Search is not available for this dataset
text
stringlengths
75
104k
def CBZ(cpu, op, dest): """ Compare and Branch on Zero compares the value in a register with zero, and conditionally branches forward a constant value. It does not affect the condition flags. :param ARMv7Operand op: Specifies the register that contains the first operand. :param ARMv7Operand dest: Specifies the label of the instruction that is to be branched to. The assembler calculates the required value of the offset from the PC value of the CBZ instruction to this label, then selects an encoding that will set imm32 to that offset. Allowed offsets are even numbers in the range 0 to 126. """ cpu.PC = Operators.ITEBV(cpu.address_bit_size, op.read(), cpu.PC, dest.read())
def TBH(cpu, dest): """ Table Branch Halfword causes a PC-relative forward branch using a table of single halfword offsets. A base register provides a pointer to the table, and a second register supplies an index into the table. The branch length is twice the value of the halfword returned from the table. :param ARMv7Operand dest: see below; register """ # Capstone merges the two registers values into one operand, so we need to extract them back # Specifies the base register. This contains the address of the table of branch lengths. This # register is allowed to be the PC. If it is, the table immediately follows this instruction. base_addr = dest.get_mem_base_addr() if dest.mem.base in ('PC', 'R15'): base_addr = cpu.PC # Specifies the index register. This contains an integer pointing to a halfword within the table. # The offset within the table is twice the value of the index. offset = cpu.read_int(base_addr + dest.get_mem_offset(), 16) offset = Operators.ZEXTEND(offset, cpu.address_bit_size) cpu.PC += (offset << 1)
def _LDM(cpu, insn_id, base, regs): """ LDM (Load Multiple) loads a non-empty subset, or possibly all, of the general-purpose registers from sequential memory locations. It is useful for block loads, stack operations and procedure exit sequences. :param int insn_id: should be one of ARM_INS_LDM, ARM_INS_LDMIB, ARM_INS_LDMDA, ARM_INS_LDMDB :param Armv7Operand base: Specifies the base register. :param list[Armv7Operand] regs: Is a list of registers. It specifies the set of registers to be loaded by the LDM instruction. The registers are loaded in sequence, the lowest-numbered register from the lowest memory address (start_address), through to the highest-numbered register from the highest memory address (end_address). If the PC is specified in the register list (opcode bit[15] is set), the instruction causes a branch to the address (data) loaded into the PC. It's technically UNKNOWN if you writeback to a register you loaded into, but we let it slide. """ if cpu.instruction.usermode: raise NotImplementedError("Use of the S bit is not supported") increment = insn_id in (cs.arm.ARM_INS_LDM, cs.arm.ARM_INS_LDMIB) after = insn_id in (cs.arm.ARM_INS_LDM, cs.arm.ARM_INS_LDMDA) address = base.read() for reg in regs: if not after: address += (1 if increment else -1) * (reg.size // 8) reg.write(cpu.read_int(address, reg.size)) if reg.reg in ('PC', 'R15'): # The general-purpose registers loaded can include the PC. If they do, the word loaded for the PC is # treated as an address and a branch occurs to that address. In ARMv5 and above, bit[0] of the loaded # value determines whether execution continues after this branch in ARM state or in Thumb state, as # though a BX instruction had been executed. cpu._set_mode_by_val(cpu.PC) cpu.PC = cpu.PC & ~1 if after: address += (1 if increment else -1) * (reg.size // 8) if cpu.instruction.writeback: base.writeback(address)
def _STM(cpu, insn_id, base, regs): """ STM (Store Multiple) stores a non-empty subset (or possibly all) of the general-purpose registers to sequential memory locations. :param int insn_id: should be one of ARM_INS_STM, ARM_INS_STMIB, ARM_INS_STMDA, ARM_INS_STMDB :param Armv7Operand base: Specifies the base register. :param list[Armv7Operand] regs: Is a list of registers. It specifies the set of registers to be stored by the STM instruction. The registers are stored in sequence, the lowest-numbered register to the lowest memory address (start_address), through to the highest-numbered register to the highest memory address (end_address). """ if cpu.instruction.usermode: raise NotImplementedError("Use of the S bit is not supported") increment = insn_id in (cs.arm.ARM_INS_STM, cs.arm.ARM_INS_STMIB) after = insn_id in (cs.arm.ARM_INS_STM, cs.arm.ARM_INS_STMDA) address = base.read() for reg in regs: if not after: address += (1 if increment else -1) * (reg.size // 8) cpu.write_int(address, reg.read(), reg.size) if after: address += (1 if increment else -1) * (reg.size // 8) if cpu.instruction.writeback: base.writeback(address)
def _SR(cpu, insn_id, dest, op, *rest): """ Notes on Capstone behavior: - In ARM mode, _SR reg has `rest`, but _SR imm does not, its baked into `op`. - In ARM mode, `lsr r1, r2` will have a `rest[0]` - In Thumb mode, `lsr r1, r2` will have an empty `rest` - In ARM mode, something like `lsr r1, 3` will not have `rest` and op will be the immediate. """ assert insn_id in (cs.arm.ARM_INS_ASR, cs.arm.ARM_INS_LSL, cs.arm.ARM_INS_LSR) if insn_id == cs.arm.ARM_INS_ASR: if rest and rest[0].type == 'immediate': srtype = cs.arm.ARM_SFT_ASR else: srtype = cs.arm.ARM_SFT_ASR_REG elif insn_id == cs.arm.ARM_INS_LSL: if rest and rest[0].type == 'immediate': srtype = cs.arm.ARM_SFT_LSL else: srtype = cs.arm.ARM_SFT_LSL_REG elif insn_id == cs.arm.ARM_INS_LSR: if rest and rest[0].type == 'immediate': srtype = cs.arm.ARM_SFT_LSR else: srtype = cs.arm.ARM_SFT_LSR_REG carry = cpu.regfile.read('APSR_C') if rest and rest[0].type == 'register': # FIXME we should make Operand.op private (and not accessible) result, carry = cpu._shift(op.read(), srtype, rest[0].op.reg, carry) elif rest and rest[0].type == 'immediate': amount = rest[0].read() result, carry = cpu._shift(op.read(), srtype, amount, carry) elif cpu.mode == cs.CS_MODE_THUMB: result, carry = cpu._shift(dest.read(), srtype, op, carry) else: result, carry = op.read(with_carry=True) dest.write(result) cpu.set_flags(N=HighBit(result), Z=(result == 0), C=carry)
def _fix_index(self, index): """ :param slice index: """ stop, start = index.stop, index.start if start is None: start = 0 if stop is None: stop = len(self) return start, stop
def _dict_diff(d1, d2): """ Produce a dict that includes all the keys in d2 that represent different values in d1, as well as values that aren't in d1. :param dict d1: First dict :param dict d2: Dict to compare with :rtype: dict """ d = {} for key in set(d1).intersection(set(d2)): if d2[key] != d1[key]: d[key] = d2[key] for key in set(d2).difference(set(d1)): d[key] = d2[key] return d
def locked_context(self, key=None, value_type=list): """ A context manager that provides safe parallel access to the global Manticore context. This should be used to access the global Manticore context when parallel analysis is activated. Code within the `with` block is executed atomically, so access of shared variables should occur within. """ plugin_context_name = str(type(self)) with self.manticore.locked_context(plugin_context_name, dict) as context: assert value_type in (list, dict, set) ctx = context.get(key, value_type()) yield ctx context[key] = ctx
def context(self): """ Convenient access to shared context """ plugin_context_name = str(type(self)) if plugin_context_name not in self.manticore.context: self.manticore.context[plugin_context_name] = {} return self.manticore.context[plugin_context_name]
def add_finding(self, state, address, pc, finding, at_init, constraint=True): """ Logs a finding at specified contract and assembler line. :param state: current state :param address: contract address of the finding :param pc: program counter of the finding :param at_init: true if executing the constructor :param finding: textual description of the finding :param constraint: finding is considered reproducible only when constraint is True """ if issymbolic(pc): pc = simplify(pc) if isinstance(pc, Constant): pc = pc.value if not isinstance(pc, int): raise ValueError("PC must be a number") self.get_findings(state).add((address, pc, finding, at_init, constraint)) with self.locked_global_findings() as gf: gf.add((address, pc, finding, at_init)) #Fixme for ever broken logger logger.warning(finding)
def add_finding_here(self, state, finding, constraint=True): """ Logs a finding in current contract and assembler line. :param state: current state :param finding: textual description of the finding :param constraint: finding is considered reproducible only when constraint is True """ address = state.platform.current_vm.address pc = state.platform.current_vm.pc if isinstance(pc, Constant): pc = pc.value if not isinstance(pc, int): raise ValueError("PC must be a number") at_init = state.platform.current_transaction.sort == 'CREATE' self.add_finding(state, address, pc, finding, at_init, constraint)
def _save_current_location(self, state, finding, condition=True): """ Save current location in the internal locations list and returns a textual id for it. This is used to save locations that could later be promoted to a finding if other conditions hold See _get_location() :param state: current state :param finding: textual description of the finding :param condition: general purpose constraint """ address = state.platform.current_vm.address pc = state.platform.current_vm.pc at_init = state.platform.current_transaction.sort == 'CREATE' location = (address, pc, finding, at_init, condition) hash_id = hashlib.sha1(str(location).encode()).hexdigest() state.context.setdefault('{:s}.locations'.format(self.name), {})[hash_id] = location return hash_id
def _get_location(self, state, hash_id): """ Get previously saved location A location is composed of: address, pc, finding, at_init, condition """ return state.context.setdefault('{:s}.locations'.format(self.name), {})[hash_id]
def _signed_sub_overflow(state, a, b): """ Sign extend the value to 512 bits and check the result can be represented in 256. Following there is a 32 bit excerpt of this condition: a - b -80000000 -3fffffff -00000001 +00000000 +00000001 +3fffffff +7fffffff +80000000 False False False False True True True +c0000001 False False False False False False True +ffffffff False False False False False False False +00000000 True False False False False False False +00000001 True False False False False False False +3fffffff True False False False False False False +7fffffff True True True False False False False """ sub = Operators.SEXTEND(a, 256, 512) - Operators.SEXTEND(b, 256, 512) cond = Operators.OR(sub < -(1 << 255), sub >= (1 << 255)) return cond
def _signed_add_overflow(state, a, b): """ Sign extend the value to 512 bits and check the result can be represented in 256. Following there is a 32 bit excerpt of this condition: a + b -80000000 -3fffffff -00000001 +00000000 +00000001 +3fffffff +7fffffff +80000000 True True True False False False False +c0000001 True False False False False False False +ffffffff True False False False False False False +00000000 False False False False False False False +00000001 False False False False False False True +3fffffff False False False False False False True +7fffffff False False False False True True True """ add = Operators.SEXTEND(a, 256, 512) + Operators.SEXTEND(b, 256, 512) cond = Operators.OR(add < -(1 << 255), add >= (1 << 255)) return cond
def _unsigned_sub_overflow(state, a, b): """ Sign extend the value to 512 bits and check the result can be represented in 256. Following there is a 32 bit excerpt of this condition: a - b ffffffff bfffffff 80000001 00000000 00000001 3ffffffff 7fffffff ffffffff True True True False True True True bfffffff True True True False False True True 80000001 True True True False False True True 00000000 False False False False False True False 00000001 True False False False False True False ffffffff True True True True True True True 7fffffff True True True False False True False """ cond = Operators.UGT(b, a) return cond
def _unsigned_add_overflow(state, a, b): """ Sign extend the value to 512 bits and check the result can be represented in 256. Following there is a 32 bit excerpt of this condition: a + b ffffffff bfffffff 80000001 00000000 00000001 3ffffffff 7fffffff ffffffff True True True False True True True bfffffff True True True False False True True 80000001 True True True False False True True 00000000 False False False False False True False 00000001 True False False False False True False ffffffff True True True True True True True 7fffffff True True True False False True False """ add = Operators.ZEXTEND(a, 512) + Operators.ZEXTEND(b, 512) cond = Operators.UGE(add, 1 << 256) return cond
