {omt_tree(), integer()}. omt_load([], _MaxNodes) -> []; omt_load(Nodes, MaxNodes) -> NodesNum = length(Nodes), % Height of the final tree Height = log_n_ceil(MaxNodes, NodesNum), Omt = omt_load(Nodes, MaxNodes, 0, Height-1, 0), {Omt, Height}. % all nodes need to be on the same level -spec omt_load(Nodes::[omt_node()], MaxNodes::integer(), Dimension::integer(), LeafDepth::integer(), Depth::integer()) -> omt_node(). omt_load(Nodes, MaxNodes, _Dimension, LeafDepth, Depth) when length(Nodes) =< MaxNodes -> lists:foldl(fun(_I, Acc) -> [Acc] end, Nodes, lists:seq(1,LeafDepth-Depth)); omt_load(Nodes, MaxNodes, Dimension, LeafDepth, Depth) -> NodesNum = length(Nodes), % Height of the final tree Height = log_n_ceil(MaxNodes, NodesNum), % Maximum number of children in a direct subtree of the root node % with the minimum depth that is needed to store all nodes ChildrenSubNum = math:pow(MaxNodes, Height-1), % Number of children RootChildrenNum = ceiling(NodesNum/ChildrenSubNum), % Entries per child EntriesNum = ceiling(NodesNum/RootChildrenNum), % NOTE vmx: currently all nodes have only 2 dimensions => "rem 2" Sorted = omt_sort_nodes(Nodes, (Dimension rem 2)+1), _Chunked = chunk_list(fun(InnerNodes) -> omt_load(InnerNodes, MaxNodes, Dimension+1, LeafDepth, Depth+1) end, Sorted, EntriesNum). % @doc Write an OMT in memory tree to disk. Returns a list of tuples that % contain the MBR and % either: % - the position of the children in the file % or: % - the actual node value (if the tree consists only of one node). -spec omt_write_tree(Fd::file:io_device(), Tree::omt_tree()) -> {ok, [{mbr(), integer()}]} | {ok, [{mbr(), tuple()}]}. omt_write_tree(Fd, Tree) -> Return = case omt_write_tree(Fd, Tree, 0, []) of {level_done, Nodes} -> Nodes; {leaf_nodes, MbrAndPos} -> [MbrAndPos] end, {ok, lists:reverse(Return)}. % no more siblings -spec omt_write_tree(Fd::file:io_device(), Tree::list(), Depth::integer(), Acc::list()) -> {ok, integer()}. omt_write_tree(_Fd, [], _Depth, Acc) -> {no_siblings, Acc}; % leaf node omt_write_tree(Fd, [H|_T]=Leafs, _Depth, _Acc) when is_tuple(H) -> % Don't write leafs nodes to disk now, they will be written later on. % Instead return a list of of tuples with the node's MBR and the node % itself Mbr = vtree:calc_nodes_mbr(Leafs), % We can not only pass in single nodes, but also subtrees. If the third % element of the tuple is a tuple, it's a leaf node, if it is a list, % it's an inner node Node = case is_tuple(element(3, H)) of true -> {Mbr, #node{type=leaf}, Leafs}; false -> PosList = lists:map(fun(N) -> {ok, Pos} = couch_file:append_term(Fd, N), ok = couch_file:sync(Fd), Pos end, Leafs), {Mbr, #node{type=inner}, PosList} end, {ok, Pos} = couch_file:append_term(Fd, Node), ok = couch_file:sync(Fd), {leaf_nodes, {Mbr, Pos}}; omt_write_tree(Fd, [H|T], Depth, Acc) -> {_, Acc2} = case omt_write_tree(Fd, H, Depth+1, []) of {no_siblings, Siblings} -> {ok, Siblings}; {leaf_nodes, MbrAndPos} -> {ok, [MbrAndPos|Acc]}; {level_done, Level} -> % NOTE vmx: reversing Level is probably not neccessary {Mbrs, Children} = lists:unzip(lists:reverse(Level)), Mbr = vtree:calc_mbr(Mbrs), Meta = #node{type=inner}, {ok, Pos} = couch_file:append_term(Fd, {Mbr, Meta, Children}), ok = couch_file:sync(Fd), {ok, [{Mbr, Pos}|Acc]} end, {_, Acc3} = omt_write_tree(Fd, T, Depth, Acc2), {level_done, Acc3}. % @doc Sort nodes by a certain dimension (which is the first element of the % node tuple) -spec omt_sort_nodes(Nodes::[omt_node()], Dimension::integer()) -> [omt_node()]. % NOTE vmx: in the future the dimensions won't be stored in tuples, but % in lists. omt_sort_nodes(Nodes, Dimension) -> lists:sort(fun(Node1, Node2) -> Mbr1 = element(1, Node1), Mbr2 = element(1, Node2), Val1 = element(Dimension, Mbr1), Val2 = element(Dimension, Mbr2), Val1 =< Val2 end, Nodes). % @doc Insert several new items into the seed tree -spec seedtree_insert_list(Tree::seedtree_root(), _Nodes::[tuple()]) -> seedtree_root(). seedtree_insert_list(Root, []) -> Root; seedtree_insert_list(Root, [H|T]=_Nodes) -> Root2 = seedtree_insert(Root, H), seedtree_insert_list(Root2, T). % @doc Insert a new item into the seed tree % Based on: % Bulk insertion for R-trees by seeded clustering -spec seedtree_insert(Tree::seedtree_root(), Node::tuple()) -> seedtree_root(). seedtree_insert(#seedtree_root{tree=Tree, outliers=Outliers}=Root, Node) -> case seedtree_insert_children([Tree], Node) of {ok, [Tree2]} -> Root#seedtree_root{tree=Tree2}; {not_inserted, _} -> Root#seedtree_root{outliers=[Node|Outliers]} end. % @doc Do the actual insertion of the nodes into the seedtree -spec seedtree_insert_children(Children::[seedtree_node()], Node::seedtree_node()) -> {ok, seedtree_node()} | {not_inserted}. seedtree_insert_children([], Node) -> {not_inserted, Node}; seedtree_insert_children(#seedtree_leaf{new=Old}=Children, Node) when not is_list(Children) -> New = [Node|Old], Children2 = Children#seedtree_leaf{new=New}, {ok, Children2}; seedtree_insert_children([H|T], Node) -> {Mbr, Meta, Children} = H, {NodeMbr, _, _Data} = Node, case vtree:within(NodeMbr, Mbr) of true -> {Status, Children2} = seedtree_insert_children(Children, Node), {Status, [{Mbr, Meta, Children2}|T]}; false -> {Status, T2} = seedtree_insert_children(T, Node), {Status, [H|T2]} end. % @doc Put an on disk tree into memory and prepare it to store new nodes in % the leafs. If MaxDepth is bigger than the tree, the lowest possible % level is returned. -spec seedtree_init(Fd::file:io_device(), RootPos::integer(), MaxDepth::integer()) -> seedtree_root(). seedtree_init(Fd, RootPos, MaxDepth) -> Tree = seedtree_init(Fd, RootPos, MaxDepth, 0), %#seedtree_root{tree=Tree, height=MaxDepth}. Height = erlang:min(MaxDepth, ?MAX_SEEDTREE_HEIGHT), #seedtree_root{tree=Tree, height=Height}. -spec seedtree_init(Fd::file:io_device(), RootPos::integer(), MaxDepth::integer(), Depth::integer()) -> seedtree_node(). % It's "Depth+1" as the root already contains several nodes (it's the nature % of a R-Tree). seedtree_init(Fd, RootPos, MaxDepth, Depth) when (Depth+1 == MaxDepth) or (Depth+1 == ?MAX_SEEDTREE_HEIGHT) -> %?debugVal(RootPos), {ok, Parent} = couch_file:pread_term(Fd, RootPos), {ParentMbr, ParentMeta, EntriesPos} = Parent, {ParentMbr, ParentMeta, #seedtree_leaf{orig=EntriesPos, pos=RootPos}}; seedtree_init(Fd, RootPos, MaxDepth, Depth) -> {ok, Parent} = couch_file:pread_term(Fd, RootPos), {ParentMbr, ParentMeta, EntriesPos} = Parent, case is_tuple(hd(EntriesPos)) of % we reached leaf level => return true -> % create seedtree leaf {ParentMbr, ParentMeta, #seedtree_leaf{orig=EntriesPos, pos=RootPos}}; false -> Children = lists:foldl(fun(EntryPos, Acc) -> Child = seedtree_init(Fd, EntryPos, MaxDepth, Depth+1), [Child|Acc] end, [], EntriesPos), {ParentMbr, ParentMeta, lists:reverse(Children)} end. % @doc Write an new vtree. InsertHeight is the height where the seedtree % should be inserted to. Returns the root nodes, the height of the tree and % the height difference to the old tree. -spec seedtree_write(Fd::file:io_device(), Seetree::seedtree_root(), InsertHeight::integer()) -> {ok, [vtree_node()], integer(), integer()}. seedtree_write(Fd, Seedtree, InsertHeight) -> Tree = Seedtree#seedtree_root.tree, TargetHeight = InsertHeight + Seedtree#seedtree_root.height, {level_done, Level} = seedtree_write(Fd, [Tree], InsertHeight, []), {RootNodes, NewHeight} = if length(Level) =< ?MAX_FILLED -> {Level, TargetHeight}; true -> {Level2, Level2Height} = omt_load(Level, ?MAX_FILLED), {ok, ParentMbrAndPos} = omt_write_tree(Fd, Level2), Parents = lists:map(fun({_, Pos}) -> {ok, Node} = couch_file:pread_term(Fd, Pos), Node end, ParentMbrAndPos), {Parents, TargetHeight+Level2Height-1} end, {RootNodes2, HeightDiff} = case Seedtree#seedtree_root.outliers of [] -> {RootNodes, 0}; Outliers -> % insert outliers by creating a temporary root node... {ok, TmpRootPos} = write_parent(Fd, RootNodes), {TmpRootPos2, TmpHeight} = lists:foldl(fun(Outlier, {CurPos, _}) -> {Mbr, Meta, {DocId, {Geom, Value}}} = Outlier, {ok, _NewMbr, CurPos2, TreeHeight} = vtree:insert( Fd, CurPos, DocId, {Mbr, Meta, Geom, Value}), {CurPos2, TreeHeight} end, {TmpRootPos, 0}, Outliers), {ok, OldRoot} = couch_file:pread_term(Fd, TmpRootPos2), % ...and get the original children back again {_, _, RootNodes2Pos} = OldRoot, LevelDiff = TmpHeight-1-NewHeight, ChildrenNodes = load_nodes(Fd, RootNodes2Pos), {ChildrenNodes, LevelDiff} end, {ok, RootNodes2, NewHeight+HeightDiff, HeightDiff}. -spec seedtree_write(Fd::file:io_device(), Seetree::seedtree_root(), InsertHeight::integer(), Acc::[vtree_node()]) -> {ok, [vtree_node()], integer(), integer()}. seedtree_write(_Fd, [], _InsertHeight, Acc) -> {no_more_siblings, Acc}; % No new nodes to insert seedtree_write(_Fd, #seedtree_leaf{orig=Orig, new=[]}, _InsertHeight, _Acc) -> {leafs, Orig}; % New nodes to insert % This function returns a new list of children, as some children were % rewritten due to repacking. The MBR doesn't change (that's the nature of % this algorithm). seedtree_write(Fd, #seedtree_leaf{orig=Orig, new=New, pos=ParentPos}, InsertHeight, _Acc) -> NewNum = length(New), OmtHeight = log_n_ceil(?MAX_FILLED, NewNum), HeightDiff = InsertHeight - (OmtHeight - 1), NewChildrenPos2 = if % Input tree can be inserted as-is into the target tree HeightDiff == 0 -> %?debugMsg("insert as is"), {OmtTree, OmtHeight} = omt_load(New, ?MAX_FILLED), NewChildren = seedtree_write_insert(Fd, Orig, OmtTree, OmtHeight), _MbrAndPos = seedtree_write_finish(NewChildren); % insert tree is too small => expand seedtree HeightDiff > 0 -> %?debugMsg("insert is too small"), % Create new seedtree % HeightDiff+1 as we like to load the level of the children Seedtree = seedtree_init(Fd, ParentPos, HeightDiff+1), Seedtree2 = seedtree_insert_list(Seedtree, New), {ok, NodesList, _NewHeight, NewHeightDiff} = seedtree_write( Fd, Seedtree2, InsertHeight-Seedtree2#seedtree_root.height+1), % NOTE vmx (2011-01-04) This makes one test fail. I have to admit I % can't see the point of this code atm. The calculation of the % insertion was based on the height of the resulting OMT tree, so % why should some levels be stripped? % There was a node overflow. Though we are currently somewhere within % the tree, so we need to return a node that has the original height. % Therefore recreate the overflowed node. NodesList2 = case NewHeightDiff > 0 of false -> NodesList; true -> lists:foldl(fun(_I, Acc) -> PosList = lists:append([Pos || {_, _, Pos} <- Acc]), load_nodes(Fd, PosList) end, NodesList, lists:seq(1, NewHeightDiff)) end, _MbrAndPos = [{Mbr, Pos} || {Mbr, _, Pos} <- NodesList2]; %MbrAndPos = [{Mbr, Pos} || {Mbr, _, Pos} <- NodesList]; % insert tree is too high => use its children HeightDiff < 0 -> %?debugMsg("insert is too high"), {OmtTree, OmtHeight} = omt_load(New, ?MAX_FILLED), % flatten the OMT tree until it has the expected height to be % inserted into the target tree (one subtree at a time). OmtTrees = lists:foldl(fun(_I, Acc) -> lists:append(Acc) end, OmtTree, lists:seq(1, abs(HeightDiff))), % (OmtHeight + HeightDiff) as OMT tree was flattened (HeighTDiff is % negative) MbrAndPos2 = seedtree_write_insert(Fd, Orig, OmtTrees, (OmtHeight + HeightDiff)), {NewMbrs, NewPos} = lists:unzip(MbrAndPos2), _MbrAndPos = seedtree_write_finish(lists:zip(NewMbrs, NewPos)) end, {new_leaf, NewChildrenPos2}; seedtree_write(Fd, [{Mbr, Meta, Children}|T], InsertHeight, Acc) -> {ok, Acc2} = case seedtree_write(Fd, Children, InsertHeight, []) of {no_more_siblings, Siblings} -> {ok, Siblings}; {leafs, Orig} -> {ok, [{Mbr, Meta, Orig}|Acc]}; % This one might return more that one new node => level_done needs % to handle a possible overflow {new_leaf, Nodes} -> Nodes2 = lists:foldl(fun({NodeMbr, NodePos}, Acc2) -> [{NodeMbr, Meta, NodePos}|Acc2] end, Acc, Nodes), {ok, Nodes2}; % Node can be written as-is, as the number of children is low enough {level_done, Level} when length(Level) =< ?MAX_FILLED-> ParentMbr = vtree:calc_nodes_mbr(Level), ChildrenPos = write_nodes(Fd, Level), % XXX vmx: Not sure if type=inner is always right Parent = {ParentMbr, #node{type=inner}, ChildrenPos}, {ok, [Parent|Acc]}; {level_done, Level} when length(Level) > ?MAX_FILLED -> {Level2, Level2Height} = omt_load(Level, ?MAX_FILLED), Level3 = if Level2Height > 2 -> % Lets's try to flatten the list in case it is too deep lists:foldl(fun(_I, ToAppend) -> lists:append(ToAppend) end, Level2, lists:seq(3, Level2Height)); true -> Level2 end, Parents = lists:foldl(fun(Level4, LevelAcc) -> ParentMbr = vtree:calc_nodes_mbr(Level4), ChildrenPos = write_nodes(Fd, Level4), % XXX vmx: Not sure if type=inner is always right Parent = {ParentMbr, #node{type=inner}, ChildrenPos}, [Parent|LevelAcc] end, Acc, Level3), {ok, Parents} end, {_, Acc3} = seedtree_write(Fd, T, InsertHeight, Acc2), {level_done, Acc3}. % @doc Writes new nodes into the existing tree. The nodes, resp. the % height of the OMT tree, must fit into the target tree. % "Orig" are the original child nodes, "New" are the nodes that should be % inserted. % Returns a tuple with the MBR and the position of the root node in the file. % The result may be a node that overflows. seedtree_write_finish/1 will fix % that. -spec seedtree_write_insert(Fd::file:io_device(), Orig::[vtree_node()], OmtTree::omt_node(), OmtHeight::integer()) -> [{mbr(), integer()}]. % Leaf node. We won't do the repacking dance, as we are on the lowest % level already. Thus we just insert the new nodes seedtree_write_insert(_Fd, Orig, OmtTree, OmtHeight) when is_tuple(hd(Orig)) -> OmtTree2 = if OmtHeight > 1 -> % Lets's try to flatten the list in case it is too deep lists:foldl(fun(_I, ToAppend) -> lists:append(ToAppend) end, OmtTree, lists:seq(2, OmtHeight)); true -> OmtTree end, NewNodes = Orig ++ OmtTree2, % create a list with tuples consisting of MBR and the actual node [{Mbr, Node} || {Mbr, _, _}=Node <- NewNodes]; % Inner node, do some repacking. seedtree_write_insert(Fd, Orig, OmtTree, _OmtHeight) -> % Write the OmtTree to to disk. % The tree might contain more than the maximum % number of allowed nodes per node. This overflow will be solved % at the end, when all child nodes get merged. {ok, OmtMbrAndPos} = omt_write_tree(Fd, OmtTree), % Get the enclosing MBR of the OMT nodes {OmtMbrs, OmtPosList} = lists:unzip(OmtMbrAndPos), OmtMbr = vtree:calc_mbr(OmtMbrs), % We do some repacking for better perfomance. First get the children % of the target node where the new nodes will be inserted in OrigNodes = load_nodes(Fd, Orig), % Transform the original nodes from normal ones to tuples where the % meta information is replaced with the position of the node in % the file. OrigMbrAndPos = lists:zipwith(fun({Mbr, _, Children}, Pos) -> {Mbr, Pos, Children} end, OrigNodes, Orig), % Get all nodes that are within or intersect with the input tree % root node(s) {Disjoint, NotDisjoint} = lists:partition(fun({Mbr, _Pos, _}) -> vtree:disjoint(Mbr, OmtMbr) end, OrigMbrAndPos), % For better performance, we now take the children of the nodes that will % be inserted, as well as all the children from the nodes that are not % disjoint. With this pool of nodes, we create new nodes that enclose % them more tightly. % First get the position of children of the not disjoint nodes NdChildrenPos = lists:append( [Children || {_Mbr, _Pos, Children} <- NotDisjoint]), % Get the position of the children of the nodes that will be inserted % as well. OmtChildren = load_nodes(Fd, OmtPosList), OmtChildrenPos = lists:append( [Children || {_Mbr, _Pos, Children} <- OmtChildren]), NewChildrenPos = NdChildrenPos ++ OmtChildrenPos, % Transform the new children's position to a tuple containing their MBR % and the position. NewChildrenMbrAndPos = case is_tuple(hd(NewChildrenPos)) of % We might have hit leaf nodes. They already contain the MBR we need to % find. true -> [{Mbr, Node} || {Mbr, _, _}=Node <- NewChildrenPos]; % Else we need to load the nodes to get their MBR false -> NewChildrenChildren = load_nodes(Fd, NewChildrenPos), lists:zipwith(fun({Mbr, _, _}, Pos) -> {Mbr, Pos} end, NewChildrenChildren, NewChildrenPos) end, % Create nodes that are packed to the maximum. This might return a several % levels deep OMT tree if there is a huge number of nodes. {NewNodesOmt, NewNodesHeight} = omt_load( NewChildrenMbrAndPos, ?MAX_FILLED), NewNodes = case NewNodesHeight of 1 -> [NewNodesOmt]; 2 -> NewNodesOmt; _ -> lists:foldl(fun(_I, Acc) -> lists:append(Acc) end, NewNodesOmt, lists:seq(3, NewNodesHeight)) end, % Write those new nodes NewChildren = lists:map(fun(Nodes) -> {Mbrs, Pos} = lists:unzip(Nodes), ParentMbr = vtree:calc_mbr(Mbrs), Meta = case is_tuple(hd(Pos)) of true -> #node{type=leaf}; false -> #node{type=inner} end, {ok, ParentPos} = couch_file:append_term(Fd, {ParentMbr, Meta, Pos}), ok = couch_file:sync(Fd), {ParentMbr, ParentPos} end, NewNodes), DisjointMbrAndPos = [{Mbr, Pos} || {Mbr, Pos, _} <- Disjoint], DisjointMbrAndPos ++ NewChildren. % @doc Creates new parent nodes out of a list of tuples containing MBRs and % positions (splits as needed). Returns MBR and position in file of the % new parent node(s) -spec seedtree_write_finish(NewChildren::[{mbr(), integer()}]) -> [{mbr(), integer()}]. seedtree_write_finish(NewChildren) -> NewChildren3 = case length(NewChildren) > ?MAX_FILLED of true -> {NewChildren2, OmtHeight} = omt_load(NewChildren, ?MAX_FILLED), if OmtHeight > 2 -> % Lets's try to flatten the list in case it is too deep lists:foldl(fun(_I, ToAppend) -> lists:append(ToAppend) end, NewChildren2, lists:seq(3, OmtHeight)); true -> NewChildren2 end; false -> [NewChildren] end, % Calculate the common MBR. The result is a tuple of the MBR and % the children positions in file it encloses NewChildrenMbrAndPos = lists:reverse(lists:foldl(fun(Nodes, Acc) -> {Mbrs, PosList} = lists:unzip(Nodes), % Poss might be actual nodes, if it is a leaf node. Then they % are already wrapped in a list. PosList2 = if is_list(hd(PosList)) -> hd(PosList); true -> PosList end, Mbr = vtree:calc_mbr(Mbrs), [{Mbr, PosList2}|Acc] end, [], NewChildren3)), % Return new nodes (might be several ones) NewChildrenMbrAndPos. % @doc Loads nodes from file -spec load_nodes(Fd::file:io_device(), Positions::[integer()]) -> [vtree_node()]. load_nodes(Fd, Positions) -> load_nodes(Fd, Positions, []). -spec load_nodes(Fd::file:io_device(), Positions::[integer()], Acc::[vtree_node()]) -> [vtree_node()]. load_nodes(_Fd, [], Acc) -> lists:reverse(Acc); load_nodes(Fd, [H|T], Acc) -> {ok, Node} = couch_file:pread_term(Fd, H), load_nodes(Fd, T, [Node|Acc]). % @doc Write nodes to file, return their positions -spec write_nodes(Fd::file:io_device(), Nodes::[vtree_node()]) -> [integer()]. write_nodes(Fd, Nodes) -> write_nodes(Fd, Nodes, []). -spec write_nodes(Fd::file:io_device(), Nodes::[vtree_node()], Acc::[vtree_node()]) -> [integer()]. write_nodes(_Fd, [], Acc) -> lists:reverse(Acc); write_nodes(Fd, [H|T], Acc) -> {ok, Pos} = couch_file:append_term(Fd, H), ok = couch_file:sync(Fd), write_nodes(Fd, T, [Pos|Acc]). % @doc Write a list of of nodes to disk with corresponding parent node. Return % the postion of the parent node in the file. -spec write_parent(Fd::file:io_device(), Nodes::[vtree_node()]) -> {ok, integer()}. write_parent(Fd, Nodes) -> ParentMbr = vtree:calc_nodes_mbr(Nodes), ChildrenPos = write_nodes(Fd, Nodes), {ok, ParentPos} = couch_file:append_term( Fd, {ParentMbr, #node{type=inner}, ChildrenPos}), ok = couch_file:sync(Fd), {ok, ParentPos}. % XXX vmx: insert_subtree and potentially other functions should be moved % from the vtree_bulk to the vtree module % @doc inserts a subtree into an vtree at a specific level. Returns the % MBR, position in the file of the new root node and the increase in height % of the tree. -spec insert_subtree(Fd::file:io_device(), RootPos::integer(), Subtree::[{mbr(), integer()}], Level::integer()) -> {ok, mbr(), integer(), integer()}. insert_subtree(Fd, RootPos, Subtree, Level) -> case insert_subtree(Fd, RootPos, Subtree, Level, 0) of {splitted, NodeMbr, {_Node1Mbr, NodePos1}, {_Node2Mbr, NodePos2}, Inc} -> Parent = {NodeMbr, #node{type=inner}, [NodePos1, NodePos2]}, {ok, Pos} = couch_file:append_term(Fd, Parent), ok = couch_file:sync(Fd), {ok, NodeMbr, Pos, Inc+1}; {ok, NewMbr, NewPos, Inc} -> {ok, NewMbr, NewPos, Inc} end. % @doc Returns either ok and MBR, position in file and the height of the tree, % or splitted and the enclosing MBR, MBR and position in file of the % individual nodes and the increase in height of the tree. -spec insert_subtree(Fd::file:io_device(), RootPos::integer(), Subtree::[{mbr(), integer()}], Level::integer(), Depth::integer()) -> {ok, mbr(), integer(), integer()} | {splitted, mbr(), {mbr(), integer()}, {mbr(), integer()}, integer()}. insert_subtree(Fd, RootPos, Subtree, Level, Depth) when Depth==Level -> {SubtreeMbrs, SubtreePosList} = lists:unzip(Subtree), {ok, Parent} = couch_file:pread_term(Fd, RootPos), {ParentMbr, ParentMeta, EntriesPos} = Parent, MergedMbr = vtree:calc_mbr([ParentMbr|SubtreeMbrs]), ChildrenPos = SubtreePosList ++ EntriesPos, if length(ChildrenPos) =< ?MAX_FILLED -> NewNode = {MergedMbr, ParentMeta, ChildrenPos}, {ok, Pos} = couch_file:append_term(Fd, NewNode), ok = couch_file:sync(Fd), {ok, MergedMbr, Pos, 0}; true -> Children = load_nodes(Fd, ChildrenPos), % Transform the the nodes, so that we don't need to write them again. % They are now a tuple with their MBR, Meta and *their* position % in the file (as opposed to the position of their children) Children2 = [{Mbr, Meta, Pos} || {{Mbr, Meta, _}, Pos} <- lists:zip(Children, ChildrenPos)], % The maximum number of nodes can be 2*MAX_FILLED, therefore we can't % use the Ang/Tan split algorithm, but use the OMT algorithm to split % the nodes into two partitions. {[Part1, Part2], _OmtHeight} = omt_load(Children2, ?MAX_FILLED), Mbr1 = vtree:calc_nodes_mbr(Part1), Mbr2 = vtree:calc_nodes_mbr(Part2), % Get the original position in file of the nodes Part1Pos = [Pos || {_Mbr, _Meta, Pos} <- Part1], Part2Pos = [Pos || {_Mbr, _Meta, Pos} <- Part2], Node1 = {Mbr1, #node{type=inner}, Part1Pos}, Node2 = {Mbr2, #node{type=inner}, Part2Pos}, {ok, Pos1} = couch_file:append_term(Fd, Node1), {ok, Pos2} = couch_file:append_term(Fd, Node2), ok = couch_file:sync(Fd), {splitted, MergedMbr, {Mbr1, Pos1}, {Mbr2, Pos2}, 0} end; insert_subtree(Fd, RootPos, Subtree, Level, Depth) -> %{SubtreeMbr, SubtreePos} = Subtree, {SubtreeMbrs, _SubtreePosList} = lists:unzip(Subtree), % Calculate the MBR that enclosed both nodes, as both nodes should end up % in the same target node SubtreeMbr = vtree:calc_mbr(SubtreeMbrs), {ok, Parent} = couch_file:pread_term(Fd, RootPos), {ParentMbr, _ParentMeta, EntriesPos} = Parent, {{_LeastMbr, LeastPos}, LeastRest} = least_expansion( Fd, SubtreeMbr, EntriesPos), LeastRestPos = [Pos || {_Mbr, Pos} <- LeastRest], case insert_subtree(Fd, LeastPos, Subtree, Level, Depth+1) of {ok, NewMbr, NewPos, Inc} -> MergedMbr = vtree:merge_mbr(ParentMbr, NewMbr), NewNode = {MergedMbr, #node{type=inner}, [NewPos|LeastRestPos]}, {ok, Pos} = couch_file:append_term(Fd, NewNode), ok = couch_file:sync(Fd), {ok, NewMbr, Pos, Inc}; {splitted, ChildMbr, {Child1Mbr, ChildPos1}, {Child2Mbr, ChildPos2}, Inc} -> MergedMbr = vtree:merge_mbr(ParentMbr, ChildMbr), LeastRestPos = [Pos || {_Mbr, Pos} <- LeastRest], if % Both nodes of the split fit in the current inner node %length(EntriesPos)+2 =< ?MAX_FILLED -> length(LeastRestPos)+2 =< ?MAX_FILLED -> ChildrenPos = [ChildPos1, ChildPos2] ++ LeastRestPos, NewNode = {MergedMbr, #node{type=inner}, ChildrenPos}, {ok, Pos} = couch_file:append_term(Fd, NewNode), ok = couch_file:sync(Fd), {ok, MergedMbr, Pos, Inc}; % We need to split the inner node true -> Child1 = {Child1Mbr, #node{type=inner}, ChildPos1}, Child2 = {Child2Mbr, #node{type=inner}, ChildPos2}, % NOTE vmx: for vtree:split/1 the nodes need to have a special % format a tuple with their MBR, Meta and *their* position in % the file (as opposed to the position of their children) Children = [{Mbr, Meta, Pos} || {{Mbr, Meta, _}, Pos} <- lists:zip(load_nodes(Fd, LeastRestPos), LeastRestPos)], Children2 = [Child1, Child2] ++ Children, {SplittedMbr, {Node1Mbr, _, _}=Node1, {Node2Mbr, _, _}=Node2} = vtree:split_node({MergedMbr, #node{type=inner}, Children2}), {ok, Pos1} = couch_file:append_term(Fd, Node1), {ok, Pos2} = couch_file:append_term(Fd, Node2), ok = couch_file:sync(Fd), {splitted, SplittedMbr, {Node1Mbr, Pos1}, {Node2Mbr, Pos2}, Inc} end end. % XXX vmx: needs tests % @doc Find the node that needs least expension for the given MBR and returns % the MBR and position of the node together with the original children MBRs and % positions in the file. -spec least_expansion(Fd::file:io_device(), NewMbr::mbr(), PosList::[integer()]) -> {{mbr, integer()}, [{mbr, integer()}]}. least_expansion(Fd, NewMbr, PosList) -> Nodes = load_nodes(Fd, PosList), NodesAndPos = lists:zip(Nodes, PosList), MbrAndPos = [{Mbr, Pos} || {{Mbr, _, _}, Pos} <- NodesAndPos], % Find node that matches best (least expansion of MBR) {_, _, Nth} = lists:foldl(fun({Mbr, _Pos}, {MinExp, Nth2, Cnt}) -> MergedMbr = vtree:merge_mbr(NewMbr, Mbr), % if there is a element which need less expansion, put the info % into the accumulator case vtree:area(MergedMbr) - vtree:area(Mbr) of Exp when Exp < MinExp orelse (MinExp==-1) -> {Exp, Cnt, Cnt+1}; _ -> {MinExp, Nth2, Cnt+1} end end, {-1, 0, 0}, MbrAndPos), % Remove the child where the node will be inserted {C1, C2} = lists:split(Nth-1, MbrAndPos), {hd(C2), C1 ++ tl(C2)}. %%%%% Helpers %%%%% % @doc Returns the ceiling of log_N(X). Returns 1 for X==1. -spec log_n_ceil(N::integer(), X::integer()) -> integer(). log_n_ceil(_N, 1) -> 1; log_n_ceil(N, X) -> ceiling(math:log(X) / math:log(N)). % From http://www.trapexit.org/Floating_Point_Rounding (2010-10-18) -spec ceiling(X::number()) -> integer(). ceiling(X) when X < 0 -> trunc(X); ceiling(X) -> T = trunc(X), case X - T == 0 of true -> T; false -> T + 1 end. % From http://www.trapexit.org/Floating_Point_Rounding (2010-10-18) -spec floor(X::number()) -> integer(). floor(X) when X < 0 -> T = trunc(X), case X - T == 0 of true -> T; false -> T - 1 end; floor(X) -> trunc(X). % @doc split a list of elements into equally sized chunks (last element might % contain less elements). Applies Fun to every chunk and inserts the % return value of this function into the result list. -spec chunk_list(Fun::fun(), List::list(), Size::integer()) -> [list()]. chunk_list(_Fun, List, Size) when Size == 0 -> [List]; chunk_list(Fun, List, Size) -> chunk_list(Fun, List, Size, 0, [], []). % List is processed, but some items are left in the last chunk chunk_list(Fun, [], _Size, _Cnt, Chunk, Result) when length(Chunk) > 0 -> Entry = Fun(lists:reverse(Chunk)), %lists:reverse(Entry ++ Result); lists:reverse([Entry|Result]); chunk_list(_Fun, [], _Size, _Cnt, _Chunk, Result) -> lists:reverse(Result); chunk_list(Fun, [H|T], Size, Cnt, Chunk, Result) when Cnt < Size -> chunk_list(Fun, T, Size, Cnt+1, [H|Chunk], Result); % Chunk is complete, add chunk to result chunk_list(Fun, [H|T], Size, _Cnt, Chunk, Result) -> Entry = Fun(lists:reverse(Chunk)), %chunk_list(Fun, T, Size, 1, [H], Entry ++ Result). chunk_list(Fun, T, Size, 1, [H], [Entry|Result]).
