def evaluate_explicit_from_fsdd(self, custom_semiring=None):
try:
parser = DefaultPrologParser(ExtendedPrologFactory())
lf = PrologFile(filename, parser=parser)
kc = _ForwardSDD.create_from(lf) # type: _ForwardSDD
kc = kc.to_explicit_encoding()
if custom_semiring is not None:
semiring = custom_semiring # forces the custom semiring code.
elif logspace:
semiring = SemiringLogProbability()
else:
semiring = SemiringProbability()
computed = kc.evaluate(semiring=semiring)
computed = {str(k): v for k, v in computed.items()}
except Exception as err:
#print("exception %s" % err)
e = err
computed = None
if computed is None:
self.assertEqual(correct, type(e).__name__)
else:
self.assertIsInstance(correct, dict)
self.assertSequenceEqual(correct, computed)
for query in correct:
self.assertAlmostEqual(correct[query],
computed[query],
msg=query)
def run_test_with_query_instance():
from problog.program import PrologString
from problog.engine import DefaultEngine
from problog.logic import Term, Var
p = PrologString("""
coin(c1). coin(c2).
0.4::heads(C); 0.6::tails(C) :- coin(C).
win :- heads(C).
""")
from .formula_wrapper import FormulaWrapper
s_qs, s_evs = ([Term("win")], [(Term("heads",
Term("c1")), False)]) # For now
engine = DefaultEngine()
probsr = SemiringProbability()
fw = FormulaWrapper(engine.prepare(p))
qobj = AMCQuery(s_qs, s_evs, fw, target_class=DDNNF, semiring=probsr)
qobj.ground(engine)
result, ground_evidence = qobj.evaluate(engine)
print("evidence: ", ground_evidence)
for r in result:
print(r)
print("---")
def run_tests_with_static_methods():
from problog.program import PrologString
from problog.engine import DefaultEngine
from problog.logic import Term, Var
p = PrologString("""
coin(c1). coin(c2).
0.4::heads(C); 0.6::tails(C) :- coin(C).
win :- heads(C).
""")
qs, evs = ([Term("win")], [(Term("heads", Term("c1")), False)]) # For now
engine = DefaultEngine()
db = engine.prepare(p)
labels = (LogicFormula.LABEL_QUERY, LogicFormula.LABEL_EVIDENCE_POS,
LogicFormula.LABEL_EVIDENCE_NEG)
lf = LogicFormula()
lf = AMCQuery.ground_query_evidence(engine, db, qs, evs, lf, labels)
circuit = AMCQuery.compile_to_circuit(lf, "ddnnf")
prob_sr = SemiringProbability()
results, ground_evidence = AMCQuery.evaluate_circuit(
circuit, labels, prob_sr)
print("evidence: ", ground_evidence)
for r in results:
print(r)
print("---")
def mpe_maxsat(dag, verbose=0, solver=None):
if dag.queries():
print('%% WARNING: ignoring queries in file', file=sys.stderr)
dag.clear_queries()
logger = init_logger(verbose)
logger.info('Ground program size: %s' % len(dag))
cnf = CNF.createFrom(dag)
for qn, qi in cnf.evidence():
if not cnf.is_true(qi):
cnf.add_constraint(TrueConstraint(qi))
queries = list(cnf.labeled())
logger.info('CNF size: %s' % cnf.clausecount)
if not cnf.is_trivial():
solver = get_solver(solver)
with Timer('Solving'):
result = frozenset(solver.evaluate(cnf))
weights = cnf.extract_weights(SemiringProbability())
output_facts = None
prob = 1.0
if result is not None:
output_facts = []
if queries:
for qn, qi, ql in queries:
if qi in result:
output_facts.append(qn)
elif -qi in result:
output_facts.append(-qn)
for i, n, t in dag:
if t == 'atom':
if i in result:
if not queries:
output_facts.append(n.name)
prob *= weights[i][0]
elif -i in result:
if not queries:
output_facts.append(-n.name)
prob *= weights[i][1]
else:
prob = 1.0
output_facts = []
return prob, output_facts
def main(inputfile, semiring, **kwdargs):
pl = PrologFile(inputfile)
if semiring == 'prob':
sm = SemiringProbability()
elif semiring == 'mpe':
sm = SemiringMPE()
elif semiring == 'mpe_state':
sm = SemiringMPEState()
result = solve(pl, semiring=sm)
for k, v in result.items():
print ('%s: %s' % (k, v))
def ad_atom_duplicate(self, eval_name=None):
"""
This test must pickup the case where during the transformation, additional _extra atoms are created because
add_atom(..., cr_extra=True) is used instead of cr_extra=False.
"""
program = """
0.2::a ; 0.8::b.
query(a).
query(b).
