neurolang.probabilistic.cplogic.program module¶

class neurolang.probabilistic.cplogic.program.CPLogicMixin(*args, **kwargs)¶

Bases: PatternWalker

Datalog extended with probabilistic facts semantics from ProbLog.

It adds a probabilistic database which is a set of probabilistic facts.

Probabilistic facts are stored in the symbol table of the program such that the key in the symbol table is the symbol of the predicate of the probabilsitic fact and the value is the probabilistic fact itself.

Attributes:

patterns: Property holding an iterator of triplets (pattern, guard, action).
pchoice_pred_symbs
pfact_pred_symbs
predicate_symbols
probabilistic_predicate_symbols

Methods

`add_probabilistic_choice_from_tuples`(symbol, ...)	Add a probabilistic choice from a predicate symbol and a set of tuples where the first element is the probability label attached to a head predicate and the other elements are the constant terms of the head predicate.
`add_probabilistic_facts_from_tuples`(symbol, ...)	Add probabilistic facts from tuples whose first element contains the probability label attached to that tuple.
`match`(expression)	Find the action for a given expression by going through the `patterns`.
`pattern_match`(pattern, expression)	Return `True` if `pattern` matches `expression`.
`pchoice_pred_symb_set_symb`
`pfact_pred_symb_set_symb`
`probabilistic_choices`()	Return probabilistic choices of the symbol table.
`probabilistic_facts`()	Return probabilistic facts of the symbol table.
`query_based_probabilistic_choice`(implication)	Construct probabilistic choices from deterministic queries.
`query_based_probabilistic_fact`(implication)	Construct probabilistic facts from deterministic queries.
`query_based_probabilistic_predicate`(implication)	Left for backward compatibility.
`type`

extensional_database
marg_implication
pattern_match_expression
pattern_match_expression_parameters
pattern_match_expression_tuple
pattern_match_tuple
probabilistic_fact
union_with_probabilistic_facts
walk
within_language_marg_query
within_language_prob_queries
within_language_succ_query

add_probabilistic_choice_from_tuples(symbol, iterable)¶

Add a probabilistic choice from a predicate symbol and a set of tuples where the first element is the probability label attached to a head predicate and the other elements are the constant terms of the head predicate.

Examples

The following

>>> P = Symbol('P')
>>> tuples = {(0.2, 'a'), (0.8, 'b')}
>>> program.add_probabilistic_choice_from_tuples(P, tuples)

adds the probabilistic choice

P(a) : 0.2 v P(b) : 0.8 <- T

to the program.

add_probabilistic_facts_from_tuples(symbol, iterable)¶

Add probabilistic facts from tuples whose first element contains the probability label attached to that tuple.

Examples

The following

>>> P = Symbol('P')
>>> tuples = {(0.9, 'a'), (0.8, 'b')}
>>> program.add_probabilistic_facts_from_tuples(P, tuples)

adds the probabilistic facts

P(a) : 0.9 <- T P(b) : 0.8 <- T

to the program.

extensional_database()¶

marg_implication(implication)¶

pchoice_pred_symb_set_symb = S{__pchoice_pred_symb_set_symb__: Unknown}¶

property pchoice_pred_symbs¶

pfact_pred_symb_set_symb = S{__pfact_pred_symb_set_symb__: Unknown}¶

property pfact_pred_symbs¶

property predicate_symbols¶

probabilistic_choices()¶: Return probabilistic choices of the symbol table.

probabilistic_fact(expression)¶

probabilistic_facts()¶: Return probabilistic facts of the symbol table.

property probabilistic_predicate_symbols¶

protected_keywords = {'PROB'}¶

query_based_probabilistic_choice(implication)¶

Construct probabilistic choices from deterministic queries.

This extends the syntax with rules such as

Or_x P(x) : f(x) :- Q(x)

where x is a set of variables, f(x) is an arithmetic expression yielding a probability between [0, 1] that may use built-ins, and where Q(x) is a conjunction of predicates. The sum of f(x) for all x must be lower than 1.

Only deterministic antecedents are allowed. Declarativity makes it impossible to enforce that at declaration time. Thus, if a query-based probabilistic fact has a dependency on a probabilistic predicate, this will be discovered at query-resolution time, after the program has been fully declared.

query_based_probabilistic_fact(implication)¶

Construct probabilistic facts from deterministic queries.

This extends the syntax with rules such as

P(x) : f(x) :- Q(x)

where x is a set of variables, f(x) is an arithmetic expression yielding a probability between [0, 1] that may use built-ins, and where Q(x) is a conjunction of predicates.

query_based_probabilistic_predicate(implication)¶: Left for backward compatibility. Should be referred to the ProbabilisticFact case

type = typing.Any¶

union_with_probabilistic_facts(code)¶

within_language_marg_query(implication)¶

within_language_prob_queries()¶

within_language_succ_query(implication)¶

class neurolang.probabilistic.cplogic.program.CPLogicProgram(*args, **kwargs)¶

Bases: CPLogicMixin, DatalogProgramNegationMixin, DatalogProgram, ExpressionWalker

Attributes:

included_constants
included_functions
patterns: Property holding an iterator of triplets (pattern, guard, action).
pchoice_pred_symbs
pfact_pred_symbs
predicate_symbols
probabilistic_predicate_symbols

Methods

`add_probabilistic_choice_from_tuples`(symbol, ...)	Add a probabilistic choice from a predicate symbol and a set of tuples where the first element is the probability label attached to a head predicate and the other elements are the constant terms of the head predicate.
`add_probabilistic_facts_from_tuples`(symbol, ...)	Add probabilistic facts from tuples whose first element contains the probability label attached to that tuple.
`constant_equals`
`infer_iterable_type`(iterable)	Infer the type of iterable elements without modifying the iterable.
`match`(expression)	Find the action for a given expression by going through the `patterns`.
`pattern_match`(pattern, expression)	Return `True` if `pattern` matches `expression`.
`pchoice_pred_symb_set_symb`
`pfact_pred_symb_set_symb`
`probabilistic_choices`()	Return probabilistic choices of the symbol table.
`probabilistic_facts`()	Return probabilistic facts of the symbol table.
`query_based_probabilistic_choice`(implication)	Construct probabilistic choices from deterministic queries.
`query_based_probabilistic_fact`(implication)	Construct probabilistic facts from deterministic queries.
`query_based_probabilistic_predicate`(implication)	Left for backward compatibility.
`type`
`union_of_cq`(expression)	This pattern is here to avoid processing the full program when one rule is rewritten.

add_extensional_predicate_from_tuples
add_included_constants_and_functions_to_symbol_table
builtins
extensional_database
fact
intensional_database
marg_implication
negation_constant
negative_fact
new_set
pattern_match_expression
pattern_match_expression_parameters
pattern_match_expression_tuple
pattern_match_tuple
pop_scope
predicate_terms
probabilistic_fact
process_expression
process_iterable_argument
push_scope
statement_intensional
symbol
union_with_probabilistic_facts
walk
within_language_marg_query
within_language_prob_queries
within_language_succ_query

protected_keywords = {'PROB', '__pchoice_pred_symb_set_symb__', '__pfact_pred_symb_set_symb__'}¶

type = typing.Any¶