14963920_（二十三）

admin

lue
        yield cls
    finally:
        cls.__pydantic_run_validation__ = original_run_validation


class DataclassProxy:
    __slots__ = '__dataclass__'

    def __init__(self, dc_cls: Type['Dataclass']) -> None:
        object.__setattr__(self, '__dataclass__', dc_cls)

    def __call__(self, *args: Any, **kwargs: Any) -> Any:
        with set_validation(self.__dataclass__, True):
            return self.__dataclass__(*args, **kwargs)

    def __getattr__(self, name: str) -> Any:
        return getattr(self.__dataclass__, name)

    def __setattr__(self, __name: str, __value: Any) -> None:
        return setattr(self.__dataclass__, __name, __value)

    def __instancecheck__(self, instance: Any) -> bool:
        return isinstance(instance, self.__dataclass__)

    def __copy__(self) -> 'DataclassProxy':
        return DataclassProxy(copy.copy(self.__dataclass__))

    def __deepcopy__(self, memo: Any) -> 'DataclassProxy':
        return DataclassProxy(copy.deepcopy(self.__dataclass__, memo))


def _add_pydantic_validation_attributes(  # noqa: C901 (ignore complexity)
    dc_cls: Type['Dataclass'],
    config: Type[BaseConfig],
    validate_on_init: bool,
    dc_cls_doc: str,
) -> None:
    """
    We need to replace the right method. If no `__post_init__` has been set in the stdlib dataclass
    it won't even exist (code is generated on the fly by `dataclasses`)
    By default, we run validation after `__init__` or `__post_init__` if defined
    """
    init = dc_cls.__init__

    @wraps(init)
    def handle_extra_init(self: 'Dataclass', *args: Any, **kwargs: Any) -> None:
        if config.extra == Extra.ignore:
            init(self, *args, **{k: v for k, v in kwargs.items() if k in self.__dataclass_fields__})

        elif config.extra == Extra.allow:
            for k, v in kwargs.items():
                self.__dict__.setdefault(k, v)
            init(self, *args, **{k: v for k, v in kwargs.items() if k in self.__dataclass_fields__})

        else:
            init(self, *args, **kwargs)

    if hasattr(dc_cls, '__post_init__'):
        try:
            post_init = dc_cls.__post_init__.__wrapped__  # type: ignore[attr-defined]
        except AttributeError:
            post_init = dc_cls.__post_init__

        @wraps(post_init)
        def new_post_init(self: 'Dataclass', *args: Any, **kwargs: Any) -> None:
            if config.post_init_call == 'before_validation':
                post_init(self, *args, **kwargs)

            if self.__class__.__pydantic_run_validation__:
                self.__pydantic_validate_values__()
                if hasattr(self, '__post_init_post_parse__'):
                    self.__post_init_post_parse__(*args, **kwargs)

            if config.post_init_call == 'after_validation':
                post_init(self, *args, **kwargs)

        setattr(dc_cls, '__init__', handle_extra_init)
        setattr(dc_cls, '__post_init__', new_post_init)

    else:

        @wraps(init)
        def new_init(self: 'Dataclass', *args: Any, **kwargs: Any) -> None:
            handle_extra_init(self, *args, **kwargs)

            if self.__class__.__pydantic_run_validation__:
                self.__pydantic_validate_values__()

            if hasattr(self, '__post_init_post_parse__'):
                # We need to find again the initvars. To do that we use `__dataclass_fields__` instead of
                # public method `dataclasses.fields`

                # get all initvars and their default values
                initvars_and_values: Dict[str, Any] = {}
                for i, f in enumerate(self.__class__.__dataclass_fields__.values()):
                    if f._field_type is dataclasses._FIELD_INITVAR:  # type: ignore[attr-defined]
                        try:
                            # set arg value by default
                            initvars_and_values[f.name] = args[i]
                        except IndexError:
                            initvars_and_values[f.name] = kwargs.get(f.name, f.default)

                self.__post_init_post_parse__(**initvars_and_values)

        setattr(dc_cls, '__init__', new_init)

    setattr(dc_cls, '__pydantic_run_validation__', ClassAttribute('__pydantic_run_validation__', validate_on_init))
    setattr(dc_cls, '__pydantic_initialised__', False)
    setattr(dc_cls, '__pydantic_model__', create_pydantic_model_from_dataclass(dc_cls, config, dc_cls_doc))
    setattr(dc_cls, '__pydantic_validate_values__', _dataclass_validate_values)
    setattr(dc_cls, '__validate__', classmethod(_validate_dataclass))
    setattr(dc_cls, '__get_validators__', classmethod(_get_validators))

    if dc_cls.__pydantic_model__.__config__.validate_assignment and not dc_cls.__dataclass_params__.frozen:
        setattr(dc_cls, '__setattr__', _dataclass_validate_assignment_setattr)


def _get_validators(cls: 'DataclassClassOrWrapper') -> 'CallableGenerator':
    yield cls.__validate__