src/vtree_bulk.erl
0.601594
0.516047
vtree_bulk.erl
starcoder
%% The contents of this file are subject to the Mozilla Public License %% Version 1.1 (the "License"); you may not use this file except in %% compliance with the License. You may obtain a copy of the License %% at http://www.mozilla.org/MPL/ %% %% Software distributed under the License is distributed on an "AS IS" %% basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See %% the License for the specific language governing rights and %% limitations under the License. %% %% The Original Code is Erlando. %% %% The Initial Developer of the Original Code is VMware, Inc. %% Copyright (c) 2011-2013 VMware, Inc. All rights reserved. %% %% This is the Omega monad which is like the list monad, but does not %% depth first, and not breadth first traversal. This implementation %% is based on <NAME>'s Control.Monad.Omega module for Haskell %% (http://hackage.haskell.org/packages/archive/control-monad-omega/latest/doc/html/Control-Monad-Omega.html). As %% the documentation there states: %% %% Warning: Omega is only a monad when the results are interpreted %% as a set; that is, a valid transformation according to the monad %% laws may change the order of the results. However, the same set %% of results will always be reachable. %% -module(omega_m). -behaviour(monad). -export(['>>='/2, return/1, fail/1]). -export([diagonal/1]). -behaviour(monad_plus). -export([mzero/0, mplus/2]). -spec '>>='([A], fun( (A) -> [B] )) -> [B]. '>>='(X, Fun) -> diagonal([Fun(E) || E <- X]). -spec return(A) -> [A]. return(X) -> [X]. -spec fail(any()) -> [_A]. fail(_X) -> []. -spec mzero() -> [_A]. mzero() -> []. -spec mplus([A], [A]) -> [A]. mplus(X, Y) -> lists:append(X, Y). %% [[a, b, c, d], %% [e, f, g, h], %% [i, j, k, l], %% [m, n, o, p]]. %% %% diagonal traverses diagonally from north-west corner, heading east %% then south-west. I.e. %% [a, b, e, c, f, i, d, g, j, m, h, k, n, l, o, p] -spec diagonal([[A]]) -> [A]. diagonal(LoL) -> lists:append(stripe(LoL)). stripe([]) -> []; stripe([[]|Xss]) -> stripe(Xss); stripe([[X|Xs]|Xss]) -> [[X] | zip_cons(Xs, stripe(Xss))]. zip_cons([], Ys) -> Ys; zip_cons(Xs, []) -> [[X] || X <- Xs]; zip_cons([X|Xs], [Y|Ys]) -> [[X|Y] | zip_cons(Xs, Ys)].
apps/erlando/src/omega_m.erl
0.696062
0.472927
omega_m.erl
starcoder
%%%------------------------------------------------------------------- %%% Copyright: (c) 2007-2010 Gemini Mobile Technologies, Inc. All rights reserved. %%% %%% Licensed under the Apache License, Version 2.0 (the "License"); %%% you may not use this file except in compliance with the License. %%% You may obtain a copy of the License at %%% %%% http://www.apache.org/licenses/LICENSE-2.0 %%% %%% Unless required by applicable law or agreed to in writing, software %%% distributed under the License is distributed on an "AS IS" BASIS, %%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %%% See the License for the specific language governing permissions and %%% limitations under the License. %%% %%% File : brick_hash.erl %%% Purpose : brick hashing upper layer. %%%------------------------------------------------------------------- %% @doc An upper (?) layer for distributed storage hashing. %% %% I've attempted to split my grand vision for distributed storage into %% two different layers: %% <ul> %% <li> A lower storage level of an individual brick. Each brick should %% be completely unaware of other bricks, to the full extent possible. </li> %% <li> An upper storage level that uses hashing techniques to determine %% which keys are stored on any particular storage brick. </li> %% </ul> %% %% This upper layer has two mostly-separate parts: %% <ul> %% <li> A mapping from key -&gt; chain name. </li> %% <li> A mapping from chain name -&gt; individual bricks. </li> %% </ul> %% %% == What is a #hash_r record? == %% %% A `#hash_r' record is used for two things: %% <ul> %% <li> The algorithm used to choose what part of the key will be used %% for hashing: e.g. fixed-length prefix, variable length prefix </li> %% <li> The algorithm for the actual hashing: e.g. naive. </li> %% </ul> %% %% NOTE: Those two tasks should be quite separate and, at the moment, %% are blurred together in (sometimes) bad ways. TODO: refactor! %% %% == Note #hash_r init functions == %% %% The #hash_r record initialization functions, e.g. naive_init() and %% var_prefix_init(), must provide a couple of well-known arities to %% external users: %% <ul> %% <li> (ChainList) ... specifying only the chains' definition. All %% other options, if any, are set to default. </li> %% <li> (via_proplist, ChainList, PropList) ... The proplist method is %% used (at first) by brick_admin:add_table(). Such (ab)use of %% proplists is a common technique for presenting a fixed-arity %% API while allowing extremely flexibility in implementing same. </li> %% </ul> %% %% == What is a #g_hash_r record? == %% %% A `#g_hash_r' record contains "global" hashing info for a %% <b>single GDSS table</b>. (Perhaps "global" is a bad word? {shrug}) %% %% The `#g_hash_r' is a collection of two `#hash_r' records, plus some %% extra stuff. The two `#hash_r' records describe: %% <ul> %% <li> The hashing scheme used in the current hash configuration. </li> %% <li> The hashing scheme used in the <b>new</b> hash configuration. </li> %% </ul> %% %% Most of the time, the current and new hash configurations are the %% same. However, if the cluster configuration changes, then the %% difference betwen current &amp; new hashes are used to "migrate" %% data from the current scheme to the new scheme. %% %% The `#g_hash_r' is data that is distributed to all the %% `brick_simple' gen_servers (one per client node), to assist with %% `brick_simple' calculation of {table + key} =&gt; brick mapping. %% The `#g_hash_r' is also distributed to all bricks, so that they can %% independently verify that that brick is truly responsible for the %% keys sent to it. If the client's `#g_hash_r' info is stale, that's %% OK: the incorrect brick will forward the request to the correct %% brick. %% %% == Note on #g_hash_r.phase State Names == %% %% The two valid state names for #g_hash_r.phase are: %% <ul> %% <li> pre ... the brick is prepared to sweep itself and to receive sweep %% updates from other bricks, but should initiate no migration %% action.</li> %% <li> migrating ... The brick is actively migrating. It can enter this %% state (from 'pre') by: %% <ul> %% <li> Being explicitly told to start sweeping itself. </li> %% <li> Receiving a sweep update from another brick B, because %% that brick got the "start sweeping" command before we did. </li> %% </ul> %% </li> %% </ul> %% Bugs found with Quickcheck: %% * discovered URI parsing prefix set to $_ (ASCII 95) instead of %% expected $/ (ASCII 47), after configuring tab1 to use 2 chains instead %% of the default 1 chain. -module(brick_hash). -include("applog.hrl"). -include("brick.hrl"). -include("brick_hash.hrl"). -include("brick_public.hrl"). -ifdef(debug_hash). -define(gmt_debug, true). -endif. -include("gmt_debug.hrl"). -define(PHASH2_BIGGEST_VAL, 134217727). % (2^27)-1 -define(SMALLEST_SIGNIFICANT_FLOAT_SIZE, 0.1e-12). %% This is one level of hash management: %% mapping from key -> chain name. %% Exports for first hash method, 'naive'. -export([naive_init/1, naive_init/2, naive_key_to_chain/3]). %% Exports for second hash method, 'var_prefix'. %% It hashes the same way as 'naive', but only operates on a %% variable-length prefix string. -export([var_prefix_init/1, var_prefix_init/2, var_prefix_init/3, var_prefix_key_to_chain/3]). %% Exports another hash method, 'fixed_prefix_numhack'. %% It hashes similarly to 'var_prefix', but only operates on a %% uses a constant-length prefix *and* -export([fixed_prefix_init/1, fixed_prefix_init/3, fixed_prefix_key_to_chain/3]). %% API for consistent hashing. We'll do what we should've done %% with previous efforts: separate: %% * mapping of key -> raw hash val %% * mapping of raw hash val -> chain name or brick name. -export([chash_init/3, chash_key_to_chain/3]). -export([chash_make_float_map/1, chash_make_float_map/2, chash_float_map_to_nextfloat_list/1, chash_nextfloat_list_to_gb_tree/1, chash_gb_next/2, chash_scale_to_int_interval/2, % debugging func chash_extract_new_float_map/1, chash_extract_old_float_map/1, chash_extract_new_chainweights/1, chash_extract_old_props/1]). %% This is the other level of hash management: %% mapping from chain names -> individual bricks. %% %% This layer does *not* make any key -> chain name mapping decisions. -export([add_migr_dict_if_missing/1, init_global_hash_state/7, init_global_hash_state/8, key_to_brick/3, key_to_chain/2, % Not for general use chain2brick/3, chain2brick/4 ]). %% API for miscellaneous/helper functions. -export([set_chain_sweep_key/3, all_chains/2, update_chain_dicts/3, desc_substitute_chain/3, verify_chain_list/2 ]). %% Useful mostly for debugging: invent a list of chains that will all %% live on the same node. -export([invent_nodelist/2, invent_nodelist/4]). %% Chain descriptor -record(chaindesc_r, { name, length, head, middles, tail }). %% Opaque record for naive method. %% TODO? Move naive functions & records into separate module? -record(naive, { num_chains, map }). %% Opaque record for the variable-length prefix method. -record(var_prefix, { num_chains, map, separator, num_separators }). %% Opaque record for the fixed-length prefix method. -record(fixed_prefix, { num_chains, map, prefix_len, prefix_is_integer_hack_p }). %% Opaque record for consistent hashing method. -record(chash, { num_chains, map, %% Key prefix mash-up prefix_method, % all | var_prefix | fixed_prefix, separator, % for var prefix num_separators, % for var prefix prefix_len, % for fixed-length prefix prefix_is_integer_hack_p, % for fixed-length prefix %% chash guts chash_map, % tuple(ChainName) old_float_map, % float_map() new_float_map, % float_map() new_chainweights, % chainweight_list() %% administrivia old_props % proplist() }). -export_type([chash_r/0, fixed_prefix_r/0, naive_r/0, var_prefix_r/0]). %% declare types for easy ref showing _r as record -type naive_r() :: #naive{}. -type var_prefix_r() :: #var_prefix{}. -type fixed_prefix_r() :: #fixed_prefix{}. -type chash_r() :: #chash{}. %% QQQ Re-write call sig! -type brick() :: {atom(),atom()}. -type bricklist() :: [brick()]. -type chainlist() :: [{atom(),bricklist()}]. -type phase() :: pre | migrating | post | phase_unused | unused. -type chain_error() :: chain2brick_error | chain_is_zero_length. -spec add_migr_dict_if_missing(#g_hash_r{}) -> #g_hash_r{}. -spec all_chains(#g_hash_r{}, current | new ) -> [{any(),any()}]. -spec chain2brick(atom(),read|write,#g_hash_r{}) -> chain_error() |brick(). -spec chain2brick(atom(),read|write,#g_hash_r{},current|new) -> chain_error()|brick(). -spec chash_extract_old_props(#hash_r{}|#g_hash_r{}) -> [{any(),any()}]. -spec chash_init(via_proplist,[{atom(),[atom()]}],[{any(),any()}]) -> #hash_r{}. -spec desc_substitute_chain(#hash_r{},atom(),[atom()]) -> #hash_r{}. -spec doo_iter_chain([{any(),float()}],integer(),integer()) -> ok. -spec fixed_prefix_init(chainlist()) -> #hash_r{}. -spec fixed_prefix_init(via_proplist,chainlist(),[{any(),any()}]) -> #hash_r{}. -spec init_global_hash_state(atom(), phase(), integer(),#hash_r{},chainlist(),#hash_r{},chainlist()) -> #g_hash_r{}. -spec init_global_hash_state(atom(), phase(), integer(), #hash_r{}, chainlist(),#hash_r{},chainlist(),true|false) -> #g_hash_r{}. -spec key_to_brick(read|write,term(),#g_hash_r{}) -> chain_error()| brick(). -spec key_to_chain(term(),#g_hash_r{}) -> atom(). -spec naive_init(chainlist(),true|false) -> #hash_r{}. -spec set_chain_sweep_key(atom(),term(),#g_hash_r{}) -> #g_hash_r{}. -spec update_chain_dicts(#g_hash_r{},atom(),[any()]) -> #g_hash_r{}. -spec verify_chain_list(chainlist()) -> ok. -spec verify_chain_list(chainlist(),true|false) -> ok. init_global_hash_state(MigratingP, Phase, CurrentRev, CurrentHashDesc, CurrentChainList, NewHashDesc, NewChainList) -> init_global_hash_state(MigratingP, Phase, CurrentRev, CurrentHashDesc, CurrentChainList, NewHashDesc, NewChainList, false). init_global_hash_state(MigratingP, Phase, CurrentRev, CurrentHashDesc, CurrentChainList, NewHashDesc, NewChainList, ZeroLengthOK) when is_integer(CurrentRev), CurrentRev > 0, is_record(CurrentHashDesc, hash_r), is_record(NewHashDesc, hash_r) -> ok = verify_chain_list(CurrentChainList, ZeroLengthOK), ok = verify_chain_list(NewChainList, ZeroLengthOK), #g_hash_r{migrating_p = MigratingP, phase = if Phase == pre; Phase == migrating; Phase == post -> Phase; true -> pre end, current_rev = CurrentRev, current_h_desc = CurrentHashDesc, new_h_desc = NewHashDesc, current_chain_dict = chain_list2dict(CurrentChainList), new_chain_dict = chain_list2dict(NewChainList), migr_dict = if MigratingP -> dict:new(); true -> undefined end }. add_migr_dict_if_missing(GH) when is_record(GH, g_hash_r) -> if GH#g_hash_r.migr_dict == undefined -> GH#g_hash_r{migr_dict = dict:new()}; true -> GH end. %% @spec (read | write, term(), g_hash_r()) -> {brick(), node()} | error | exit key_to_brick(ReadOrWrite, Key, GH) -> key_to_chainsweep(ReadOrWrite, Key, GH, fun chain2brick/4). key_to_chainsweep(ReadOrWrite, Key, GH, FinalFun) when ReadOrWrite == read; ReadOrWrite == write -> #g_hash_r{current_rev = CurRev, minor_rev = MinorRev, current_h_desc = H} = GH, ChainName = (H#hash_r.mod):(H#hash_r.func)( ReadOrWrite, Key, H#hash_r.opaque), if GH#g_hash_r.migrating_p == false orelse (GH#g_hash_r.migrating_p == true andalso GH#g_hash_r.phase == pre) -> B = FinalFun(ChainName, ReadOrWrite, GH, current), Debs = [{migrating_p, GH#g_hash_r.migrating_p}, {phase, GH#g_hash_r.phase}], ?DBG_HASHx({ChainName, Key, no_mig, CurRev, MinorRev, B, Debs}), B; true -> HNew = GH#g_hash_r.new_h_desc, ChainNameNew = (HNew#hash_r.mod):(HNew#hash_r.func)( ReadOrWrite, Key, HNew#hash_r.opaque), if ChainName == ChainNameNew -> B = FinalFun(ChainName, ReadOrWrite, GH, current), ?DBG_HASHx({Key, mig_same, CurRev, MinorRev, B}), B; true -> case where_is_key_relative_to_sweep(Key, ChainName, GH) of in_front -> B = FinalFun(ChainName, ReadOrWrite, GH, current), ?DBG_HASHx({Key, mig_current, CurRev, MinorRev, B}), B; behind -> B = FinalFun(ChainNameNew, ReadOrWrite, GH, new), ?DBG_HASHx({Key, mig_new, CurRev, MinorRev, B}), B end end end. %% @spec (term(), g_hash_r()) -> atom() | error | exit key_to_chain(Key, GH) -> key_to_chainsweep(read, Key, GH, fun (ChainName, _, _, _) -> ChainName end). chain2brick(ChainName, ReadOrWrite, GH) -> chain2brick(ChainName, ReadOrWrite, GH, current). chain2brick(ChainName, ReadOrWrite, GH, WhichDict) -> Dict = if WhichDict == current -> GH#g_hash_r.current_chain_dict; WhichDict == new -> GH#g_hash_r.new_chain_dict end, case dict:find(ChainName, Dict) of {ok, CH} -> %% TODO: What to do if CH#chaindesc_r.length == 0? if ReadOrWrite == read -> CH#chaindesc_r.tail; true -> CH#chaindesc_r.head end; _ -> chain2brick_error end. %% @spec (atom(), term(), g_hash_r()) -> g_hash_r() %% @doc Store the new migration sweep key Key for chain ChainName in the %% global hash's migration dictionary. set_chain_sweep_key(ChainName, Key, GH) when is_record(GH, g_hash_r) -> NewD = dict:store(ChainName, Key, GH#g_hash_r.migr_dict), GH#g_hash_r{migr_dict = NewD}; set_chain_sweep_key(_ChainName, _Key, GH) -> %% Just return GH as-is. Odds are high that GH = undefined. %% If GH = undefined, we're in a pathological situation where we %% don't know what the global hash is. If that's true, we definitely %% cannot update it, but crashing here isn't a good idea either, so %% we will just pass it through. GH. %% @spec (g_hash_r(), current | new) -> list({atom(), list()}) %% @doc Return the list of all {chain name, brick list} (active bricks only!) %% for the current/new hash descriptor in a global hash record. all_chains(GH, WhichDescriptor) -> Dict = if WhichDescriptor == current -> GH#g_hash_r.current_chain_dict; WhichDescriptor == new -> GH#g_hash_r.new_chain_dict end, [{ChainName, if D#chaindesc_r.head == D#chaindesc_r.tail -> [D#chaindesc_r.head]; true -> [D#chaindesc_r.head] ++ D#chaindesc_r.middles ++ [D#chaindesc_r.tail] end} || {ChainName, D} <- dict:to_list(Dict)]. update_chain_dicts(GH, ChainName, BrickList) -> Desc = make_chaindesc(ChainName, BrickList), NewCur = case dict:find(ChainName, GH#g_hash_r.current_chain_dict) of {ok, _D1} -> dict:store(ChainName,Desc,GH#g_hash_r.current_chain_dict); _ -> GH#g_hash_r.current_chain_dict end, NewNew = case dict:find(ChainName, GH#g_hash_r.new_chain_dict) of {ok, _D2} -> dict:store(ChainName,Desc,GH#g_hash_r.new_chain_dict); _ -> GH#g_hash_r.new_chain_dict end, MinorRev = GH#g_hash_r.minor_rev + 1, GH#g_hash_r{minor_rev = MinorRev, current_chain_dict = NewCur, new_chain_dict = NewNew}. desc_substitute_chain(Desc, ChainName, BrickList) -> case proplists:get_value(ChainName, Desc#hash_r.healthy_chainlist) of undefined -> Desc; _ -> NewCL = lists:keyreplace( ChainName, 1, Desc#hash_r.healthy_chainlist, {ChainName, BrickList}), if Desc#hash_r.method == naive -> naive_init(NewCL); Desc#hash_r.method == var_prefix -> Separator = (Desc#hash_r.opaque)#var_prefix.separator, NumSeps = (Desc#hash_r.opaque)#var_prefix.num_separators, var_prefix_init(NewCL, false, Separator, NumSeps); Desc#hash_r.method == fixed_prefix -> Len = (Desc#hash_r.opaque)#fixed_prefix.prefix_len, HackP = (Desc#hash_r.opaque)#fixed_prefix.prefix_is_integer_hack_p, fixed_prefix_init(via_proplist, NewCL, [{prefix_length, Len}, {prefix_is_integer_hack, HackP}]); Desc#hash_r.method == chash -> Props = (Desc#hash_r.opaque)#chash.old_props, chash_init(via_proplist, NewCL, Props) end end. where_is_key_relative_to_sweep(Key, ChainName, GH) -> case dict:find(ChainName, GH#g_hash_r.migr_dict) of {ok, SweepKey} -> ?DBG_HASHx({relative_to_sweep, Key, ChainName, found, SweepKey}), if SweepKey == ?BRICK__GET_MANY_LAST -> behind; SweepKey == ?BRICK__GET_MANY_FIRST -> in_front; Key =< SweepKey -> behind; true -> in_front end; _ -> ?DBG_HASHx({relative_to_sweep, Key, ChainName, not_found}), in_front end. %% QQQ Hrm, I'm thinking that it's going to be difficult to standardize the %% API for a specific hash algorithm's init() func? %% @spec (list({brick_name(), list(node_name())})) -> hash_r() naive_init(ChainList) -> naive_init(ChainList, false). naive_init(ChainList, ZeroLengthOK) -> ok = verify_chain_list(ChainList, ZeroLengthOK), Opaque = #naive{num_chains = length(ChainList), map = list_to_tuple(ChainList)}, #hash_r{method = naive, mod = ?MODULE, %%REAL VALUE: func = naive_key_to_chain, %%DEBUG VALUE: func = naive_key_to_chain_DEBUG, %DEBUGGING ONLY!!! func = naive_key_to_chain, healthy_chainlist = ChainList, opaque = Opaque}. naive_key_to_chain(_ReadOrWrite, Key, Opaque) -> N = naive_hash(Key), ChainI = (N rem Opaque#naive.num_chains) + 1, {ChainName, _ChainMembers} = element(ChainI, Opaque#naive.map), ChainName. naive_hash(Key) -> erlang:phash2(Key). naive_key_to_chain_DEBUG(_ReadOrWrite, Key, Opaque) -> IgnoreBytes = size(Key) - 1, <<_:IgnoreBytes/binary, Last:8>> = Key, N = Last - $0, ChainI = (N rem Opaque#naive.num_chains) + 1, {ChainName, _ChainMembers} = element(ChainI, Opaque#naive.map), ChainName. %% @spec (list({brick_name(), list(node_name())})) -> hash_r() var_prefix_init(ChainList) -> var_prefix_init(ChainList, false). var_prefix_init(ChainList, ZeroLengthOK) -> var_prefix_init(ChainList, ZeroLengthOK, ?HASH_PREFIX_SEPARATOR, 1). %% @spec (via_proplist, ChainList::chain_list(), Props::prop_list()) -> hash_r() %% @doc Return a #hash_r record for a variable-length prefix hashing scheme, %% throwing an exception if sanity checks fail. %% %% When using the via_proplist version of this function, the following %% properties are examined: %% <ul> %% <li> zero_length_ok ... should never be true when defining a %% new table. </li> %% <li> prefix_separator ... the single byte ASCII value of the byte %% that separates the key's prefix from the rest of the key. %% Default is ?HASH_PREFIX_SEPARATOR. </li> %% <li> num_separators ... number of prefix_separator components that are %% included in the key's prefix. %% Default is 2 (assuming key looks like "/prefix/not-prefix-stuff").