"""
pl = PrologString(program)
lf = LogicFormula.create_from(pl, label_all=True, avoid_name_clash=True)
semiring = SemiringProbability()
kc_class = get_evaluatable(name=eval_name, semiring=semiring)
kc = kc_class.create_from(lf) # type: LogicFormula
self.assertEqual(3, kc.atomcount)
def mpe_maxsat(dag, verbose=0, solver=None, minpe=False):
logger = init_logger(verbose)
logger.info("Ground program size: %s" % len(dag))
cnf = CNF.createFrom(dag, force_atoms=True)
for qn, qi in cnf.evidence():
if not cnf.is_true(qi):
cnf.add_constraint(TrueConstraint(qi))
queries = list(cnf.labeled())
logger.info("CNF size: %s" % cnf.clausecount)
if not cnf.is_trivial():
solver = get_solver(solver)
with Timer("Solving"):
result = frozenset(solver.evaluate(cnf, invert_weights=minpe))
weights = cnf.extract_weights(SemiringProbability())
output_facts = None
prob = 1.0
if result is not None:
output_facts = []
if queries:
for qn, qi, ql in queries:
if qi in result:
output_facts.append(qn)
elif -qi in result:
output_facts.append(-qn)
for i, n, t in dag:
if t == "atom":
if i in result:
if not queries:
output_facts.append(n.name)
prob *= weights[i][0]
elif -i in result:
if not queries:
output_facts.append(-n.name)
prob *= weights[i][1]
else:
prob = 1.0
output_facts = []
return prob, output_facts
def test(self) :
try :
parser = DefaultPrologParser(ExtendedPrologFactory())
sdd = SDD.createFrom(PrologFile(filename, parser=parser))
if logspace :
semiring = SemiringLogProbability()
else :
semiring = SemiringProbability()
computed = sdd.evaluate(semiring=semiring)
computed = { str(k) : v for k,v in computed.items() }
except Exception as err :
e = err
computed = None
if computed is None :
self.assertEqual(correct, type(e).__name__)
else :
self.assertIsInstance( correct, dict )
self.assertSequenceEqual(correct, computed)
for query in correct :
self.assertAlmostEqual(correct[query], computed[query], msg=query)
def create_query(query_type, queries, evidence, formula_wrapper, **kwargs):
if query_type == QueryFactory.QueryType.PROBABILITY:
return AMCQuery(queries,
evidence,
formula_wrapper,
semiring=SemiringProbability(),
**kwargs)
elif query_type == QueryFactory.QueryType.MPE:
return AMCQuery(queries,
evidence,
formula_wrapper,
semiring=SemiringMPEState(),
**kwargs)
elif query_type == QueryFactory.QueryType.MINPE:
return AMCQuery(queries,
evidence,
formula_wrapper,
semiring=SemiringMinPEState(),
**kwargs)
elif query_type == QueryFactory.QueryType.SAMPLE:
return SampleQuery(queries, evidence, formula_wrapper, **kwargs)
else:
raise NotImplementedError("query_type not yet supported: %s" %
str(query_type))
def __init__(self):
self._semiring = SemiringProbability()
class SemiringProbabilityCopy(Semiring):
"""Mocking SemiringProbability to force the 'custom semiring' code -> Must not extend SemiringProbability."""
def __init__(self):
self._semiring = SemiringProbability()
def one(self):
return self._semiring.one()
def zero(self):
return self._semiring.zero()
def is_one(self, value):
return self._semiring.is_one(value)
def is_zero(self, value):
return self._semiring.is_zero(value)
def plus(self, a, b):
return self._semiring.plus(a, b)
def times(self, a, b):
return self._semiring.times(a, b)
def negate(self, a):
return self._semiring.negate(a)
def normalize(self, a, z):
return self._semiring.normalize(a, z)
def value(self, a):
return self._semiring.value(a)
def is_dsp(self):
return self._semiring.is_dsp()
def in_domain(self, a):
return self._semiring.in_domain(a)
def __init__(self,
source,
examples,
max_iter=10000,
min_improv=1e-10,
verbose=0,
knowledge=None,
leakprob=None,
propagate_evidence=True,
normalize=False,
**extra):
"""
:param source: filename of file containing input model
:type source: str
:param examples: list of observed terms / value
:type examples: list[tuple(Term, bool)]
:param max_iter: maximum number of iterations to run
:type max_iter: int
:param min_improv: minimum improvement in log-likelihood for convergence detection
:type min_improv: float
:param verbose: verbosity level
:type verbose: int
:param knowledge: class to use for knowledge compilation
:type knowledge: class
:param leakprob: Add all true evidence atoms with the given probability
to avoid 'inconsistent evidence' errors. This also
allows to learn a program without constants and
retrieve the constants from the evidence file.
(default: None)
:type leakprob: float or None
:param extra: catch all for additional parameters (not used)
"""
SemiringProbability.__init__(self)
LogicProgram.__init__(self)
self.source = source
# The names of the atom for which we want to learn weights.
self.names = []
# The weights to learn.
# The initial weights are of type 'float'.
# When necessary they are replaced by a dictionary [t(arg1, arg2, ...) -> float]
# for weights of form t(SV, arg1, arg2, ...).
self._weights = []
self.examples = examples
self.leakprob = leakprob
self.leakprobatoms = None
self.propagate_evidence = propagate_evidence
self._compiled_examples = None
self.max_iter = max_iter
self.min_improv = min_improv
self.verbose = verbose
self.iteration = 0
if knowledge is None:
knowledge = get_evaluatable()
self.knowledge = knowledge
self.output_mode = False
self.extra = extra
self._enable_normalize = normalize
self._adatoms = []
def __init__(self, weights):
SemiringProbability.__init__(self)
self._weights = weights
