24546459_第62章 请把我的诗雨放开

admin

mport annotated_types as at

    FieldInfo = import_cached_field_info()

    for annotation in annotations:
        if isinstance(annotation, at.GroupedMetadata):
            yield from annotation
        elif isinstance(annotation, FieldInfo):
            yield from annotation.metadata
            # this is a bit problematic in that it results in duplicate metadata
            # all of our "consumers" can handle it, but it is not ideal
            # we probably should split up FieldInfo into:
            # - annotated types metadata
            # - individual metadata known only to Pydantic
            annotation = copy(annotation)
            annotation.metadata = []
            yield annotation
        else:
            yield annotation


@lru_cache
def _get_at_to_constraint_map() -> dict[type, str]:
    """Return a mapping of annotated types to constraints.

    Normally, we would define a mapping like this in the module scope, but we can't do that
    because we don't permit module level imports of `annotated_types`, in an attempt to speed up
    the import time of `pydantic`. We still only want to have this dictionary defined in one place,
    so we use this function to cache the result.
    """
    import annotated_types as at

    return {
        at.Gt: 'gt',
        at.Ge: 'ge',
        at.Lt: 'lt',
        at.Le: 'le',
        at.MultipleOf: 'multiple_of',
        at.MinLen: 'min_length',
        at.MaxLen: 'max_length',
    }


def apply_known_metadata(annotation: Any, schema: CoreSchema) -> CoreSchema | None:  # noqa: C901
    """Apply `annotation` to `schema` if it is an annotation we know about (Gt, Le, etc.).
    Otherwise return `None`.

    This does not handle all known annotations. If / when it does, it can always
    return a CoreSchema and return the unmodified schema if the annotation should be ignored.

    Assumes that GroupedMetadata has already been expanded via `expand_grouped_metadata`.

    Args:
        annotation: The annotation.
        schema: The schema.

    Returns:
        An updated schema with annotation if it is an annotation we know about, `None` otherwise.

    Raises:
        RuntimeError: If a constraint can't be applied to a specific schema type.
        ValueError: If an unknown constraint is encountered.
    """
    import annotated_types as at

    from ._validators import NUMERIC_VALIDATOR_LOOKUP, forbid_inf_nan_check

    schema = schema.copy()
    schema_update, other_metadata = collect_known_metadata([annotation])
    schema_type = schema['type']

    chain_schema_constraints: set[str] = {
        'pattern',
        'strip_whitespace',
        'to_lower',
        'to_upper',
        'coerce_numbers_to_str',
        'ascii_only',
    }
    chain_schema_steps: list[CoreSchema] = []

    for constraint, value in schema_update.items():
        if constraint not in CONSTRAINTS_TO_ALLOWED_SCHEMAS:
            raise ValueError(f'Unknown constraint {constraint}')
        allowed_schemas = CONSTRAINTS_TO_ALLOWED_SCHEMAS[constraint]

        # if it becomes necessary to handle more than one constraint
        # in this recursive case with function-after or function-wrap, we should refactor
        # this is a bit challenging because we sometimes want to apply constraints to the inner schema,
        # whereas other times we want to wrap the existing schema with a new one that enforces a new constraint.
        if schema_type in {'function-before', 'function-wrap', 'function-after'} and constraint == 'strict':
            schema['schema'] = apply_known_metadata(annotation, schema['schema'])  # type: ignore  # schema is function schema
            return schema

        # if we're allowed to apply constraint directly to the schema, like le to int, do that
        if schema_type in allowed_schemas:
            if constraint == 'union_mode' and schema_type == 'union':
                schema['mode'] = value  # type: ignore  # schema is UnionSchema
            else:
                schema[constraint] = value
            continue

        #  else, apply a function after validator to the schema to enforce the corresponding constraint
        if constraint in chain_schema_constraints:

            def _apply_constraint_with_incompatibility_info(
                value: Any, handler: cs.ValidatorFunctionWrapHandler
            ) -> Any:
                try:
                    x = handler(value)
                except ValidationError as ve:
                    # if the error is about the type, it's likely that the constraint is incompatible the type of the field
                    # for example, the following invalid schema wouldn't be caught during schema build, but rather at this point
                    # with a cryptic 'string_type' error coming from the string validator,
                    # that we'd rather express as a constraint incompatibility error (TypeError)
                    # Annotated[list[int], Field(pattern='abc')]
                    if 'type' in ve.errors()[0]['type']:
                        raise TypeError(
                            f"Unable to apply constraint '{constraint}' to supplied value {value} for schema of type '{schema_type}'"  # noqa: B023
                        )
                    raise ve
                return x

            chain_schema_steps.append(
                cs.no_info_wrap_validator_function(
                    _apply_constraint_with_incompatibility_info, cs.str_schema(**{constraint: value})
                )
            )
        elif constraint in NUMERIC_VALIDATOR_LOOKUP:
            if constraint in LENGTH_CONSTRAINTS:
                inner_schema = schema
                while inner_schema['type'] in {'function-before', 'function-wrap', 'function-after'}:
                    inner_schema = inner_schema['schema']  # type: ignore
                inner_schema_type = inner_schema['type']
                if inner_schema_type == 'list' or (
                    inner_schema_type == 'json-or-python' and inner_schema['json_schema']['type'] == 'list'  # type: ignore
                ):
                    js_constraint_key = 'minItems' if constraint == 'min_length' else 'maxItems'
                else:
                    js_constraint_key = 'minLength' if constraint == 'min_length' else 'maxLength'
            else:
                js_constraint_key = constraint