</li> %% </ul> var_prefix_init(via_proplist, ChainList, Props) -> ZLok = proplists:get_value(zero_length_ok, Props, false), Sep = proplists:get_value(prefix_separator, Props, ?HASH_PREFIX_SEPARATOR), NumSeps = proplists:get_value(num_separators, Props, 2), var_prefix_init(ChainList, ZLok, Sep, NumSeps). var_prefix_init(ChainList, ZeroLengthOK, PrefixSeparator, NumSeps) -> ok = verify_chain_list(ChainList, ZeroLengthOK), Opaque = #var_prefix{num_chains = length(ChainList), map = list_to_tuple(ChainList), separator = PrefixSeparator, num_separators = NumSeps}, #hash_r{method = var_prefix, mod = ?MODULE, func = var_prefix_key_to_chain, healthy_chainlist = ChainList, opaque = Opaque}. var_prefix_key_to_chain(_ReadOrWrite, Key, Opaque) -> N = var_prefix_hash(Key, Opaque#var_prefix.separator, Opaque#var_prefix.num_separators), ChainI = (N rem Opaque#var_prefix.num_chains) + 1, {ChainName, _ChainMembers} = element(ChainI, Opaque#var_prefix.map), ChainName. var_prefix_hash(Key, Separator, NumSeps) -> HashKey = find_var_prefix(Key, Separator, 0, NumSeps), erlang:phash2(HashKey). %% @spec (list({brick_name(), list(node_name())})) -> hash_r() fixed_prefix_init(ChainList) -> fixed_prefix_init(via_proplist, ChainList, []). %% @spec (via_proplist, chain_list(), prop_list()) -> hash_r() %% @doc Return a #hash_r record for a variable-length prefix hashing scheme, %% throwing an exception if sanity checks fail. %% %% When using the via_proplist version of this function, the following %% properties are examined: %% <ul> %% <li> zero_length_ok ... should never be true when defining a %% new table. </li> %% <li> prefix_length ... length of the prefix, default = 4 bytes. </li> %% <li> prefix_is_integer_hack ... if true, the prefix should be interpreted %% as an ASCII representation of a base 10 integer for use as the %% hash calculation. </li> %% </ul> fixed_prefix_init(via_proplist, ChainList, Props) -> ZeroLengthOK = proplists:get_value(zero_length_ok, Props, false), Len = proplists:get_value(prefix_length, Props, 4), PrefixIsIntegerP = proplists:get_value(prefix_is_integer_hack, Props, false), ok = verify_chain_list(ChainList, ZeroLengthOK), Opaque = #fixed_prefix{num_chains = length(ChainList), map = list_to_tuple(ChainList), prefix_len = Len, prefix_is_integer_hack_p = PrefixIsIntegerP}, #hash_r{method = fixed_prefix, mod = ?MODULE, func = fixed_prefix_key_to_chain, healthy_chainlist = ChainList, opaque = Opaque}. fixed_prefix_key_to_chain(_ReadOrWrite, Key, Opaque) -> N = fixed_prefix_hash(Key, Opaque#fixed_prefix.prefix_len, Opaque#fixed_prefix.prefix_is_integer_hack_p), ChainI = (N rem Opaque#fixed_prefix.num_chains) + 1, {ChainName, _ChainMembers} = element(ChainI, Opaque#fixed_prefix.map), ChainName. fixed_prefix_hash(Key, PrefixLen, IntegerHackP) -> Len = PrefixLen, Prefix = if is_binary(Key), size(Key) > Len -> <<P:Len/binary, _/binary>> = Key, P; is_list(Key), length(Key) > Len -> lists:sublist(Key, Len); true -> Key end, if IntegerHackP -> list_to_integer(gmt_util:list_ify(Prefix)); true -> erlang:phash2(Prefix) end. %% @spec (via_proplist, chain_list(), prop_list()) -> hash_r() %% @doc Return a #hash_r record for the consistent hashing scheme, %% throwing an exception if sanity checks fail. %% %% The following properties are examined for key preprocessing: %% <ul> %% <li> prefix_method ... all | var_prefix | fixed_prefix. This property %% is mandatory. Its value will affect whether or not the other %% properties in this section must also be present. </li> %% <li> zero_length_ok ... should never be true when defining a %% new table. If undefined, assumed to be false. </li> %% <li> prefix_length ... length of the prefix, default = 4 bytes. </li> %% <li> prefix_is_integer_hack ... if true, the prefix should be interpreted %% as an ASCII representation of a base 10 integer for use as the %% hash calculation. </li> %% <li> prefix_separator ... the single byte ASCII value of the byte %% that separates the key's prefix from the rest of the key. %% Default is ?HASH_PREFIX_SEPARATOR. </li> %% <li> num_separators ... number of prefix_separator components that are %% included in the key's prefix. %% Default is 2 (assuming key looks like "/prefix/not-prefix-stuff").</li> %% </ul> %% %% The following properties are used for the consistent hash map creation: %% <ul> %% <li> old_float_map ... The old/current float_map. If missing, we assume %% we're starting from scratch and will use [{unused, 1.0}]. </li> %% <li> new_chainweights ... list({ChainName::atom(), Weight::integer()}). %% This item is mandatory: an obscure exception will be thrown if %% it is missing. </li> %% </ul> %% chash_init(via_proplist, ChainList, Props) -> ZeroLengthOK = proplists:get_value(zero_length_ok, Props, false), ok = verify_chain_list(ChainList, ZeroLengthOK), ChainListChains = lists:usort([Ch || {Ch, _Brs} <- ChainList]), NewWeights = proplists:get_value(new_chainweights, Props, 'missing_bad!'), WeightChains = lists:usort([Ch || {Ch, _Wt} <- NewWeights]), if ChainListChains == WeightChains -> ok end, % sanity %% Key prefix config items Method = case proplists:get_value(prefix_method, Props) of M when M == all; M == var_prefix; M == fixed_prefix -> M end, Sep = proplists:get_value(prefix_separator, Props, ?HASH_PREFIX_SEPARATOR), NumSeps = proplists:get_value(num_separators, Props, 2), Len = proplists:get_value(prefix_length, Props, 4), PrefixIsIntegerP = proplists:get_value(prefix_is_integer_hack, Props, false), %% Start the real work. OldFloatMap = proplists:get_value(old_float_map, Props, []), NewFloatMap = chash_make_float_map(OldFloatMap, NewWeights), NextFloatList = chash_float_map_to_nextfloat_list(NewFloatMap), Opaque = #chash{num_chains = length(ChainList), map = list_to_tuple(ChainList), prefix_method = Method, separator = Sep, num_separators = NumSeps, prefix_len = Len, prefix_is_integer_hack_p = PrefixIsIntegerP, chash_map = chash_nextfloat_list_to_gb_tree(NextFloatList), old_float_map = OldFloatMap, new_float_map = NewFloatMap, new_chainweights = NewWeights, old_props = Props }, #hash_r{method = chash, mod = ?MODULE, func = chash_key_to_chain, healthy_chainlist = ChainList, opaque = Opaque}. chash_key_to_chain(_ReadOrWrite, Key, Opaque) -> N = if Opaque#chash.prefix_method == all -> naive_hash(Key) / ?PHASH2_BIGGEST_VAL; Opaque#chash.prefix_method == var_prefix -> var_prefix_hash(Key, Opaque#chash.separator, Opaque#chash.num_separators) / ?PHASH2_BIGGEST_VAL; Opaque#chash.prefix_method == fixed_prefix -> X = fixed_prefix_hash(Key, Opaque#chash.prefix_len, Opaque#chash.prefix_is_integer_hack_p), X / ?PHASH2_BIGGEST_VAL; true -> ?APPLOG_ALERT(?APPLOG_APPM_050,"pid ~p: Unknown Opaque: ~p (~p)\n", [self(), element(1, Opaque), Opaque]), timer:sleep(1000), exit({unknown_prefix_method, Opaque#chash.prefix_method}) end, {_Float, Chain} = chash_gb_next(N, Opaque#chash.chash_map), Chain. %% @doc Not used directly, but can give a developer an idea of how well %% chash_float_map_to_nextfloat_list will do for a given value of Max. %% %% For example: %% <verbatim> %% NewFloatMap = chash_make_float_map([{unused, 1.0}], %% [{a,100}, {b, 100}, {c, 10}]), %% ChashMap = chash_scale_to_int_interval(NewFloatMap, 100), %% io:format("QQQ: int int = ~p\n", [ChashIntInterval]), %% -> [{a,1,47},{b,48,94},{c,94,100}] %% </verbatim> %% %% Interpretation: out of the 100 slots: %% <ul> %% <li> 'a' uses the slots 1-47 </li> %% <li> 'b' uses the slots 48-94 </li> %% <li> 'c' uses the slots 95-100 </li> %% </ul> chash_scale_to_int_interval(NewFloatMap, Max) -> chash_scale_to_int_interval(NewFloatMap, 0, Max). chash_scale_to_int_interval([{Ch, _Wt}], Cur, Max) -> [{Ch, Cur, Max}]; chash_scale_to_int_interval([{Ch, Wt}|T], Cur, Max) -> Int = trunc(Wt * Max), [{Ch, Cur + 1, Cur + Int}|chash_scale_to_int_interval(T, Cur + Int, Max)]. %% chash_scale_to_int_interval(NewFloatMap, Max) -> %% chash_scale_to_int_interval(NewFloatMap, 0, Max). %% chash_scale_to_int_interval([{Ch, _Wt}], _Cur, Max) -> %% [{Max, Ch}]; %% chash_scale_to_int_interval([{Ch, Wt}|T], Cur, Max) -> %% Int = trunc(Wt * Max), %% [{Cur + Int, Ch}|chash_scale_to_int_interval(T, Cur + Int, Max)]. %% TODO: Would it ever be necessary to extract new_h_desc from a g_hash_r()? %% @spec (hash_r() | g_hash_r()) -> float_map() %% @doc Given a consistent-hash-flavored #hash_r or #g_hash_r, return the %% new float_map list (i.e. the float_map used for weighting calculations). chash_extract_new_float_map(LH) when is_record(LH, hash_r) -> (LH#hash_r.opaque)#chash.new_float_map; chash_extract_new_float_map(GH) when is_record(GH, g_hash_r) -> chash_extract_new_float_map(GH#g_hash_r.current_h_desc). %% @spec (hash_r() | g_hash_r()) -> float_map() %% @doc Given a consistent-hash-flavored #hash_r or #g_hash_r, return the %% old float_map list (i.e. the "prior" float_map that was used as an input %% to chash_make_float_map/2 to create the "new" float_map). chash_extract_old_float_map(LH) when is_record(LH, hash_r) -> (LH#hash_r.opaque)#chash.old_float_map; chash_extract_old_float_map(GH) when is_record(GH, g_hash_r) -> chash_extract_old_float_map(GH#g_hash_r.current_h_desc). %% @spec (hash_r() | g_hash_r()) -> float_map() %% @doc Given a consistent-hash-flavored #hash_r or #g_hash_r, return the %% new chain weights list that was used to create the new float_map list %% (i.e. the float_map used for weighting calculations). chash_extract_new_chainweights(LH) when is_record(LH, hash_r) -> (LH#hash_r.opaque)#chash.new_chainweights; chash_extract_new_chainweights(GH) when is_record(GH, g_hash_r) -> chash_extract_new_chainweights(GH#g_hash_r.current_h_desc). %% @spec (hash_r() | g_hash_r()) -> proplist() %% @doc Given a consistent-hash-flavored #hash_r or #g_hash_r, return the %% old_props properties list, used for initializing the #hash_r. chash_extract_old_props(LH) when is_record(LH, hash_r) -> (LH#hash_r.opaque)#chash.old_props; chash_extract_old_props(GH) when is_record(GH, g_hash_r) -> chash_extract_old_props(GH#g_hash_r.current_h_desc). %% @spec (list({chain_name, list({brick_name(), node_name()})})) %% -> ok | {error, term()} %% @doc This is a naive chain single (!) list checker. verify_chain_list(ChainList) -> verify_chain_list(ChainList, false). verify_chain_list(ChainList, ZeroLengthOK) -> case (catch verify_chain_list_2(ChainList, ZeroLengthOK)) of {'EXIT', Err} -> {error, Err}; Res -> Res end. verify_chain_list_2(ChainList, ZeroLengthOK) when is_list(ChainList), length(ChainList) > 0 -> lists:map( fun({ChainName, ChainMembers}) when is_atom(ChainName) -> SortedChainMembers = lists:sort(ChainMembers), if length(ChainMembers) < 1, not ZeroLengthOK -> exit({error, ChainName}); true -> ok end, case {SortedChainMembers, list_uniq(SortedChainMembers)} of {Sl, Sl} -> ok; _ -> exit({error, duplicate_bricks_in_chain}) end, %% Check for valid 2-tuples for brick name. lists:map( fun({Br, Nd}) when is_atom(Br), is_atom(Nd) -> ok; (X) -> exit({error, X}) end, ChainMembers), ok; (X) -> exit({error, X}) end, ChainList), ChainNames = [Ch || {Ch, _} <- ChainList], SortedChainNames = lists:sort(ChainNames), case {SortedChainNames, list_uniq(SortedChainNames)} of {Sl0, Sl0} -> ok; _ -> exit({error, duplicate_chain_names}) end, BrickNames = [Br || {_, Brs} <- ChainList, Br <- Brs], SortedBrickNames = lists:sort(BrickNames), case {SortedBrickNames, list_uniq(SortedBrickNames)} of {Sl1, Sl1} -> ok; _ -> exit({error, duplicate_brick_names}) end, ok; verify_chain_list_2(ChainList, true) when is_list(ChainList), length(ChainList) =:= 0 -> %% Added this case when brick_admin server is started very early, %% and none of the chains are currently available. %% TODO: This is harmless/a good idea, right? ok; verify_chain_list_2(Bad, _) -> {error, {bad_list, Bad}}. chain_list2dict(ChainList) -> lists:foldl( fun({ChainName, []}, Dict) -> V = #chaindesc_r{name = ChainName, length = 0, head = no_such_head, middles = [], tail = no_such_tail}, dict:store(ChainName, V, Dict); ({ChainName, ChainMembers}, Dict) -> V = make_chaindesc(ChainName, ChainMembers), dict:store(ChainName, V, Dict) end, dict:new(), ChainList). make_chaindesc(ChainName, []) -> #chaindesc_r{name = ChainName, length = 0, head = chain_is_zero_length, middles = [], tail = chain_is_zero_length}; make_chaindesc(ChainName, ChainMembers) -> Head = hd(ChainMembers), Tail = lists:last(ChainMembers), L = length(ChainMembers), Middles = if L > 2 -> Num = L - 2, lists:sublist(ChainMembers, 2, Num); true -> [] end, #chaindesc_r{name = ChainName, length = L, head = Head, middles = Middles, tail = Tail}. invent_nodelist(NumChains, ChainLen) -> invent_nodelist(NumChains, ChainLen, node(), 0). invent_nodelist(NumChains, ChainLen, Node, BaseChainNum) -> [invent_nodelist2(ChainLen, "test_ch" ++ integer_to_list(I), Node) || I <- lists:seq(BaseChainNum, BaseChainNum + NumChains - 1)]. invent_nodelist2(1, NameBase, Node) -> ChainName = list_to_atom(NameBase), Brick1 = list_to_atom(NameBase ++ "_stand"), {ChainName, [{Brick1, Node}]}; invent_nodelist2(N, NameBase, Node) -> ChainName = list_to_atom(NameBase), Head = list_to_atom(NameBase ++ "_head"), Tail = list_to_atom(NameBase ++ "_tail"), NumMiddles = N - 2, Middles = if NumMiddles == 0 -> []; true -> lists:map( fun(I) -> Mid = list_to_atom(NameBase ++ "_middle" ++ integer_to_list(I)), {Mid, Node} end, lists:seq(1, NumMiddles)) end, {ChainName, [{Head, Node}] ++ Middles ++ [{Tail, Node}]}. find_var_prefix(Key, Separator, N, Num) when N < size(Key), Num > 0 -> case Key of <<Prefix:N/binary, Separator, _/binary>> -> if Num == 1 -> Prefix; true -> find_var_prefix(Key, Separator, N + 1, Num - 1) end; _ -> find_var_prefix(Key, Separator, N + 1, Num) end; find_var_prefix(Key, _, _, _) -> Key. list_uniq([X,X|Xs]) -> list_uniq([X|Xs]); list_uniq([X|Xs]) -> [X|list_uniq(Xs)]; list_uniq([]) -> []. %% %% Consistent hashing API and helper funcs %% chash_make_float_map(NewChainWeights) -> chash_make_float_map([], NewChainWeights). chash_make_float_map([], NewChainWeights) -> Sum = add_all_weights(NewChainWeights), DiffMap = [{Ch, Wt/Sum} || {Ch, Wt} <- NewChainWeights], chash_make_float_map2([{unused, 1.0}], DiffMap, NewChainWeights); chash_make_float_map(OldFloatMap, NewChainWeights) -> NewSum = add_all_weights(NewChainWeights), %% Normalize to unit interval %% NewChainWeights2 = [{Ch, Wt / NewSum} || {Ch, Wt} <- NewChainWeights], %% Reconstruct old chain weights (will be normalized to unit interval) SumOldFloatsDict = lists:foldl(fun({Ch, Wt}, OrdDict) -> orddict:update_counter(Ch, Wt, OrdDict) end, orddict:new(), OldFloatMap), OldChainWeights = orddict:to_list(SumOldFloatsDict), OldSum = add_all_weights(OldChainWeights), OldChs = [Ch || {Ch, _} <- OldChainWeights], NewChs = [Ch || {Ch, _} <- NewChainWeights], OldChsOnly = OldChs -- NewChs, %% Mark any space in by a deleted chain as unused. OldFloatMap2 = lists:map( fun({Ch, Wt} = ChWt) -> case lists:member(Ch, OldChsOnly) of true -> {unused, Wt}; false -> ChWt end end, OldFloatMap), %% Create a diff map of changing chains and added chains DiffMap = lists:map(fun({Ch, NewWt}) -> case orddict:find(Ch, SumOldFloatsDict) of {ok, OldWt} -> {Ch, (NewWt / NewSum) - (OldWt / OldSum)}; error -> {Ch, NewWt / NewSum} end end, NewChainWeights), chash_make_float_map2(OldFloatMap2, DiffMap, NewChainWeights). chash_make_float_map2(OldFloatMap, DiffMap, _NewChainWeights) -> FloatMap = apply_diffmap(DiffMap, OldFloatMap), XX = combine_neighbors(collapse_unused_in_float_map(FloatMap)), XX. apply_diffmap(DiffMap, FloatMap) -> SubtractDiff = [{Ch, abs(Diff)} || {Ch, Diff} <- DiffMap, Diff < 0], AddDiff = [D || {_Ch, Diff} = D <- DiffMap, Diff > 0], TmpFloatMap = iter_diffmap_subtract(SubtractDiff, FloatMap), iter_diffmap_add(AddDiff, TmpFloatMap). add_all_weights(ChainWeights) -> lists:foldl(fun({_Ch, Weight}, Sum) -> Sum + Weight end, 0.0, ChainWeights). iter_diffmap_subtract([{Ch, Diff}|T], FloatMap) -> iter_diffmap_subtract(T, apply_diffmap_subtract(Ch, Diff, FloatMap)); iter_diffmap_subtract([], FloatMap) -> FloatMap. iter_diffmap_add([{Ch, Diff}|T], FloatMap) -> iter_diffmap_add(T, apply_diffmap_add(Ch, Diff, FloatMap)); iter_diffmap_add([], FloatMap) -> FloatMap. apply_diffmap_subtract(Ch, Diff, [{Ch, Wt}|T]) -> if Wt == Diff -> [{unused, Wt}|T]; Wt > Diff -> [{Ch, Wt - Diff}, {unused, Diff}|T]; Wt < Diff -> [{unused, Wt}|apply_diffmap_subtract(Ch, Diff - Wt, T)] end; apply_diffmap_subtract(Ch, Diff, [H|T]) -> [H|apply_diffmap_subtract(Ch, Diff, T)]; apply_diffmap_subtract(_Ch, _Diff, []) -> []. apply_diffmap_add(Ch, Diff, [{unused, Wt}|T]) -> if Wt == Diff -> [{Ch, Wt}|T]; Wt > Diff -> [{Ch, Diff}, {unused, Wt - Diff}|T]; Wt < Diff -> [{Ch, Wt}|apply_diffmap_add(Ch, Diff - Wt, T)] end; apply_diffmap_add(Ch, Diff, [H|T]) -> [H|apply_diffmap_add(Ch, Diff, T)]; apply_diffmap_add(_Ch, _Diff, []) -> []. combine_neighbors([{Ch, Wt1}, {Ch, Wt2}|T]) -> combine_neighbors([{Ch, Wt1 + Wt2}|T]); combine_neighbors([H|T]) -> [H|combine_neighbors(T)]; combine_neighbors([]) -> []. collapse_unused_in_float_map([{Ch, Wt1}, {unused, Wt2}|T]) -> collapse_unused_in_float_map([{Ch, Wt1 + Wt2}|T]); collapse_unused_in_float_map([{unused, _}] = L) -> L; % Degenerate case only collapse_unused_in_float_map([H|T]) -> [H|collapse_unused_in_float_map(T)]; collapse_unused_in_float_map([]) -> []. chash_float_map_to_nextfloat_list(FloatMap) when length(FloatMap) > 0 -> %% QuickCheck found a bug ... need to weed out stuff smaller than %% ?SMALLEST_SIGNIFICANT_FLOAT_SIZE here. FM1 = [P || {_X, Y} = P <- FloatMap, Y > ?SMALLEST_SIGNIFICANT_FLOAT_SIZE], {_Sum, NFs0} = lists:foldl(fun({Name, Amount}, {Sum, List}) -> {Sum+Amount, [{Sum+Amount, Name}|List]} end, {0, []}, FM1), lists:reverse(NFs0). chash_nextfloat_list_to_gb_tree([]) -> gb_trees:balance(gb_trees:from_orddict([])); chash_nextfloat_list_to_gb_tree(NextFloatList) -> {_FloatPos, Name} = lists:last(NextFloatList), %% QuickCheck found a bug ... it really helps to add a catch-all item %% at the far "right" of the list ... 42.0 is much greater than 1.0. NFs = NextFloatList ++ [{42.0, Name}], gb_trees:balance(gb_trees:from_orddict(orddict:from_list(NFs))). chash_gb_next(X, {_, GbTree}) -> chash_gb_next1(X, GbTree). chash_gb_next1(X, {Key, Val, Left, _Right}) when X < Key -> case chash_gb_next1(X, Left) of nil -> {Key, Val}; Res -> Res end; chash_gb_next1(X, {Key, _Val, _Left, Right}) when X >= Key -> chash_gb_next1(X, Right); chash_gb_next1(_X, nil) -> nil. %% @type float_map() = list({brick(), float()}). A float_map is a %% definition of brick assignments over the unit interval [0.0, 1.0]. %% The sum of all floats must be 1.0. %% For example, [{{br1, nd1}, 0.25}, {{br2, nd1}, 0.5}, {{br3, nd1}, 0.25}]. %% @type nextfloat_list() = list({float(), brick()}). A nextfloat_list %% differs from a float_map in two respects: 1) nextfloat_list contains %% tuples with the brick name in 2nd position, 2) the float() at each %% position I_n > I_m, for all n, m such that n > m. %% For example, a nextfloat_list of the float_map example above, %% [{0.25, {br1, nd1}}, {0.75, {br2, nd1}}, {1.0, {br3, nd1}]. %% %% Doodling %% doo_make_chain_weights(Len) -> [{list_to_atom([Name]), 100} || Name <- lists:seq($a, $a + Len - 1)]. %% doo_make_chain(Weights) -> %% [{Name, [{x,y}]} || {Name, _Wt} <- Weights]. doo_iter_chain(StartFloatMap, StartLen, Iters) -> lists:foldl( fun(ChainLen, FloatMap) -> NewWeights = doo_make_chain_weights(ChainLen), %% NewChain = doo_make_chain(NewWeights), chash_make_float_map(FloatMap, NewWeights) end, StartFloatMap, lists:seq(StartLen + 1, StartLen + Iters)). %% length(brick_hash:doo_iter_chain([], 0, 100)) -> 4265 %% And, to see the smallest subdivisions: %% io:format("~P\n", [lists:keysort(2, brick_hash:doo_iter_chain([], 0, 100)), 150]). %% [{'\257',3.46945e-18}, %% {'\230',3.46945e-18}, %% {'\223',1.04083e-17}, %% ... %% {'\246',9.36751e-17}, %% {'\221',9.71445e-17}, %% {'\276',1.04083e-16}, %% ... %% {'\257',7.26849e-16}, %% {'\300',3.01631e-7}, %% {'\301',3.38759e-7}, %% %% So, even after 100 migrations of equal size, the smallest %% significant partition is still as large as 3.01e-7. The sum of all %% partitions smaller than that (less or equal to 1.04083e-16) is only %% 1.32810e-14, so they're pretty safe to ignore. I've added the %% constant ?SMALLEST_SIGNIFICANT_FLOAT_SIZE to keep track of this. doo_weight1() -> [{a, 0.1}, {b, 0.025}, {c, 0.125}, {d, 0.125}, {e, 0.025}, {f, 0.1}, {g, 0.125}, {h, 0.125}, {i, 0.05}, {j, 0.075}, {k, 0.125}].