def _signed_mul_overflow(state, a, b): """ Sign extend the value to 512 bits and check the result can be represented in 256. Following there is a 32 bit excerpt of this condition: a * b +00000000000000000 +00000000000000001 +0000000003fffffff +0000000007fffffff +00000000080000001 +000000000bfffffff +000000000ffffffff +0000000000000000 +0000000000000000 +0000000000000000 +0000000000000000 +0000000000000000 +0000000000000000 +0000000000000000 +0000000000000000 +0000000000000001 +0000000000000000 +0000000000000001 +000000003fffffff +000000007fffffff +0000000080000001 +00000000bfffffff +00000000ffffffff +000000003fffffff +0000000000000000 +000000003fffffff *+0fffffff80000001 *+1fffffff40000001 *+1fffffffbfffffff *+2fffffff00000001 *+3ffffffec0000001 +000000007fffffff +0000000000000000 +000000007fffffff *+1fffffff40000001 *+3fffffff00000001 *+3fffffffffffffff *+5ffffffec0000001 *+7ffffffe80000001 +0000000080000001 +0000000000000000 +0000000080000001 *+1fffffffbfffffff *+3fffffffffffffff *+4000000100000001 *+600000003fffffff *+800000007fffffff +00000000bfffffff +0000000000000000 +00000000bfffffff *+2fffffff00000001 *+5ffffffec0000001 *+600000003fffffff *+8ffffffe80000001 *+bffffffe40000001 +00000000ffffffff +0000000000000000 +00000000ffffffff *+3ffffffec0000001 *+7ffffffe80000001 *+800000007fffffff *+bffffffe40000001 *+fffffffe00000001 """ mul = Operators.SEXTEND(a, 256, 512) * Operators.SEXTEND(b, 256, 512) cond = Operators.OR(mul < -(1 << 255), mul >= (1 << 255)) return cond
def _unsigned_mul_overflow(state, a, b): """ Sign extend the value to 512 bits and check the result can be represented in 256. Following there is a 32 bit excerpt of this condition: a * b +00000000000000000 +00000000000000001 +0000000003fffffff +0000000007fffffff +00000000080000001 +000000000bfffffff +000000000ffffffff +0000000000000000 +0000000000000000 +0000000000000000 +0000000000000000 +0000000000000000 +0000000000000000 +0000000000000000 +0000000000000000 +0000000000000001 +0000000000000000 +0000000000000001 +000000003fffffff +000000007fffffff +0000000080000001 +00000000bfffffff +00000000ffffffff +000000003fffffff +0000000000000000 +000000003fffffff *+0fffffff80000001 *+1fffffff40000001 *+1fffffffbfffffff *+2fffffff00000001 *+3ffffffec0000001 +000000007fffffff +0000000000000000 +000000007fffffff *+1fffffff40000001 *+3fffffff00000001 *+3fffffffffffffff *+5ffffffec0000001 *+7ffffffe80000001 +0000000080000001 +0000000000000000 +0000000080000001 *+1fffffffbfffffff *+3fffffffffffffff *+4000000100000001 *+600000003fffffff *+800000007fffffff +00000000bfffffff +0000000000000000 +00000000bfffffff *+2fffffff00000001 *+5ffffffec0000001 *+600000003fffffff *+8ffffffe80000001 *+bffffffe40000001 +00000000ffffffff +0000000000000000 +00000000ffffffff *+3ffffffec0000001 *+7ffffffe80000001 *+800000007fffffff *+bffffffe40000001 *+fffffffe00000001 """ mul = Operators.SEXTEND(a, 256, 512) * Operators.SEXTEND(b, 256, 512) cond = Operators.UGE(mul, 1 << 256) return cond
def _in_user_func(state): """ :param state: current state :return: whether the current execution is in a user-defined function or not. NOTE / TODO / FIXME: As this may produce false postives, this is not in the base `Detector` class. It should be fixed at some point and moved there. See below. The first 4 bytes of tx data is keccak256 hash of the function signature that is called by given tx. All transactions start within Solidity dispatcher function: it takes passed hash and dispatches the execution to given function based on it. So: if we are in the dispatcher, *and contract have some functions* one of the first four tx data bytes will effectively have more than one solutions. BUT if contract have only a fallback function, the equation below may return more solutions when we are in a dispatcher function. <--- because of that, we warn that the detector is not that stable for contracts with only a fallback function. """ # If we are already in user function (we cached it) let's just return True in_function = state.context.get('in_function', False) prev_tx_count = state.context.get('prev_tx_count', 0) curr_tx_count = len(state.platform.transactions) new_human_tx = prev_tx_count != curr_tx_count if in_function and not new_human_tx: return True # This is expensive call, so we cache it in_function = len(state.solve_n(state.platform.current_transaction.data[:4], 2)) == 1 state.context['in_function'] = in_function state.context['prev_tx_count'] = curr_tx_count return in_function
def disassemble_instruction(self, code, pc): """Get next instruction using the Capstone disassembler :param str code: binary blob to be disassembled :param long pc: program counter """ return next(self.disasm.disasm(code, pc))
def istainted(arg, taint=None): """ Helper to determine whether an object if tainted. :param arg: a value or Expression :param taint: a regular expression matching a taint value (eg. 'IMPORTANT.*'). If None, this function checks for any taint value. """ if not issymbolic(arg): return False if taint is None: return len(arg.taint) != 0 for arg_taint in arg.taint: m = re.match(taint, arg_taint, re.DOTALL | re.IGNORECASE) if m: return True return False
def get_taints(arg, taint=None): """ Helper to list an object taints. :param arg: a value or Expression :param taint: a regular expression matching a taint value (eg. 'IMPORTANT.*'). If None, this function checks for any taint value. """ if not issymbolic(arg): return for arg_taint in arg.taint: if taint is not None: m = re.match(taint, arg_taint, re.DOTALL | re.IGNORECASE) if m: yield arg_taint else: yield arg_taint return
def taint_with(arg, taint, value_bits=256, index_bits=256): """ Helper to taint a value. :param arg: a value or Expression :param taint: a regular expression matching a taint value (eg. 'IMPORTANT.*'). If None, this function checks for any taint value. """ from ..core.smtlib import BitVecConstant # prevent circular imports tainted_fset = frozenset((taint,)) if not issymbolic(arg): if isinstance(arg, int): arg = BitVecConstant(value_bits, arg) arg._taint = tainted_fset else: raise ValueError("type not supported") else: arg = copy.copy(arg) arg._taint |= tainted_fset return arg
def interval_intersection(min1, max1, min2, max2): """ Given two intervals, (min1, max1) and (min2, max2) return their intersecting interval, or None if they do not overlap. """ left, right = max(min1, min2), min(max1, max2) if left < right: return left, right return None
def add(self, constraint, check=False): """ Add a constraint to the set :param constraint: The constraint to add to the set. :param check: Currently unused. :return: """ if isinstance(constraint, bool): constraint = BoolConstant(constraint) assert isinstance(constraint, Bool) constraint = simplify(constraint) # If self._child is not None this constraint set has been forked and a # a derived constraintset may be using this. So we can't add any more # constraints to this one. After the child constraintSet is deleted # we regain the ability to add constraints. if self._child is not None: raise Exception('ConstraintSet is frozen') if isinstance(constraint, BoolConstant): if not constraint.value: logger.info("Adding an impossible constant constraint") self._constraints = [constraint] else: return self._constraints.append(constraint) if check: from ...core.smtlib import solver if not solver.check(self): raise ValueError("Added an impossible constraint")
def _declare(self, var): """ Declare the variable `var` """ if var.name in self._declarations: raise ValueError('Variable already declared') self._declarations[var.name] = var return var
def declarations(self): """ Returns the variable expressions of this constraint set """ declarations = GetDeclarations() for a in self.constraints: try: declarations.visit(a) except RuntimeError: # TODO: (defunct) move recursion management out of PickleSerializer if sys.getrecursionlimit() >= PickleSerializer.MAX_RECURSION: raise Exception(f'declarations recursion limit surpassed {PickleSerializer.MAX_RECURSION}, aborting') new_limit = sys.getrecursionlimit() + PickleSerializer.DEFAULT_RECURSION if new_limit <= PickleSerializer.DEFAULT_RECURSION: sys.setrecursionlimit(new_limit) return self.declarations return declarations.result
def constraints(self): """ :rtype tuple :return: All constraints represented by this and parent sets. """ if self._parent is not None: return tuple(self._constraints) + self._parent.constraints return tuple(self._constraints)
def is_declared(self, expression_var): """ True if expression_var is declared in this constraint set """ if not isinstance(expression_var, Variable): raise ValueError(f'Expression must be a Variable (not a {type(expression_var)})') return any(expression_var is x for x in self.get_declared_variables())
def migrate(self, expression, name_migration_map=None): """ Migrate an expression created for a different constraint set to self. Returns an expression that can be used with this constraintSet All the foreign variables used in the expression are replaced by variables of this constraint set. If the variable was replaced before the replacement is taken from the provided migration map. The migration mapping is updated with new replacements. :param expression: the potentially foreign expression :param name_migration_map: mapping of already migrated variables. maps from string name of foreign variable to its currently existing migrated string name. this is updated during this migration. :return: a migrated expression where all the variables are local. name_migration_map is updated """ if name_migration_map is None: name_migration_map = {} # name_migration_map -> object_migration_map # Based on the name mapping in name_migration_map build an object to # object mapping to be used in the replacing of variables # inv: object_migration_map's keys should ALWAYS be external/foreign # expressions, and its values should ALWAYS be internal/local expressions object_migration_map = {} #List of foreign vars used in expression foreign_vars = itertools.filterfalse(self.is_declared, get_variables(expression)) for foreign_var in foreign_vars: # If a variable with the same name was previously migrated if foreign_var.name in name_migration_map: migrated_name = name_migration_map[foreign_var.name] native_var = self.get_variable(migrated_name) assert native_var is not None, "name_migration_map contains a variable that does not exist in this ConstraintSet" object_migration_map[foreign_var] = native_var else: # foreign_var was not found in the local declared variables nor # any variable with the same name was previously migrated # let's make a new unique internal name for it migrated_name = foreign_var.name if migrated_name in self._declarations: migrated_name = self._make_unique_name(f'{foreign_var.name}_migrated') # Create and declare a new variable of given type if isinstance(foreign_var, Bool): new_var = self.new_bool(name=migrated_name) elif isinstance(foreign_var, BitVec): new_var = self.new_bitvec(foreign_var.size, name=migrated_name) elif isinstance(foreign_var, Array): # Note that we are discarding the ArrayProxy encapsulation new_var = self.new_array(index_max=foreign_var.index_max, index_bits=foreign_var.index_bits, value_bits=foreign_var.value_bits, name=migrated_name).array else: raise NotImplemented(f"Unknown expression type {type(var)} encountered during expression migration") # Update the var to var mapping object_migration_map[foreign_var] = new_var # Update the name to name mapping name_migration_map[foreign_var.name] = new_var.name # Actually replace each appearance of migrated variables by the new ones migrated_expression = replace(expression, object_migration_map) return migrated_expression
def new_bool(self, name=None, taint=frozenset(), avoid_collisions=False): """ Declares a free symbolic boolean in the constraint store :param name: try to assign name to internal variable representation, if not unique, a numeric nonce will be appended :param avoid_collisions: potentially avoid_collisions the variable to avoid name collisions if True :return: a fresh BoolVariable """ if name is None: name = 'B' avoid_collisions = True if avoid_collisions: name = self._make_unique_name(name) if not avoid_collisions and name in self._declarations: raise ValueError(f'Name {name} already used') var = BoolVariable(name, taint=taint) return self._declare(var)
def new_bitvec(self, size, name=None, taint=frozenset(), avoid_collisions=False): """ Declares a free symbolic bitvector in the constraint store :param size: size in bits for the bitvector :param name: try to assign name to internal variable representation, if not unique, a numeric nonce will be appended :param avoid_collisions: potentially avoid_collisions the variable to avoid name collisions if True :return: a fresh BitVecVariable """ if not (size == 1 or size % 8 == 0): raise Exception(f'Invalid bitvec size {size}') if name is None: name = 'BV' avoid_collisions = True if avoid_collisions: name = self._make_unique_name(name) if not avoid_collisions and name in self._declarations: raise ValueError(f'Name {name} already used') var = BitVecVariable(size, name, taint=taint) return self._declare(var)
def new_array(self, index_bits=32, name=None, index_max=None, value_bits=8, taint=frozenset(), avoid_collisions=False, default=None): """ Declares a free symbolic array of value_bits long bitvectors in the constraint store. :param index_bits: size in bits for the array indexes one of [32, 64] :param value_bits: size in bits for the array values :param name: try to assign name to internal variable representation, if not unique, a numeric nonce will be appended :param index_max: upper limit for indexes on this array (#FIXME) :param avoid_collisions: potentially avoid_collisions the variable to avoid name collisions if True :param default: default for not initialized values :return: a fresh ArrayProxy """ if name is None: name = 'A' avoid_collisions = True if avoid_collisions: name = self._make_unique_name(name) if not avoid_collisions and name in self._declarations: raise ValueError(f'Name {name} already used') var = self._declare(ArrayVariable(index_bits, index_max, value_bits, name, taint=taint)) return ArrayProxy(var, default=default)
