In DataFrame.select_dtypes type values like np.number or np.datetime64 should be allowed

[benediktsatalia]

Describe the bug
The documentation of DataFrame.select_dtypes clearly states that it is allowed to use values like np.number or np.datetime64 for the include and exclude parameters, but the current stubs don't allow it, they enforce include: _str | list[_str] | None and exclude: _str | list[_str] | None. This leads to typecheckers like mypy reporting false positives on such calls.

To Reproduce
If I run mypy on the following short snippet I get an error:

import numpy as np
import pandas as pd

df = pd.DataFrame()
df.select_dtypes(np.number)

the error is:

min_example.py:5: error: Argument 1 to "select_dtypes" of "DataFrame" has incompatible type "type[number[Any]]"; expected "str | list[str] | None"  [arg-type]
Found 1 error in 1 file (checked 1 source file)

Please complete the following information:

• OS: Linux
• OS Version: Arch Linux
• python version: 3.11
• version of type checker: mypy 1.8.0
• version of installed pandas-stubs: 2.2.0.240218

[Dr-Irv]

Thanks for the report. Definitely needs improvement on that particular stub.

Should use AsTypeArg from pandas-stubs/_typing.pyi instead of str PLUS add the other possible strings (e.g., "number") and object types (e.g., np.number) listed in the docs.

PR plus tests welcome.

[mutricyl]

I made an attempt for this issue. It is not that simple to cover all the restrictions stated in the documentation