src/brick_hash.erl
0.525125
0.462109
brick_hash.erl
starcoder
%%% @doc This module implements a simple registry to keep track of the number of FSMs. %%% %%% The `aesc_limits' server monitors each FSM, and inserts an entry into an ETS table. %%% The size of the ETS table reflects the number of active FSMs in the node. %%% We define two entry points in order to distinguish between reestablish attempts %%% (`register_returning()'), for channels which we have already agreed to serve, and %%% attempts to create new channels (`allow_new()'), where a limit check is actually %%% imposed. This means that in certain corner cases, if lots of channels happen to %%% leave and later reestablish roughly simultaneously, we might overshoot and end %%% up with more channels than the configured limit (theoretically, many times as %%% many). To avoid this, a count of channels which are currently 'on leave' would %%% need to be kept. This could perhaps be kept current by the state cache. %%% -module(aesc_limits). -behavior(gen_server). -export([ allow_new/0 , register_returning/0 ]). -export([ start_link/0 , init/1 , handle_call/3 , handle_cast/2 , handle_info/2 , terminate/2 , code_change/3 ]). -record(st, { }). -define(PIDS, aesc_limits_pids). allow_new() -> gen_server:call(?MODULE, allow_new). register_returning() -> gen_server:call(?MODULE, register_returning). start_link() -> %% We make the ets table public, since it's created by the supervisor. %% An alternative would be to set an `heir' option and use `ets:give_away/3` %% to give the newly (re-)started server control; this way, the ets table %% could be private. In either case, if the server restarts, it will have to %% either establish new monitors for all pids in the surviving table, or %% find all running FSMs, e.g. via gproc, to set up new monitors. ets:new(?PIDS, [set, public, named_table]), gen_server:start_link({local, ?MODULE}, ?MODULE, [], []). init([]) -> %% If this is a restart, ensure that all existing pids have working mrefs ets:foldl(fun refresh_monitor/2, ok, ?PIDS), {ok, #st{}}. handle_call(register_returning, {Pid, _}, S) -> MRef = monitor(process, Pid), case ets:insert_new(?PIDS, {Pid, MRef}) of true -> lager:debug("Returning session (~p) allowed", [Pid]), {reply, ok, S}; false -> lager:debug("Returning session (~p) denied: already exists", [Pid]), demonitor(MRef), {reply, {error, exists}, S} end; handle_call(allow_new, {Pid,_}, S) -> Limit = get_limit(), case ets:info(?PIDS, size) of Sz when Sz >= Limit -> lager:debug("New session (~p) denied; Sz = ~p, Limit = ~p", [Pid, Sz, Limit]), {reply, {error, channel_count_limit_exceeded}, S}; Sz -> MRef = monitor(process, Pid), case ets:insert_new(?PIDS, {Pid, MRef}) of true -> lager:debug("New session (~p) allowed; Size = ~p, Limit = ~p", [Pid, Sz, Limit]), {reply, ok, S}; false -> lager:debug("New session (~p) denied: already exists", [Pid]), demonitor(MRef), {reply, {error, exists}, S} end end; handle_call(_, _, S) -> {reply, {error, unknown_call}, S}. handle_cast(_Msg, S) -> {noreply, S}. handle_info({'DOWN', MRef, process, Pid, _Reason}, S) -> NDeleted = ets:select_delete(?PIDS, [{ {Pid, MRef}, [], [true] }]), lager:debug("'DOWN' received; ~p entries deleted for ~p", [NDeleted, Pid]), {noreply, S}; handle_info(_, S) -> {noreply, S}. terminate(_, _) -> ok. code_change(_FromVsn, S, _Extra) -> {ok, S}. %% This could possibly be optimized. The `find_config/2` function normally doesn't %% do more than fetching the whole user config (as a map) and extracting an element %% from it. If no limit has been specified (which may be a common scenario here), the %% `schema_config' option will be pursued, which involves fetching and traversing the %% JSON-Schema (hopefully not from disk, as it should be pre-loaded). %% An issue with caching the value would be that we don't know if it gets changed %% dynamically. An API function in this module for dynamically updating the limit %% would be one way of addressing that. get_limit() -> {ok, Max} = aeu_env:find_config([ <<"channels">>, <<"max_count">> ] , [ user_config , schema_default , {value, 1000} ]), Max. refresh_monitor({Pid, _OldMRef}, ok) -> MRef = monitor(process, Pid), ets:update_element(?PIDS, Pid, {2, MRef}), ok.
apps/aechannel/src/aesc_limits.erl
0.528047
0.59796
aesc_limits.erl
starcoder
%%----------------------------------------------------------------------------- %% @copyright (C) 2019, <NAME> %% @author <NAME> %% @doc Interface for the xor16 filter. %% %% Shorthand for the `exor_filter' module. For indepth documentation, see %% that module. %% %% Example usage: %% ``` %% Filter = xor16:new(["cat", "dog", "mouse"]), %% true = xor16:contain(Filter, "cat"), %% false = xor16:contain(Filter, "goose"), %% ''' %% @end %%----------------------------------------------------------------------------- -module(xor16). -export([ new/1, new/2, new_buffered/1, new_buffered/2, new_empty/0, add/2, finalize/1, contain/2, contain/3, to_bin/1, from_bin/1 ]). %%----------------------------------------------------------------------------- %% @doc Initializes the xor filter, and runs the default hash function on %% each of the elements in the list. This should be fine for the general case. %% @end %%----------------------------------------------------------------------------- -spec new(list()) -> {reference(), atom()} | {error, atom()}. new(List) -> exor_filter:xor16(List). %%----------------------------------------------------------------------------- %% @doc Initializes the xor filter, and runs the specified hash on each of %% the elements in the list. %% %% The option `default_hash' uses the `erlang:phash2/1' function. %% The option `none' is for prehashed data. %% A fun can be passed that will be applied to each element. %% @end %%----------------------------------------------------------------------------- -spec new(list(), exor_filter:hash_function()) -> {reference(), exor_filter:hash_function()} | {error, atom()}. new(List, HashFunction) -> exor_filter:xor16(List, HashFunction). %%----------------------------------------------------------------------------- %% @doc Initializes the xor filter, and runs the default hash function on %% each of the elements in the list. This is the buffered version, meant for %% large filters. %% @end %%----------------------------------------------------------------------------- -spec new_buffered(list()) -> {reference(), atom()} | {error, atom()}. new_buffered(List) -> exor_filter:xor16_buffered(List). %%----------------------------------------------------------------------------- %% @doc Initializes an empty filter. Can be filled incrementally, and is %% more memory efficient than storing entire data set in the Erlang VM. %% Initializes the filter to 64 elements, but will be dynamically expanded %% if more elements are added. %% @end %%----------------------------------------------------------------------------- -spec new_empty() -> {builder, reference()}. new_empty() -> exor_filter:exor_empty(). %%----------------------------------------------------------------------------- %% @doc Adds elements to filter, and applys the default hashing mechanism. %% Dynamically re-sizes filter if needed. %% @end %%----------------------------------------------------------------------------- -spec add({builder, reference()}, list()) -> {builder, reference()}. add(Filter, Elements) -> SortedElements = lists:sort(Elements), exor_filter:exor_add(Filter, SortedElements). %%----------------------------------------------------------------------------- %% @doc Initializes filter internally, and frees data buffer. Equivalent to %% calling `xor16:new'. %% Deduplication is not done, `finalize' will fail if duplicates are inserted. %% @end %%----------------------------------------------------------------------------- -spec finalize({builder, reference()}) -> reference(). finalize(Filter) -> exor_filter:xor16_finalize(Filter). %%----------------------------------------------------------------------------- %% @doc Initializes the xor filter, and runs the default hash function on %% each of the elements in the list. This is the buffered version, meant for %% large filters. See the `xor16:new/2' or `exor_filter:xor16_new/2' funtions %% for more indepth documentaiton. %% @end %%----------------------------------------------------------------------------- -spec new_buffered(list(), exor_filter:hash_function()) -> {reference(), exor_filter:hash_function()} | {error, atom()}. new_buffered(List, HashFunction) -> exor_filter:xor16_buffered(List, HashFunction). %%----------------------------------------------------------------------------- %% @doc Tests to see if the passed argument is in the filter. The first %% argument must be the pre-initialized filter. %% %% A filter previously serialized by `to_bin' is allowed %% @end %%----------------------------------------------------------------------------- -spec contain({reference() | binary(), exor_filter:hash_function()}, term()) -> true | false. contain(Filter, Key) -> exor_filter:xor16_contain(Filter, Key). %%----------------------------------------------------------------------------- %% @doc Tests to see if the passed argument is in the filter. The first %% argument must be the pre-initialized filter. %% %% A filter previously serialized by `to_bin' is allowed %% %% Will return the third argument if the element doesn't exist in the filter. %% @end %%----------------------------------------------------------------------------- -spec contain({reference() | binary(), exor_filter:hash_function()}, term(), any()) -> true | any(). contain(Filter, Key, ReturnValue) -> exor_filter:xor16_contain(Filter, Key, ReturnValue). %%----------------------------------------------------------------------------- %% @doc Serializes the filter to a binary that can be later be deserialized with %% `from_bin/1'. %% %% Returns `{binary(), hash_function()}'. %% @end %%----------------------------------------------------------------------------- -spec to_bin({reference(), exor_filter:hash_function()}) -> {binary(), exor_filter:hash_function()}. to_bin(Filter) -> exor_filter:xor16_to_bin(Filter). %%----------------------------------------------------------------------------- %% @doc Deserializes a filter previously serialized with `to_bin'. %% %% @end %%----------------------------------------------------------------------------- -spec from_bin({binary(), exor_filter:hash_function()}) -> {reference(), exor_filter:hash_function()}. from_bin({Filter, Hash}) -> exor_filter:xor16_from_bin({Filter, Hash}).
src/xor16.erl
0.636918
0.463626
xor16.erl
starcoder
%% @author Couchbase <<EMAIL>> %% @copyright 2017-Present Couchbase, Inc. %% %% Use of this software is governed by the Business Source License included %% in the file licenses/BSL-Couchbase.txt. As of the Change Date specified %% in that file, in accordance with the Business Source License, use of this %% software will be governed by the Apache License, Version 2.0, included in %% the file licenses/APL2.txt. %% %% Generic programming framework loosely based on the paper "Scrap Your %% Boilerplate: A Practical Design Pattern for Generic Programming" by Ralf %% Lammel and <NAME>. %% %% Link to the paper: %% https://www.microsoft.com/en-us/research/wp-content/uploads/2003/01/hmap.pdf %% -module(generic). -include("generic.hrl"). -ifdef(TEST). -include("triq.hrl"). -endif. -export([transformb/2, transformb/3, transformt/2, transformt/3, matching/2, matching/3, universe/1, universe/2, maybe_transform/2, maybe_transform/3, query/3]). %% Apply a transformation everywhere in bottom-up manner. transformb(Fun, State, Term) -> {NewTerm, NewState} = gmap(fun (T, S) -> transformb(Fun, S, T) end, State, Term), Fun(NewTerm, NewState). transformb(Fun, Term) -> ignoring_state(fun transformb/3, Fun, Term). %% Apply a transformation everywhere in top-down manner. transformt(Fun, State, Term) -> maybe_transform(fun (T, S) -> {NewT, NewS} = Fun(T, S), {continue, NewT, NewS} end, State, Term). transformt(Fun, Term) -> ignoring_state(fun transformt/3, Fun, Term). %% Return the subterms matching a predicate. matching(Pred, Term) -> matching(Pred, Term, fun transformb/3). matching(Pred, Term, Traversal) -> {_, Result} = Traversal(fun (T, Acc) -> NewAcc = case Pred(T) of true -> [T | Acc]; false -> Acc end, {T, NewAcc} end, [], Term), lists:reverse(Result). %% Return all possible subterms. universe(Term) -> universe(Term, fun transformb/3). universe(Term, Traversal) -> matching(functools:const(true), Term, Traversal). %% Apply a transformation everywhere in top-down manner. The 'Fun' %% function may choose to stop the recursive descent early by %% returning {stop, ResultTerm, ResultState}. Note, there's not 't' %% suffix here, because short-cutting doesn't make much sense in %% bottom-up traversal. maybe_transform(Fun, State, Term) -> case Fun(Term, State) of {continue, NewTerm, NewState} -> gmap(fun (T, S) -> maybe_transform(Fun, S, T) end, NewState, NewTerm); {stop, NewTerm, NewState} -> {NewTerm, NewState} end. maybe_transform(Fun, Term) -> do_ignoring_state(fun maybe_transform/3, fun (T, S) -> {Action, NewT} = Fun(T), {Action, NewT, S} end, Term). %% Run a query on the term. The 'Fun' is called on each element in the term %% and these values are then recombined by 'K'. The traversal order is %% top-down, left-to-right. query(K, Fun, Term) -> lists:foldl(K, Fun(Term), gmapq(fun (T) -> query(K, Fun, T) end, Term)). %% internal gfold(Fun, State, Term) -> {Type, Children} = term_destructure(Term), Fun(Children, State, fun (NewChildren) -> try term_recover(Type, NewChildren) catch T:E:S -> error({term_recover_failed, {T, E, S}, {term, Term}, {type, Type}, {children, Children}, {new_children, NewChildren}}) end end). term_destructure([H|T]) -> {cons, [H,T]}; term_destructure(Tuple) when is_tuple(Tuple) -> {{tuple, tuple_size(Tuple)}, tuple_to_list(Tuple)}; term_destructure(Map) when is_map(Map) -> {map, lists:append([[K, V] || {K, V} <- maps:to_list(Map)])}; term_destructure(Term) -> {{simple, Term}, []}. term_recover(cons, [H,T]) -> [H|T]; term_recover({tuple, Size}, List) -> Tuple = list_to_tuple(List), Size = tuple_size(Tuple), Tuple; term_recover(map, Values) -> maps:from_list(pairs(Values)); term_recover({simple, Term}, []) -> Term. pairs([]) -> []; pairs([K, V | Rest]) -> [{K, V} | pairs(Rest)]. %% Apply a transformation to direct children of a term. gmap(Fun, State, Term) -> gfold(fun (Children, S, Recover) -> {NewChildren, NewState} = lists:foldl( fun (Child, {AccChildren, AccS}) -> {NewChild, NewAccS} = Fun(Child, AccS), {[NewChild | AccChildren], NewAccS} end, {[], S}, Children), {Recover(lists:reverse(NewChildren)), NewState} end, State, Term). %% Run a query on all direct children of a term. Return results as a list. gmapq(Fun, Term) -> {Result, unused} = gfold(fun (Children, State, _Recover) -> {lists:map(Fun, Children), State} end, unused, Term), Result. ignoring_state(BaseFun, Fun, Term) -> do_ignoring_state(BaseFun, fun (T, S) -> {Fun(T), S} end, Term). do_ignoring_state(BaseFun, WrappedFun, Term) -> {NewTerm, unused} = BaseFun(WrappedFun, unused, Term), NewTerm. -ifdef(TEST). %% test-related helpers random_term([]) -> oneof([{}, [], #{}]); random_term([X]) -> oneof([[X], {X}, X, #{key => X}]); random_term([X, Y | Rest] = Items) -> frequency([{2, Items}, {2, list_to_tuple(Items)}, {2, glue_terms(#{X => Y}, random_term(Rest))}, {6, random_term_split(Items)}]). random_term_split(Items) -> ?LET(N, choose(0, length(Items)), begin {Front, Rear} = lists:split(N, Items), glue_terms(random_term(Front), random_term(Rear)) end). glue_terms(X, Y) -> oneof([{X, Y}, [X, Y], [X | Y], singleton_map(X, Y)]). singleton_map(KSpec, VSpec) -> %% Typically triq instantiates the random specs lazily by walking through %% the spec at the very end. But at the time of this writing, it doesn't %% know how to deal with maps. So we need to explicitly instantiate the %% subterms when constructing maps. ?LET(K, KSpec, ?LET(V, VSpec, #{K => V})). %% triq properties prop_transform_id(Transform) -> ?FORALL(Term, any(), Transform(fun functools:id/1, Term) =:= Term). prop_transformt_id() -> prop_transform_id(fun transformt/2). prop_transformb_id() -> prop_transform_id(fun transformb/2). %% traversal order is left to right, so the order of original elements must be %% the same as in Items list prop_transform_items_order(Transform) -> forall_terms(fun (Items, Term) -> Items =:= matching(fun is_integer/1, Term, Transform) end). prop_transformt_items_order() -> prop_transform_items_order(fun transformt/3). prop_transformb_items_order() -> prop_transform_items_order(fun transformb/3). prop_transforms_same_subterms() -> forall_terms(fun (_Items, Term) -> AllT = universe(Term, fun transformt/3), AllB = universe(Term, fun transformb/3), lists:sort(AllT) =:= lists:sort(AllB) end). prop_transforms_result(Transform) -> Props = ?FORALL(Fun, triq_utils:random_integer_fun(), forall_terms( fun (Items, Term) -> TransFun = ?transform(I when is_integer(I), Fun(I)), Items1 = lists:map(Fun, Items), Term1 = Transform(TransFun, Term), Items1 =:= matching(fun is_integer/1, Term1) end)), %% each forall multiplies number of tests by 100 (by default), so we'd %% have to run 10^6 number of tests which is a bit too much; here we lower %% it to 22^3 (yes, it's somewhat confusing) which is approximately 10000 triq:numtests(22, Props). prop_transformt_result() -> prop_transforms_result(fun transformt/2). prop_transformb_result() -> prop_transforms_result(fun transformb/2). prop_query_result(QueryK, QueryFun, ListFun) -> forall_terms(fun (Items, Term) -> ListFun(Items) =:= query(QueryK, QueryFun, Term) end). prop_query_count() -> prop_query_result(fun functools:add/2, ?query(I when is_integer(I), 1, 0), fun erlang:length/1). prop_query_sum() -> prop_query_result(fun functools:add/2, ?query(I when is_integer(I), I, 0), fun lists:sum/1). forall_terms(Prop) -> ?FORALL(Items, list(int()), ?FORALL(Term, random_term(Items), Prop(Items, Term))). -endif.
src/generic.erl
0.633183
0.450843
generic.erl
starcoder
%% %% %CopyrightBegin% %% %% Copyright Ericsson AB 1999-2018. All Rights Reserved. %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %% %% %CopyrightEnd% %% %% Purpose : Converts intermediate assembly code to final format. -module(beam_flatten). -export([module/2]). -import(lists, [reverse/1,reverse/2]). -spec module(beam_utils:module_code(), [compile:option()]) -> {'ok',beam_utils:module_code()}. module({Mod,Exp,Attr,Fs,Lc}, _Opt) -> {ok,{Mod,Exp,Attr,[function(F) || F <- Fs],Lc}}. function({function,Name,Arity,CLabel,Is0}) -> Is = block(Is0), {function,Name,Arity,CLabel,Is}. block(Is) -> block(Is, []). block([{block,Is0}|Is1], Acc) -> block(Is1, norm_block(Is0, Acc)); block([I|Is], Acc) -> block(Is, [I|Acc]); block([], Acc) -> reverse(Acc). norm_block([{set,[],[],{alloc,R,Alloc}}|Is], Acc0) -> norm_block(Is, reverse(norm_allocate(Alloc, R), Acc0)); norm_block([I|Is], Acc) -> norm_block(Is, [norm(I)|Acc]); norm_block([], Acc) -> Acc. norm({set,[D],As,{bif,N,F}}) -> {bif,N,F,As,D}; norm({set,[D],As,{alloc,R,{gc_bif,N,F}}}) -> {gc_bif,N,F,R,As,D}; norm({set,[D],[S],move}) -> {move,S,D}; norm({set,[D],[S],fmove}) -> {fmove,S,D}; norm({set,[D],[S],fconv}) -> {fconv,S,D}; norm({set,[D],[S1,S2],put_list}) -> {put_list,S1,S2,D}; norm({set,[D],Els,put_tuple2}) -> {put_tuple2,D,{list,Els}}; norm({set,[D],[],{put_tuple,A}}) -> {put_tuple,A,D}; norm({set,[],[S],put}) -> {put,S}; norm({set,[D],[S],{get_tuple_element,I}}) -> {get_tuple_element,S,I,D}; norm({set,[],[S,D],{set_tuple_element,I}}) -> {set_tuple_element,S,D,I}; norm({set,[D],[S],get_hd}) -> {get_hd,S,D}; norm({set,[D],[S],get_tl}) -> {get_tl,S,D}; norm({set,[D],[S|Puts],{alloc,R,{put_map,Op,F}}}) -> {put_map,F,Op,S,D,R,{list,Puts}}; norm({set,[],[],remove_message}) -> remove_message; norm({set,[],[],{line,_}=Line}) -> Line. norm_allocate({_Zero,nostack,Nh,[]}, Regs) -> [{test_heap,Nh,Regs}]; norm_allocate({nozero,Ns,0,Inits}, Regs) -> [{allocate,Ns,Regs}|Inits]; norm_allocate({nozero,Ns,Nh,Inits}, Regs) -> [{allocate_heap,Ns,Nh,Regs}|Inits].
lib/compiler/src/beam_flatten.erl
0.586168
0.410993
beam_flatten.erl
starcoder
%% The contents of this file are subject to the Erlang Public License, Version %% 1.1, (the "License"); you may not use this file except in compliance with %% the License. You should have received a copy of the Erlang Public License %% along with this software. If not, it can be retrieved via the world wide web %% at http://www.erlang.org/. %% %% Software distributed under the License is distributed on an "AS IS" basis, %% WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License for %% the specific language governing rights and limitations under the License. %% Based on: Scalable Bloom Filters %% <NAME>, <NAME>, <NAME>, <NAME> %% Information Processing Letters %% Volume 101, Issue 6, 31 March 2007, Pages 255-261 %% %% Provides scalable bloom filters that can grow indefinitely while ensuring a %% desired maximum false positive probability. Also provides standard %% partitioned bloom filters with a maximum capacity. Bit arrays are %% dimensioned as a power of 2 to enable reusing hash values across filters %% through bit operations. Double hashing is used (no need for enhanced double %% hashing for partitioned bloom filters). %% Modified slightly by Justin Sheehy to make it a single file (incorporated %% the array-based bitarray internally). -module(hanoidb_bloom). -author("<NAME> <<EMAIL>>"). -export([sbf/1, sbf/2, sbf/3, sbf/4, bloom/1, bloom/2, member/2, add/2, size/1, capacity/1, encode/1, decode/1]). -import(math, [log/1, pow/2]). -ifdef(TEST). -ifdef(EQC). -include_lib("eqc/include/eqc.hrl"). -endif. -include_lib("eunit/include/eunit.hrl"). -endif. -define(W, 27). -type bitmask() :: array() | any(). -record(bloom, { e :: float(), % error probability n :: non_neg_integer(), % maximum number of elements mb :: non_neg_integer(), % 2^mb = m, the size of each slice (bitvector) size :: non_neg_integer(), % number of elements a :: [bitmask()] % list of bitvectors }). -record(sbf, { e :: float(), % error probability r :: float(), % error probability ratio s :: non_neg_integer(), % log 2 of growth ratio size :: non_neg_integer(), % number of elements b :: [#bloom{}] % list of plain bloom filters }). %% Constructors for (fixed capacity) bloom filters %% %% N - capacity %% E - error probability bloom(N) -> bloom(N, 0.001). bloom(N, E) when is_number(N), N > 0, is_float(E), E > 0, E < 1, N >= 4/E -> % rule of thumb; due to double hashing bloom(size, N, E); bloom(N, E) when is_number(N), N >= 0, is_float(E), E > 0, E < 1 -> bloom(bits, 32, E). bloom(Mode, N, E) -> K = case Mode of size -> 1 + trunc(log2(1/E)); bits -> 1 end, P = pow(E, 1 / K), Mb = case Mode of size -> 1 + trunc(-log2(1 - pow(1 - P, 1 / N))); bits -> N end, M = 1 bsl Mb, D = trunc(log(1-P) / log(1-1/M)), #bloom{e=E, n=D, mb=Mb, size = 0, a = [bitmask_new(Mb) || _ <- lists:seq(1, K)]}. log2(X) -> log(X) / log(2). %% Constructors for scalable bloom filters %% %% N - initial capacity before expanding %% E - error probability %% S - growth ratio when full (log 2) can be 1, 2 or 3 %% R - tightening ratio of error probability sbf(N) -> sbf(N, 0.001). sbf(N, E) -> sbf(N, E, 1). sbf(N, E, 1) -> sbf(N, E, 1, 0.85); sbf(N, E, 2) -> sbf(N, E, 2, 0.75); sbf(N, E, 3) -> sbf(N, E, 3, 0.65). sbf(N, E, S, R) when is_number(N), N > 0, is_float(E), E > 0, E < 1, is_integer(S), S > 0, S < 4, is_float(R), R > 0, R < 1, N >= 4/(E*(1-R)) -> % rule of thumb; due to double hashing #sbf{e=E, s=S, r=R, size=0, b=[bloom(N, E*(1-R))]}. %% Returns number of elements %% size(#bloom{size=Size}) -> Size; size(#sbf{size=Size}) -> Size. %% Returns capacity %% capacity(#bloom{n=N}) -> N; capacity(#sbf{}) -> infinity. %% Test for membership %% member(Elem, #bloom{mb=Mb}=B) -> Hashes = make_hashes(Mb, Elem), hash_member(Hashes, B); member(Elem, #sbf{b=[H|_]}=Sbf) -> Hashes = make_hashes(H#bloom.mb, Elem), hash_member(Hashes, Sbf). hash_member(Hashes, #bloom{mb=Mb, a=A}) -> Mask = 1 bsl Mb -1, {I1, I0} = make_indexes(Mask, Hashes), all_set(Mask, I1, I0, A); hash_member(Hashes, #sbf{b=B}) -> lists:any(fun(X) -> hash_member(Hashes, X) end, B). make_hashes(Mb, E) when Mb =< 16 -> erlang:phash2({E}, 1 bsl 32); make_hashes(Mb, E) when Mb =< 32 -> {erlang:phash2({E}, 1 bsl 32), erlang:phash2([E], 1 bsl 32)}. make_indexes(Mask, {H0, H1}) when Mask > 1 bsl 16 -> masked_pair(Mask, H0, H1); make_indexes(Mask, {H0, _}) -> make_indexes(Mask, H0); make_indexes(Mask, H0) -> masked_pair(Mask, H0 bsr 16, H0). masked_pair(Mask, X, Y) -> {X band Mask, Y band Mask}. all_set(_Mask, _I1, _I, []) -> true; all_set(Mask, I1, I, [H|T]) -> bitmask_get(I, H) andalso all_set(Mask, I1, (I+I1) band Mask, T). %% Adds element to set %% add(Elem, #bloom{mb=Mb} = B) -> Hashes = make_hashes(Mb, Elem), hash_add(Hashes, B); add(Elem, #sbf{size=Size, r=R, s=S, b=[H|T]=Bs}=Sbf) -> #bloom{mb=Mb, e=E, n=N, size=HSize} = H, Hashes = make_hashes(Mb, Elem), case hash_member(Hashes, Sbf) of true -> Sbf; false -> case HSize < N of true -> Sbf#sbf{size=Size+1, b=[hash_add(Hashes, H)|T]}; false -> B = add(Elem, bloom(bits, Mb + S, E * R)), Sbf#sbf{size=Size+1, b=[B|Bs]} end end. hash_add(Hashes, #bloom{mb=Mb, a=A, size=Size} = B) -> Mask = 1 bsl Mb -1, {I1, I0} = make_indexes(Mask, Hashes), B#bloom{size=Size+1, a=set_bits(Mask, I1, I0, A, [])}. set_bits(_Mask, _I1, _I, [], Acc) -> lists:reverse(Acc); set_bits(Mask, I1, I, [H|T], Acc) -> set_bits(Mask, I1, (I+I1) band Mask, T, [bitmask_set(I, H) | Acc]). %%%========== Dispatch to appropriate representation: bitmask_new(LogN) -> if LogN >= 20 -> % Use sparse representation. hanoidb_sparse_bitmap:new(LogN); true -> % Use dense representation. hanoidb_dense_bitmap:new(1 bsl LogN) end. bitmask_set(I, BM) -> case element(1,BM) of array -> bitarray_set(I, as_array(BM)); sparse_bitmap -> hanoidb_sparse_bitmap:set(I, BM); dense_bitmap_ets -> hanoidb_dense_bitmap:set(I, BM); dense_bitmap -> %% Surprise - we need to mutate a built representation: hanoidb_dense_bitmap:set(I, hanoidb_dense_bitmap:unbuild(BM)) end. %%% Convert to external form. bitmask_build(BM) -> case element(1,BM) of array -> BM; sparse_bitmap -> BM; dense_bitmap_ets -> hanoidb_dense_bitmap:build(BM) end. bitmask_get(I, BM) -> case element(1,BM) of array -> bitarray_get(I, as_array(BM)); sparse_bitmap -> hanoidb_sparse_bitmap:member(I, BM); dense_bitmap_ets -> hanoidb_dense_bitmap:member(I, BM); dense_bitmap -> hanoidb_dense_bitmap:member(I, BM) end. -spec as_array(bitmask()) -> array(). as_array(BM) -> case array:is_array(BM) of true -> BM end. %%%========== Bitarray representation - suitable for sparse arrays ========== bitarray_new(N) -> array:new((N-1) div ?W + 1, {default, 0}). -spec bitarray_set( non_neg_integer(), array() ) -> array(). bitarray_set(I, A1) -> A = as_array(A1), AI = I div ?W, V = array:get(AI, A), V1 = V bor (1 bsl (I rem ?W)), if V =:= V1 -> A; % The bit is already set true -> array:set(AI, V1, A) end. -spec bitarray_get( non_neg_integer(), array() ) -> boolean(). bitarray_get(I, A) -> AI = I div ?W, V = array:get(AI, A), (V band (1 bsl (I rem ?W))) =/= 0. %%%^^^^^^^^^^ Bitarray representation - suitable for sparse arrays ^^^^^^^^^^ encode(Bloom) -> zlib:gzip(term_to_binary(bloom_build(Bloom))). decode(Bin) -> binary_to_term(zlib:gunzip(Bin)). %%% Convert to external form. bloom_build(Bloom=#bloom{a=Bitmasks}) -> Bloom#bloom{a=[bitmask_build(X) || X <- Bitmasks]}; bloom_build(Sbf=#sbf{b=Blooms}) -> Sbf#sbf{b=[bloom_build(X) || X <- Blooms]}. %% UNIT TESTS -ifdef(TEST). -ifdef(EQC). prop_bloom_test_() -> {timeout, 60, fun() -> ?assert(eqc:quickcheck(prop_bloom())) end}. g_keys() -> non_empty(list(non_empty(binary()))). prop_bloom() -> ?FORALL(Keys, g_keys(), begin Bloom = ?MODULE:bloom(Keys), F = fun(X) -> member(X, Bloom) end, lists:all(F, Keys) end). -endif. -endif.