def train_encoder(X, y, fold_count, encoder): """ Defines folds and performs the data preprocessing (categorical encoding, NaN imputation, normalization) Returns a list with {X_train, y_train, X_test, y_test}, average fit_encoder_time and average score_encoder_time Note: We normalize all features (not only numerical features) because otherwise SVM would get stuck for hours on ordinal encoded cylinder.bands.arff dataset due to presence of unproportionally high values. Note: The fold count is variable because there are datasets, which have less than 10 samples in the minority class. Note: We do not use pipelines because of: https://github.com/scikit-learn/scikit-learn/issues/11832 """ kf = StratifiedKFold(n_splits=fold_count, shuffle=True, random_state=2001) encoder = deepcopy(encoder) # Because of https://github.com/scikit-learn-contrib/categorical-encoding/issues/106 imputer = SimpleImputer(strategy='mean') scaler = StandardScaler() folds = [] fit_encoder_time = 0 score_encoder_time = 0 for train_index, test_index in kf.split(X, y): # Split data X_train, X_test = X.iloc[train_index, :].reset_index(drop=True), X.iloc[test_index, :].reset_index(drop=True) y_train, y_test = y[train_index].reset_index(drop=True), y[test_index].reset_index(drop=True) # Training start_time = time.time() X_train = encoder.fit_transform(X_train, y_train) fit_encoder_time += time.time() - start_time X_train = imputer.fit_transform(X_train) X_train = scaler.fit_transform(X_train) # Testing start_time = time.time() X_test = encoder.transform(X_test) score_encoder_time += time.time() - start_time X_test = imputer.transform(X_test) X_test = scaler.transform(X_test) folds.append([X_train, y_train, X_test, y_test]) return folds, fit_encoder_time/fold_count, score_encoder_time/fold_count
def train_model(folds, model): """ Evaluation with: Matthews correlation coefficient: represents thresholding measures AUC: represents ranking measures Brier score: represents calibration measures """ scores = [] fit_model_time = 0 # Sum of all the time spend on fitting the training data, later on normalized score_model_time = 0 # Sum of all the time spend on scoring the testing data, later on normalized for X_train, y_train, X_test, y_test in folds: # Training start_time = time.time() with ignore_warnings(category=ConvergenceWarning): # Yes, neural networks do not always converge model.fit(X_train, y_train) fit_model_time += time.time() - start_time prediction_train_proba = model.predict_proba(X_train)[:, 1] prediction_train = (prediction_train_proba >= 0.5).astype('uint8') # Testing start_time = time.time() prediction_test_proba = model.predict_proba(X_test)[:, 1] score_model_time += time.time() - start_time prediction_test = (prediction_test_proba >= 0.5).astype('uint8') # When all the predictions are of a single class, we get a RuntimeWarning in matthews_corr with warnings.catch_warnings(): warnings.simplefilter("ignore") scores.append([ sklearn.metrics.matthews_corrcoef(y_test, prediction_test), sklearn.metrics.matthews_corrcoef(y_train, prediction_train), sklearn.metrics.roc_auc_score(y_test, prediction_test_proba), sklearn.metrics.roc_auc_score(y_train, prediction_train_proba), sklearn.metrics.brier_score_loss(y_test, prediction_test_proba), sklearn.metrics.brier_score_loss(y_train, prediction_train_proba) ]) return np.mean(scores, axis=0), fit_model_time/len(folds), score_model_time/len(folds)
def fit(self, X, y=None, **kwargs): """Fit encoder according to X and y. Parameters ---------- X : array-like, shape = [n_samples, n_features] Training vectors, where n_samples is the number of samples and n_features is the number of features. y : array-like, shape = [n_samples] Target values. Returns ------- self : encoder Returns self. """ self.base_n_encoder.fit(X, y, **kwargs) return self
def fit(self, X, y=None, **kwargs): """Fit encoder according to X and y. Parameters ---------- X : array-like, shape = [n_samples, n_features] Training vectors, where n_samples is the number of samples and n_features is the number of features. y : array-like, shape = [n_samples] Target values. Returns ------- self : encoder Returns self. """ # if the input dataset isn't already a dataframe, convert it to one (using default column names) # first check the type X = util.convert_input(X) self._dim = X.shape[1] # if columns aren't passed, just use every string column if self.cols is None: self.cols = util.get_obj_cols(X) else: self.cols = util.convert_cols_to_list(self.cols) if self.handle_missing == 'error': if X[self.cols].isnull().any().bool(): raise ValueError('Columns to be encoded can not contain null') # train an ordinal pre-encoder self.ordinal_encoder = OrdinalEncoder( verbose=self.verbose, cols=self.cols, handle_unknown='value', handle_missing='value' ) self.ordinal_encoder = self.ordinal_encoder.fit(X) ordinal_mapping = self.ordinal_encoder.category_mapping mappings_out = [] for switch in ordinal_mapping: values = switch.get('mapping') col = switch.get('col') column_mapping = self.fit_sum_coding(col, values, self.handle_missing, self.handle_unknown) mappings_out.append({'col': switch.get('col'), 'mapping': column_mapping, }) self.mapping = mappings_out X_temp = self.transform(X, override_return_df=True) self.feature_names = X_temp.columns.tolist() # drop all output columns with 0 variance. if self.drop_invariant: self.drop_cols = [] generated_cols = util.get_generated_cols(X, X_temp, self.cols) self.drop_cols = [x for x in generated_cols if X_temp[x].var() <= 10e-5] try: [self.feature_names.remove(x) for x in self.drop_cols] except KeyError as e: if self.verbose > 0: print("Could not remove column from feature names." "Not found in generated cols.\n{}".format(e)) return self
def inverse_transform(self, X_in): """ Perform the inverse transformation to encoded data. Will attempt best case reconstruction, which means it will return nan for handle_missing and handle_unknown settings that break the bijection. We issue warnings when some of those cases occur. Parameters ---------- X_in : array-like, shape = [n_samples, n_features] Returns ------- p: array, the same size of X_in """ X = X_in.copy(deep=True) # first check the type X = util.convert_input(X) if self._dim is None: raise ValueError( 'Must train encoder before it can be used to inverse_transform data') # then make sure that it is the right size if X.shape[1] != self._dim: if self.drop_invariant: raise ValueError("Unexpected input dimension %d, the attribute drop_invariant should " "set as False when transform data" % (X.shape[1],)) else: raise ValueError('Unexpected input dimension %d, expected %d' % (X.shape[1], self._dim,)) if not self.cols: return X if self.return_df else X.values if self.handle_unknown == 'value': for col in self.cols: if any(X[col] == -1): warnings.warn("inverse_transform is not supported because transform impute " "the unknown category -1 when encode %s" % (col,)) if self.handle_unknown == 'return_nan' and self.handle_missing == 'return_nan': for col in self.cols: if X[col].isnull().any(): warnings.warn("inverse_transform is not supported because transform impute " "the unknown category nan when encode %s" % (col,)) for switch in self.mapping: column_mapping = switch.get('mapping') inverse = pd.Series(data=column_mapping.index, index=column_mapping.get_values()) X[switch.get('col')] = X[switch.get('col')].map(inverse).astype(switch.get('data_type')) return X if self.return_df else X.values
def ordinal_encoding(X_in, mapping=None, cols=None, handle_unknown='value', handle_missing='value'): """ Ordinal encoding uses a single column of integers to represent the classes. An optional mapping dict can be passed in, in this case we use the knowledge that there is some true order to the classes themselves. Otherwise, the classes are assumed to have no true order and integers are selected at random. """ return_nan_series = pd.Series(data=[np.nan], index=[-2]) X = X_in.copy(deep=True) if cols is None: cols = X.columns.values if mapping is not None: mapping_out = mapping for switch in mapping: column = switch.get('col') X[column] = X[column].map(switch['mapping']) try: X[column] = X[column].astype(int) except ValueError as e: X[column] = X[column].astype(float) if handle_unknown == 'value': X[column].fillna(-1, inplace=True) elif handle_unknown == 'error': missing = X[column].isnull() if any(missing): raise ValueError('Unexpected categories found in column %s' % column) if handle_missing == 'return_nan': X[column] = X[column].map(return_nan_series).where(X[column] == -2, X[column]) else: mapping_out = [] for col in cols: nan_identity = np.nan if util.is_category(X[col].dtype): categories = X[col].cat.categories else: categories = X[col].unique() index = pd.Series(categories).fillna(nan_identity).unique() data = pd.Series(index=index, data=range(1, len(index) + 1)) if handle_missing == 'value' and ~data.index.isnull().any(): data.loc[nan_identity] = -2 elif handle_missing == 'return_nan': data.loc[nan_identity] = -2 mapping_out.append({'col': col, 'mapping': data, 'data_type': X[col].dtype}, ) return X, mapping_out
def transform(self, X, override_return_df=False): """Perform the transformation to new categorical data. Parameters ---------- X : array-like, shape = [n_samples, n_features] Returns ------- p : array, shape = [n_samples, n_numeric + N] Transformed values with encoding applied. """ if self.handle_missing == 'error': if X[self.cols].isnull().any().bool(): raise ValueError('Columns to be encoded can not contain null') if self._dim is None: raise ValueError( 'Must train encoder before it can be used to transform data.') # first check the type X = util.convert_input(X) # then make sure that it is the right size if X.shape[1] != self._dim: raise ValueError('Unexpected input dimension %d, expected %d' % ( X.shape[1], self._dim, )) if not self.cols: return X if self.return_df else X.values X = self.ordinal_encoder.transform(X) if self.handle_unknown == 'error': if X[self.cols].isin([-1]).any().any(): raise ValueError('Columns to be encoded can not contain new values') X = self.get_dummies(X) if self.drop_invariant: for col in self.drop_cols: X.drop(col, 1, inplace=True) if self.return_df or override_return_df: return X else: return X.values
def reverse_dummies(self, X, mapping): """ Convert dummy variable into numerical variables Parameters ---------- X : DataFrame mapping: list-like Contains mappings of column to be transformed to it's new columns and value represented Returns ------- numerical: DataFrame """ out_cols = X.columns.values.tolist() mapped_columns = [] for switch in mapping: col = switch.get('col') mod = switch.get('mapping') insert_at = out_cols.index(mod.columns[0]) X.insert(insert_at, col, 0) positive_indexes = mod.index[mod.index > 0] for i in range(positive_indexes.shape[0]): existing_col = mod.columns[i] val = positive_indexes[i] X.loc[X[existing_col] == 1, col] = val mapped_columns.append(existing_col) X.drop(mod.columns, axis=1, inplace=True) out_cols = X.columns.values.tolist() return X
def get_cars_data(): """ Load the cars dataset, split it into X and y, and then call the label encoder to get an integer y column. :return: """ df = pd.read_csv('source_data/cars/car.data.txt') X = df.reindex(columns=[x for x in df.columns.values if x != 'class']) y = df.reindex(columns=['class']) y = preprocessing.LabelEncoder().fit_transform(y.values.reshape(-1, )) mapping = [ {'col': 'buying', 'mapping': [('vhigh', 0), ('high', 1), ('med', 2), ('low', 3)]}, {'col': 'maint', 'mapping': [('vhigh', 0), ('high', 1), ('med', 2), ('low', 3)]}, {'col': 'doors', 'mapping': [('2', 0), ('3', 1), ('4', 2), ('5more', 3)]}, {'col': 'persons', 'mapping': [('2', 0), ('4', 1), ('more', 2)]}, {'col': 'lug_boot', 'mapping': [('small', 0), ('med', 1), ('big', 2)]}, {'col': 'safety', 'mapping': [('high', 0), ('med', 1), ('low', 2)]}, ] return X, y, mapping
def get_splice_data(): """ Load the mushroom dataset, split it into X and y, and then call the label encoder to get an integer y column. :return: """ df = pd.read_csv('source_data/splice/splice.csv') X = df.reindex(columns=[x for x in df.columns.values if x != 'class']) X['dna'] = X['dna'].map(lambda x: list(str(x).strip())) for idx in range(60): X['dna_%d' % (idx, )] = X['dna'].map(lambda x: x[idx]) del X['dna'] y = df.reindex(columns=['class']) y = preprocessing.LabelEncoder().fit_transform(y.values.reshape(-1, )) # this data is truly categorical, with no known concept of ordering mapping = None return X, y, mapping
def basen_to_integer(self, X, cols, base): """ Convert basen code as integers. Parameters ---------- X : DataFrame encoded data cols : list-like Column names in the DataFrame that be encoded base : int The base of transform Returns ------- numerical: DataFrame """ out_cols = X.columns.values.tolist() for col in cols: col_list = [col0 for col0 in out_cols if str(col0).startswith(str(col))] insert_at = out_cols.index(col_list[0]) if base == 1: value_array = np.array([int(col0.split('_')[-1]) for col0 in col_list]) else: len0 = len(col_list) value_array = np.array([base ** (len0 - 1 - i) for i in range(len0)]) X.insert(insert_at, col, np.dot(X[col_list].values, value_array.T)) X.drop(col_list, axis=1, inplace=True) out_cols = X.columns.values.tolist() return X
def col_transform(self, col, digits): """ The lambda body to transform the column values """ if col is None or float(col) < 0.0: return None else: col = self.number_to_base(int(col), self.base, digits) if len(col) == digits: return col else: return [0 for _ in range(digits - len(col))] + col
