def preload_for_predicates(literal):
"""
A custom caching strategy.
The idea is that a query for a specific subject with some predicate is often
done in the context of a query that retrieves all subjects with that
predicate. Alternatively, the literal is passed through to the caching
mechanism.
"""
if literal.terms[1].is_const():
new_terms = [datalog.make_fresh_var(), literal.terms[1], literal.terms[2]]
return datalog.Literal(literal.pred, new_terms)
else:
return literal
def make_term(token):
"""
Helper to convert a token into either a variable or constant term.
Constant terms are prefixed with a type indicator to ensure intuitive
handling of strings, names and numbers.
"""
kind, spelling, location = token
# See if this is a variable
if kind == NAME and spelling[:1].isupper():
return datalog.Variable(spelling)
# See if this is a do not care variable
if kind == NAME and spelling == '_':
return datalog.make_fresh_var()
if kind == NAME:
return datalog.Constant(spelling)
elif kind == STRING:
return datalog.Constant(str(spelling), kind='string', data=spelling)
elif kind == NUMBER:
return datalog.Constant(str(spelling), kind='number', data=spelling)
def transform_literal(literal):
new_terms = [datalog.make_fresh_var() if i in consider_free else t for i,t in enumerate(literal.terms)]
return datalog.Literal(literal.pred, new_terms)
def transform_literal(literal):
new_terms = [datalog.make_fresh_var() for t in literal.terms]
return datalog.Literal(literal.pred, new_terms)