src/hanoidb_bloom.erl
0.648911
0.489992
hanoidb_bloom.erl
starcoder
% Licensed under the Apache License, Version 2.0 (the "License"); you may not % use this file except in compliance with the License. You may obtain a copy of % the License at % % http://www.apache.org/licenses/LICENSE-2.0 % % Unless required by applicable law or agreed to in writing, software % distributed under the License is distributed on an "AS IS" BASIS, WITHOUT % WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the % License for the specific language governing permissions and limitations under % the License. -module(ets_lru). -behaviour(gen_server). -vsn(2). -export([ start_link/2, stop/1, insert/3, lookup/2, match/3, match_object/3, remove/2, clear/1, % Dirty functions read straight from % the ETS tables which means there are % race conditions with concurrent access. lookup_d/2 ]). -export([ init/1, terminate/2, handle_call/3, handle_cast/2, handle_info/2, code_change/3 ]). -define(DEFAULT_TIME_UNIT, millisecond). -type time_value() :: integer(). -type strict_monotonic_time() :: {time_value(), integer()}. -record(entry, { key :: term(), val :: term(), atime :: strict_monotonic_time(), ctime :: strict_monotonic_time() }). -record(st, { objects, atimes, ctimes, max_objs :: non_neg_integer() | undefined, max_size :: non_neg_integer() | undefined, max_lifetime :: non_neg_integer() | undefined, time_unit = ?DEFAULT_TIME_UNIT :: atom() }). start_link(Name, Options) when is_atom(Name) -> gen_server:start_link({local, Name}, ?MODULE, {Name, Options}, []). stop(LRU) -> gen_server:cast(LRU, stop). lookup(LRU, Key) -> gen_server:call(LRU, {lookup, Key}). insert(LRU, Key, Val) -> gen_server:call(LRU, {insert, Key, Val}). remove(LRU, Key) -> gen_server:call(LRU, {remove, Key}). %% @doc match/3 provides an efficient way to retrieve parts of the %% keys and values without copying or requiring circumvention of the %% ets_lru API. The KeySpec and ValueSpec parameters are used as part %% of one larger match spec so keep in mind that all capturing %% placeholders will be aliased between the key and value parts. -spec match(atom() | pid(), term(), term()) -> [[any()]]. match(LRU, KeySpec, ValueSpec) -> gen_server:call(LRU, {match, KeySpec, ValueSpec}). %% @doc match_object/3 provides an efficient way to retrieve multiple %% values using match conditions. The KeySpec and ValueSpec parameters %% are used as part of one larger match spec so keep in mind that all %% capturing placeholders will be aliased between the key and value %% parts. -spec match_object(atom() | pid(), term(), term()) -> [any()]. match_object(Name, KeySpec, ValueSpec) when is_atom(Name) -> Pattern = #entry{key=KeySpec, val=ValueSpec, _='_'}, Entries = ets:match_object(obj_table(Name), Pattern), lists:map(fun(#entry{key=Key,val=Val}) -> gen_server:cast(Name, {accessed, Key}), Val end, Entries); match_object(LRU, KeySpec, ValueSpec) -> gen_server:call(LRU, {match_object, KeySpec, ValueSpec}). clear(LRU) -> gen_server:call(LRU, clear). lookup_d(Name, Key) when is_atom(Name) -> case ets:lookup(obj_table(Name), Key) of [#entry{val=Val}] -> gen_server:cast(Name, {accessed, Key}), {ok, Val}; [] -> not_found end. init({Name, Options}) -> St = set_options(#st{}, Options), ObjOpts = [set, named_table, protected, {keypos, #entry.key}], TimeOpts = [ordered_set, named_table, protected], {ok, St#st{ objects = ets:new(obj_table(Name), ObjOpts), atimes = ets:new(at_table(Name), TimeOpts), ctimes = ets:new(ct_table(Name), TimeOpts) }}. terminate(_Reason, St) -> true = ets:delete(St#st.objects), true = ets:delete(St#st.atimes), true = ets:delete(St#st.ctimes), ok. handle_call({lookup, Key}, _From, St) -> Reply = case ets:lookup(St#st.objects, Key) of [#entry{val=Val} | _] -> accessed(Key, St), {ok, Val}; [] -> not_found end, {reply, Reply, St, 0}; handle_call({match_object, KeySpec, ValueSpec}, _From, St) -> Pattern = #entry{key=KeySpec, val=ValueSpec, _='_'}, Entries = ets:match_object(St#st.objects, Pattern), Values = lists:map(fun(#entry{key=Key,val=Val}) -> accessed(Key, St), Val end, Entries), {reply, Values, St, 0}; handle_call({match, KeySpec, ValueSpec}, _From, St) -> Pattern = #entry{key=KeySpec, val=ValueSpec, _='_'}, Values = ets:match(St#st.objects, Pattern), {reply, Values, St, 0}; handle_call({insert, Key, Val}, _From, St) -> NewATime = strict_monotonic_time(St#st.time_unit), Pattern = #entry{key=Key, atime='$1', _='_'}, case ets:match(St#st.objects, Pattern) of [[ATime]] -> Update = {#entry.val, Val}, true = ets:update_element(St#st.objects, Key, Update), true = ets:delete(St#st.atimes, ATime), true = ets:insert(St#st.atimes, {NewATime, Key}); [] -> Entry = #entry{key=Key, val=Val, atime=NewATime, ctime=NewATime}, true = ets:insert(St#st.objects, Entry), true = ets:insert(St#st.atimes, {NewATime, Key}), true = ets:insert(St#st.ctimes, {NewATime, Key}) end, {reply, ok, St, 0}; handle_call({remove, Key}, _From, St) -> Pattern = #entry{key=Key, atime='$1', ctime='$2', _='_'}, Reply = case ets:match(St#st.objects, Pattern) of [[ATime, CTime]] -> true = ets:delete(St#st.objects, Key), true = ets:delete(St#st.atimes, ATime), true = ets:delete(St#st.ctimes, CTime), ok; [] -> not_found end, {reply, Reply, St, 0}; handle_call(clear, _From, St) -> true = ets:delete_all_objects(St#st.objects), true = ets:delete_all_objects(St#st.atimes), true = ets:delete_all_objects(St#st.ctimes), % No need to timeout here and evict cache % entries because its now empty. {reply, ok, St}; handle_call(Msg, _From, St) -> {stop, {invalid_call, Msg}, {invalid_call, Msg}, St}. handle_cast({accessed, Key}, St) -> accessed(Key, St), {noreply, St, 0}; handle_cast(stop, St) -> {stop, normal, St}; handle_cast(Msg, St) -> {stop, {invalid_cast, Msg}, St}. handle_info(timeout, St) -> trim(St), {noreply, St, next_timeout(St)}; handle_info(Msg, St) -> {stop, {invalid_info, Msg}, St}. code_change(_OldVsn, St, _Extra) -> {ok, St}. accessed(Key, St) -> Pattern = #entry{key=Key, atime='$1', _='_'}, case ets:match(St#st.objects, Pattern) of [[ATime]] -> NewATime = strict_monotonic_time(St#st.time_unit), Update = {#entry.atime, NewATime}, true = ets:update_element(St#st.objects, Key, Update), true = ets:delete(St#st.atimes, ATime), true = ets:insert(St#st.atimes, {NewATime, Key}), ok; [] -> ok end. trim(St) -> trim_count(St), trim_size(St), trim_lifetime(St). trim_count(#st{max_objs=undefined}) -> ok; trim_count(#st{max_objs=Max}=St) -> case ets:info(St#st.objects, size) > Max of true -> drop_lru(St, fun trim_count/1); false -> ok end. trim_size(#st{max_size=undefined}) -> ok; trim_size(#st{max_size=Max}=St) -> case ets:info(St#st.objects, memory) > Max of true -> drop_lru(St, fun trim_size/1); false -> ok end. trim_lifetime(#st{max_lifetime=undefined}) -> ok; trim_lifetime(#st{max_lifetime=Max}=St) -> Now = erlang:monotonic_time(St#st.time_unit), case ets:first(St#st.ctimes) of '$end_of_table' -> ok; CTime = {Time, _} -> case Now - Time > Max of true -> [{CTime, Key}] = ets:lookup(St#st.ctimes, CTime), Pattern = #entry{key=Key, atime='$1', _='_'}, [[ATime]] = ets:match(St#st.objects, Pattern), true = ets:delete(St#st.objects, Key), true = ets:delete(St#st.atimes, ATime), true = ets:delete(St#st.ctimes, CTime), trim_lifetime(St); false -> ok end end. drop_lru(St, Continue) -> case ets:first(St#st.atimes) of '$end_of_table' -> empty; ATime -> [{ATime, Key}] = ets:lookup(St#st.atimes, ATime), Pattern = #entry{key=Key, ctime='$1', _='_'}, [[CTime]] = ets:match(St#st.objects, Pattern), true = ets:delete(St#st.objects, Key), true = ets:delete(St#st.atimes, ATime), true = ets:delete(St#st.ctimes, CTime), Continue(St) end. next_timeout(#st{max_lifetime=undefined}) -> infinity; next_timeout(St) -> case ets:first(St#st.ctimes) of '$end_of_table' -> infinity; {Time, _} -> Now = erlang:monotonic_time(St#st.time_unit), TimeDiff = Now - Time, erlang:max(St#st.max_lifetime - TimeDiff, 0) end. set_options(St, []) -> St; set_options(St, [{max_objects, N} | Rest]) when is_integer(N), N >= 0 -> set_options(St#st{max_objs=N}, Rest); set_options(St, [{max_size, N} | Rest]) when is_integer(N), N >= 0 -> set_options(St#st{max_size=N}, Rest); set_options(St, [{max_lifetime, N} | Rest]) when is_integer(N), N >= 0 -> set_options(St#st{max_lifetime=N}, Rest); set_options(St, [{time_unit, T} | Rest]) when is_atom(T) -> set_options(St#st{time_unit=T}, Rest); set_options(_, [Opt | _]) -> throw({invalid_option, Opt}). obj_table(Name) -> table_name(Name, "_objects"). at_table(Name) -> table_name(Name, "_atimes"). ct_table(Name) -> table_name(Name, "_ctimes"). table_name(Name, Ext) -> list_to_atom(atom_to_list(Name) ++ Ext). -spec strict_monotonic_time(atom()) -> strict_monotonic_time(). strict_monotonic_time(TimeUnit) -> {erlang:monotonic_time(TimeUnit), erlang:unique_integer([monotonic])}.
src/ets_lru.erl
0.690663
0.46132
ets_lru.erl
starcoder
%% @doc %% Summary metric, to track the size of events and report quantiles %% Based on prometheus_summary %% %% Example use cases for Summaries: %% - Response latency; %% - Request size; %% - Response size. %% %% Example: %% <pre lang="erlang"> %% -module(my_proxy_instrumenter). %% %% setup() -> %% prometheus_quantile_summary:declare([{name, request_size_bytes}, %% {help, "Request size in bytes."}]), %% prometheus_quantile_summary:declare([{name, response_size_bytes}, %% {help, "Response size in bytes."}]). %% %% observe_request(Size) -> %% prometheus_quantile_summary:observe(request_size_bytes, Size). %% %% observe_response(Size) -> %% prometheus_quantile_summary:observe(response_size_bytes, Size). %% </pre> %% %% Reports: %% request_size_bytes_size %% request_size_bytes_count %% request_size_bytes{quantile="0.5"} %% request_size_bytes{quantile="0.9"} %% request_size_bytes{quantile="0.95"} %% @end -module(prometheus_quantile_summary). %%% metric -export([new/1, declare/1, deregister/1, deregister/2, set_default/2, observe/2, observe/3, observe/4, observe_duration/2, observe_duration/3, observe_duration/4, remove/1, remove/2, remove/3, reset/1, reset/2, reset/3, value/1, value/2, value/3, values/2]). %%% collector -export([deregister_cleanup/1, collect_mf/2, collect_metrics/2]). -include("prometheus.hrl"). -include_lib("quantile_estimator/include/quantile_estimator.hrl"). -behaviour(prometheus_metric). -behaviour(prometheus_collector). %%==================================================================== %% Macros %%==================================================================== -define(TABLE, ?PROMETHEUS_QUANTILE_SUMMARY_TABLE). -define(SUM_POS, 3). -define(COUNTER_POS, 2). -define(QUANTILE_POS, 4). -define(WIDTH, 16). %%==================================================================== %% Metric API %%==================================================================== %% @doc Creates a summary using `Spec'. %% %% Raises `{missing_metric_spec_key, Key, Spec}' error if required `Soec' key %% is missing.<br/> %% Raises `{invalid_metric_name, Name, Message}' error if metric `Name' %% is invalid.<br/> %% Raises `{invalid_metric_help, Help, Message}' error if metric `Help' %% is invalid.<br/> %% Raises `{invalid_metric_labels, Labels, Message}' error if `Labels' %% isn't a list.<br/> %% Raises `{invalid_label_name, Name, Message}' error if `Name' isn't a valid %% label name.<br/> %% Raises `{invalid_value_error, Value, Message}' error if `duration_unit' is %% unknown or doesn't match metric name.<br/> %% Raises `{mf_already_exists, {Registry, Name}, Message}' error if a summary %% with the same `Spec' already exists. %% @end new(Spec) -> Spec1 = validate_summary_spec(Spec), prometheus_metric:insert_new_mf(?TABLE, ?MODULE, Spec1). %% @doc Creates a summary using `Spec'. %% If a summary with the same `Spec' exists returns `false'. %% %% Raises `{missing_metric_spec_key, Key, Spec}' error if required `Soec' key %% is missing.<br/> %% Raises `{invalid_metric_name, Name, Message}' error if metric `Name' %% is invalid.<br/> %% Raises `{invalid_metric_help, Help, Message}' error if metric `Help' %% is invalid.<br/> %% Raises `{invalid_metric_labels, Labels, Message}' error if `Labels' %% isn't a list.<br/> %% Raises `{invalid_label_name, Name, Message}' error if `Name' isn't a valid %% label name.<br/> %% Raises `{invalid_value_error, Value, MessagE}' error if `duration_unit' is %% unknown or doesn't match metric name.<br/> %% @end declare(Spec) -> Spec1 = validate_summary_spec(Spec), prometheus_metric:insert_mf(?TABLE, ?MODULE, Spec1). %% @equiv deregister(default, Name) deregister(Name) -> deregister(default, Name). %% @doc %% Removes all summary series with name `Name' and %% removes Metric Family from `Registry'. %% %% After this call new/1 for `Name' and `Registry' will succeed. %% %% Returns `{true, _}' if `Name' was a registered summary. %% Otherwise returns `{false, _}'. %% @end deregister(Registry, Name) -> MFR = prometheus_metric:deregister_mf(?TABLE, Registry, Name), NumDeleted = ets:select_delete(?TABLE, deregister_select(Registry, Name)), {MFR, NumDeleted > 0}. %% @private set_default(Registry, Name) -> Configuration = get_configuration(Registry, Name), #{compress_limit := CompressLimit} = Configuration, ets:insert_new(?TABLE, { key(Registry, Name, []), 0, 0, quantile(Configuration), CompressLimit}). %% @equiv observe(default, Name, [], Value) observe(Name, Value) -> observe(default, Name, [], Value). %% @equiv observe(default, Name, LabelValues, Value) observe(Name, LabelValues, Value) -> observe(default, Name, LabelValues, Value). %% @doc Observes the given `Value'. %% %% Raises `{invalid_value, Value, Message}' if `Value' %% isn't an integer.<br/> %% Raises `{unknown_metric, Registry, Name}' error if summary with named `Name' %% can't be found in `Registry'.<br/> %% Raises `{invalid_metric_arity, Present, Expected}' error if labels count %% mismatch. %% @end observe(Registry, Name, LabelValues, Value) when is_number(Value) -> Key = key(Registry, Name, LabelValues), case ets:lookup(?TABLE, Key) of [] -> insert_metric(Registry, Name, LabelValues, Value, fun observe/4); [{Key, Count, S, Q, CompressLimit}] -> ets:insert(?TABLE, {Key, Count + 1, S + Value, quantile_add(Q, Value, CompressLimit), CompressLimit}) end, ok; observe(_Registry, _Name, _LabelValues, Value) -> erlang:error({invalid_value, Value, "observe accepts only numbers"}). %% @equiv observe_duration(default, Name, [], Fun) observe_duration(Name, Fun) -> observe_duration(default, Name, [], Fun). %% @equiv observe_duration(default, Name, LabelValues, Fun) observe_duration(Name, LabelValues, Fun) -> observe_duration(default, Name, LabelValues, Fun). %% @doc Tracks the amount of time spent executing `Fun'. %% %% Raises `{unknown_metric, Registry, Name}' error if summary with named `Name' %% can't be found in `Registry'.<br/> %% Raises `{invalid_metric_arity, Present, Expected}' error if labels count %% mismatch. %% Raises `{invalid_value, Value, Message}' if `Fun' %% isn't a function.<br/> %% @end observe_duration(Registry, Name, LabelValues, Fun) when is_function(Fun)-> Start = erlang:monotonic_time(), try Fun() after observe(Registry, Name, LabelValues, erlang:monotonic_time() - Start) end; observe_duration(_Regsitry, _Name, _LabelValues, Fun) -> erlang:error({invalid_value, Fun, "observe_duration accepts only functions"}). %% @equiv remove(default, Name, []) remove(Name) -> remove(default, Name, []). %% @equiv remove(default, Name, LabelValues) remove(Name, LabelValues) -> remove(default, Name, LabelValues). %% @doc Removes summary series identified by `Registry', `Name' %% and `LabelValues'. %% %% Raises `{unknown_metric, Registry, Name}' error if summary with name `Name' %% can't be found in `Registry'.<br/> %% Raises `{invalid_metric_arity, Present, Expected}' error if labels count %% mismatch. %% @end remove(Registry, Name, LabelValues) -> prometheus_metric:check_mf_exists(?TABLE, Registry, Name, LabelValues), case lists:flatten([ets:take(?TABLE, {Registry, Name, LabelValues, Scheduler}) || Scheduler <- schedulers_seq()]) of [] -> false; _ -> true end. %% @equiv reset(default, Name, []) reset(Name) -> reset(default, Name, []). %% @equiv reset(default, Name, LabelValues) reset(Name, LabelValues) -> reset(default, Name, LabelValues). %% @doc Resets the value of the summary identified by `Registry', `Name' %% and `LabelValues'. %% %% Raises `{unknown_metric, Registry, Name}' error if summary with name `Name' %% can't be found in `Registry'.<br/> %% Raises `{invalid_metric_arity, Present, Expected}' error if labels count %% mismatch. %% @end reset(Registry, Name, LabelValues) -> MF = prometheus_metric:check_mf_exists(?TABLE, Registry, Name, LabelValues), Configuration = prometheus_metric:mf_data(MF), case lists:usort([ets:update_element(?TABLE, {Registry, Name, LabelValues, Scheduler}, [{?COUNTER_POS, 0}, {?SUM_POS, 0}, {?QUANTILE_POS, quantile(Configuration)}]) || Scheduler <- schedulers_seq()]) of [_, _] -> true; [true] -> true; _ -> false end. %% @equiv value(default, Name, []) value(Name) -> value(default, Name, []). %% @equiv value(default, Name, LabelValues) value(Name, LabelValues) -> value(default, Name, LabelValues). %% @doc Returns the value of the summary identified by `Registry', `Name' %% and `LabelValues'. If there is no summary for `LabelValues', %% returns `undefined'. %% %% If duration unit set, sum will be converted to the duration unit. %% {@link prometheus_time. Read more here.} %% %% Raises `{unknown_metric, Registry, Name}' error if summary named `Name' %% can't be found in `Registry'.<br/> %% Raises `{invalid_metric_arity, Present, Expected}' error if labels count %% mismatch. %% @end value(Registry, Name, LabelValues) -> MF = prometheus_metric:check_mf_exists(?TABLE, Registry, Name, LabelValues), DU = prometheus_metric:mf_duration_unit(MF), #{quantiles := QNs} = prometheus_metric:mf_data(MF), case ets:select(?TABLE, [{{{Registry, Name, LabelValues, '_'}, '$1', '$2', '$3', '_'}, [], ['$$']}]) of [] -> undefined; Values -> {Count, Sum, QE} = reduce_values(Values), {Count, prometheus_time:maybe_convert_to_du(DU, Sum), quantile_values(QE, QNs)} end. values(Registry, Name) -> case prometheus_metric:check_mf_exists(?TABLE, Registry, Name) of false -> []; MF -> DU = prometheus_metric:mf_duration_unit(MF), Labels = prometheus_metric:mf_labels(MF), #{quantiles := QNs} = Configuration = prometheus_metric:mf_data(MF), MFValues = load_all_values(Registry, Name), ReducedMap = lists:foldl( fun([L, C, S, QE], ResAcc) -> {PrevCount, PrevSum, PrevQE} = maps:get(L, ResAcc, {0, 0, quantile(Configuration)}), ResAcc#{L => {PrevCount + C, PrevSum + S, quantile_merge(PrevQE, QE)}} end, #{}, MFValues), ReducedMapList = lists:sort(maps:to_list(ReducedMap)), lists:foldr( fun({LabelValues, {Count, Sum, QE}}, Acc) -> [{lists:zip(Labels, LabelValues), Count, prometheus_time:maybe_convert_to_du(DU, Sum), quantile_values(QE, QNs)} | Acc] end, [], ReducedMapList) end. %%==================================================================== %% Collector API %%==================================================================== %% @private deregister_cleanup(Registry) -> prometheus_metric:deregister_mf(?TABLE, Registry), true = ets:match_delete(?TABLE, {{Registry, '_', '_', '_'}, '_', '_', '_', '_'}), ok. %% @private collect_mf(Registry, Callback) -> [Callback(create_summary(Name, Help, {CLabels, Labels, Registry, DU, Data})) || [Name, {Labels, Help}, CLabels, DU, Data] <- prometheus_metric:metrics(?TABLE, Registry)], ok. %% @private collect_metrics(Name, {CLabels, Labels, Registry, DU, Configuration}) -> #{quantiles := QNs} = Configuration, MFValues = load_all_values(Registry, Name), ReducedMap = lists:foldl( fun([L, C, S, QE], ResAcc) -> {PrevCount, PrevSum, PrevQE} = maps:get(L, ResAcc, {0, 0, quantile(Configuration)}), ResAcc#{L => {PrevCount + C, PrevSum + S, quantile_merge(PrevQE, QE)}} end, #{}, MFValues), ReducedMapList = lists:sort(maps:to_list(ReducedMap)), lists:foldr( fun({LabelValues, {Count, Sum, QE}}, Acc) -> [prometheus_model_helpers:summary_metric( CLabels ++ lists:zip(Labels, LabelValues), Count, prometheus_time:maybe_convert_to_du(DU, Sum), quantile_values(QE, QNs)) | Acc] end, [], ReducedMapList). %%==================================================================== %% Private Parts %%==================================================================== deregister_select(Registry, Name) -> [{{{Registry, Name, '_', '_'}, '_', '_', '_', '_'}, [], [true]}]. validate_summary_spec(Spec) -> Labels = prometheus_metric_spec:labels(Spec), validate_summary_labels(Labels), {Invariant, QNs} = invariant_and_quantiles_from_spec(Spec), CompressLimit = compress_limit_from_spec(Spec), [ {data, #{quantiles => QNs, invariant => Invariant, compress_limit => CompressLimit}} | Spec ]. validate_summary_labels(Labels) -> [raise_error_if_quantile_label_found(Label) || Label <- Labels]. raise_error_if_quantile_label_found("quantile") -> erlang:error({invalid_metric_label_name, "quantile", "summary cannot have a label named \"quantile\""}); raise_error_if_quantile_label_found(Label) -> Label. insert_metric(Registry, Name, LabelValues, Value, ConflictCB) -> MF = prometheus_metric:check_mf_exists(?TABLE, Registry, Name, LabelValues), Configuration = prometheus_metric:mf_data(MF), #{compress_limit := CompressLimit} = Configuration, Quantile = quantile(Configuration, Value), case ets:insert_new(?TABLE, {key(Registry, Name, LabelValues), 1, Value, Quantile, CompressLimit}) of false -> %% some sneaky process already inserted ConflictCB(Registry, Name, LabelValues, Value); true -> ok end. load_all_values(Registry, Name) -> ets:match(?TABLE, {{Registry, Name, '$1', '_'}, '$2', '$3', '$4', '_'}). get_configuration(Registry, Name) -> MF = prometheus_metric:check_mf_exists(?TABLE, Registry, Name), prometheus_metric:mf_data(MF). schedulers_seq() -> lists:seq(0, ?WIDTH-1). key(Registry, Name, LabelValues) -> X = erlang:system_info(scheduler_id), Rnd = X band (?WIDTH-1), {Registry, Name, LabelValues, Rnd}. reduce_values(Values) -> {lists:sum([C || [C, _, _] <- Values]), lists:sum([S || [_, S, _] <- Values]), fold_quantiles([Q || [_C, _S, Q] <- Values])}. create_summary(Name, Help, Data) -> prometheus_model_helpers:create_mf(Name, Help, summary, ?MODULE, Data). default_compress_limit() -> 100. invariant_and_quantiles_from_spec(Spec) -> Targets = prometheus_metric_spec:get_value(targets, Spec, default_targets()), validate_targets(Targets), {QNs, _} = lists:unzip(Targets), Invariant = quantile_estimator:f_targeted(Targets), {Invariant, QNs}. compress_limit_from_spec(Spec) -> prometheus_metric_spec:get_value(compress_limit, Spec, default_compress_limit()). validate_targets(Targets) when is_list(Targets) -> lists:foreach( fun ({Q, _E}) when not is_float(Q) -> erlang:error({invalid_targets, "target quantile value should be float"}); ({_Q, E}) when not is_float(E) -> erlang:error({invalid_targets, "target error value should be float"}); ({_, _}) -> ok; (_) -> erlang:error({invalid_targets, "targets should be tuples of quantile and error"}) end, Targets); validate_targets(_Targets) -> erlang:error({invalid_targets, "targets should be a list of tuples"}). default_targets() -> [{0.5, 0.02}, {0.9, 0.01}, {0.95, 0.005}]. quantile(#{invariant := Invariant}) -> quantile_estimator:new(Invariant). quantile(Configuration, Val) -> quantile_estimator:insert(Val, quantile(Configuration)). quantile_add(Q = #quantile_estimator{inserts_since_compression = ISS}, Val, CompressLimit) -> Q1 = case ISS > CompressLimit of true -> quantile_estimator:compress(Q); false -> Q end, quantile_estimator:insert(Val, Q1). %% Quantile estimator throws on empty stats quantile_values(#quantile_estimator{data = []}, _QNs) -> []; quantile_values(Q, QNs) -> [{QN, quantile_estimator:quantile(QN, Q)} || QN <- QNs]. fold_quantiles(QList) -> lists:foldl( fun (Q, init) -> Q; (Q1, Q2) -> quantile_merge(Q1, Q2) end, init, QList). quantile_merge(QE1, QE2) -> #quantile_estimator{samples_count = N1, data = Data1, invariant = Invariant} = QE1, #quantile_estimator{samples_count = N2, data = Data2} = QE2, quantile_estimator:compress(#quantile_estimator{ samples_count = N1 + N2, data = Data1 ++ Data2, invariant = Invariant }).
