绑定抽象接口的TypeVar引发mypy类型不兼容错误求助

问题背景 #

我在Python中定义了协变TypeVar PersonDetails，绑定到抽象接口PersonDetailsAbstractInterface，该接口有PersonalInfo、Employment、UpdateStatus、Vehicle四个实现类。随后实现了泛型类Person，其类方法from_iterables接收上述实现类的可迭代对象生成Person实例，代码运行正常，但mypy报错类型不兼容。

错误信息 #

error: Argument 1 to "Person" has incompatible type "*list[PersonDetails]"; expected "PersonalInfo"  [arg-type]
error: Argument 1 to "Person" has incompatible type "*list[PersonDetails]"; expected "Employment"  [arg-type]
error: Argument 1 to "Person" has incompatible type "*list[PersonDetails]"; expected "UpdateStatus"  [arg-type]
error: Argument 1 to "Person" has incompatible type "*list[PersonDetails]"; expected "Vehicle"  [arg-type]

相关代码片段 #

自定义类型定义 #

# define custom types
PersonDetails = TypeVar("PersonDetails", bound="PersonDetailsAbstractInterface", covariant=True)

抽象接口实现 #

@dataclass
class PersonDetailsAbstractInterface(ABC, PersonDetailsBase):
    """Person Personal Details Abstract Interface"""

Person类实现 #

@dataclass
class Person(Generic[PersonDetails]):
    personal_info: PersonalInfo
    employment: Employment
    update_status: UpdateStatus
    vehicle: Vehicle

    @classmethod
    def from_iterables(cls, *iterables: Iterable[PersonDetails]) -> Generator[Person, None, None]:
        iterators = [iter(iterable) for iterable in iterables]

        while True:
            try:
                construction_values = [next(iterator) for iterator in iterators]
                yield Person(*construction_values)

            except StopIteration:
                break

调用代码 #

if __name__ == '__main__':
    personal_info_data = PersonalInfo.from_file("personal_info.csv")
    employment_data = Employment.from_file("employment.csv")
    update_status_data = UpdateStatus.from_file("update_status.csv")
    vehicles_data = Vehicle.from_file("vehicles.csv")

    person_data = Person.from_iterables(
        personal_info_data, employment_data, update_status_data, vehicles_data
    )

问题原因 #

1. 泛型参数未关联类字段：Person类虽继承Generic[PersonDetails]，但四个字段都是固定的具体子类类型，并未使用泛型参数PersonDetails。这导致泛型参数完全多余，mypy无法将传入的不同子类型与固定字段类型对应。
2. 协变场景误用：协变仅允许子类类型替换父类类型，但这里试图用单一PersonDetails类型覆盖四个不同子类型，协变无法解决“多子类型匹配多固定字段”的问题。from_iterables的*iterables: Iterable[PersonDetails]会被mypy推断为所有迭代器是同一类型，与实际传入的四个不同子类型冲突，构造Person时自然触发类型不匹配报错。

解决方案 #

方案1：移除冗余泛型，明确参数类型 #

由于Person字段类型固定，直接在from_iterables中声明每个参数的具体类型，无需泛型：

from typing import Generator, Iterable

@dataclass
class Person:
    personal_info: PersonalInfo
    employment: Employment
    update_status: UpdateStatus
    vehicle: Vehicle

    @classmethod
    def from_iterables(
        cls,
        personal_info_iter: Iterable[PersonalInfo],
        employment_iter: Iterable[Employment],
        update_status_iter: Iterable[UpdateStatus],
        vehicle_iter: Iterable[Vehicle]
    ) -> Generator["Person", None, None]:
        iterators = [
            iter(personal_info_iter),
            iter(employment_iter),
            iter(update_status_iter),
            iter(vehicle_iter)
        ]

        while True:
            try:
                construction_values = [next(it) for it in iterators]
                yield cls(*construction_values)
            except StopIteration:
                break

方案2：使用多绑定TypeVar（保留灵活性） #

若后续需扩展Person字段类型，可定义多个绑定到抽象接口的TypeVar，确保参数与字段类型一一匹配：

from typing import TypeVar, Generic, Generator, Iterable

PD1 = TypeVar("PD1", bound=PersonDetailsAbstractInterface)
PD2 = TypeVar("PD2", bound=PersonDetailsAbstractInterface)
PD3 = TypeVar("PD3", bound=PersonDetailsAbstractInterface)
PD4 = TypeVar("PD4", bound=PersonDetailsAbstractInterface)

@dataclass
class Person(Generic[PD1, PD2, PD3, PD4]):
    personal_info: PD1
    employment: PD2
    update_status: PD3
    vehicle: PD4

    @classmethod
    def from_iterables(
        cls,
        iter1: Iterable[PD1],
        iter2: Iterable[PD2],
        iter3: Iterable[PD3],
        iter4: Iterable[PD4]
    ) -> Generator["Person[PD1, PD2, PD3, PD4]", None, None]:
        iterators = [iter(iter1), iter(iter2), iter(iter3), iter(iter4)]
        while True:
            try:
                yield cls(*[next(it) for it in iterators])
            except StopIteration:
                break

说明 #

方案1适配当前固定字段的场景，简洁直接；方案2保留泛型灵活性，支持后续替换不同实现类，同时让mypy能正确推断类型。

内容的提问来源于stack exchange，提问作者64rl0