Source code for neuralogic.core.constructs.predicate

from typing import Sequence

from neuralogic.core.constructs.metadata import Metadata




class Predicate:
    __slots__ = "name", "arity", "hidden", "special"

    def __init__(self, name, arity, hidden=False, special=False):
        if name.startswith("_"):
            name = name[1:]
            hidden = True

        self.name = name
        self.arity = arity
        self.hidden = hidden
        self.special = special



    def set_arity(self, arity):
        if self.arity == arity:
            return self




    def to_str(self):
        if not self.special and not self.hidden:
            return self.name

        special = "@" if self.special else ""
        hidden = "*" if self.hidden else ""
        return f"{hidden}{special}{self.name}"


    def __str__(self) -> str:
        special = "@" if self.special else ""
        hidden = "*" if self.hidden else ""
        return f"{hidden}{special}{self.name}/{self.arity}"

    def __repr__(self) -> str:
        return self.__str__()

    def __or__(self, other) -> "PredicateMetadata":
        if isinstance(other, Sequence):
            other = Metadata.from_iterable(other)
        elif not isinstance(other, Metadata):
            raise NotImplementedError
        return PredicateMetadata(self, other)





class PredicateMetadata:
    __slots__ = "predicate", "metadata"

    def __init__(self, predicate: Predicate, metadata: Metadata):
        if metadata.aggregation is not None:
            raise ValueError(f"Cannot set 'aggregation' parameter on predicate ({predicate}) metadata")

        if metadata.learnable is not None:
            raise ValueError(f"Cannot set 'learnable' parameter on predicate ({predicate}) metadata")

        self.predicate = predicate
        self.metadata = metadata

    def __str__(self) -> str:
        return f"{self.predicate} {self.metadata}"

    def __repr__(self) -> str:
        return self.__str__()