src/metrics/prometheus_quantile_summary.erl
0.7641
0.416975
prometheus_quantile_summary.erl
starcoder
%% vim: set ai et sw=4 sts=4: %% See LICENSE for licensing information. % This module has been derived from: % https://github.com/marco-m/mock_io % Copyright (c) 2015, <NAME> -module(solarized_capture). -export([ output/1 , output/3 , result_and_output/1 , result_and_output/3 ]). -export_type([ geometry/0 ]). -type geometry() :: pos_integer() | enotsup | {ok, pos_integer()} | {error, enotsup}. -type internal_geometry() :: pos_integer() | {error, enotsup}. -ifdef(TEST). -include_lib("eunit/include/eunit.hrl"). -endif. %======================================================================= -record(state, { output = <<>> , binary = false , columns = 80 , rows = 25 }). %======================================================================= -spec output(Test) -> Output when Test :: fun(() -> Result), Result :: term(), Output :: binary(). output(Test) when is_function(Test, 0) -> output(Test, io:columns(), io:rows()). %======================================================================= -spec output(Test, Columns, Rows) -> Output when Test :: fun(() -> Result), Columns :: geometry(), Rows :: geometry(), Result :: term(), Output :: binary(). output(Test, Columns, Rows) when is_function(Test, 0) -> {_, Output} = result_and_output(Test, Columns, Rows), Output. %======================================================================= -spec result_and_output(Test) -> {Result, Output} when Test :: fun(() -> Result), Result :: term(), Output :: binary(). result_and_output(Test) when is_function(Test, 0) -> result_and_output(Test, io:columns(), io:rows()). %======================================================================= -spec result_and_output(Test, Columns, Rows) -> {Result, Output} when Test :: fun(() -> Result), Columns :: geometry(), Rows :: geometry(), Result :: term(), Output :: binary(). result_and_output(Test, Columns, Rows) when is_function(Test, 0) -> State = #state{columns = geometry(Columns), rows = geometry(Rows)}, mock(Test, State). %======================================================================= -spec geometry(geometry()) -> internal_geometry(). geometry(N) when is_integer(N) andalso N > 0 -> N; geometry(enotsup) -> {error, enotsup}; geometry({ok, N}) when is_integer(N) andalso N > 0 -> N; geometry(Error = {error, enotsup}) -> Error; geometry(_) -> throw(badarg). %----------------------------------------------------------------------- mock(Test, State) -> Was = erlang:group_leader(), Mock = spawn_link(fun () -> mock_loop(State) end), true = erlang:group_leader(Mock, self()), Result = try Test() after true = erlang:group_leader(Was, self()) end, {ok, Output} = mock_call(Mock, stop), {Result, Output}. %======================================================================= mock_loop(State) -> receive {mock, From, Request} -> mock_request(From, Request, State); {io_request, From, Opaque, Request} -> io_request(From, Opaque, Request, State) end. %----------------------------------------------------------------------- mock_request(From, stop, State) -> mock_reply(From, {ok, State#state.output}), stop. %----------------------------------------------------------------------- io_request(From, Opaque, {put_chars, Encoding, Chars}, State) -> io_put_chars(From, Opaque, Encoding, Chars, State); io_request(From, Opaque, {put_chars, Encoding, M, F, A}, State) -> io_put_chars(From, Opaque, Encoding, M, F, A, State); io_request(From, Opaque, {get_geometry, Geometry}, State) -> io_get_geometry(From, Opaque, Geometry, State); io_request(From, Opaque, _Request, State) -> io_reply(From, Opaque, {error, enotsup}), mock_loop(State). %----------------------------------------------------------------------- io_put_chars(From, Opaque, _Encoding, Chars, State) when is_binary(Chars) -> io_reply(From, Opaque, ok), Output = <<(State#state.output)/binary, Chars/binary>>, mock_loop(State#state{output = Output}); io_put_chars(From, Opaque, Encoding, Chars, State) -> io_put_chars(From, Opaque, Encoding, iolist_to_binary(Chars), State). %----------------------------------------------------------------------- io_put_chars(From, Opaque, unicode, io_lib, format, [Format, Data], State) -> Chars = io_lib:format(Format, Data), io_put_chars(From, Opaque, unicode, Chars, State). %----------------------------------------------------------------------- io_get_geometry(From, Opaque, columns, State = #state{columns = Columns}) -> io_reply(From, Opaque, Columns), mock_loop(State); io_get_geometry(From, Opaque, rows, State = #state{rows = Rows}) -> io_reply(From, Opaque, Rows), mock_loop(State); io_get_geometry(From, Opaque, _Geometry, State) -> io_reply(From, Opaque, {error, enotsup}), mock_loop(State). %----------------------------------------------------------------------- io_reply(To, Opaque, Reply) -> To ! {io_reply, Opaque, Reply}, ok. %----------------------------------------------------------------------- mock_reply(To, Reply) -> To ! {mock, self(), Reply}, ok. %----------------------------------------------------------------------- mock_call(Mock, Request) -> Mock ! {mock, self(), Request}, receive {mock, Mock, Response} -> Response after 1000 -> erlang:error({?MODULE, timeout}) end. %======================================================================= -ifdef(TEST). %----------------------------------------------------------------------- basic_test() -> Test = fun () -> io:put_chars(<<"Hello">>) end, Expect = <<"Hello">>, ?assertEqual(Expect, output(Test)). %----------------------------------------------------------------------- mixed_test() -> Test = fun () -> io:put_chars([$H, ["el", $l], <<"o">>]), io:format(" ~p!~n", [42]) end, Expect = <<"Hello 42!\n">>, ?assertEqual(Expect, output(Test)). %----------------------------------------------------------------------- columns_test() -> Test = fun () -> {ok, Columns} = io:columns(), {error, Rows} = io:rows(), {Columns, Rows} end, Expect = {{40, enotsup}, <<>>}, ?assertEqual(Expect, result_and_output(Test, {ok, 40}, enotsup)). %----------------------------------------------------------------------- rows_test() -> Test = fun () -> {error, Columns} = io:columns(), {ok, Rows} = io:rows(), {Columns, Rows} end, Expect = {{enotsup, 12}, <<>>}, ?assertEqual(Expect, result_and_output(Test, {error, enotsup}, 12)). %----------------------------------------------------------------------- input_test() -> Test = fun () -> io:get_chars(prompt, 10) end, Expect = {{error, enotsup}, <<>>}, ?assertEqual(Expect, result_and_output(Test)). %----------------------------------------------------------------------- -endif.
src/solarized_capture.erl
0.62395
0.4917
solarized_capture.erl
starcoder
%% ------------------------------------------------------------------- %% %% cuttlefish_validator: models a cuttlefish validator %% %% Copyright (c) 2013 Basho Technologies, Inc. All Rights Reserved. %% %% This file is provided to you under the Apache License, %% Version 2.0 (the "License"); you may not use this file %% except in compliance with the License. You may obtain %% a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, %% software distributed under the License is distributed on an %% "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY %% KIND, either express or implied. See the License for the %% specific language governing permissions and limitations %% under the License. %% %% ------------------------------------------------------------------- -module(cuttlefish_validator). -ifdef(TEST). -include_lib("eunit/include/eunit.hrl"). -compile(export_all). -endif. -record(validator, { name::string(), description::string(), func::fun() }). -type validator() :: #validator{}. -type validator_fun() :: fun((any()) -> boolean()). -type raw_validator() :: {validator, string(), string(), validator_fun()}. -export_type([validator/0]). -export([ parse/1, parse_and_merge/2, is_validator/1, name/1, description/1, func/1, replace/2]). -spec parse(raw_validator()) -> validator() | cuttlefish_error:error(). parse({validator, Name, Description, Fun}) -> #validator{ name = Name, description = Description, func = Fun }; parse(X) -> {error, {validator_parse, X}}. %% This assumes it's run as part of a foldl over new schema elements %% in which case, there's only ever one instance of a key in the list %% so keyreplace works fine. -spec parse_and_merge( raw_validator(), [validator()]) -> [validator()|cuttlefish_error:error()]. parse_and_merge({validator, ValidatorName, _, _} = ValidatorSource, Validators) -> NewValidator = parse(ValidatorSource), case lists:keyfind(ValidatorName, #validator.name, Validators) of false -> [ NewValidator | Validators]; _OldMapping -> lists:keyreplace(ValidatorName, #validator.name, Validators, NewValidator) end. -spec is_validator(any()) -> boolean(). is_validator(V) -> is_tuple(V) andalso element(1, V) =:= validator. -spec name(validator()) -> string(). name(V) -> V#validator.name. -spec description(validator()) -> string(). description(V) -> V#validator.description. -spec func(validator()) -> fun(). func(V) -> V#validator.func. -spec replace(validator(), [validator()]) -> [validator()]. replace(Validator, ListOfValidators) -> Exists = lists:keymember(name(Validator), #validator.name, ListOfValidators), case Exists of true -> lists:keyreplace(name(Validator), #validator.name, ListOfValidators, Validator); _ -> [Validator | ListOfValidators] end. -ifdef(TEST). -define(XLATE(X), lists:flatten(cuttlefish_error:xlate(X))). parse_test() -> ValidatorDataStruct = { validator, "name", "description", fun(X) -> X*2 end }, Validator = parse(ValidatorDataStruct), ?assertEqual("name", Validator#validator.name), ?assertEqual("description", Validator#validator.description), F = Validator#validator.func, ?assertEqual(4, F(2)), ok. getter_test() -> Validator = #validator{ name = "name", description = "description", func = fun(X) -> X*2 end }, ?assertEqual("name", name(Validator)), ?assertEqual("description", description(Validator)), Fun = func(Validator), ?assertEqual(4, Fun(2)), ok. replace_test() -> Element1 = #validator{ name = "name18", description = "description18", func = fun(X) -> X*2 end }, ?assertEqual(4, (Element1#validator.func)(2)), Element2 = #validator{ name = "name1", description = "description1", func = fun(X) -> X*4 end }, ?assertEqual(8, (Element2#validator.func)(2)), SampleValidators = [Element1, Element2], Override = #validator{ name = "name1", description = "description1", func = fun(X) -> X*5 end }, ?assertEqual(25, (Override#validator.func)(5)), NewValidators = replace(Override, SampleValidators), ?assertEqual([Element1, Override], NewValidators), ok. remove_duplicates_test() -> Sample1 = #validator{ name = "name1", description = "description1", func = fun(X) -> X*3 end }, ?assertEqual(6, (Sample1#validator.func)(2)), Sample2 = #validator{ name = "name1", description = "description1", func = fun(X) -> X*4 end }, ?assertEqual(8, (Sample2#validator.func)(2)), SampleValidators = [Sample1, Sample2], [NewValidator|_] = parse_and_merge( {validator, "name1", "description2", fun(X) -> X*10 end}, SampleValidators), F = func(NewValidator), ?assertEqual(50, F(5)), ?assertEqual("description2", description(NewValidator)), ?assertEqual("name1", name(NewValidator)), ok. parse_error_test() -> {ErrorAtom, ErrorTerm} = parse(not_a_raw_validator), ?assertEqual(error, ErrorAtom), ?assertEqual( "Poorly formatted input to cuttlefish_validator:parse/1 : not_a_raw_validator", ?XLATE(ErrorTerm)), ok. is_validator_test() -> ?assert(not(is_validator(not_a_validator))), V = #validator{ name = "name1", description = "description1", func = fun(X) -> X*4 end }, ?assertEqual(8, (V#validator.func)(2)), ?assert(is_validator(V)), ok. -endif.
deps/cuttlefish/src/cuttlefish_validator.erl
0.664976
0.41947
cuttlefish_validator.erl
starcoder
%% @author <NAME> <<EMAIL>> %% @copyright 2014 <NAME> %% @version 0.2.0 %% %% @doc %% %% The HDR histogram library is an Erlang native interface function wrapper of %% Mike Barker's C port of Gil Tene's HDR Histogram utility. %% %% %% A high dynamic range histogram is one that supports recording and analyzing %% sampled data points across a configurable range with configurable precision %% within that range. The precision is expressed as a number of significant %% figures in the recording. %% %% This HDR histogram implementation is designed for recording histograms of %% value measurements in latency sensitive environments. Although the native %% recording times can be as low as single digit nanoseconds there is added %% overhead in this wrapper/binding due to both the frontend overhead of converting %% from native C to the NIF interface, and the erlang overhead incurred calling %% into the NIFs. C'est la vie, I suppose. %% %% A distinct advantage of this histogram implementation is constant space and %% recording (time) overhead with an ability to recycle and reset instances whilst %% reclaiming already allocated space for reuse thereby reducing allocation cost %% and garbage collection overhead in the BEAM where repeated or continuous usage %% is likely. For example, a gen_server recording metrics continuously and resetting %% and logging histogram dumps on a periodic or other windowed basis. %% %% The code is released to the public domain, under the same terms as its %% sibling projects, as explained in the LICENSE.txt and COPYING.txt in the %% root of this repository, but normatively at: %% %% http://creativecommons.org/publicdomain/zero/1.0/ %% %% For users of this code who wish to consume it under the "BSD" license %% rather than under the public domain or CC0 contribution text mentioned %% above, the code found under this directory is *also* provided under the %% following license (commonly referred to as the BSD 2-Clause License). This %% license does not detract from the above stated release of the code into %% the public domain, and simply represents an additional license granted by %% http://creativecommons.org/publicdomain/zero/1.0/ %% %% ----------------------------------------------------------------------------- %% ** Beginning of "BSD 2-Clause License" text. ** %% %% Copyright (c) 2012, 2013, 2014 <NAME> %% Copyright (c) 2014 <NAME> %% Copyright (c) 2014 <NAME> %% All rights reserved. %% %% Redistribution and use in source and binary forms, with or without %% modification, are permitted provided that the following conditions are met: %% %% 1. Redistributions of source code must retain the above copyright notice, %% this list of conditions and the following disclaimer. %% %% 2. Redistributions in binary form must reproduce the above copyright notice, %% this list of conditions and the following disclaimer in the documentation %% and/or other materials provided with the distribution. %% %% THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" %% AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE %% IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE %% ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE %% LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR %% CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF %% SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS %% INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN %% CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) %% ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF %% THE POSSIBILITY OF SUCH DAMAGE. %% %% @end -module(hdr_iter). -export([open/3]). -export([each/2]). -export([each/3]). -export([close/1]). -type ref() :: binary(). %% NIF private data (looks like empty binary) -spec open(IterationType, HdrRef, Options) -> {ok,Ref} | {error,Reason} when IterationType :: record | linear | logarithmic | percentile, HdrRef :: ref(), Ref :: ref(), Options :: [proplists:property()], Reason :: term(). %% @doc Open a fresh instance of a high dynamic range (HDR) histogram iterator open(record,HdrRef,Options) -> open_and_init(1, HdrRef, Options); open(linear,HdrRef,Options) -> open_and_init(2, HdrRef, Options); open(logarithmic,HdrRef,Options) -> open_and_init(4, HdrRef, Options); open(percentile,HdrRef,Options) -> open_and_init(8, HdrRef, Options). %% @private open_and_init(IterType,HdrRef,Options) -> {ok,IterRef} = hdr_histogram:iter_open(IterType), case hdr_histogram:iter_init(IterRef,HdrRef,Options) of ok -> {ok,IterRef}; Error -> throw({bad_init,Error}) end. -spec each(IteratorRef,EachFun) -> ok | {error,Reason} when EachFun :: fun((Data) -> any()), IteratorRef :: ref(), Data :: term(), Reason :: term(). %% @doc Iterate over histogram applying a function to each data point each(IteratorRef,EachFun) -> each(IteratorRef,fun(Data,_) -> EachFun(Data), ok end, ok). -spec each(IteratorRef,EachFun,Initial) -> {ok,Accum} | {error,Reason} when EachFun :: fun(({IteratorType,Data},Acc) -> any()), IteratorType :: record | linear | logarithmic | percentile, Data :: list({atom(),term()}), %% TODO FIXME type spec Acc :: term(), IteratorRef :: ref(), Reason :: term(), Initial :: term(), Accum :: term(). %% @doc Iterate over histogram applying a function to each data point accumulating a result each(IteratorRef,EachFun,InitialAcc) -> scan(IteratorRef,EachFun,InitialAcc). scan(IteratorRef,EachFun,Acc) -> case hdr_histogram:iter_next(IteratorRef) of {false,_} -> Acc; {Type,Data} -> NewAcc = EachFun({Type,Data},Acc), scan(IteratorRef,EachFun,NewAcc) end. -spec close(Ref) -> ok | {error,term()} when Ref :: ref(). %% @doc Close this HDR histogram instance and free any system resources close(Ref) -> hdr_histogram:iter_close(Ref).
src/hdr_iter.erl
0.57093
0.600042
hdr_iter.erl
starcoder
% @copyright 2007-2014 Zuse Institute Berlin % Licensed under the Apache License, Version 2.0 (the "License"); % you may not use this file except in compliance with the License. % You may obtain a copy of the License at % % http://www.apache.org/licenses/LICENSE-2.0 % % Unless required by applicable law or agreed to in writing, software % distributed under the License is distributed on an "AS IS" BASIS, % WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. % See the License for the specific language governing permissions and % limitations under the License. %% @author <NAME> <<EMAIL>> %% TODO change from camel case to prefix_infix_postfix() %% @doc Math utility functions. %% @version $Id$ -module(mathlib). -author('<EMAIL>'). -vsn('$Id$'). -export([closestPoints/1, euclideanDistance/1, euclideanDistance/2, u/1, vecAdd/2, vecSub/2, vecMult/2, vecWeightedAvg/4, zeros/1, median/1, nearestCentroid/2, aggloClustering/2]). -export([factorial/1, binomial_coeff/2, gcd/2]). %% for type_check_SUITE -export([binomial_coeff_feeder/2, factorial_feeder/1, zeros_feeder/1]). -type(vector() :: [number(),...]). %% @doc Median of an unsorted non-empty list of numbers, i.e. a vector. -spec median(vector()) -> number(). median(L) -> L1 = lists:sort(L), N = length(L1), case N rem 2 of 1 -> lists:nth(round(N / 2), L1); 0 -> (lists:nth(trunc(N / 2), L1) + lists:nth(trunc(N / 2) + 1, L1)) / 2 end. %% @doc Add two vectors X,Y, i.e. X + Y. -spec vecAdd(X::vector(), Y::vector()) -> vector(). vecAdd(X, Y) -> lists:zipwith(fun(Xi, Yi) -> Xi + Yi end, X, Y). %% @doc Substract two vectors X,Y, i.e. X - Y. -spec vecSub(X::vector(), Y::vector()) -> vector(). vecSub(X, Y) -> lists:zipwith(fun(Xi, Yi) -> Xi - Yi end, X, Y). %% @doc Multiply vector V with a scalar S. -spec vecMult(V::vector(), S::float()) -> vector(). vecMult(V, S) -> lists:map(fun(X) -> S*X end, V). -spec vecWeightedAvg(V1::vector(), V2::vector(), W1::float(), W2::float()) -> vector(). vecWeightedAvg(V1, V2, W1, W2) -> vecMult(vecAdd(vecMult(V1, W1), vecMult(V2, W2)), 1 / (W1 + W2)). %% @doc Euclidean distance between origin and V. -spec euclideanDistance(V::vector()) -> Distance::float(). euclideanDistance(V) -> math:sqrt(lists:foldl(fun(Vi, OldDist) -> OldDist + math:pow(Vi, 2) end, 0.0, V)). %% @doc Euclidean distance between two vectors. -spec euclideanDistance(V::vector(), W::vector()) -> Distance::float(). euclideanDistance(V, W) -> math:sqrt(util:zipfoldl(fun(Vi, Wi) -> math:pow(Vi - Wi, 2) end, fun(Dist, OldDist) -> OldDist + Dist end, V, W, 0.0)). %% @doc Unit vector u(v) = v/||v|| -spec u(V::vector()) -> UV::vector(). u(V) -> vecMult(V, 1 / euclideanDistance(V)). %% @doc Get the nearest centroid to U from the list Centroids, including the euclidian % distance. The function returns 'none' if no nearest centroid can be found. Ambiguity is % resolved by picking the first one of the nearest centroids. -spec nearestCentroid(U::dc_centroids:centroid(), dc_centroids:centroids()) -> {Distance::float(), NearestCentroid::dc_centroids:centroid()} | none. nearestCentroid(_U, []) -> none; nearestCentroid(U, [U|T]) -> nearestCentroid(U, T); nearestCentroid(U, [X|Centroids]) -> CoordU = dc_centroids:get_coordinate(U), CoordsX = dc_centroids:get_coordinate(X), First = {euclideanDistance(CoordU, CoordsX), X}, lists:foldl(fun (C, {CurrentDistance, _CurrentMin} = Current) -> case C of U -> Current; _ -> CoordC = dc_centroids:get_coordinate(C), NewDistance = euclideanDistance(CoordU, CoordC), case NewDistance < CurrentDistance of true -> {NewDistance, C}; false -> Current end end end, First, Centroids). %% @doc Find indices of closest centroids. -spec closestPoints(dc_centroids:centroids()) -> {float(), dc_centroids:centroid(), dc_centroids:centroid()} | none. closestPoints([]) -> none; closestPoints([_]) -> none; closestPoints([First, Second|_] = Centroids) -> FirstDist = dc_centroids:distance(First, Second), lists:foldl( fun (Centroid, {CurrentMinDist, _, _} = Acc) -> % get the centroid with minimum distance to Centroid. If this is less than % the distance of the centroids in Acc, exchange case nearestCentroid(Centroid, Centroids) of none -> Acc; {Dist, CentroidMin} -> case Dist < CurrentMinDist of true -> {Dist, Centroid, CentroidMin}; false -> Acc end end end, {FirstDist, First, Second}, Centroids). -spec zeros_feeder(0..10000) -> {0..10000}. zeros_feeder(N) -> {N}. %% @doc Create a list with N zeros. -spec zeros(N::0) -> []; (N::pos_integer()) -> [0,...]. zeros(N) -> lists:duplicate(N, 0). %% @doc Get closest centroids and merge them if their distance is within Radius. -spec aggloClustering(Centroids::dc_centroids:centroids(), Radius::number()) -> dc_centroids:centroids(). aggloClustering(Centroids, Radius) when Radius >= 0 -> case closestPoints(Centroids) of none -> Centroids; {Min, I, J} -> aggloClusteringHelper(Centroids, Radius, Min, I, J) end. -spec aggloClusteringHelper (Centroids::[dc_centroids:centroid(),...], Radius::number(), Min::float(), I::dc_centroids:centroid(), J::dc_centroids:centroid()) -> dc_centroids:centroids(). % Note: closestPoints/1 creates Min, I, J and only returns {-1, -1, -1} if % Centroids contains less than two elements. This is not the case in the first % pattern and we can thus assume these values are pos_integer(). aggloClusteringHelper(Centroids, _Radius, 0.0, _, _) -> Centroids; aggloClusteringHelper(Centroids, Radius, Min, I, J) when Min =< Radius -> {C1, S1} = dc_centroids:get_coordinate_and_relative_size(I), {C2, S2} = dc_centroids:get_coordinate_and_relative_size(J), NewCoordinate = vecWeightedAvg(C1, C2, S1, S2), NewCentroid = dc_centroids:new(NewCoordinate, S1+S2), NewCentroids = [NewCentroid | lists:subtract(Centroids, [I,J])], case closestPoints(NewCentroids) of none -> NewCentroids; {Min1, I1, J1} -> aggloClusteringHelper(NewCentroids, Radius, Min1, I1, J1) end; aggloClusteringHelper(Centroids, _Radius, _Min, _I, _J) -> Centroids. % @doc Calculates the binomial coefficient of n over k for n >= k. % see http://rosettacode.org/wiki/Evaluate_binomial_coefficients#Erlang -spec binomial_coeff(non_neg_integer(), non_neg_integer()) -> integer(). binomial_coeff(_, 0) -> 1; binomial_coeff(N, K) when N >= K -> choose(N, K, 1, 1). -spec binomial_coeff_feeder(0..100, 0..100) -> {non_neg_integer(), non_neg_integer()}. binomial_coeff_feeder(X, Y) -> {erlang:max(X, Y), erlang:min(X, Y)}. -spec choose(non_neg_integer(), non_neg_integer(), non_neg_integer(), non_neg_integer()) -> non_neg_integer(). choose(N, K, K, Acc) -> (Acc * (N-K+1)) div K; choose(N, K, I, Acc) -> choose(N, K, I+1, (Acc * (N-I+1)) div I). -spec factorial_feeder(0..20) -> {0..20}. factorial_feeder(N) -> {N}. % @doc calculates N! -spec factorial(non_neg_integer()) -> pos_integer(). factorial(N) -> factorial(N, 1). -compile({nowarn_unused_function, {factorial_feeder, 2}}). -spec factorial_feeder(0..20, pos_integer()) -> {0..20, pos_integer()}. factorial_feeder(N, Acc) -> {N, Acc}. -spec factorial(non_neg_integer(), pos_integer()) -> pos_integer(). factorial(0, Acc) -> Acc; factorial(N, Acc) -> factorial(N - 1, N * Acc). %% @doc Calculates the greatest common divisor of two integers. -spec gcd(non_neg_integer(), non_neg_integer()) -> non_neg_integer(). gcd(A, 0) -> A; gcd(A, B) -> gcd(B, A rem B).
