Add test/hk_search_job_tests.erl

test/hk_search_job_tests.erl

@@ -0,0 +1,68 @@
+%%%-------------------------------------------------------------------
+%%% @doc Covers "search orchestration" from `<test>': runs the full
+%%% `<search_pipeline>' end-to-end against the bundled mock provider
+%%% (no network required) and asserts every stage's event fires, in
+%%% pipeline order, and the job reaches `done' with synthesized
+%%% results.
+%%% @end
+%%%-------------------------------------------------------------------
+-module(hk_search_job_tests).
+-include_lib("eunit/include/eunit.hrl").
+
+setup() -> hk_test_helper:start_app().
+teardown(Ok) -> hk_test_helper:stop_app(Ok).
+
+search_fixture_test_() ->
+    {timeout, 20, {setup, fun setup/0, fun teardown/1, fun(_) -> [
+        {"a search job runs every pipeline stage in order and completes",
+         fun full_pipeline_runs_in_order/0}
+    ] end}}.
+
+full_pipeline_runs_in_order() ->
+    ok = hk_event_bus:subscribe(#{category_prefix => <<"search.">>}),
+    {ok, JobId} = hk_search:start(undefined, <<"erlang otp supervision">>, #{max_iterations => 2}),
+    Categories = collect_until(<<"search.completed">>, JobId, [], 10000),
+    ok = hk_event_bus:unsubscribe(),
+
+    ?assertEqual(<<"search.started">>, hd(Categories)),
+    ?assertEqual(<<"search.completed">>, lists:last(Categories)),
+    %% The pipeline stage events must appear in the same relative order
+    %% every time (deterministic orchestration) -- assert each of these
+    %% appears, and that it appears before the next one in the list.
+    ExpectedOrder = [<<"search.started">>, <<"search.query_generated">>,
+                      <<"search.provider_called">>, <<"search.results_received">>,
+                      <<"search.urls_deduplicated">>, <<"search.document_retrieved">>,
+                      <<"search.result_synthesized">>, <<"search.completed">>],
+    assert_relative_order(ExpectedOrder, Categories),
+
+    {ok, Pid} = hk_search_registry:lookup(JobId),
+    {ok, Result} = hk_search_job:await_result(Pid, 1000),
+    ?assert(maps:is_key(results, Result)),
+    ?assert(maps:is_key(sources, Result)).
+
+%%% internal
+
+collect_until(TargetCategory, JobId, Acc, TimeoutMs) ->
+    receive
+        {hk_event, #{category := TargetCategory, data := #{job_id := JobId}}} ->
+            lists:reverse([TargetCategory | Acc]);
+        {hk_event, #{category := Cat, data := #{job_id := JobId}}} ->
+            collect_until(TargetCategory, JobId, [Cat | Acc], TimeoutMs);
+        {hk_event, _Other} ->
+            collect_until(TargetCategory, JobId, Acc, TimeoutMs)
+    after TimeoutMs ->
+        lists:reverse(Acc)
+    end.
+
+assert_relative_order([_Last], _Categories) ->
+    ok;
+assert_relative_order([A, B | Rest], Categories) ->
+    IdxA = index_of(A, Categories),
+    IdxB = index_of(B, Categories),
+    ?assert(IdxA =/= not_found andalso IdxB =/= not_found andalso IdxA < IdxB),
+    assert_relative_order([B | Rest], Categories).
+
+index_of(X, List) -> index_of(X, List, 1).
+index_of(_X, [], _N) -> not_found;
+index_of(X, [X | _], N) -> N;
+index_of(X, [_ | T], N) -> index_of(X, T, N + 1).