def fit(self, X, y=None, **kwargs): """Fit encoder according to X and y. Parameters ---------- X : array-like, shape = [n_samples, n_features] Training vectors, where n_samples is the number of samples and n_features is the number of features. y : array-like, shape = [n_samples] Target values. Returns ------- self : encoder Returns self. """ # first check the type X = util.convert_input(X) self._dim = X.shape[1] # if columns aren't passed, just use every string column if self.cols is None: self.cols = util.get_obj_cols(X) else: self.cols = util.convert_cols_to_list(self.cols) X_temp = self.transform(X, override_return_df=True) self.feature_names = X_temp.columns.tolist() # drop all output columns with 0 variance. if self.drop_invariant: self.drop_cols = [] generated_cols = util.get_generated_cols(X, X_temp, self.cols) self.drop_cols = [x for x in generated_cols if X_temp[x].var() <= 10e-5] try: [self.feature_names.remove(x) for x in self.drop_cols] except KeyError as e: if self.verbose > 0: print("Could not remove column from feature names." "Not found in generated cols.\n{}".format(e)) return self
def transform(self, X, override_return_df=False): """Perform the transformation to new categorical data. Parameters ---------- X : array-like, shape = [n_samples, n_features] Returns ------- p : array, shape = [n_samples, n_numeric + N] Transformed values with encoding applied. """ if self._dim is None: raise ValueError('Must train encoder before it can be used to transform data.') # first check the type X = util.convert_input(X) # then make sure that it is the right size if X.shape[1] != self._dim: raise ValueError('Unexpected input dimension %d, expected %d' % (X.shape[1], self._dim, )) if not self.cols: return X X = self.hashing_trick(X, hashing_method=self.hash_method, N=self.n_components, cols=self.cols) if self.drop_invariant: for col in self.drop_cols: X.drop(col, 1, inplace=True) if self.return_df or override_return_df: return X else: return X.values
def hashing_trick(X_in, hashing_method='md5', N=2, cols=None, make_copy=False): """A basic hashing implementation with configurable dimensionality/precision Performs the hashing trick on a pandas dataframe, `X`, using the hashing method from hashlib identified by `hashing_method`. The number of output dimensions (`N`), and columns to hash (`cols`) are also configurable. Parameters ---------- X_in: pandas dataframe description text hashing_method: string, optional description text N: int, optional description text cols: list, optional description text make_copy: bool, optional description text Returns ------- out : dataframe A hashing encoded dataframe. References ---------- Cite the relevant literature, e.g. [1]_. You may also cite these references in the notes section above. .. [1] Kilian Weinberger; Anirban Dasgupta; John Langford; Alex Smola; Josh Attenberg (2009). Feature Hashing for Large Scale Multitask Learning. Proc. ICML. """ try: if hashing_method not in hashlib.algorithms_available: raise ValueError('Hashing Method: %s Not Available. Please use one from: [%s]' % ( hashing_method, ', '.join([str(x) for x in hashlib.algorithms_available]) )) except Exception as e: try: _ = hashlib.new(hashing_method) except Exception as e: raise ValueError('Hashing Method: %s Not Found.') if make_copy: X = X_in.copy(deep=True) else: X = X_in if cols is None: cols = X.columns.values def hash_fn(x): tmp = [0 for _ in range(N)] for val in x.values: if val is not None: hasher = hashlib.new(hashing_method) if sys.version_info[0] == 2: hasher.update(str(val)) else: hasher.update(bytes(str(val), 'utf-8')) tmp[int(hasher.hexdigest(), 16) % N] += 1 return pd.Series(tmp, index=new_cols) new_cols = ['col_%d' % d for d in range(N)] X_cat = X.loc[:, cols] X_num = X.loc[:, [x for x in X.columns.values if x not in cols]] X_cat = X_cat.apply(hash_fn, axis=1) X_cat.columns = new_cols X = pd.concat([X_cat, X_num], axis=1) return X
def transform(self, X, y=None, override_return_df=False): """Perform the transformation to new categorical data. Parameters ---------- X : array-like, shape = [n_samples, n_features] y : array-like, shape = [n_samples] when transform by leave one out None, when transform without target information (such as transform test set) Returns ------- p : array, shape = [n_samples, n_numeric + N] Transformed values with encoding applied. """ if self.handle_missing == 'error': if X[self.cols].isnull().any().bool(): raise ValueError('Columns to be encoded can not contain null') if self._dim is None: raise ValueError('Must train encoder before it can be used to transform data.') # unite the input into pandas types X = util.convert_input(X) # then make sure that it is the right size if X.shape[1] != self._dim: raise ValueError('Unexpected input dimension %d, expected %d' % (X.shape[1], self._dim,)) # if we are encoding the training data, we have to check the target if y is not None: y = util.convert_input_vector(y, X.index).astype(float) if X.shape[0] != y.shape[0]: raise ValueError("The length of X is " + str(X.shape[0]) + " but length of y is " + str(y.shape[0]) + ".") if not self.cols: return X X = self.transform_leave_one_out( X, y, mapping=self.mapping ) if self.drop_invariant: for col in self.drop_cols: X.drop(col, 1, inplace=True) if self.return_df or override_return_df: return X else: return X.values
def transform_leave_one_out(self, X_in, y, mapping=None): """ Leave one out encoding uses a single column of floats to represent the means of the target variables. """ X = X_in.copy(deep=True) random_state_ = check_random_state(self.random_state) # Prepare the data if y is not None: # Convert bools to numbers (the target must be summable) y = y.astype('double') # Cumsum and cumcount do not work nicely with None. # This is a terrible workaround that will fail, when the # categorical input contains -999.9 for cat_col in X.select_dtypes('category').columns.values: X[cat_col] = X[cat_col].cat.add_categories(-999.9) X = X.fillna(-999.9) for col, colmap in mapping.items(): level_notunique = colmap['count'] > 1 unique_train = colmap.index unseen_values = pd.Series([x for x in X_in[col].unique() if x not in unique_train]) is_nan = X_in[col].isnull() is_unknown_value = X_in[col].isin(unseen_values.dropna()) if self.handle_unknown == 'error' and is_unknown_value.any(): raise ValueError('Columns to be encoded can not contain new values') if y is None: # Replace level with its mean target; if level occurs only once, use global mean level_means = ((colmap['sum'] + self._mean) / (colmap['count'] + 1)).where(level_notunique, self._mean) X[col] = X[col].map(level_means) else: # Simulation of CatBoost implementation, which calculates leave-one-out on the fly. # The nice thing about this is that it helps to prevent overfitting. The bad thing # is that CatBoost uses many iterations over the data. But we run just one iteration. # Still, it works better than leave-one-out without any noise. # See: # https://tech.yandex.com/catboost/doc/dg/concepts/algorithm-main-stages_cat-to-numberic-docpage/ temp = y.groupby(X[col]).agg(['cumsum', 'cumcount']) X[col] = (temp['cumsum'] - y + self._mean) / (temp['cumcount'] + 1) if self.handle_unknown == 'value': X.loc[is_unknown_value, col] = self._mean elif self.handle_unknown == 'return_nan': X.loc[is_unknown_value, col] = np.nan if self.handle_missing == 'value': X.loc[is_nan & unseen_values.isnull().any(), col] = self._mean elif self.handle_missing == 'return_nan': X.loc[is_nan, col] = np.nan if self.sigma is not None and y is not None: X[col] = X[col] * random_state_.normal(1., self.sigma, X[col].shape[0]) return X
def get_obj_cols(df): """ Returns names of 'object' columns in the DataFrame. """ obj_cols = [] for idx, dt in enumerate(df.dtypes): if dt == 'object' or is_category(dt): obj_cols.append(df.columns.values[idx]) return obj_cols
def convert_input(X): """ Unite data into a DataFrame. """ if not isinstance(X, pd.DataFrame): if isinstance(X, list): X = pd.DataFrame(X) elif isinstance(X, (np.generic, np.ndarray)): X = pd.DataFrame(X) elif isinstance(X, csr_matrix): X = pd.DataFrame(X.todense()) elif isinstance(X, pd.Series): X = pd.DataFrame(X) else: raise ValueError('Unexpected input type: %s' % (str(type(X)))) X = X.apply(lambda x: pd.to_numeric(x, errors='ignore')) return X
def convert_input_vector(y, index): """ Unite target data type into a Series. If the target is a Series or a DataFrame, we preserve its index. But if the target does not contain index attribute, we use the index from the argument. """ if y is None: return None if isinstance(y, pd.Series): return y elif isinstance(y, np.ndarray): if len(np.shape(y))==1: # vector return pd.Series(y, name='target', index=index) elif len(np.shape(y))==2 and np.shape(y)[0]==1: # single row in a matrix return pd.Series(y[0, :], name='target', index=index) elif len(np.shape(y))==2 and np.shape(y)[1]==1: # single column in a matrix return pd.Series(y[:, 0], name='target', index=index) else: raise ValueError('Unexpected input shape: %s' % (str(np.shape(y)))) elif np.isscalar(y): return pd.Series([y], name='target', index=index) elif isinstance(y, list): if len(y)==0 or (len(y)>0 and not isinstance(y[0], list)): # empty list or a vector return pd.Series(y, name='target', index=index) elif len(y)>0 and isinstance(y[0], list) and len(y[0])==1: # single row in a matrix flatten = lambda y: [item for sublist in y for item in sublist] return pd.Series(flatten(y), name='target', index=index) elif len(y)==1 and isinstance(y[0], list): # single column in a matrix return pd.Series(y[0], name='target', index=index) else: raise ValueError('Unexpected input shape') elif isinstance(y, pd.DataFrame): if len(list(y))==0: # empty DataFrame return pd.Series(y, name='target') if len(list(y))==1: # a single column return y.iloc[:, 0] else: raise ValueError('Unexpected input shape: %s' % (str(y.shape))) else: return pd.Series(y, name='target', index=index)
def get_generated_cols(X_original, X_transformed, to_transform): """ Returns a list of the generated/transformed columns. Arguments: X_original: df the original (input) DataFrame. X_transformed: df the transformed (current) DataFrame. to_transform: [str] a list of columns that were transformed (as in the original DataFrame), commonly self.cols. Output: a list of columns that were transformed (as in the current DataFrame). """ original_cols = list(X_original.columns) if len(to_transform) > 0: [original_cols.remove(c) for c in to_transform] current_cols = list(X_transformed.columns) if len(original_cols) > 0: [current_cols.remove(c) for c in original_cols] return current_cols
def fit(self, X, y, **kwargs): """Fit encoder according to X and y. Parameters ---------- X : array-like, shape = [n_samples, n_features] Training vectors, where n_samples is the number of samples and n_features is the number of features. y : array-like, shape = [n_samples] Target values. Returns ------- self : encoder Returns self. """ # unite the input into pandas types X = util.convert_input(X) y = util.convert_input_vector(y, X.index) if X.shape[0] != y.shape[0]: raise ValueError("The length of X is " + str(X.shape[0]) + " but length of y is " + str(y.shape[0]) + ".") self._dim = X.shape[1] # if columns aren't passed, just use every string column if self.cols is None: self.cols = util.get_obj_cols(X) else: self.cols = util.convert_cols_to_list(self.cols) if self.handle_missing == 'error': if X[self.cols].isnull().any().bool(): raise ValueError('Columns to be encoded can not contain null') self.ordinal_encoder = OrdinalEncoder( verbose=self.verbose, cols=self.cols, handle_unknown='value', handle_missing='value' ) self.ordinal_encoder = self.ordinal_encoder.fit(X) X_ordinal = self.ordinal_encoder.transform(X) self.mapping = self.fit_target_encoding(X_ordinal, y) X_temp = self.transform(X, override_return_df=True) self.feature_names = list(X_temp.columns) if self.drop_invariant: self.drop_cols = [] X_temp = self.transform(X) generated_cols = util.get_generated_cols(X, X_temp, self.cols) self.drop_cols = [x for x in generated_cols if X_temp[x].var() <= 10e-5] try: [self.feature_names.remove(x) for x in self.drop_cols] except KeyError as e: if self.verbose > 0: print("Could not remove column from feature names." "Not found in generated cols.\n{}".format(e)) return self
def transform_leave_one_out(self, X_in, y, mapping=None): """ Leave one out encoding uses a single column of floats to represent the means of the target variables. """ X = X_in.copy(deep=True) random_state_ = check_random_state(self.random_state) for col, colmap in mapping.items(): level_notunique = colmap['count'] > 1 unique_train = colmap.index unseen_values = pd.Series([x for x in X[col].unique() if x not in unique_train]) is_nan = X[col].isnull() is_unknown_value = X[col].isin(unseen_values.dropna()) if self.handle_unknown == 'error' and is_unknown_value.any(): raise ValueError('Columns to be encoded can not contain new values') if y is None: # Replace level with its mean target; if level occurs only once, use global mean level_means = (colmap['sum'] / colmap['count']).where(level_notunique, self._mean) X[col] = X[col].map(level_means) else: # Replace level with its mean target, calculated excluding this row's target # The y (target) mean for this level is normally just the sum/count; # excluding this row's y, it's (sum - y) / (count - 1) level_means = (X[col].map(colmap['sum']) - y) / (X[col].map(colmap['count']) - 1) # The 'where' fills in singleton levels (count = 1 -> div by 0) with the global mean X[col] = level_means.where(X[col].map(colmap['count'][level_notunique]).notnull(), self._mean) if self.handle_unknown == 'value': X.loc[is_unknown_value, col] = self._mean elif self.handle_unknown == 'return_nan': X.loc[is_unknown_value, col] = np.nan if self.handle_missing == 'value': X.loc[is_nan & unseen_values.isnull().any(), col] = self._mean elif self.handle_missing == 'return_nan': X.loc[is_nan, col] = np.nan if self.sigma is not None and y is not None: X[col] = X[col] * random_state_.normal(1., self.sigma, X[col].shape[0]) return X