            schema = cs.no_info_after_validator_function(
                partial(NUMERIC_VALIDATOR_LOOKUP[constraint], **{constraint: value}), schema
            )
            metadata = schema.get('metadata', {})
            if (existing_json_schema_updates := metadata.get('pydantic_js_updates')) is not None:
                metadata['pydantic_js_updates'] = {
                    **existing_json_schema_updates,
                    **{js_constraint_key: as_jsonable_value(value)},
                }
            else:
                metadata['pydantic_js_updates'] = {js_constraint_key: as_jsonable_value(value)}
            schema['metadata'] = metadata
        elif constraint == 'allow_inf_nan' and value is False:
            schema = cs.no_info_after_validator_function(
                forbid_inf_nan_check,
                schema,
            )
        else:
            # It's rare that we'd get here, but it's possible if we add a new constraint and forget to handle it
            # Most constraint errors are caught at runtime during attempted application
            raise RuntimeError(f"Unable to apply constraint '{constraint}' to schema of type '{schema_type}'")

    for annotation in other_metadata:
        if (annotation_type := type(annotation)) in (at_to_constraint_map := _get_at_to_constraint_map()):
            constraint = at_to_constraint_map[annotation_type]
            validator = NUMERIC_VALIDATOR_LOOKUP.get(constraint)
            if validator is None:
                raise ValueError(f'Unknown constraint {constraint}')
            schema = cs.no_info_after_validator_function(
                partial(validator, {constraint: getattr(annotation, constraint)}), schema
            )
            continue
        elif isinstance(annotation, (at.Predicate, at.Not)):
            predicate_name = f'{annotation.func.__qualname__!r} ' if hasattr(annotation.func, '__qualname__') else ''

            # Note: B023 is ignored because even though we iterate over `other_metadata`, it is guaranteed
            # to be of length 1. `apply_known_metadata()` is called from `GenerateSchema`, where annotations
            # were already expanded via `expand_grouped_metadata()`. Confusing, but this falls into the annotations
            # refactor.
            if isinstance(annotation, at.Predicate):

                def val_func(v: Any) -> Any:
                    predicate_satisfied = annotation.func(v)  # noqa: B023
                    if not predicate_satisfied:
                        raise PydanticCustomError(
                            'predicate_failed',
                            f'Predicate {predicate_name}failed',  # pyright: ignore[reportArgumentType]  # noqa: B023
                        )
                    return v

            else:

                def val_func(v: Any) -> Any:
                    predicate_satisfied = annotation.func(v)  # noqa: B023
                    if predicate_satisfied:
                        raise PydanticCustomError(
                            'not_operation_failed',
                            f'Not of {predicate_name}failed',  # pyright: ignore[reportArgumentType]  # noqa: B023
                        )
                    return v

            schema = cs.no_info_after_validator_function(val_func, schema)
        else:
            # ignore any other unknown metadata
            return None

    if chain_schema_steps:
        chain_schema_steps = [schema] + chain_schema_steps
        return cs.chain_schema(chain_schema_steps)

    return schema


def collect_known_metadata(annotations: Iterable[Any]) -> tuple[dict[str, Any], list[Any]]:
    """Split `annotations` into known metadata and unknown annotations.

    Args:
        annotations: An iterable of annotations.

    Returns:
        A tuple contains a dict of known metadata and a list of unknown annotations.

    Example:
        ```python
        from annotated_types import Gt, Len

        from pydantic._internal._known_annotated_metadata import collect_known_metadata

        print(collect_known_metadata([Gt(1), Len(42), ...]))
        #> ({'gt': 1, 'min_length': 42}, [Ellipsis])
        ```
    """
    annotations = expand_grouped_metadata(annotations)

    res: dict[str, Any] = {}
    remaining: list[Any] = []

    for annotation in annotations:
        # isinstance(annotation, PydanticMetadata) also covers ._fields:_PydanticGeneralMetadata
        if isinstance(annotation, PydanticMetadata):
            res.update(annotation.__dict__)
        # we don't use dataclasses.asdict because that recursively calls asdict on the field values
        elif (annotation_type := type(annotation)) in (at_to_constraint_map := _get_at_to_constraint_map()):
            constraint = at_to_constraint_map[annotation_type]
            res[constraint] = getattr(annotation, constraint)
        elif isinstance(annotation, type) and issubclass(annotation, PydanticMetadata):
            # also support PydanticMetadata classes being used without initialisation,
            # e.g. `Annotated[int, Strict]` as well as `Annotated[int, Strict()]`
            res.update({k: v for k, v in vars(annotation).items() if not k.startswith('_')})
        else:
            remaining.append(annotation)
    # Nones can sneak in but pydantic-core will reject them
    # it'd be nice to clean things up so we don't put in None (we probably don't _need_ to, it was just easier)
    # but this is simple enough to kick that can down the road
    res = {k: v for k, v in res.items() if v is not None}
    return res, remaining


def check_metadata(metadata: dict[str, Any], allowed: Iterable[str], source_type: Any) -> None:
    """A small utility function to validate that the given metadata can be applied to the target.
    More than saving lines of code, this gives us a consistent error message for all of our internal implementations.

    Args:
        metadata: A dict of metadata.
        allowed: An iterable of allowed metadata.
        source_type: The source type.

    Raises:
        TypeError: If there is metadatas that can't be applied on source type.
    """
    unknown = metadata.keys() - set(allowed)
    if unknown:
        raise TypeError(
            f'The following constraints cannot be applied to {source_type!r}: {", ".join([f"{k!r}" for k in unknown])}'
        )