def _validate_dataclass(cls: Type['DataclassT'], v: Any) -> 'DataclassT':
    with set_validation(cls, True):
        if isinstance(v, cls):
            v.__pydantic_validate_values__()
            return v
        elif isinstance(v, (list, tuple)):
            return cls(*v)
        elif isinstance(v, dict):
            return cls(**v)
        else:
            raise DataclassTypeError(class_name=cls.__name__)


def create_pydantic_model_from_dataclass(
    dc_cls: Type['Dataclass'],
    config: Type[Any] = BaseConfig,
    dc_cls_doc: Optional[str] = None,
) -> Type['BaseModel']:
    field_definitions: Dict[str, Any] = {}
    for field in dataclasses.fields(dc_cls):
        default: Any = Undefined
        default_factory: Optional['NoArgAnyCallable'] = None
        field_info: FieldInfo

        if field.default is not dataclasses.MISSING:
            default = field.default
        elif field.default_factory is not dataclasses.MISSING:
            default_factory = field.default_factory
        else:
            default = Required

        if isinstance(default, FieldInfo):
            field_info = default
            dc_cls.__pydantic_has_field_info_default__ = True
        else:
            field_info = Field(default=default, default_factory=default_factory, **field.metadata)

        field_definitions[field.name] = (field.type, field_info)

    validators = gather_all_validators(dc_cls)
    model: Type['BaseModel'] = create_model(
        dc_cls.__name__,
        __config__=config,
        __module__=dc_cls.__module__,
        __validators__=validators,
        __cls_kwargs__={'__resolve_forward_refs__': False},
        **field_definitions,
    )
    model.__doc__ = dc_cls_doc if dc_cls_doc is not None else dc_cls.__doc__ or ''
    return model


if sys.version_info >= (3, 8):

    def _is_field_cached_property(obj: 'Dataclass', k: str) -> bool:
        return isinstance(getattr(type(obj), k, None), cached_property)

else:

    def _is_field_cached_property(obj: 'Dataclass', k: str) -> bool:
        return False


def _dataclass_validate_values(self: 'Dataclass') -> None:
    # validation errors can occur if this function is called twice on an already initialised dataclass.
    # for example if Extra.forbid is enabled, it would consider __pydantic_initialised__ an invalid extra property
    if getattr(self, '__pydantic_initialised__'):
        return
    if getattr(self, '__pydantic_has_field_info_default__', False):
        # We need to remove `FieldInfo` values since they are not valid as input
        # It's ok to do that because they are obviously the default values!
        input_data = {
            k: v
            for k, v in self.__dict__.items()
            if not (isinstance(v, FieldInfo) or _is_field_cached_property(self, k))
        }
    else:
        input_data = {k: v for k, v in self.__dict__.items() if not _is_field_cached_property(self, k)}
    d, _, validation_error = validate_model(self.__pydantic_model__, input_data, cls=self.__class__)
    if validation_error:
        raise validation_error
    self.__dict__.update(d)
    object.__setattr__(self, '__pydantic_initialised__', True)


def _dataclass_validate_assignment_setattr(self: 'Dataclass', name: str, value: Any) -> None:
    if self.__pydantic_initialised__:
        d = dict(self.__dict__)
        d.pop(name, None)
        known_field = self.__pydantic_model__.__fields__.get(name, None)
        if known_field:
            value, error_ = known_field.validate(value, d, loc=name, cls=self.__class__)
            if error_:
                raise ValidationError([error_], self.__class__)

    object.__setattr__(self, name, value)


def is_builtin_dataclass(_cls: Type[Any]) -> bool:
    """
    Whether a class is a stdlib dataclass
    (useful to discriminated a pydantic dataclass that is actually a wrapper around a stdlib dataclass)

    we check that
    - `_cls` is a dataclass
    - `_cls` is not a processed pydantic dataclass (with a basemodel attached)
    - `_cls` is not a pydantic dataclass inheriting directly from a stdlib dataclass
    e.g.
    ```
    @dataclasses.dataclass
    class A:
        x: int

    @pydantic.dataclasses.dataclass
    class B(A):
        y: int
    ```
    In this case, when we first check `B`, we make an extra check and look at the annotations ('y'),
    which won't be a superset of all the dataclass fields (only the stdlib fields i.e. 'x')
    """
    return (
        dataclasses.is_dataclass(_cls)
        and not hasattr(_cls, '__pydantic_model__')
        and set(_cls.__dataclass_fields__).issuperset(set(getattr(_cls, '__annotations__', {})))
    )


def make_dataclass_validator(dc_cls: Type['Dataclass'], config: Type[BaseConfig]) -> 'CallableGenerator':
    """
    Create a pydantic.dataclass from a builtin dataclass to add type validation
    and yield the validators
    It retrieves the parameters of the dataclass and forwards them to the newly created dataclass
    """
    yield from _get_validators(dataclass(dc_cls, config=config, use_proxy=True))