def score_models(clf, X, y, encoder, runs=1): """ Takes in a classifier that supports multiclass classification, and X and a y, and returns a cross validation score. """ scores = [] X_test = None for _ in range(runs): X_test = encoder().fit_transform(X, y) # Some models, like logistic regression, like normalized features otherwise they underperform and/or take a long time to converge. # To be rigorous, we should have trained the normalization on each fold individually via pipelines. # See grid_search_example to learn how to do it. X_test = StandardScaler().fit_transform(X_test) scores.append(cross_validate(clf, X_test, y, n_jobs=1, cv=5)['test_score']) gc.collect() scores = [y for z in [x for x in scores] for y in z] return float(np.mean(scores)), float(np.std(scores)), scores, X_test.shape[1]
def main(loader, name): """ Here we iterate through the datasets and score them with a classifier using different encodings. """ scores = [] raw_scores_ds = {} # first get the dataset X, y, mapping = loader() clf = linear_model.LogisticRegression(solver='lbfgs', multi_class='auto', max_iter=200, random_state=0) # try each encoding method available, which works on multiclass problems encoders = (set(category_encoders.__all__) - {'WOEEncoder'}) # WoE is currently only for binary targets for encoder_name in encoders: encoder = getattr(category_encoders, encoder_name) start_time = time.time() score, stds, raw_scores, dim = score_models(clf, X, y, encoder) scores.append([encoder_name, name, dim, score, stds, time.time() - start_time]) raw_scores_ds[encoder_name] = raw_scores gc.collect() results = pd.DataFrame(scores, columns=['Encoding', 'Dataset', 'Dimensionality', 'Avg. Score', 'Score StDev', 'Elapsed Time']) raw = pd.DataFrame.from_dict(raw_scores_ds) ax = raw.plot(kind='box', return_type='axes') plt.title('Scores for Encodings on %s Dataset' % (name,)) plt.ylabel('Score (higher is better)') for tick in ax.get_xticklabels(): tick.set_rotation(90) plt.grid() plt.tight_layout() plt.show() return results, raw
def secho(message, **kwargs): """A wrapper around click.secho that disables any coloring being used if colors have been disabled. """ # If colors are disabled, remove any color or other style data # from keyword arguments. if not settings.color: for key in ('fg', 'bg', 'bold', 'blink'): kwargs.pop(key, None) # Okay, now call click.secho normally. return click.secho(message, **kwargs)
def associate_notification_template(self, job_template, notification_template, status): """Associate a notification template from this job template. =====API DOCS===== Associate a notification template from this job template. :param job_template: The job template to associate to. :type job_template: str :param notification_template: The notification template to be associated. :type notification_template: str :param status: type of notification this notification template should be associated to. :type status: str :returns: Dictionary of only one key "changed", which indicates whether the association succeeded. :rtype: dict =====API DOCS===== """ return self._assoc('notification_templates_%s' % status, job_template, notification_template)
def disassociate_notification_template(self, job_template, notification_template, status): """Disassociate a notification template from this job template. =====API DOCS===== Disassociate a notification template from this job template. :param job_template: The job template to disassociate from. :type job_template: str :param notification_template: The notification template to be disassociated. :type notification_template: str :param status: type of notification this notification template should be disassociated from. :type status: str :returns: Dictionary of only one key "changed", which indicates whether the disassociation succeeded. :rtype: dict =====API DOCS===== """ return self._disassoc('notification_templates_%s' % status, job_template, notification_template)
def callback(self, pk=None, host_config_key='', extra_vars=None): """Contact Tower and request a configuration update using this job template. =====API DOCS===== Contact Tower and request a provisioning callback using this job template. :param pk: Primary key of the job template to run provisioning callback against. :type pk: int :param host_config_key: Key string used to authenticate the callback host. :type host_config_key: str :param extra_vars: Extra variables that are passed to provisioning callback. :type extra_vars: array of str :returns: A dictionary of a single key "changed", which indicates whether the provisioning callback is successful. :rtype: dict =====API DOCS===== """ url = self.endpoint + '%s/callback/' % pk if not host_config_key: host_config_key = client.get(url).json()['host_config_key'] post_data = {'host_config_key': host_config_key} if extra_vars: post_data['extra_vars'] = parser.process_extra_vars(list(extra_vars), force_json=True) r = client.post(url, data=post_data, auth=None) if r.status_code == 201: return {'changed': True}
def jt_aggregate(func, is_create=False, has_pk=False): """Decorator to aggregate unified_jt-related fields. Args: func: The CURD method to be decorated. is_create: Boolean flag showing whether this method is create. has_pk: Boolean flag showing whether this method uses pk as argument. Returns: A function with necessary click-related attributes whose keyworded arguments are aggregated. Raises: exc.UsageError: Either more than one unified jt fields are provided, or none is provided when is_create flag is set. """ def helper(kwargs, obj): """The helper function preceding actual function that aggregates unified jt fields. """ unified_job_template = None for item in UNIFIED_JT: if kwargs.get(item, None) is not None: jt_id = kwargs.pop(item) if unified_job_template is None: unified_job_template = (item, jt_id) else: raise exc.UsageError( 'More than one unified job template fields provided, ' 'please tighten your criteria.' ) if unified_job_template is not None: kwargs['unified_job_template'] = unified_job_template[1] obj.identity = tuple(list(obj.identity) + ['unified_job_template']) return '/'.join([UNIFIED_JT[unified_job_template[0]], str(unified_job_template[1]), 'schedules/']) elif is_create: raise exc.UsageError('You must provide exactly one unified job' ' template field during creation.') def decorator_without_pk(obj, *args, **kwargs): old_endpoint = obj.endpoint new_endpoint = helper(kwargs, obj) if is_create: obj.endpoint = new_endpoint result = func(obj, *args, **kwargs) obj.endpoint = old_endpoint return result def decorator_with_pk(obj, pk=None, *args, **kwargs): old_endpoint = obj.endpoint new_endpoint = helper(kwargs, obj) if is_create: obj.endpoint = new_endpoint result = func(obj, pk=pk, *args, **kwargs) obj.endpoint = old_endpoint return result decorator = decorator_with_pk if has_pk else decorator_without_pk for item in CLICK_ATTRS: setattr(decorator, item, getattr(func, item, [])) decorator.__doc__ = func.__doc__ return decorator
def lookup_stdout(self, pk=None, start_line=None, end_line=None, full=True): """ Internal method that lies to our `monitor` method by returning a scorecard for the workflow job where the standard out would have been expected. """ uj_res = get_resource('unified_job') # Filters # - limit search to jobs spawned as part of this workflow job # - order in the order in which they should add to the list # - only include final job states query_params = (('unified_job_node__workflow_job', pk), ('order_by', 'finished'), ('status__in', 'successful,failed,error')) jobs_list = uj_res.list(all_pages=True, query=query_params) if jobs_list['count'] == 0: return '' return_content = ResSubcommand(uj_res)._format_human(jobs_list) lines = return_content.split('\n') if not full: lines = lines[:-1] N = len(lines) start_range = start_line if start_line is None: start_range = 0 elif start_line > N: start_range = N end_range = end_line if end_line is None or end_line > N: end_range = N lines = lines[start_range:end_range] return_content = '\n'.join(lines) if len(lines) > 0: return_content += '\n' return return_content
def launch(self, workflow_job_template=None, monitor=False, wait=False, timeout=None, extra_vars=None, **kwargs): """Launch a new workflow job based on a workflow job template. Creates a new workflow job in Ansible Tower, starts it, and returns back an ID in order for its status to be monitored. =====API DOCS===== Launch a new workflow job based on a workflow job template. :param workflow_job_template: Primary key or name of the workflow job template to launch new job. :type workflow_job_template: str :param monitor: Flag that if set, immediately calls ``monitor`` on the newly launched workflow job rather than exiting with a success. :type monitor: bool :param wait: Flag that if set, monitor the status of the workflow job, but do not print while job is in progress. :type wait: bool :param timeout: If provided with ``monitor`` flag set, this attempt will time out after the given number of seconds. :type timeout: int :param extra_vars: yaml formatted texts that contains extra variables to pass on. :type extra_vars: array of strings :param `**kwargs`: Fields needed to create and launch a workflow job. :returns: Result of subsequent ``monitor`` call if ``monitor`` flag is on; Result of subsequent ``wait`` call if ``wait`` flag is on; loaded JSON output of the job launch if none of the two flags are on. :rtype: dict =====API DOCS===== """ if extra_vars is not None and len(extra_vars) > 0: kwargs['extra_vars'] = parser.process_extra_vars(extra_vars) debug.log('Launching the workflow job.', header='details') self._pop_none(kwargs) post_response = client.post('workflow_job_templates/{0}/launch/'.format( workflow_job_template), data=kwargs).json() workflow_job_id = post_response['id'] post_response['changed'] = True if monitor: return self.monitor(workflow_job_id, timeout=timeout) elif wait: return self.wait(workflow_job_id, timeout=timeout) return post_response
def parse_args(self, ctx, args): """Parse arguments sent to this command. The code for this method is taken from MultiCommand: https://github.com/mitsuhiko/click/blob/master/click/core.py It is Copyright (c) 2014 by Armin Ronacher. See the license: https://github.com/mitsuhiko/click/blob/master/LICENSE """ if not args and self.no_args_is_help and not ctx.resilient_parsing: click.echo(ctx.get_help()) ctx.exit() return super(ActionSubcommand, self).parse_args(ctx, args)
def format_options(self, ctx, formatter): """Monkey-patch click's format_options method to support option categorization. """ field_opts = [] global_opts = [] local_opts = [] other_opts = [] for param in self.params: if param.name in SETTINGS_PARMS: opts = global_opts elif getattr(param, 'help', None) and param.help.startswith('[FIELD]'): opts = field_opts param.help = param.help[len('[FIELD]'):] else: opts = local_opts rv = param.get_help_record(ctx) if rv is None: continue else: opts.append(rv) if self.add_help_option: help_options = self.get_help_option_names(ctx) if help_options: other_opts.append([join_options(help_options)[0], 'Show this message and exit.']) if field_opts: with formatter.section('Field Options'): formatter.write_dl(field_opts) if local_opts: with formatter.section('Local Options'): formatter.write_dl(local_opts) if global_opts: with formatter.section('Global Options'): formatter.write_dl(global_opts) if other_opts: with formatter.section('Other Options'): formatter.write_dl(other_opts)
def list(self, **kwargs): """Return a list of objects. =====API DOCS===== Retrieve a list of Tower settings. :param category: The category slug in which to look up indevidual settings. :type category: str :param `**kwargs`: Keyword arguments list of available fields used for searching resource objects. :returns: A JSON object containing details of all resource objects returned by Tower backend. :rtype: dict =====API DOCS===== """ self.custom_category = kwargs.get('category', 'all') try: result = super(Resource, self).list(**kwargs) except exc.NotFound as e: categories = map( lambda category: category['slug'], client.get('/settings/').json()['results'] ) e.message = '%s is not a valid category. Choose from [%s]' % ( kwargs['category'], ', '.join(categories) ) raise e finally: self.custom_category = None return { 'results': [{'id': k, 'value': v} for k, v in result.items()] }
def get(self, pk): """Return one and exactly one object =====API DOCS===== Return one and exactly one Tower setting. :param pk: Primary key of the Tower setting to retrieve :type pk: int :returns: loaded JSON of the retrieved Tower setting object. :rtype: dict :raises tower_cli.exceptions.NotFound: When no specified Tower setting exists. =====API DOCS===== """ # The Tower API doesn't provide a mechanism for retrieving a single # setting value at a time, so fetch them all and filter try: return next(s for s in self.list()['results'] if s['id'] == pk) except StopIteration: raise exc.NotFound('The requested object could not be found.')