src/mathlib.erl
0.53437
0.469703
mathlib.erl
starcoder
%%% @doc A Flat Tree is a deterministic way of using a list as an index %%% for nodes in a tree. Essentially a simpler way of representing the %%% position of nodes. %%% %%% A Flat Tree is also refered to as 'bin numbers' described here %%% in RFC 7574: https://datatracker.ietf.org/doc/html/rfc7574#section-4.2 %%% %%% As an example (from the RFC), here's a tree with a width of 8 leafs %%% and a depth of 3: %%%``` %%% 3 7 %%% / \ %%% / \ %%% / \ %%% / \ %%% 2 3 11 %%% / \ / \ %%% / \ / \ %%% / \ / \ %%% 1 1 5 9 13 %%% / \ / \ / \ / \ %%% Depth 0 0 2 4 6 8 10 12 14 %%% C0 C1 C2 C3 C4 C5 C6 C7 %%% %%% The flat tree is the list [0..14] %%% The content (leafs) is C0..C7 %%% %%% Using the flat tree, we can see index: %%% - 7 represents all the content (C0..C7) %%% - 1 represents C0 and C1 %%% - 3 represent C0..C3 %%% ... etc ... %%% ''' %%% %%% Even numbers are always leafs at depth 0 %%% Odd numbers are parents at depths > 0 %%% %%% This work is almost a direct port of the DAT rust project: %%% https://github.com/datrs/flat-tree %%% %%% Along with the content described here: %%% https://datprotocol.github.io/book/ch01-01-flat-tree.html %%% @end -module(flat_tree). -export([ index/2, depth/1, offset/1, parent/1, sibling/1, uncle/1, children/1, left_child/1, right_child/1, left_span/1, right_span/1, spans/1, count/1, full_roots/1 ]). %% @doc Find the index value at a given depth and offset from the left of the tree. -spec index(Depth :: pos_integer(), Offset :: pos_integer()) -> Index :: pos_integer(). index(Depth, Offset) -> Offset bsl (Depth + 1) bor ((1 bsl Depth) - 1). %% @doc Calculate the depth of the tree for a given index. %% The root of the tree is the highest depth. %% The leafs are always at depth 0. %% %% To find the depth count the number of trailing '1' bits (LSBs) %% for the given Index. For example, the value: %%``` %% 11000001, %% ^- has a depth of 1. %% 11000011 %% ^- has a depth of 2. %%''' %% %% Interestingly, even numbers have no trailing '1s' and %% therefore are always at depth 0. %% %% Side note: Funky business converting an integer to binary... %% the built in function (integer_to_binary) returns something different. %% ``` %% <<Index>> != integer_to_binary(Index) %%''' %% %% This is not the most optimal solution. A better approach would be %% with a lookup table. %% Here we: %% - reverse the binary and push the bits on to the head of a list (Acc) %% - then, we count '1s' till we hit a '0'. -spec depth(Index :: pos_integer()) -> Depth :: pos_integer(). depth(Index) -> dep(<<Index>>, <<>>). dep(<<>>, Acc) -> count_ones(Acc, 0); dep(<<X:1, R/bitstring>>, Acc) -> dep(R, [X | Acc]). count_ones([], Count) -> Count; count_ones([H | T], Count) -> case H =:= 1 of true -> count_ones(T, Count + 1); _ -> Count end. %% @doc Return the offset for an index from the left side of the tree. %% %% For example: (0, 1, 3, 7) have an offset of 0 %% (Tree is rotated to right in diagram) %%``` %% (0)┐ %% (1)┐ %% 2─┘ │ %% (3)┐ %% 4─┐ │ │ %% 5─┘ │ %% 6─┘ │ %% (7) %% %% While (2, 5, 11) have an offset of 1: %% 0──┐ %% 1──┐ %% (2)─┘ │ %% 3──┐ %% 4──┐ │ │ %% (5)─┘ │ %% 6──┘ │ %% 7 %% 8──┐ │ %% 9──┐ │ %% 10──┘ │ │ %% (11)─┘ %% 12──┐ │ %% 13──┘ %% 14──┘ %%''' -spec offset(Index :: pos_integer()) -> Offset :: pos_integer(). offset(Index) -> Depth = depth(Index), case is_even(Index) of true -> Index div 2; _ -> Index bsr (Depth + 1) end. %% @doc Return the index of the parent for the given index. -spec parent(Index :: pos_integer()) -> Parent :: pos_integer(). parent(Index) -> Depth = depth(Index), index(Depth + 1, offset(Index) bsr 1). %% @doc Return the index of node that shares a parent -spec sibling(Index :: pos_integer()) -> Sibling :: pos_integer(). sibling(Index) -> Depth = depth(Index), index(Depth, offset(Index) bxor 1). %% @doc Return a parent's sibiling -spec uncle(Index :: pos_integer()) -> Uncle :: pos_integer(). uncle(Index) -> Depth = depth(Index), index(Depth + 1, offset(parent(Index)) bxor 1). %% @doc Find the Indices of the children for a given index. -spec children(Index :: pos_integer()) -> none | {LeftIndex :: pos_integer(), RightIndex :: pos_integer}. children(Index) -> Depth = depth(Index), find_chillin(Index, Depth). find_chillin(Index, _Depth) when (Index band 1) =:= 0 -> %% The index is an even number, it's a leaf none; find_chillin(Index, Depth) when Depth =:= 0 -> %% Still at a leaf, return the Index {Index, Index}; find_chillin(Index, Depth) -> Offset = offset(Index) * 2, {index(Depth - 1, Offset), index(Depth - 1, Offset + 1)}. %% @doc Get the child to the left -spec left_child(Index :: pos_integer()) -> none | Index :: pos_integer(). left_child(Index) -> Depth = depth(Index), find_left_child(Index, Depth). find_left_child(Index, _Depth) when (Index band 1) =:= 0 -> none; find_left_child(Index, Depth) when Depth =:= 0 -> Index; find_left_child(Index, Depth) -> index(Depth - 1, offset(Index) bsl 1). %% @doc Get the child to the right -spec right_child(Index :: pos_integer()) -> none | Index :: pos_integer(). right_child(Index) -> Depth = depth(Index), find_right_child(Index, Depth). find_right_child(Index, _Depth) when (Index band 1) =:= 0 -> none; find_right_child(Index, Depth) when Depth =:= 0 -> Index; find_right_child(Index, Depth) -> index(Depth - 1, (offset(Index) bsl 1) + 1). %% @doc The index of the left most node in the span -spec left_span(Index :: pos_integer()) -> Index :: pos_integer(). left_span(Index) -> Depth = depth(Index), l_span(Index, Depth). l_span(Index, Depth) when Depth =:= 0 -> Index; l_span(Index, Depth) -> offset(Index) * (2 bsl Depth). %% @doc The index of the right most node in the span -spec right_span(Index :: pos_integer()) -> Index :: pos_integer(). right_span(Index) -> Depth = depth(Index), r_span(Index, Depth). r_span(Index, Depth) when Depth =:= 0 -> Index; r_span(Index, Depth) -> (offset(Index) + 1) * (2 bsl Depth) - 2. %% @doc return the span of the left and right nodes for the given index. -spec spans(Index :: pos_integer()) -> {LeftIndex :: pos_integer(), RightIndex :: pos_integer()}. spans(Index) -> {left_span(Index), right_span(Index)}. %% Return how many nodes are in the subtee at the given index including %% the node of the index in the count. -spec count(Index :: pos_integer()) -> Count :: pos_integer(). count(Index) -> Depth = depth(Index), (2 bsl Depth) - 1. %% @doc Return a list of indices that represent the full nodes and subtrees %% to the left of the given index. For example, given this (partial) tree: %%``` %% 3 %% / \ %% / \ %% 1 5 %% / \ / \ %% 0 2 4 6 %% %% You get: %% [1,4] = flat_tree:full_roots(6) %% As index 1 is full to the left, and 4 is the node to the left %% %% [1] = flat_tree:full_roots(4) %% 1 is the only full node to the left of 4 %% %% and... %% [] = flat_tree:full_roots(0). %% as there are no nodes to the left of 0 %%''' %% The input Index must be a leaf index (even number), otherwise you get an error. -spec full_roots(Index :: pos_integer()) -> {error, only_even_indices_allowed} | [Index :: pos_integer()]. full_roots(Index) when Index band 1 =:= 1 -> {error, only_even_indices_allowed}; full_roots(Index) -> find_roots(Index bsr 1, 0, 1, []). calculate_factor(Factor, Index) when Factor * 2 =< Index -> calculate_factor(Factor * 2, Index); calculate_factor(Factor, _) -> Factor. find_roots(0, _, _, Nodes) -> lists:reverse(Nodes); find_roots(Index, Offset, Factor, Nodes) -> NextFactor = calculate_factor(Factor, Index), find_roots( Index - NextFactor, Offset + 2 * NextFactor, 1, [Offset + NextFactor - 1 | Nodes] ). %%% %%% Private %%% is_even(Index) when is_integer(Index) -> (Index band 1) =:= 0.
src/flat_tree.erl
0.791096
0.909265
flat_tree.erl
starcoder
%% %% %CopyrightBegin% %% %% Copyright Ericsson AB 1996-2010. All Rights Reserved. %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %% %% %CopyrightEnd% %% %% -module(othello_adt). -compile(export_all). %%------------------------------------------------------- %% Use three main states for the strategy: %% %% BeginPlay: Stay in the inner square as long as possible. %% Use the possible_draws/3. %% %% MiddlePlay: Try to choose stable markers (?) %% Use stable/3 %% %% EndPlay: Try to flip as many markers as possible %% %% The transition from Begin to Middle is obvious. From Middle %% to End however, is can be discussed. %%------------------------------------------------------- test(N,B) -> X=new(B), statistics(wall_clock), test0(N,X), {_,T} = statistics(wall_clock), {time_was,T/N}. test0(0,_) -> true; test0(N,X) -> possible_draws(black,X,begin_play), test0(N-1,X). %%------------------------------------------------------- %% new/1 - returns a new board %% %% Uses a tuple for storing the board %%------------------------------------------------------- new(B) -> Board = mk_board(B), {ordsets:from_list([18,19,20,21,26,29,34,37,42,43,44,45]),Board}. mk_board(t) -> Tup = list_to_tuple(gen_list(64,grey)), Tup1 = setelement(28+1, Tup, white), Tup2 = setelement(35+1, Tup1, white), Tup3 = setelement(27+1, Tup2, black), gen_score_board(), setelement(36+1, Tup3, black). gen_list(0,_) -> []; gen_list(I,Def) -> [Def|gen_list(I-1,Def)]. gen_score_board() -> put(score,list_to_tuple(gen_list(64,0))). %%------------------------------------------------------- %% pos(Col,Row) - returns a position describing column %% and row. %% Col and Row have the range 1 - 8. %%------------------------------------------------------- pos(Col,Row) -> ((Row - 1) bsl 3) + (Col - 1). %%------------------------------------------------------- %% col(Pos) - returns the column of the Pos position %%------------------------------------------------------- col(Pos) -> (Pos band 7) + 1. %%------------------------------------------------------- %% row(Pos) - returns the row of the Pos position %%------------------------------------------------------- row(Pos) -> (Pos bsr 3) + 1. %%------------------------------------------------------- %% is_draw(Pos,Colour,Board) - returns true if Pos is a %% correct draw. %%------------------------------------------------------- is_draw(Pos,Colour,{Bset,Board}) -> case ordsets:is_element(Pos,Bset) of true -> case catch is_good(Colour,Pos,Board) of true -> true; _ -> false end; _ -> false end. %%------------------------------------------------------- %% set(Pos,Colour,Board) - returns an updated board %%------------------------------------------------------- set(Pos,Colour,{Bset,Board}) -> case ordsets:is_element(Pos,Bset) of true -> NewBoard = setelement(Pos+1,Board,Colour), Empty = empty_neighbour(Pos,NewBoard), NewBset = ordsets:union(Empty,ordsets:del_element(Pos,Bset)), turn(Colour,Pos,{NewBset,NewBoard}); _ -> {error,invalid_position} end. empty_neighbour(Pos,Board) -> ordsets:from_list(empty_neighbour(Pos,Board,deltas())). empty_neighbour(_,_,[]) -> []; empty_neighbour(Pos,Board,[H|T]) -> case is_empty(Pos+H,dir(Pos,H),Board) of true -> [Pos+H|empty_neighbour(Pos,Board,T)]; _ -> empty_neighbour(Pos,Board,T) end. is_empty(_,false,_) -> false; is_empty(X,_,_Board) when X<0 -> false; is_empty(X,_,_Board) when X>63 -> false; is_empty(X,_,Board) -> case element(X+1,Board) of grey -> true; % Empty _ -> false end. %%------------------------------------------------------- %% get(Pos,Board) - returns the contents in Pos %%------------------------------------------------------- get(Pos,{_Bset,Board}) -> element(Pos+1,Board). %%------------------------------------------------------- %% pieces(Colour,Board) - returns the number of Colour %% pieces. %%------------------------------------------------------- pieces(Colour,{_Bset,Board}) -> pieces(Colour,Board,0,0). pieces(Colour,Board,Pos,Count) when Pos < 64 -> case element(Pos+1,Board) of Colour -> pieces(Colour,Board,Pos+1,Count+1); _ -> pieces(Colour,Board,Pos+1,Count) end; pieces(_,_,_,Count) -> Count. %%------------------------------------------------------- %% possible_draws(Colour, Board, State) %% %% Returns a list of possible draws regarding the current %% strategy state. %%------------------------------------------------------- possible_draws(Colour,{Bset,Board},begin_play) -> Dset = ordsets:intersection(Bset,inner_square()), possible_draws_0(Colour,Dset,Board); possible_draws(Colour,{Bset,Board},_) -> possible_draws_0(Colour,Bset,Board). possible_draws(Colour,{Bset,Board}) -> possible_draws_0(Colour,Bset,Board). possible_draws_0(_,[],_) -> []; possible_draws_0(Colour,[H|T],Board) -> case catch is_good(Colour,H,Board) of true -> [H|possible_draws_0(Colour,T,Board)]; false -> possible_draws_0(Colour,T,Board) end. %%------------------------------------------------------- %% evaluate_board(Colour,Board) - returns the value of %% the board from Colours %% point of view. %%------------------------------------------------------- evaluate_board(Colour,{_Bset,Board}) -> Score = get(score), % Initialized (zeroed) score board !! Colour1 = swap(Colour), Score1 = eval_board(Colour,Colour1,Score,Board,0), Score2 = cnt_corner(0,Score1,Board,Colour,Colour1), Score3 = cnt_corner(7,Score2,Board,Colour,Colour1), Score4 = cnt_corner(56,Score3,Board,Colour,Colour1), Score5 = cnt_corner(63,Score4,Board,Colour,Colour1), count(Score5,0). % A = count(Score5,0), % io:format('Score = ~w~n',[A]), % A. eval_board(MyCol,OtCol,Score,Board,Pos) when Pos < 64 -> case element(Pos+1,Board) of MyCol -> Score1 = setelement(Pos+1,Score,score(Pos)), eval_board(MyCol,OtCol,Score1,Board,Pos+1); OtCol -> Score1 = setelement(Pos+1,Score,-score(Pos)), eval_board(MyCol,OtCol,Score1,Board,Pos+1); _ -> eval_board(MyCol,OtCol,Score,Board,Pos+1) end; eval_board(_,_,Score,_,_) -> Score. cnt_corner(Corner,Score,Board,MyCol,OtCol) -> case element(Corner+1,Board) of MyCol -> cnt_corn(Corner,setelement(Corner+1,Score,50), Board,50,MyCol); OtCol -> cnt_corn(Corner,setelement(Corner+1,Score,-50), Board,-50,OtCol); _ -> Score end. cnt_corn(0,Score,Board,Value,Colour) -> Score1 = cnt_corn(0,1,8,Score,Board,Value,Colour), cnt_corn(0,8,1,Score1,Board,Value,Colour); cnt_corn(7,Score,Board,Value,Colour) -> Score1 = cnt_corn(7,-1,8,Score,Board,Value,Colour), cnt_corn(7,8,-1,Score1,Board,Value,Colour); cnt_corn(56,Score,Board,Value,Colour) -> Score1 = cnt_corn(56,1,-8,Score,Board,Value,Colour), cnt_corn(56,-8,1,Score1,Board,Value,Colour); cnt_corn(63,Score,Board,Value,Colour) -> Score1 = cnt_corn(63,-1,-8,Score,Board,Value,Colour), cnt_corn(63,-8,-1,Score1,Board,Value,Colour). cnt_corn(Pos,Dir,LineDir,Score,Board,Value,Colour) -> case dir(Pos,Dir) of Dir -> NextEdge = Pos+Dir, case element(NextEdge+1,Board) of Colour -> Score1 = setelement(NextEdge+1,Score,Value), Score2 = cnt_line(NextEdge,LineDir,Score1,Board, Colour,Value), cnt_corn(NextEdge,Dir,LineDir,Score2,Board,Value,Colour); _ -> Score end; _ -> Score end. cnt_line(Pos,Dir,Score,Board,Colour,Value) -> case dir(Pos,Dir) of Dir -> OnLinePos = Pos+Dir, case element(OnLinePos+1,Board) of Colour -> Score1 = setelement(OnLinePos+1,Score,Value), cnt_line(OnLinePos,Dir,Score1,Board,Colour,Value); _ -> Score end; _ -> Score end. count(Score,Pos) when Pos < 64 -> element(Pos+1,Score) + count(Score,Pos+1); count(_,_) -> 0. swap(white) -> black; swap(black) -> white. %%------------------------------------------------------- %% stable(Colour,Pos,Board) - returns a value 0-8 %% %% A high value is regarded as more stable than a lower one. %% The stability means how many "friendly" neighbours there %% are, i.e markers of the same colour. Neighbours positions %% outside the board are regarded as friendly. %%------------------------------------------------------- stable(Colour,Pos,{_,Board}) -> stable(deltas(),Colour,Pos,Board). stable([],_,_,_) -> 0; stable([H|T],Colour,Pos,Board) -> stable_val(Colour,Pos,H,Board) + stable(T,Colour,Pos,Board). stable_val(_,H,D,_) when H+D<0 -> 1; stable_val(_,H,D,_) when H+D>63 -> 1; stable_val(black,H,D,Board) -> case element((H+D)+1,Board) of black -> 1; _ -> 0 end; stable_val(white,H,D,Board) -> case element((H+D)+1,Board) of white -> 1; _ -> 0 end. %%------------------------------------------------------- %% diff(Board,OldBoard) - return a list of the positions %% with changed pieces. %% [{Pos1,Colour1},...] %%------------------------------------------------------- diff(Board,OldBoard) -> diff(0,Board,OldBoard). diff(Pos,Board,OldBoard) when Pos < 64 -> OldP = get(Pos,OldBoard), case get(Pos,Board) of OldP -> diff(Pos+1,Board,OldBoard); NewP -> [{Pos,NewP}|diff(Pos+1,Board,OldBoard)] end; diff(_,_,_) -> []. %%------------------------------------------------------- %% all_pos(Board) - return a list of the positions colour. %% [{Pos1,Colour1},...] %%------------------------------------------------------- all_pos(Board) -> all_pos(0,Board). all_pos(Pos,Board) when Pos < 64 -> [{Pos,get(Pos,Board)}|all_pos(Pos+1,Board)]; all_pos(_,_) -> []. %%------------------------------------------------------- %% Internal stuff %%------------------------------------------------------- deltas() -> [9,8,7,1,-1,-7,-8,-9]. inner_square() -> [18,19,20,21,26,27,28,29,34,35,36,37,42,43,44,45]. % Is already an ordset % Save list traversing. % ordsets:list_to_set([18,19,20,21,26,27,28,29,34,35,36,37,42,43,44,45]). inv(black) -> white; inv(white) -> black. is_good(Colour,H,Board) -> is_good_0(Colour,H,dir(H,-9),Board), is_good_0(Colour,H,dir(H,-8),Board), is_good_0(Colour,H,dir(H,-7),Board), is_good_0(Colour,H,dir(H,-1),Board), is_good_0(Colour,H,dir(H,1),Board), is_good_0(Colour,H,dir(H,7),Board), is_good_0(Colour,H,dir(H,8),Board), is_good_0(Colour,H,dir(H,9),Board), false. is_good_0(_,_,false,_) -> false; is_good_0(_,H,D,_) when is_integer(H), is_integer(D), H+D<0 -> false; is_good_0(_,H,D,_) when is_integer(H), is_integer(D), H+D>63 -> false; is_good_0(black,H,D,Board) when is_integer(H), is_integer(D) -> case element((H+D)+1,Board) of white -> is_good_1(black,H+D,dir(H+D,D),Board); _ -> false end; is_good_0(white,H,D,Board) when is_integer(H), is_integer(D) -> case element((H+D)+1,Board) of black -> is_good_1(white,H+D,dir(H+D,D),Board); _ -> false end. is_good_1(_,_,false,_) -> false; is_good_1(_,H,D,_) when is_integer(H), is_integer(D), H+D<0 -> false; is_good_1(_,H,D,_) when is_integer(H), is_integer(D), H+D>63 -> false; is_good_1(black,H,D,Board) when is_integer(H), is_integer(D) -> case element((H+D)+1,Board) of white -> is_good_1(black,H+D,dir(H+D,D),Board); black -> throw(true); _ -> false end; is_good_1(white,H,D,Board) when is_integer(H), is_integer(D) -> case element((H+D)+1,Board) of black -> is_good_1(white,H+D,dir(H+D,D),Board); white -> throw(true); _ -> false end. %%------------------------------------------------------- %% turn(Colour,Draw,Board) - returns an updated board %% turn all possible pieces %% on the board %% Neighbours are not changed !! %%------------------------------------------------------- turn(Colour,Draw,{Bset,Board}) -> {Bset,turn(Colour,Draw,-9, turn(Colour,Draw,-8, turn(Colour,Draw,-7, turn(Colour,Draw,-1, turn(Colour,Draw,1, turn(Colour,Draw,7, turn(Colour,Draw,8, turn(Colour,Draw,9,Board))))))))}. turn(Colour,H,D,Board) -> case catch is_good_0(Colour,H,dir(H,D),Board) of true -> turn_0(Colour,H,D,Board); false -> Board end. turn_0(_,H,D,B) when is_integer(H), is_integer(D), H+D<0 -> B; turn_0(_,H,D,B) when is_integer(H), is_integer(D), H+D>63 -> B; turn_0(black,H,D,Board) when is_integer(H), is_integer(D) -> E = H+D, case element(E+1,Board) of white -> turn_0(black,H+D,D,swap(black,E,Board)); _ -> Board end; turn_0(white,H,D,Board) when is_integer(H), is_integer(D) -> E = H+D, case element(E+1,Board) of black -> turn_0(white,H+D,D,swap(white,E,Board)); _ -> Board end. %%------------------------------------------------------- %% swap(Colour,Pos,Board) - returns an updated board %% turn a piece on the board %% Neighbours are not changed !! %%------------------------------------------------------- swap(Colour,Pos,Board) when is_integer(Pos) -> setelement(Pos+1,Board,Colour). score(Pos) -> score1({col(Pos),row(Pos)}). score1({Column,1}) when Column >= 3, Column =< 6 -> 20; score1({Column,8}) when Column >= 3, Column =< 6 -> 20; score1({1,Line}) when Line >= 3, Line =< 6 -> 20; score1({8,Line}) when Line >= 3, Line =< 6 -> 20; score1({Column,2}) when Column >= 3, Column =< 6 -> -7; score1({Column,7}) when Column >= 3, Column =< 6 -> -7; score1({2,Line}) when Line >= 3, Line =< 6 -> -7; score1({7,Line}) when Line >= 3, Line =< 6 -> -7; score1({Column,Line}) when Column >= 3, Column =< 6, Line >= 3, Line =< 6 -> 1; score1({1,1}) -> 100; score1({1,8}) -> 100; score1({8,1}) -> 100; score1({8,8}) -> 100; score1({2,1}) -> -30; score1({7,1}) -> -30; score1({1,2}) -> -30; score1({8,2}) -> -30; score1({1,7}) -> -30; score1({8,7}) -> -30; score1({2,8}) -> -30; score1({7,8}) -> -30; score1({2,2}) -> -50; score1({7,2}) -> -50; score1({2,7}) -> -50; score1({7,7}) -> -50. %%------------------------------------------------------- %% dir(Pos,Dir) - return Dir if allowed direction at Pos. %% else return false. %%------------------------------------------------------- dir(0,1) -> 1; % {1,1} dir(0,8) -> 8; dir(0,9) -> 9; dir(0,_) -> false; dir(7,-1) -> -1; % {8,1} dir(7,7) -> 7; dir(7,8) -> 8; dir(7,_) -> false; dir(56,-8) -> -8; % {1,8} dir(56,-7) -> -7; dir(56,1) -> 1; dir(56,_) -> false; dir(63,-9) -> -9; % {8,8} dir(63,-8) -> -8; dir(63,-1) -> -1; dir(63,_) -> false; dir(Pos,-1) when (Pos bsr 3) == 0 -> -1; % {_,1} dir(Pos,1) when (Pos bsr 3) == 0 -> 1; dir(Pos,7) when (Pos bsr 3) == 0 -> 7; dir(Pos,8) when (Pos bsr 3) == 0 -> 8; dir(Pos,9) when (Pos bsr 3) == 0 -> 9; dir(Pos,_) when (Pos bsr 3) == 0 -> false; dir(Pos,-9) when (Pos bsr 3) == 7 -> -9; % {_,8} dir(Pos,-8) when (Pos bsr 3) == 7 -> -8; dir(Pos,-7) when (Pos bsr 3) == 7 -> -7; dir(Pos,-1) when (Pos bsr 3) == 7 -> -1; dir(Pos,1) when (Pos bsr 3) == 7 -> 1; dir(Pos,_) when (Pos bsr 3) == 7 -> false; dir(Pos,-8) when (Pos band 7) == 0 -> -8; % {1,_} dir(Pos,-7) when (Pos band 7) == 0 -> -7; dir(Pos,1) when (Pos band 7) == 0 -> 1; dir(Pos,8) when (Pos band 7) == 0 -> 8; dir(Pos,9) when (Pos band 7) == 0 -> 9; dir(Pos,_) when (Pos band 7) == 0 -> false; dir(Pos,-9) when (Pos band 7) == 7 -> -9; % {8,_} dir(Pos,-8) when (Pos band 7) == 7 -> -8; dir(Pos,-1) when (Pos band 7) == 7 -> -1; dir(Pos,7) when (Pos band 7) == 7 -> 7; dir(Pos,8) when (Pos band 7) == 7 -> 8; dir(Pos,_) when (Pos band 7) == 7 -> false; dir(_Pos,Dir) -> Dir.
