类型保护 function 第一个参数的属性类型由第二个字符串键参数确定（例如 key="Dict"，然后 -> TypeGuard["Dict"]）？

Question

I tried to parse a JSON data into a Python Pydantic class object. The problem is that the data is like

class Data(BaseModel):
    val: Union[Dict, SubData1, SubData2]
    key: str # e.g. "dict_type", "sub_data_type", "sub_data_type2"

I'm thinking to make a type guard function ( check_type ) and use it like below:

if check_type(data, "dict_type"):
   value = data.val ## type checker will automatically know it is Dict
elif check_type(data, "sub_data_type"):
   value = data.val ## type checker will automatically know it is SubData1

It will allow my IDE to have better maintainability. Is it doable?

Answer 1

import typing

@typing.overload
def check_type_1(data: typing.Any, typename: typing.Literal["dict"]) -> typing.TypeGuard[dict]:
    ...

@typing.overload
def check_type_1(data: typing.Any, typename: typing.Literal["list"]) -> typing.TypeGuard[list]:
    ...

def check_type_1(data: typing.Any, typename: str) -> bool:
    if typename == "dict":
        return isinstance(data, dict)
    elif typename == "list":
        return isinstance(data, list)
    else:
        raise NotImplementedError

def test2():
    x: typing.Any = 1
    if check_type_1(x, "dict"):
        print(f"x={x} is dict")
    elif check_type_1(x, "list"):
        print(f"x={x} is list")
    else:
        print(f"x={x} is {type(x)}")

if __name__ == "__main__":
    test2()

Try this code in Python 3.10.