def modify(self, setting, value): """Modify an already existing object. Positional argument SETTING is the setting name and VALUE is its value, which can be provided directly or obtained from a file name if prefixed with '@'. =====API DOCS===== Modify an already existing Tower setting. :param setting: The name of the Tower setting to be modified. :type setting: str :param value: The new value of the Tower setting. :type value: str :returns: A dictionary combining the JSON output of the modified resource, as well as two extra fields: "changed", a flag indicating if the resource is successfully updated; "id", an integer which is the primary key of the updated object. :rtype: dict =====API DOCS===== """ prev_value = new_value = self.get(setting)['value'] answer = OrderedDict() encrypted = '$encrypted$' in six.text_type(prev_value) if encrypted or six.text_type(prev_value) != six.text_type(value): if setting == 'LICENSE': r = client.post('/config/', data=self.coerce_type(setting, value)) new_value = r.json() else: r = client.patch( self.endpoint, data={setting: self.coerce_type(setting, value)} ) new_value = r.json()[setting] answer.update(r.json()) changed = encrypted or (prev_value != new_value) answer.update({ 'changed': changed, 'id': setting, 'value': new_value, }) return answer
def _pop_none(self, kwargs): """Remove default values (anything where the value is None). click is unfortunately bad at the way it sends through unspecified defaults.""" for key, value in copy(kwargs).items(): # options with multiple=True return a tuple if value is None or value == (): kwargs.pop(key) if hasattr(value, 'read'): kwargs[key] = value.read()
def _lookup(self, fail_on_missing=False, fail_on_found=False, include_debug_header=True, **kwargs): """ =====API DOCS===== Attempt to perform a lookup that is expected to return a single result, and return the record. This method is a wrapper around `get` that strips out non-unique keys, and is used internally by `write` and `delete`. :param fail_on_missing: Flag that raise exception if no resource is found. :type fail_on_missing: bool :param fail_on_found: Flag that raise exception if a resource is found. :type fail_on_found: bool :param include_debug_header: Flag determining whether to print debug messages when querying Tower backend. :type include_debug_header: bool :param `**kwargs`: Keyword arguments list of available fields used for searching resource. :returns: A JSON object containing details of the resource returned by Tower backend. :rtype: dict :raises tower_cli.exceptions.BadRequest: When no field are provided in kwargs. :raises tower_cli.exceptions.Found: When a resource is found and fail_on_found flag is on. :raises tower_cli.exceptions.NotFound: When no resource is found and fail_on_missing flag is on. =====API DOCS===== """ read_params = {} for field_name in self.identity: if field_name in kwargs: read_params[field_name] = kwargs[field_name] if 'id' in self.identity and len(self.identity) == 1: return {} if not read_params: raise exc.BadRequest('Cannot reliably determine which record to write. Include an ID or unique ' 'fields.') try: existing_data = self.get(include_debug_header=include_debug_header, **read_params) if fail_on_found: raise exc.Found('A record matching %s already exists, and you requested a failure in that case.' % read_params) return existing_data except exc.NotFound: if fail_on_missing: raise exc.NotFound('A record matching %s does not exist, and you requested a failure in that case.' % read_params) return {}
def _convert_pagenum(self, kwargs): """ Convert next and previous from URLs to integers """ for key in ('next', 'previous'): if not kwargs.get(key): continue match = re.search(r'page=(?P<num>[\d]+)', kwargs[key]) if match is None and key == 'previous': kwargs[key] = 1 continue kwargs[key] = int(match.groupdict()['num'])
def read(self, pk=None, fail_on_no_results=False, fail_on_multiple_results=False, **kwargs): """ =====API DOCS===== Retrieve and return objects from the Ansible Tower API. :param pk: Primary key of the resource to be read. Tower CLI will only attempt to read that object if ``pk`` is provided (not ``None``). :type pk: int :param fail_on_no_results: Flag that if set, zero results is considered a failure case and raises an exception; otherwise, empty list is returned. (Note: This is always True if a primary key is included.) :type fail_on_no_results: bool :param fail_on_multiple_results: Flag that if set, at most one result is expected, and more results constitutes a failure case. (Note: This is meaningless if a primary key is included, as there can never be multiple results.) :type fail_on_multiple_results: bool :param query: Contains 2-tuples used as query parameters to filter resulting resource objects. :type query: list :param `**kwargs`: Keyword arguments which, all together, will be used as query parameters to filter resulting resource objects. :returns: loaded JSON from Tower backend response body. :rtype: dict :raises tower_cli.exceptions.BadRequest: When 2-tuples in ``query`` overlaps key-value pairs in ``**kwargs``. :raises tower_cli.exceptions.NotFound: When no objects are found and ``fail_on_no_results`` flag is on. :raises tower_cli.exceptions.MultipleResults: When multiple objects are found and ``fail_on_multiple_results`` flag is on. =====API DOCS===== """ # Piece together the URL we will be hitting. url = self.endpoint if pk: url += '%s/' % pk # Pop the query parameter off of the keyword arguments; it will # require special handling (below). queries = kwargs.pop('query', []) # Remove default values (anything where the value is None). self._pop_none(kwargs) # Remove fields that are specifically excluded from lookup for field in self.fields: if field.no_lookup and field.name in kwargs: kwargs.pop(field.name) # If queries were provided, process them. params = list(kwargs.items()) for query in queries: params.append((query[0], query[1])) # Make the request to the Ansible Tower API. r = client.get(url, params=params) resp = r.json() # If this was a request with a primary key included, then at the # point that we got a good result, we know that we're done and can # return the result. if pk: # Make the results all look the same, for easier parsing # by other methods. # # Note that the `get` method will effectively undo this operation, # but that's a good thing, because we might use `get` without a # primary key. return {'count': 1, 'results': [resp]} # Did we get zero results back when we shouldn't? # If so, this is an error, and we need to complain. if fail_on_no_results and resp['count'] == 0: raise exc.NotFound('The requested object could not be found.') # Did we get more than one result back? # If so, this is also an error, and we need to complain. if fail_on_multiple_results and resp['count'] >= 2: raise exc.MultipleResults('Expected one result, got %d. Possibly caused by not providing required ' 'fields. Please tighten your criteria.' % resp['count']) # Return the response. return resp
def write(self, pk=None, create_on_missing=False, fail_on_found=False, force_on_exists=True, **kwargs): """ =====API DOCS===== Modify the given object using the Ansible Tower API. :param pk: Primary key of the resource to be read. Tower CLI will only attempt to read that object if ``pk`` is provided (not ``None``). :type pk: int :param create_on_missing: Flag that if set, a new object is created if ``pk`` is not set and objects matching the appropriate unique criteria is not found. :type create_on_missing: bool :param fail_on_found: Flag that if set, the operation fails if an object matching the unique criteria already exists. :type fail_on_found: bool :param force_on_exists: Flag that if set, then if an object is modified based on matching via unique fields (as opposed to the primary key), other fields are updated based on data sent; If unset, then the non-unique values are only written in a creation case. :type force_on_exists: bool :param `**kwargs`: Keyword arguments which, all together, will be used as POST/PATCH body to create/modify the resource object. if ``pk`` is not set, key-value pairs of ``**kwargs`` which are also in resource's identity will be used to lookup existing reosource. :returns: A dictionary combining the JSON output of the resource, as well as two extra fields: "changed", a flag indicating if the resource is created or successfully updated; "id", an integer which is the primary key of the specified object. :rtype: dict :raises tower_cli.exceptions.BadRequest: When required fields are missing in ``**kwargs`` when creating a new resource object. =====API DOCS===== """ existing_data = {} # Remove default values (anything where the value is None). self._pop_none(kwargs) # Determine which record we are writing, if we weren't given a primary key. if not pk: debug.log('Checking for an existing record.', header='details') existing_data = self._lookup( fail_on_found=fail_on_found, fail_on_missing=not create_on_missing, include_debug_header=False, **kwargs ) if existing_data: pk = existing_data['id'] else: # We already know the primary key, but get the existing data. # This allows us to know whether the write made any changes. debug.log('Getting existing record.', header='details') existing_data = self.get(pk) # Sanity check: Are we missing required values? # If we don't have a primary key, then all required values must be set, and if they're not, it's an error. missing_fields = [] for i in self.fields: if i.key not in kwargs and i.name not in kwargs and i.required: missing_fields.append(i.key or i.name) if missing_fields and not pk: raise exc.BadRequest('Missing required fields: %s' % ', '.join(missing_fields).replace('_', '-')) # Sanity check: Do we need to do a write at all? # If `force_on_exists` is False and the record was, in fact, found, then no action is required. if pk and not force_on_exists: debug.log('Record already exists, and --force-on-exists is off; do nothing.', header='decision', nl=2) answer = OrderedDict((('changed', False), ('id', pk))) answer.update(existing_data) return answer # Similarly, if all existing data matches our write parameters, there's no need to do anything. if all([kwargs[k] == existing_data.get(k, None) for k in kwargs.keys()]): debug.log('All provided fields match existing data; do nothing.', header='decision', nl=2) answer = OrderedDict((('changed', False), ('id', pk))) answer.update(existing_data) return answer # Reinsert None for special case of null association for key in kwargs: if kwargs[key] == 'null': kwargs[key] = None # Get the URL and method to use for the write. url = self.endpoint method = 'POST' if pk: url = self._get_patch_url(url, pk) method = 'PATCH' # If debugging is on, print the URL and data being sent. debug.log('Writing the record.', header='details') # Actually perform the write. r = getattr(client, method.lower())(url, data=kwargs) # At this point, we know the write succeeded, and we know that data was changed in the process. answer = OrderedDict((('changed', True), ('id', r.json()['id']))) answer.update(r.json()) return answer
def delete(self, pk=None, fail_on_missing=False, **kwargs): """Remove the given object. If `fail_on_missing` is True, then the object's not being found is considered a failure; otherwise, a success with no change is reported. =====API DOCS===== Remove the given object. :param pk: Primary key of the resource to be deleted. :type pk: int :param fail_on_missing: Flag that if set, the object's not being found is considered a failure; otherwise, a success with no change is reported. :type fail_on_missing: bool :param `**kwargs`: Keyword arguments used to look up resource object to delete if ``pk`` is not provided. :returns: dictionary of only one field "changed", which is a flag indicating whether the specified resource is successfully deleted. :rtype: dict =====API DOCS===== """ # If we weren't given a primary key, determine which record we're deleting. if not pk: existing_data = self._lookup(fail_on_missing=fail_on_missing, **kwargs) if not existing_data: return {'changed': False} pk = existing_data['id'] # Attempt to delete the record. If it turns out the record doesn't exist, handle the 404 appropriately # (this is an okay response if `fail_on_missing` is False). url = '%s%s/' % (self.endpoint, pk) debug.log('DELETE %s' % url, fg='blue', bold=True) try: client.delete(url) return {'changed': True} except exc.NotFound: if fail_on_missing: raise return {'changed': False}
def get(self, pk=None, **kwargs): """Return one and exactly one object. Lookups may be through a primary key, specified as a positional argument, and/or through filters specified through keyword arguments. If the number of results does not equal one, raise an exception. =====API DOCS===== Retrieve one and exactly one object. :param pk: Primary key of the resource to be read. Tower CLI will only attempt to read *that* object if ``pk`` is provided (not ``None``). :type pk: int :param `**kwargs`: Keyword arguments used to look up resource object to retrieve if ``pk`` is not provided. :returns: loaded JSON of the retrieved resource object. :rtype: dict =====API DOCS===== """ if kwargs.pop('include_debug_header', True): debug.log('Getting the record.', header='details') response = self.read(pk=pk, fail_on_no_results=True, fail_on_multiple_results=True, **kwargs) return response['results'][0]
def list(self, all_pages=False, **kwargs): """Return a list of objects. If one or more filters are provided through keyword arguments, filter the results accordingly. If no filters are provided, return all results. =====API DOCS===== Retrieve a list of objects. :param all_pages: Flag that if set, collect all pages of content from the API when returning results. :type all_pages: bool :param page: The page to show. Ignored if all_pages is set. :type page: int :param query: Contains 2-tuples used as query parameters to filter resulting resource objects. :type query: list :param `**kwargs`: Keyword arguments list of available fields used for searching resource objects. :returns: A JSON object containing details of all resource objects returned by Tower backend. :rtype: dict =====API DOCS===== """ # TODO: Move to a field callback method to make it generic # If multiple statuses where given, add OR queries for each of them if kwargs.get('status', None) and ',' in kwargs['status']: all_status = kwargs.pop('status').strip(',').split(',') queries = list(kwargs.pop('query', ())) for status in all_status: if status in STATUS_CHOICES: queries.append(('or__status', status)) else: raise exc.TowerCLIError('This status does not exist: {}'.format(status)) kwargs['query'] = tuple(queries) # If the `all_pages` flag is set, then ignore any page that might also be sent. if all_pages: kwargs.pop('page', None) kwargs.pop('page_size', None) # Get the response. debug.log('Getting records.', header='details') response = self.read(**kwargs) # Convert next and previous to int self._convert_pagenum(response) # If we were asked for all pages, keep retrieving pages until we have them all. if all_pages and response['next']: cursor = copy(response) while cursor['next']: cursor = self.read(**dict(kwargs, page=cursor['next'])) self._convert_pagenum(cursor) response['results'] += cursor['results'] response['count'] += cursor['count'] response['next'] = None # Done; return the response return response
def _disassoc(self, url_fragment, me, other): """Disassociate the `other` record from the `me` record.""" # Get the endpoint for foreign records within this object. url = self.endpoint + '%d/%s/' % (me, url_fragment) # Attempt to determine whether the other record already is absent, for the "changed" moniker. r = client.get(url, params={'id': other}).json() if r['count'] == 0: return {'changed': False} # Send a request removing the foreign record from this one. r = client.post(url, data={'disassociate': True, 'id': other}) return {'changed': True}
def copy(self, pk=None, new_name=None, **kwargs): """Copy an object. Only the ID is used for the lookup. All provided fields are used to override the old data from the copied resource. =====API DOCS===== Copy an object. :param pk: Primary key of the resource object to be copied :param new_name: The new name to give the resource if deep copying via the API :type pk: int :param `**kwargs`: Keyword arguments of fields whose given value will override the original value. :returns: loaded JSON of the copied new resource object. :rtype: dict =====API DOCS===== """ orig = self.read(pk, fail_on_no_results=True, fail_on_multiple_results=True) orig = orig['results'][0] # Remove default values (anything where the value is None). self._pop_none(kwargs) newresource = copy(orig) newresource.pop('id') basename = newresource['name'].split('@', 1)[0].strip() # Modify data to fit the call pattern of the tower-cli method for field in self.fields: if field.multiple and field.name in newresource: newresource[field.name] = (newresource.get(field.name),) if new_name is None: # copy client-side, the old mechanism newresource['name'] = "%s @ %s" % (basename, time.strftime('%X')) newresource.update(kwargs) return self.write(create_on_missing=True, fail_on_found=True, **newresource) else: # copy server-side, the new mechanism if kwargs: raise exc.TowerCLIError('Cannot override {} and also use --new-name.'.format(kwargs.keys())) copy_endpoint = '{}/{}/copy/'.format(self.endpoint.strip('/'), pk) return client.post(copy_endpoint, data={'name': new_name}).json()
def modify(self, pk=None, create_on_missing=False, **kwargs): """Modify an already existing object. Fields in the resource's `identity` tuple can be used in lieu of a primary key for a lookup; in such a case, only other fields are written. To modify unique fields, you must use the primary key for the lookup. =====API DOCS===== Modify an already existing object. :param pk: Primary key of the resource to be modified. :type pk: int :param create_on_missing: Flag that if set, a new object is created if ``pk`` is not set and objects matching the appropriate unique criteria is not found. :type create_on_missing: bool :param `**kwargs`: Keyword arguments which, all together, will be used as PATCH body to modify the resource object. if ``pk`` is not set, key-value pairs of ``**kwargs`` which are also in resource's identity will be used to lookup existing reosource. :returns: A dictionary combining the JSON output of the modified resource, as well as two extra fields: "changed", a flag indicating if the resource is successfully updated; "id", an integer which is the primary key of the updated object. :rtype: dict =====API DOCS===== """ return self.write(pk, create_on_missing=create_on_missing, force_on_exists=True, **kwargs)
def last_job_data(self, pk=None, **kwargs): """ Internal utility function for Unified Job Templates. Returns data about the last job run off of that UJT """ ujt = self.get(pk, include_debug_header=True, **kwargs) # Determine the appropriate inventory source update. if 'current_update' in ujt['related']: debug.log('A current job; retrieving it.', header='details') return client.get(ujt['related']['current_update'][7:]).json() elif ujt['related'].get('last_update', None): debug.log('No current job or update exists; retrieving the most recent.', header='details') return client.get(ujt['related']['last_update'][7:]).json() else: raise exc.NotFound('No related jobs or updates exist.')