lib/gs/contribs/othello/othello_adt.erl
0.521959
0.487063
othello_adt.erl
starcoder
-module(erlang_tc_poly). -export([ %% polynomial API from_coeffs/1, eval/2, eval_from_fr/2, cmp/2, interpolate/1, interpolate_from_fr/1, gen_monomial/1, degree/1, random/1, zero/0, zeroize/1, is_zero/1, constant/1, mul/2, add/2, sub/2, mul_scalar/2, add_scalar/2, sub_scalar/2, reveal/1, commitment/1 ]). -type coeffs() :: [integer()]. -type samples() :: [{integer(), integer()}, ...]. -type fr_samples() :: [{reference(), reference()}, ...]. -type poly() :: reference(). -export_type([poly/0]). -spec from_coeffs(Coeffs :: coeffs()) -> poly(). from_coeffs(Coeffs) -> erlang_tc:poly_from_coeffs(Coeffs). -spec eval(Poly :: poly(), Point :: integer()) -> erlang_tc_fr:fr(). eval(Poly, Point) -> erlang_tc:eval_uni_poly(Poly, Point). -spec eval_from_fr(Poly :: poly(), Point :: erlang_tc_fr:fr()) -> erlang_tc_fr:fr(). eval_from_fr(Poly, Point) -> erlang_tc:eval_uni_poly_from_fr(Poly, Point). -spec cmp(P1 :: poly(), P2 :: poly()) -> boolean(). cmp(P1, P2) -> erlang_tc:cmp_poly(P1, P2). %% NOTE: only works if the number of samples is `degree + 1` minimum -spec interpolate(Samples :: samples()) -> poly(). interpolate(Samples) -> erlang_tc:interpolate_uni_poly(Samples). %% NOTE: only works if the number of samples is `degree + 1` minimum -spec interpolate_from_fr(Samples :: fr_samples()) -> poly(). interpolate_from_fr(Samples) -> erlang_tc:interpolate_uni_poly_from_fr(Samples). -spec degree(Poly :: poly()) -> non_neg_integer(). degree(Poly) -> erlang_tc:degree_poly(Poly). -spec gen_monomial(Degree :: non_neg_integer()) -> poly(). gen_monomial(Degree) -> erlang_tc:gen_monomial(Degree). -spec random(Degree :: non_neg_integer()) -> poly(). random(Degree) -> erlang_tc:random_poly(Degree). -spec constant(C :: number()) -> poly(). constant(C) -> erlang_tc:constant_poly(C). -spec zero() -> poly(). zero() -> erlang_tc:zero_poly(). -spec zeroize(P :: poly()) -> poly(). zeroize(P) -> erlang_tc:zeroize_poly(P). -spec is_zero(P :: poly()) -> boolean(). is_zero(P) -> erlang_tc:is_zero_poly(P). -spec add(P1 :: poly(), P2 :: poly()) -> poly(). add(P1, P2) -> erlang_tc:add_poly(P1, P2). -spec sub(P1 :: poly(), P2 :: poly()) -> poly(). sub(P1, P2) -> erlang_tc:sub_poly(P1, P2). -spec mul(P1 :: poly(), P2 :: poly()) -> poly(). mul(P1, P2) -> erlang_tc:mul_poly(P1, P2). -spec add_scalar(Scalar :: number(), P :: poly()) -> poly(). add_scalar(Scalar, P) -> case Scalar < 0 of false -> erlang_tc:add_scalar_poly(Scalar, P); true -> sub_scalar(Scalar, P) end. -spec sub_scalar(Scalar :: number(), P :: poly()) -> poly(). sub_scalar(Scalar, P) -> erlang_tc:sub_scalar_poly(Scalar, P). -spec mul_scalar(Scalar :: number(), P :: poly()) -> poly(). mul_scalar(Scalar, P) -> erlang_tc:mul_scalar_poly(Scalar, P). -spec reveal(P :: poly()) -> string(). reveal(P) -> erlang_tc:reveal_poly(P). -spec commitment(P :: poly()) -> erlang_tc_commitment:commitment(). commitment(P) -> erlang_tc:commitment_poly(P).
src/erlang_tc_poly.erl
0.709221
0.554591
erlang_tc_poly.erl
starcoder
%%============================================================================== %% Copyright 2015 Erlang Solutions Ltd. %% %% Licensed under the Apache License, Version 2.0 (the "License"); %% you may not use this file except in compliance with the License. %% You may obtain a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, software %% distributed under the License is distributed on an "AS IS" BASIS, %% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. %% See the License for the specific language governing permissions and %% limitations under the License. %%============================================================================== -module(escalus_mongooseim). -behaviour(escalus_server). %% API -export([pre_story/1, post_story/1, name/0]). %% for testing -export([check_metric_change/2]). -spec pre_story(escalus:config()) -> escalus:config(). pre_story(Config) -> maybe_read_initial_metric_values(Config). maybe_read_initial_metric_values(Config) -> Metrics = proplists:get_value(mongoose_metrics, Config, undefined), case Metrics of undefined -> Config; _ -> InitialMetrics = lists:foldl(fun read_metric_initial_value/2, [], Metrics), [{mongoose_metrics_initial, InitialMetrics} | Config] end. -spec post_story(escalus:config()) -> escalus:config(). post_story(Config) -> maybe_check_metrics_post_story(Config). maybe_check_metrics_post_story(Config) -> case proplists:get_value(mongoose_metrics_initial, Config) of undefined -> ok; InitialMetrics -> post_story_check_metrics(InitialMetrics) end, Config. -spec name() -> mongooseim. name() -> mongooseim. read_metric_initial_value({Metric, _} = MetricSpec, Acc) -> Type = metric_type(Metric), maybe_reset_metric(Metric, Type), Value = get_value(Metric, Type), [{MetricSpec, Type, Value} | Acc]; read_metric_initial_value({Precond, Metric, Change}, Acc) -> case Precond() of true -> read_metric_initial_value({Metric, Change}, Acc); _ -> Acc end. post_story_check_metrics(CountersToCheck) -> After = [{MetricSpec, OldValue, get_value(Metric, Type)} || {{Metric, _} = MetricSpec, Type, OldValue } <- CountersToCheck], [] = lists:foldl(fun check_metric_change/2, [], After), ok. maybe_reset_metric(Metric, histogram) -> escalus_ejabberd:rpc(exometer, reset, [Metric]); maybe_reset_metric(_, _) -> ok. get_value(Metric, spiral) -> Values = get_values(Metric), lists:foldl(fun({_, [{count, X}, _]}, Sum) -> Sum + X end, 0, Values); get_value(Metric, histogram) -> get_value(Metric); get_value(Metric, _) -> get_value(Metric). get_value(Metric) -> {ok, Value} = escalus_ejabberd:rpc(mongoose_metrics, get_metric_value, [Metric]), Value. get_values(Metric) -> escalus_ejabberd:rpc(mongoose_metrics, get_metric_values, [Metric]). metric_type(Metric) -> [{_, Type, _} | _] = escalus_ejabberd:rpc(exometer, find_entries, [Metric]), Type. check_metric_change({{Metric, {MinChange, MaxChange}}, Before, After}, Acc) -> Change = After - Before, case {Change < MinChange, Change > MaxChange} of {true, _} -> [{Metric, {minimum_change, MinChange}, {before_story, Before}, {after_story, After}} | Acc]; {_, true} -> [{Metric, {maximum_change, MinChange}, {before_story, Before}, {after_story, After}} | Acc]; {_, _} -> Acc end; check_metric_change({{Metric, Change}, Before, After}, Acc) when is_integer(Change) -> case Before + Change =:= After of true -> Acc; _ -> [{Metric, {expected_diff, Change}, {before_story, Before}, {after_story, After}} | Acc] end; check_metric_change({{Metric, changed}, Before, After}, Acc) -> case After of Before -> [{Metric, expected_change, {before_story, Before}, {after_story, After}} | Acc]; _ -> Acc end; check_metric_change({{Metric, Changes}, Before, After}, Acc) when is_list(Changes)-> Check = fun({Property, Change}) -> {Property, BeforeProp} = lists:keyfind(Property, 1, Before), {Property, AfterProp} = lists:keyfind(Property, 1, After), check_change(BeforeProp, AfterProp, Change) end, case lists:all(Check, Changes) of false -> [{Metric, {expected_change, Changes}, {before_story, Before}, {after_story, After}} | Acc]; _ -> Acc end. check_change(Before, After, Change) when is_atom(Change) -> erlang:apply(erlang, Change, [After, Before]); check_change(Before, After, Change) when is_integer(Change) -> After == Before + Change.
src/escalus_mongooseim.erl
0.689619
0.422147
escalus_mongooseim.erl
starcoder
%% @doc Integer counter based on an ordered list of counter events. %% %% A counter is stored as an orddict of counter events. Each counter %% event has a unique key based on the timestamp and some entropy, and it %% stores the delta from the inc operation. The value of a counter is the %% sum of all these deltas. %% %% As an optimization, counter events older than a given age are coalesced %% to a single counter event with a key in the form of %% <code>{timestamp(), 'acc'}</code>. -module(statebox_counter). -export([value/1, merge/1, accumulate/2, inc/3]). -export([f_inc_acc/2, f_inc_acc/3, op_inc_acc/4]). -type op() :: statebox:op(). -type timestamp() :: statebox_clock:timestamp(). -type timedelta() :: statebox:timedelta(). -type counter_id() :: statebox_identity:entropy() | acc. -type counter_key() :: {timestamp(), counter_id()}. -type counter_op() :: {counter_key(), integer()}. -type counter() :: [counter_op()]. %% @doc Return the value of the counter (the sum of all counter event deltas). -spec value(counter()) -> integer(). value([]) -> 0; value([{_Key, Value} | Rest]) -> Value + value(Rest). %% @doc Merge the given list of counters and return a new counter %% with the union of that history. -spec merge([counter()]) -> counter(). merge([Counter]) -> Counter; merge(Counters) -> orddict:from_list(merge_prune(Counters)). %% @doc Accumulate all counter events older than <code>Timestamp</code> to %% the key <code>{Timestamp, acc}</code>. If there is already an %% <code>acc</code> at or before <code>Timestamp</code> this is a no-op. -spec accumulate(timestamp(), counter()) -> counter(). accumulate(Timestamp, Counter=[{{T0, acc}, _} | _]) when Timestamp =< T0 -> Counter; accumulate(Timestamp, Counter) -> accumulate(Timestamp, Counter, 0). %% @doc Return a new counter with the given counter event. If there is %% an <code>acc</code> at or before the timestamp of the given key then %% this is a no-op. -spec inc(counter_key(), integer(), counter()) -> counter(). inc({T1, _Id1}, _Value, Counter=[{{T0, acc}, _} | _Rest]) when T1 =< T0 -> Counter; inc(Key, Value, Counter) -> orddict:store(Key, Value, Counter). %% @equiv f_inc_acc(Value, Age, {statebox_clock:timestamp(), %% statebox_identity:entropy()}) -spec f_inc_acc(integer(), timedelta()) -> op(). f_inc_acc(Value, Age) -> Key = {statebox_clock:timestamp(), statebox_identity:entropy()}, f_inc_acc(Value, Age, Key). %% @doc Return a statebox event to increment and accumulate the counter. %% <code>Value</code> is the delta, %% <code>Age</code> is the maximum age of counter events in milliseconds %% (this should be longer than the amount of time you expect your cluster to %% reach a consistent state), %% <code>Key</code> is the counter event key. -spec f_inc_acc(integer(), timedelta(), counter_key()) -> op(). f_inc_acc(Value, Age, Key={Timestamp, _Id}) -> {fun ?MODULE:op_inc_acc/4, [Timestamp - Age, Key, Value]}. %% @private op_inc_acc(Timestamp, Key, Value, Counter) -> accumulate(Timestamp, inc(Key, Value, Counter)). %% Internal API merge_prune(Counters) -> %% Merge of all of the counters and prune all entries older than the %% newest {_, acc}. prune(lists:umerge(Counters)). prune(All) -> prune(All, All). prune(Here=[{{_Ts, acc}, _V} | Rest], _Last) -> prune(Rest, Here); prune([_ | Rest], Last) -> prune(Rest, Last); prune([], Last) -> Last. accumulate(Timestamp, [{{T1, _Id}, Value} | Rest], Sum) when T1 =< Timestamp -> %% Roll up old counter events accumulate(Timestamp, Rest, Value + Sum); accumulate(Timestamp, Counter, Sum) -> %% Return the new counter inc({Timestamp, acc}, Sum, Counter).
src/statebox_counter.erl
0.657978
0.878523
statebox_counter.erl
starcoder
-module(ask_area). -export([area/0]). -spec(area() -> number()). area() -> Answer = io:get_line("R)ectangle, T)riangle, or E)llipse > "), Shape = char_to_shape(hd(Answer)), case Shape of rectangle -> Numbers = get_dimensions("width", "height"); triangle -> Numbers = get_dimensions("base", "height"); ellipse -> Numbers = get_dimensions("major axis", "minor axis"); unknown -> Numbers = {error, "Unknown shape " ++ [hd(Answer)]} end, Area = calculate(Shape, element(1, Numbers), element(2, Numbers)), Area. %% @doc Given a character, returns an atom representing the %% specified shape (or the atom unknown if a bad character is given). -spec(char_to_shape(char()) -> atom()). char_to_shape(Char) -> case Char of $R -> rectangle; $r -> rectangle; $T -> triangle; $t -> triangle; $E -> ellipse; $e -> ellipse; _ -> unknown end. %% @doc Present a prompt and get a number from the %% user. Allow either integers or floats. -spec(get_number(string()) -> number()). get_number(Prompt) -> Str = io:get_line("Enter " ++ Prompt ++ " > "), {Test, _} = string:to_float(Str), case Test of error -> {N, _} = string:to_integer(Str); _ -> N = Test end, N. %% @doc Get dimensions for a shape. Input are the two prompts, %% output is a tuple {Dimension1, Dimension2}. -spec(get_dimensions(string(), string()) -> {number(), number()}). get_dimensions(Prompt1, Prompt2) -> N1 = get_number(Prompt1), N2 = get_number(Prompt2), {N1, N2}. %% @doc Calculate area of a shape, given its shape and dimensions. %% Handle errors appropriately. -spec(calculate(atom(), number(), number()) -> number()). calculate(unknown, _, Err) -> io:format("~s~n", [Err]); calculate(_, error, _) -> io:format("Error in first number.~n"); calculate(_, _, error) -> io:format("Error in second number.~n"); calculate(_, A, B) when A < 0; B < 0 -> io:format("Both numbers must be greater than or equal to zero~n"); calculate(Shape, A, B) -> geom:area(Shape, A, B).
src/ask_area.erl
0.583915
0.734976
ask_area.erl
starcoder
-module('99-problems'). -compile(export_all). % P01: Find the last element of a list. last(L) -> lists:last(L). % P02: Find the last but one element of a list. lastButOne(L) -> hd(lists:nthtail(length(L) - 2, L)). % P03: Find the k-th element of a list. kthElement(K, L) -> lists:nth(K, L). % P04: Find the number of elements of a list. listLength(L) -> length(L). % P05: Reverse a list. reverse(L) -> lists:reverse(L). % P06: Find out whether a list is a palindrome. isPalindrome(L) -> lists:reverse(L) == L. % P07: Flatten a nested list structure. flatten(L) -> lists:flatten(L). % P08: Eliminate consecutive duplicates of list elements. compress(H, [ H | T ]) -> [ H | T ]; compress(H, L) -> [ H | L ]. compress(L) -> lists:foldr(fun compress/2, [], L). % P09: Pack consecutive duplicates of list elements into sublists. pack(H, [ [ H | T ] | TT ]) -> [ [ H, H | T ] | TT ]; pack(H, L) -> [ [ H ] | L ]. pack(L) -> lists:foldr(fun pack/2, [], L). % P10: Run-length encoding of a list. encode(H, []) -> [ [ 1, H ] ]; encode(H, [ [ N, H ] | T ]) -> [ [ N + 1, H ] | T ]; encode(H, [ [ _, _ ] | _ ] = L) -> [ [ 1, H ] | L ]. encode(L) -> lists:foldr(fun encode/2, [], L). % P11: Modified run-length encoding. encodeWithoutDuplicates(H, []) -> [ H ]; encodeWithoutDuplicates(H, [ H | T ]) -> [ [ 2, H ] | T ]; encodeWithoutDuplicates(H, [ [ N, H ] | T ]) -> [ [ N + 1, H ] | T ]; encodeWithoutDuplicates(H, [ [ _, _ ] | _ ] = L) -> [ H | L ]; encodeWithoutDuplicates(H, [ _ | _ ] = L) -> [ H | L ]. encodeWithoutDuplicates(L) -> lists:foldr(fun encodeWithoutDuplicates/2, [], L). % P12: Decode a run-length encoded list. decode([ N, C ], L) -> lists:duplicate(N, C) ++ L; decode(C, L) -> [ C | L ]. decode(L) -> lists:foldr(fun decode/2, [], L). % P13: Run-length encoding of a list (direct solution). encodeDirect(L) -> encodeWithoutDuplicates(L). % P14: Duplicate the elements of a list. duplicate(H, []) -> [ H, H ]; duplicate(H, L) -> [ H, H ] ++ L. duplicate(L) -> lists:foldr(fun duplicate/2, [], L). % P15: Duplicate the elements of a list a given number of times. multiplicate(N, H, []) -> lists:duplicate(N, H); multiplicate(N, H, L) -> lists:duplicate(N, H) ++ L. multiplicate(L, N) -> lists:foldr(fun(H, T) -> multiplicate(N, H, T) end, [], L). % P16: Drop every n-th element from a list. drop([], _) -> []; drop(L, N) -> drop(L, N, 1, []). drop([], _, _, Result) -> lists:reverse(Result); drop([ _ | T ], N, Count, Result) when Count rem N == 0 -> drop(T, N, Count + 1, Result); drop([ H | T ], N, Count, Result) -> drop(T, N, Count + 1, [ H | Result ]). % P17: Split a list into two parts - the length of the first part is given. split([], _) -> []; split(L, 0) -> [L,[]]; split(L, N) when length(L) < N -> []; split([ H | T ], N) -> split(T, N, 1, [ H ]). split(L, N, Count, Result) when Count >= N -> [ lists:reverse(Result), L ]; split([ H | T ], N, Count, Result) -> split(T, N, Count + 1, [ H | Result ]). % P18: Extract a slice from a list, by given indexes (result should contain values from both indexes). slice(L, I, K) when I =< K, I >= 1, K >= 1 -> slice(L, 1, 1, I, K, []). slice(_, _, _, I, K, Result) when length(Result) >= K - I + 1 -> lists:reverse(Result); slice([ _ | T ], II, KK, I, K, Result) when II < I -> slice(T, II + 1, KK + 1, I, K, Result); slice([ H | T ], II, KK, I, K, Result) when II >= I, KK =< K -> slice(T, II + 1, KK + 1, I, K, [ H | Result ]). % P19: Rotate a list N places to the right (negative number means rotation to the left). rotate(L, N) when N =:= 0 -> L; rotate(L, N) when N < 0 -> rotateLeft(L, -N, []); rotate(L, N) when N > 0 -> rotateRight(lists:reverse(L), N, []). rotateLeft([ H | T ], N, Result) when N > 0 -> rotateLeft(T, N - 1, Result ++ [ H ]); rotateLeft(L, N, Result) when N =:= 0 -> L ++ Result. rotateRight([ H | T ], N, Result) when N > 0 -> rotateRight(T, N - 1, [ H | Result ]); rotateRight(L, N, Result) when N =:= 0 -> Result ++ lists:reverse(L). % P20: Remove the k-th element from a list (index of first element is 1). remove(L, K) -> remove(L, 1, K, []). remove([ _ | T ], I, K, Result) when I =:= K -> Result ++ T; remove([ H | T ], I, K, Result) when I < K -> remove(T, I + 1, K, Result ++ [ H ]). % P21: Insert an element at a given position into a list. insert(L, K, E) -> insert(L, 1, K, E, []). insert([], I, K, E, Result) when I =:= K -> Result ++ [ E ]; insert([ H | T ], I, K, E, Result) when I =:= K -> Result ++ [ E | [ H | T ] ]; insert([ H | T ], I, K, E, Result) when I < K, I > 0 -> insert(T, I + 1, K, E, Result ++ [ H ]). % P22: Create a list containing all integers within a given range (and will contain both limits). sequence(Start, End) when Start =< End -> sequence(Start, End, []). sequence(I, E, Result) when I =< E -> sequence(I + 1, E, Result ++ [ I ]); sequence(I, E, Result) when I > E -> Result. % P23: Extract a given number of randomly selected elements from a list (without using the same value many times). selectRandom(_, N) when N =:= 0 -> []; selectRandom(L, N) when N > 0, N =< length(L) -> selectRandom(L, N, []). selectRandom(_, N, Result) when N =:= 0 -> Result; selectRandom(L, N, Result) when N > 0 -> RandomizedIndex = random:uniform(length(L)), Element = lists:nth(RandomizedIndex, L), selectRandom(remove(L, RandomizedIndex), N - 1, Result ++ [ Element ]). % P24: Lotto: Draw 'N' different random numbers from the set '1 .. M'. lotto(N, M) when N =< M -> selectRandom(sequence(1, M), N). % P25: Generate a random permutation of the elements of a list. permutation(L) -> selectRandom(L, length(L)). % P26: Generate the combinations of 'K' distinct objects chosen from the 'N' elements of a list. combinations(0, _) -> []; combinations(1, L) -> [ [ X ] || X <- L ]; combinations(N, L) -> [ [ H | Res ] || H <- L, Res <- combinations(N - 1, L), false == lists:member(H, Res) ]. % P27: Group the elements of a set into disjoint subsets. % % a) In how many ways can a group of 9 people work in 3 disjoint subgroups % of 2, 3 and 4 persons? Write a predicate that generates all the % possibilities via backtracking. combinationWithoutRepetition(0, Xs) -> [ { [], Xs } ]; combinationWithoutRepetition(_, []) -> []; combinationWithoutRepetition(N, [ X | Xs ]) -> Ts = [ { [ X | Ys ], Zs } || { Ys, Zs } <- combinationWithoutRepetition(N - 1, Xs) ], Ds = [ { Ys, [ X | Zs ] } || { Ys, Zs } <- combinationWithoutRepetition(N, Xs) ], Ts ++ Ds. group3(L) -> [ [ Twos, Threes, Fours ] || { Twos, L2 } <- combinationWithoutRepetition(2, L), { Threes, L3 } <- combinationWithoutRepetition(3, L2), { Fours, _ } <- combinationWithoutRepetition(4, L3) ]. % b) Generalize the above predicate in a way that we can specify a % list of group sizes and the predicate will return a list of groups. group([], _) -> [[]]; group([ N | Ns ], Xs) -> [ [ G | Gs ] || { G, Rs } <- combinationWithoutRepetition(N, Xs), Gs <- group(Ns, Rs) ]. % P28: Sorting a list of lists according to length of sublists. % % a) We suppose that a list (InList) contains elements that are lists themselves. % The objective is to sort the elements of InList according to their length. % E.g. short lists first, longer lists later, or vice versa. lsort([]) -> []; lsort([ H | T ]) -> lsort(H, T, [], []). lsort(P, [], Left, Right) -> lsort(Left) ++ [ P ] ++ lsort(Right); lsort(P, [ H | T ], Left, Right) when length(H) >= length(P) -> lsort(P, T, Left, [ H | Right ]); lsort(P, [ H | T ], Left, Right) -> lsort(P, T, [ H | Left ], Right). % b) Again, we suppose that a list (InList) contains elements that are lists % themselves. But this time the objective is to sort the elements of InList % according to their length frequency; i.e. in the default, where sorting % is done ascendingly, lists with rare lengths are placed first, others % with a more frequent length come later. groupBy(F, L) -> lists:foldr(fun({ K, V }, D) -> dict:append(K, V, D) end, dict:new(), [ { F(X), X } || X <- L ]). lfsort(L) -> Sorted = lsort(L), GrouppedEqualLengthDict = groupBy(fun (A) -> length(A) end, Sorted), GrouppedEqualLength = dict:fold(fun (_K, V, Acc) -> [ V | Acc ] end, [], GrouppedEqualLengthDict), lists:append(lsort(GrouppedEqualLength)).
assignments/solutions/99-problems/erlang/99-problems.erl
0.673943
0.600159
99-problems.erl
starcoder
%% ------------------------------------------------------------------- %% %% Copyright (c) 2012 Basho Technologies, Inc. %% %% This file is provided to you under the Apache License, %% Version 2.0 (the "License"); you may not use this file %% except in compliance with the License. You may obtain %% a copy of the License at %% %% http://www.apache.org/licenses/LICENSE-2.0 %% %% Unless required by applicable law or agreed to in writing, %% software distributed under the License is distributed on an %% "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY %% KIND, either express or implied. See the License for the %% specific language governing permissions and limitations %% under the License. %% %% ------------------------------------------------------------------- -module(verify_build_cluster). -behavior(riak_test). -export([confirm/0]). -include_lib("eunit/include/eunit.hrl"). confirm() -> %% Deploy a set of new nodes lager:info("Deploying 4 nodes"), %% handoff_concurrency needs to be raised to make the leave operation faster. %% most clusters go up to 10, but this one is one louder, isn't it? [Node1, Node2, Node3, Node4] = Nodes = rt:deploy_nodes(4, [{riak_core, [{handoff_concurrency, 11}]}]), %% Ensure each node owns 100% of it's own ring lager:info("Ensure each nodes 100% of it's own ring"), [rt:wait_until_owners_according_to(Node, [Node]) || Node <- Nodes], lager:info("joining Node 2 to the cluster... It takes two to make a thing go right"), rt:join(Node2, Node1), wait_and_validate([Node1, Node2]), lager:info("joining Node 3 to the cluster"), rt:join(Node3, Node1), wait_and_validate([Node1, Node2, Node3]), lager:info("joining Node 4 to the cluster"), rt:join(Node4, Node1), wait_and_validate(Nodes), lager:info("taking Node 1 down"), rt:stop(Node1), ?assertEqual(ok, rt:wait_until_unpingable(Node1)), wait_and_validate(Nodes, [Node2, Node3, Node4]), lager:info("taking Node 2 down"), rt:stop(Node2), ?assertEqual(ok, rt:wait_until_unpingable(Node2)), wait_and_validate(Nodes, [Node3, Node4]), lager:info("bringing Node 1 up"), rt:start(Node1), ok = rt:wait_until_pingable(Node1), wait_and_validate(Nodes, [Node1, Node3, Node4]), lager:info("bringing Node 2 up"), rt:start(Node2), ok = rt:wait_until_pingable(Node2), wait_and_validate(Nodes), % leave 1, 2, and 3 lager:info("leaving Node 1"), rt:leave(Node1), ?assertEqual(ok, rt:wait_until_unpingable(Node1)), wait_and_validate([Node2, Node3, Node4]), lager:info("leaving Node 2"), rt:leave(Node2), ?assertEqual(ok, rt:wait_until_unpingable(Node2)), wait_and_validate([Node3, Node4]), lager:info("leaving Node 3"), rt:leave(Node3), ?assertEqual(ok, rt:wait_until_unpingable(Node3)), % verify 4 wait_and_validate([Node4]), pass. wait_and_validate(Nodes) -> wait_and_validate(Nodes, Nodes). wait_and_validate(RingNodes, UpNodes) -> lager:info("Wait until all nodes are ready and there are no pending changes"), ?assertEqual(ok, rt:wait_until_nodes_ready(UpNodes)), ?assertEqual(ok, rt:wait_until_all_members(UpNodes)), ?assertEqual(ok, rt:wait_until_no_pending_changes(UpNodes)), lager:info("Ensure each node owns a portion of the ring"), [rt:wait_until_owners_according_to(Node, RingNodes) || Node <- UpNodes], [rt:wait_for_service(Node, freya_stats) || Node <- UpNodes], done.
dist-tests/verify_build_cluster.erl
0.61173
0.566498
verify_build_cluster.erl
starcoder