def lookup_stdout(self, pk=None, start_line=None, end_line=None, full=True): """ Internal utility function to return standard out. Requires the pk of a unified job. """ stdout_url = '%s%s/stdout/' % (self.unified_job_type, pk) payload = {'format': 'json', 'content_encoding': 'base64', 'content_format': 'ansi'} if start_line: payload['start_line'] = start_line if end_line: payload['end_line'] = end_line debug.log('Requesting a copy of job standard output', header='details') resp = client.get(stdout_url, params=payload).json() content = b64decode(resp['content']) return content.decode('utf-8', 'replace')
def stdout(self, pk, start_line=None, end_line=None, outfile=sys.stdout, **kwargs): """ Print out the standard out of a unified job to the command line or output file. For Projects, print the standard out of most recent update. For Inventory Sources, print standard out of most recent sync. For Jobs, print the job's standard out. For Workflow Jobs, print a status table of its jobs. =====API DOCS===== Print out the standard out of a unified job to the command line or output file. For Projects, print the standard out of most recent update. For Inventory Sources, print standard out of most recent sync. For Jobs, print the job's standard out. For Workflow Jobs, print a status table of its jobs. :param pk: Primary key of the job resource object to be monitored. :type pk: int :param start_line: Line at which to start printing job output :param end_line: Line at which to end printing job output :param outfile: Alternative file than stdout to write job stdout to. :type outfile: file :param `**kwargs`: Keyword arguments used to look up job resource object to monitor if ``pk`` is not provided. :returns: A dictionary containing changed=False :rtype: dict =====API DOCS===== """ # resource is Unified Job Template if self.unified_job_type != self.endpoint: unified_job = self.last_job_data(pk, **kwargs) pk = unified_job['id'] # resource is Unified Job, but pk not given elif not pk: unified_job = self.get(**kwargs) pk = unified_job['id'] content = self.lookup_stdout(pk, start_line, end_line) opened = False if isinstance(outfile, six.string_types): outfile = open(outfile, 'w') opened = True if len(content) > 0: click.echo(content, nl=1, file=outfile) if opened: outfile.close() return {"changed": False}
def monitor(self, pk, parent_pk=None, timeout=None, interval=0.5, outfile=sys.stdout, **kwargs): """ Stream the standard output from a job, project update, or inventory udpate. =====API DOCS===== Stream the standard output from a job run to stdout. :param pk: Primary key of the job resource object to be monitored. :type pk: int :param parent_pk: Primary key of the unified job template resource object whose latest job run will be monitored if ``pk`` is not set. :type parent_pk: int :param timeout: Number in seconds after which this method will time out. :type timeout: float :param interval: Polling interval to refresh content from Tower. :type interval: float :param outfile: Alternative file than stdout to write job stdout to. :type outfile: file :param `**kwargs`: Keyword arguments used to look up job resource object to monitor if ``pk`` is not provided. :returns: A dictionary combining the JSON output of the finished job resource object, as well as two extra fields: "changed", a flag indicating if the job resource object is finished as expected; "id", an integer which is the primary key of the job resource object being monitored. :rtype: dict :raises tower_cli.exceptions.Timeout: When monitor time reaches time out. :raises tower_cli.exceptions.JobFailure: When the job being monitored runs into failure. =====API DOCS===== """ # If we do not have the unified job info, infer it from parent if pk is None: pk = self.last_job_data(parent_pk, **kwargs)['id'] job_endpoint = '%s%s/' % (self.unified_job_type, pk) # Pause until job is in running state self.wait(pk, exit_on=['running', 'successful'], outfile=outfile) # Loop initialization start = time.time() start_line = 0 result = client.get(job_endpoint).json() click.echo('\033[0;91m------Starting Standard Out Stream------\033[0m', nl=2, file=outfile) # Poll the Ansible Tower instance for status and content, and print standard out to the out file while not result['failed'] and result['status'] != 'successful': result = client.get(job_endpoint).json() # Put the process to sleep briefly. time.sleep(interval) # Make request to get standard out content = self.lookup_stdout(pk, start_line, full=False) # In the first moments of running the job, the standard out # may not be available yet if not content.startswith("Waiting for results"): line_count = len(content.splitlines()) start_line += line_count click.echo(content, nl=0, file=outfile) if timeout and time.time() - start > timeout: raise exc.Timeout('Monitoring aborted due to timeout.') # Special final line for closure with workflow jobs if self.endpoint == '/workflow_jobs/': click.echo(self.lookup_stdout(pk, start_line, full=True), nl=1) click.echo('\033[0;91m------End of Standard Out Stream--------\033[0m', nl=2, file=outfile) if result['failed']: raise exc.JobFailure('Job failed.') # Return the job ID and other response data answer = OrderedDict((('changed', True), ('id', pk))) answer.update(result) # Make sure to return ID of resource and not update number relevant for project creation and update if parent_pk: answer['id'] = parent_pk else: answer['id'] = pk return answer
def wait(self, pk, parent_pk=None, min_interval=1, max_interval=30, timeout=None, outfile=sys.stdout, exit_on=['successful'], **kwargs): """ Wait for a running job to finish. Blocks further input until the job completes (whether successfully or unsuccessfully) and a final status can be given. =====API DOCS===== Wait for a job resource object to enter certain status. :param pk: Primary key of the job resource object to wait. :type pk: int :param parent_pk: Primary key of the unified job template resource object whose latest job run will be waited if ``pk`` is not set. :type parent_pk: int :param timeout: Number in seconds after which this method will time out. :type timeout: float :param min_interval: Minimum polling interval to request an update from Tower. :type min_interval: float :param max_interval: Maximum polling interval to request an update from Tower. :type max_interval: float :param outfile: Alternative file than stdout to write job status updates on. :type outfile: file :param exit_on: Job resource object statuses to wait on. :type exit_on: array :param `**kwargs`: Keyword arguments used to look up job resource object to wait if ``pk`` is not provided. :returns: A dictionary combining the JSON output of the status-changed job resource object, as well as two extra fields: "changed", a flag indicating if the job resource object is status-changed as expected; "id", an integer which is the primary key of the job resource object being status-changed. :rtype: dict :raises tower_cli.exceptions.Timeout: When wait time reaches time out. :raises tower_cli.exceptions.JobFailure: When the job being waited on runs into failure. =====API DOCS===== """ # If we do not have the unified job info, infer it from parent if pk is None: pk = self.last_job_data(parent_pk, **kwargs)['id'] job_endpoint = '%s%s/' % (self.unified_job_type, pk) dots = itertools.cycle([0, 1, 2, 3]) longest_string = 0 interval = min_interval start = time.time() # Poll the Ansible Tower instance for status, and print the status to the outfile (usually standard out). # # Note that this is one of the few places where we use `secho` even though we're in a function that might # theoretically be imported and run in Python. This seems fine; outfile can be set to /dev/null and very # much the normal use for this method should be CLI monitoring. result = client.get(job_endpoint).json() last_poll = time.time() timeout_check = 0 while result['status'] not in exit_on: # If the job has failed, we want to raise an Exception for that so we get a non-zero response. if result['failed']: if is_tty(outfile) and not settings.verbose: secho('\r' + ' ' * longest_string + '\n', file=outfile) raise exc.JobFailure('Job failed.') # Sanity check: Have we officially timed out? # The timeout check is incremented below, so this is checking to see if we were timed out as of # the previous iteration. If we are timed out, abort. if timeout and timeout_check - start > timeout: raise exc.Timeout('Monitoring aborted due to timeout.') # If the outfile is a TTY, print the current status. output = '\rCurrent status: %s%s' % (result['status'], '.' * next(dots)) if longest_string > len(output): output += ' ' * (longest_string - len(output)) else: longest_string = len(output) if is_tty(outfile) and not settings.verbose: secho(output, nl=False, file=outfile) # Put the process to sleep briefly. time.sleep(0.2) # Sanity check: Have we reached our timeout? # If we're about to time out, then we need to ensure that we do one last check. # # Note that the actual timeout will be performed at the start of the **next** iteration, # so there's a chance for the job's completion to be noted first. timeout_check = time.time() if timeout and timeout_check - start > timeout: last_poll -= interval # If enough time has elapsed, ask the server for a new status. # # Note that this doesn't actually do a status check every single time; we want the "spinner" to # spin even if we're not actively doing a check. # # So, what happens is that we are "counting down" (actually up) to the next time that we intend # to do a check, and once that time hits, we do the status check as part of the normal cycle. if time.time() - last_poll > interval: result = client.get(job_endpoint).json() last_poll = time.time() interval = min(interval * 1.5, max_interval) # If the outfile is *not* a TTY, print a status update when and only when we make an actual # check to job status. if not is_tty(outfile) or settings.verbose: click.echo('Current status: %s' % result['status'], file=outfile) # Wipe out the previous output if is_tty(outfile) and not settings.verbose: secho('\r' + ' ' * longest_string, file=outfile, nl=False) secho('\r', file=outfile, nl=False) # Return the job ID and other response data answer = OrderedDict((('changed', True), ('id', pk))) answer.update(result) # Make sure to return ID of resource and not update number relevant for project creation and update if parent_pk: answer['id'] = parent_pk else: answer['id'] = pk return answer
def status(self, pk=None, detail=False, **kwargs): """Print the current job status. This is used to check a running job. You can look up the job with the same parameters used for a get request. =====API DOCS===== Retrieve the current job status. :param pk: Primary key of the resource to retrieve status from. :type pk: int :param detail: Flag that if set, return the full JSON of the job resource rather than a status summary. :type detail: bool :param `**kwargs`: Keyword arguments used to look up resource object to retrieve status from if ``pk`` is not provided. :returns: full loaded JSON of the specified unified job if ``detail`` flag is on; trimed JSON containing only "elapsed", "failed" and "status" fields of the unified job if ``detail`` flag is off. :rtype: dict =====API DOCS===== """ # Remove default values (anything where the value is None). self._pop_none(kwargs) # Search for the record if pk not given if not pk: job = self.get(include_debug_header=True, **kwargs) # Get the job from Ansible Tower if pk given else: debug.log('Asking for job status.', header='details') finished_endpoint = '%s%s/' % (self.endpoint, pk) job = client.get(finished_endpoint).json() # In most cases, we probably only want to know the status of the job and the amount of time elapsed. # However, if we were asked for verbose information, provide it. if detail: return job # Print just the information we need. return { 'elapsed': job['elapsed'], 'failed': job['failed'], 'status': job['status'], }
def cancel(self, pk=None, fail_if_not_running=False, **kwargs): """Cancel a currently running job. Fails with a non-zero exit status if the job cannot be canceled. You must provide either a pk or parameters in the job's identity. =====API DOCS===== Cancel a currently running job. :param pk: Primary key of the job resource to restart. :type pk: int :param fail_if_not_running: Flag that if set, raise exception if the job resource cannot be canceled. :type fail_if_not_running: bool :param `**kwargs`: Keyword arguments used to look up job resource object to restart if ``pk`` is not provided. :returns: A dictionary of two keys: "status", which is "canceled", and "changed", which indicates if the job resource has been successfully canceled. :rtype: dict :raises tower_cli.exceptions.TowerCLIError: When the job resource cannot be canceled and ``fail_if_not_running`` flag is on. =====API DOCS===== """ # Search for the record if pk not given if not pk: existing_data = self.get(**kwargs) pk = existing_data['id'] cancel_endpoint = '%s%s/cancel/' % (self.endpoint, pk) # Attempt to cancel the job. try: client.post(cancel_endpoint) changed = True except exc.MethodNotAllowed: changed = False if fail_if_not_running: raise exc.TowerCLIError('Job not running.') # Return a success. return {'status': 'canceled', 'changed': changed}
def relaunch(self, pk=None, **kwargs): """Relaunch a stopped job. Fails with a non-zero exit status if the job cannot be relaunched. You must provide either a pk or parameters in the job's identity. =====API DOCS===== Relaunch a stopped job resource. :param pk: Primary key of the job resource to relaunch. :type pk: int :param `**kwargs`: Keyword arguments used to look up job resource object to relaunch if ``pk`` is not provided. :returns: A dictionary combining the JSON output of the relaunched job resource object, as well as an extra field "changed", a flag indicating if the job resource object is status-changed as expected. :rtype: dict =====API DOCS===== """ # Search for the record if pk not given if not pk: existing_data = self.get(**kwargs) pk = existing_data['id'] relaunch_endpoint = '%s%s/relaunch/' % (self.endpoint, pk) data = {} # Attempt to relaunch the job. answer = {} try: result = client.post(relaunch_endpoint, data=data).json() if 'id' in result: answer.update(result) answer['changed'] = True except exc.MethodNotAllowed: answer['changed'] = False # Return the answer. return answer
def survey(self, pk=None, **kwargs): """Get the survey_spec for the job template. To write a survey, use the modify command with the --survey-spec parameter. =====API DOCS===== Get the survey specification of a resource object. :param pk: Primary key of the resource to retrieve survey from. Tower CLI will only attempt to read *that* object if ``pk`` is provided (not ``None``). :type pk: int :param `**kwargs`: Keyword arguments used to look up resource object to retrieve survey if ``pk`` is not provided. :returns: loaded JSON of the retrieved survey specification of the resource object. :rtype: dict =====API DOCS===== """ job_template = self.get(pk=pk, **kwargs) if settings.format == 'human': settings.format = 'json' return client.get(self._survey_endpoint(job_template['id'])).json()
def batch_update(self, pk=None, **kwargs): """Update all related inventory sources of the given inventory. Note global option --format is not available here, as the output would always be JSON-formatted. =====API DOCS===== Update all related inventory sources of the given inventory. :param pk: Primary key of the given inventory. :type pk: int :param `**kwargs`: Keyword arguments list of available fields used for searching resource objects. :returns: A JSON object of update status of the given inventory. :rtype: dict =====API DOCS===== """ res = self.get(pk=pk, **kwargs) url = self.endpoint + '%d/%s/' % (res['id'], 'update_inventory_sources') return client.post(url, data={}).json()
def read(self, *args, **kwargs): ''' Do extra processing so we can display the actor field as a top-level field ''' if 'actor' in kwargs: kwargs['actor'] = kwargs.pop('actor') r = super(Resource, self).read(*args, **kwargs) if 'results' in r: for d in r['results']: self._promote_actor(d) else: self._promote_actor(d) return r
def log(s, header='', file=sys.stderr, nl=1, **kwargs): """Log the given output to stderr if and only if we are in verbose mode. If we are not in verbose mode, this is a no-op. """ # Sanity check: If we are not in verbose mode, this is a no-op. if not settings.verbose: return # Construct multi-line string to stderr if header is provided. if header: word_arr = s.split(' ') multi = [] word_arr.insert(0, '%s:' % header.upper()) i = 0 while i < len(word_arr): to_add = ['***'] count = 3 while count <= 79: count += len(word_arr[i]) + 1 if count <= 79: to_add.append(word_arr[i]) i += 1 if i == len(word_arr): break # Handle corner case of extra-long word longer than 75 characters. if len(to_add) == 1: to_add.append(word_arr[i]) i += 1 if i != len(word_arr): count -= len(word_arr[i]) + 1 to_add.append('*' * (78 - count)) multi.append(' '.join(to_add)) s = '\n'.join(multi) lines = len(multi) else: lines = 1 # If `nl` is an int greater than the number of rows of a message, # add the appropriate newlines to the output. if isinstance(nl, int) and nl > lines: s += '\n' * (nl - lines) # Output to stderr. return secho(s, file=file, **kwargs)
def configure_model(self, attrs, field_name): ''' Hook for ResourceMeta class to call when initializing model class. Saves fields obtained from resource class backlinks ''' self.relationship = field_name self._set_method_names(relationship=field_name) if self.res_name is None: self.res_name = grammar.singularize(attrs.get('endpoint', 'unknown').strip('/'))
def _produce_raw_method(self): ''' Returns a callable which becomes the associate or disassociate method for the related field. Method can be overridden to add additional functionality, but `_produce_method` may also need to be subclassed to decorate it appropriately. ''' def method(res_self, **kwargs): obj_pk = kwargs.get(method._res_name) other_obj_pk = kwargs.get(method._other_name) internal_method = getattr(res_self, method._internal_name) return internal_method(method._relationship, obj_pk, other_obj_pk) return method
def create(self, fail_on_found=False, force_on_exists=False, **kwargs): """Create a new label. There are two types of label creation: isolatedly creating a new label and creating a new label under a job template. Here the two types are discriminated by whether to provide --job-template option. Fields in the resource's `identity` tuple are used for a lookup; if a match is found, then no-op (unless `force_on_exists` is set) but do not fail (unless `fail_on_found` is set). =====API DOCS===== Create a label. :param job_template: Primary key or name of the job template for the created label to associate to. :type job_template: str :param fail_on_found: Flag that if set, the operation fails if an object matching the unique criteria already exists. :type fail_on_found: bool :param force_on_exists: Flag that if set, then if a match is found on unique fields, other fields will be updated to the provided values.; If unset, a match causes the request to be a no-op. :type force_on_exists: bool :param `**kwargs`: Keyword arguments which, all together, will be used as POST body to create the resource object. :returns: A dictionary combining the JSON output of the created resource, as well as two extra fields: "changed", a flag indicating if the resource is created successfully; "id", an integer which is the primary key of the created object. :rtype: dict :raises tower_cli.exceptions.TowerCLIError: When the label already exists and ``fail_on_found`` flag is on. =====API DOCS===== """ jt_id = kwargs.pop('job_template', None) old_endpoint = self.endpoint if jt_id is not None: jt = get_resource('job_template') jt.get(pk=jt_id) try: label_id = self.get(name=kwargs.get('name', None), organization=kwargs.get('organization', None))['id'] except exc.NotFound: pass else: if fail_on_found: raise exc.TowerCLIError('Label already exists and fail-on-found is switched on. Please use' ' "associate_label" method of job_template instead.') else: debug.log('Label already exists, associating with job template.', header='details') return jt.associate_label(job_template=jt_id, label=label_id) self.endpoint = '/job_templates/%d/labels/' % jt_id result = super(Resource, self).create(fail_on_found=fail_on_found, force_on_exists=force_on_exists, **kwargs) self.endpoint = old_endpoint return result
def version(): """Display full version information.""" # Print out the current version of Tower CLI. click.echo('Tower CLI %s' % __version__) # Print out the current API version of the current code base. click.echo('API %s' % CUR_API_VERSION) # Attempt to connect to the Ansible Tower server. # If we succeed, print a version; if not, generate a failure. try: r = client.get('/config/') except RequestException as ex: raise exc.TowerCLIError('Could not connect to Ansible Tower.\n%s' % six.text_type(ex)) config = r.json() license = config.get('license_info', {}).get('license_type', 'open') if license == 'open': server_type = 'AWX' else: server_type = 'Ansible Tower' click.echo('%s %s' % (server_type, config['version'])) # Print out Ansible version of server click.echo('Ansible %s' % config['ansible_version'])
def _echo_setting(key): """Echo a setting to the CLI.""" value = getattr(settings, key) secho('%s: ' % key, fg='magenta', bold=True, nl=False) secho( six.text_type(value), bold=True, fg='white' if isinstance(value, six.text_type) else 'cyan', )
def config(key=None, value=None, scope='user', global_=False, unset=False): """Read or write tower-cli configuration. `tower config` saves the given setting to the appropriate Tower CLI; either the user's ~/.tower_cli.cfg file, or the /etc/tower/tower_cli.cfg file if --global is used. Writing to /etc/tower/tower_cli.cfg is likely to require heightened permissions (in other words, sudo). """ # If the old-style `global_` option is set, issue a deprecation notice. if global_: scope = 'global' warnings.warn('The `--global` option is deprecated and will be ' 'removed. Use `--scope=global` to get the same effect.', DeprecationWarning) # If no key was provided, print out the current configuration # in play. if not key: seen = set() parser_desc = { 'runtime': 'Runtime options.', 'environment': 'Options from environment variables.', 'local': 'Local options (set with `tower-cli config ' '--scope=local`; stored in .tower_cli.cfg of this ' 'directory or a parent)', 'user': 'User options (set with `tower-cli config`; stored in ' '~/.tower_cli.cfg).', 'global': 'Global options (set with `tower-cli config ' '--scope=global`, stored in /etc/tower/tower_cli.cfg).', 'defaults': 'Defaults.', } # Iterate over each parser (English: location we can get settings from) # and print any settings that we haven't already seen. # # We iterate over settings from highest precedence to lowest, so any # seen settings are overridden by the version we iterated over already. click.echo('') for name, parser in zip(settings._parser_names, settings._parsers): # Determine if we're going to see any options in this # parser that get echoed. will_echo = False for option in parser.options('general'): if option in seen: continue will_echo = True # Print a segment header if will_echo: secho('# %s' % parser_desc[name], fg='green', bold=True) # Iterate over each option in the parser and, if we haven't # already seen an option at higher precedence, print it. for option in parser.options('general'): if option in seen: continue _echo_setting(option) seen.add(option) # Print a nice newline, for formatting. if will_echo: click.echo('') return # Sanity check: Is this a valid configuration option? If it's not # a key we recognize, abort. if not hasattr(settings, key): raise exc.TowerCLIError('Invalid configuration option "%s".' % key) # Sanity check: The combination of a value and --unset makes no # sense. if value and unset: raise exc.UsageError('Cannot provide both a value and --unset.') # If a key was provided but no value was provided, then just # print the current value for that key. if key and not value and not unset: _echo_setting(key) return # Okay, so we're *writing* a key. Let's do this. # First, we need the appropriate file. filename = os.path.expanduser('~/.tower_cli.cfg') if scope == 'global': if not os.path.isdir('/etc/tower/'): raise exc.TowerCLIError('/etc/tower/ does not exist, and this ' 'command cowardly declines to create it.') filename = '/etc/tower/tower_cli.cfg' elif scope == 'local': filename = '.tower_cli.cfg' # Read in the appropriate config file, write this value, and save # the result back to the file. parser = Parser() parser.add_section('general') parser.read(filename) if unset: parser.remove_option('general', key) else: parser.set('general', key, value) with open(filename, 'w') as config_file: parser.write(config_file) # Give rw permissions to user only fix for issue number 48 try: os.chmod(filename, stat.S_IRUSR | stat.S_IWUSR) except Exception as e: warnings.warn( 'Unable to set permissions on {0} - {1} '.format(filename, e), UserWarning ) click.echo('Configuration updated successfully.')