na_value in to_numpy does not accept NumPy number types

[vnmabus]

Describe the bug
Calling to_numpy with na_value set to a NumPy type gives a type error. This works at runtime.

To Reproduce

1. Provide a minimal runnable pandas example that is not properly checked by the stubs.

   import numpy as np
   import pandas as pd
   
   data_array = pd.arrays.IntegerArray(
       np.array([1, 2, -1], dtype=np.int32),
       mask=np.array([0, 0, 1], dtype=np.bool),
   )
   data = data_array.to_numpy(
       dtype=np.int32,
       na_value=np.int32(42),
   )

2. Indicate which type checker you are using (mypy or pyright).

   I am using MyPy.

3. Show the error message received from that type checker while checking your example.

   Argument "na_value" to "to_numpy" of "BaseMaskedArray" has incompatible type "signedinteger[_32Bit]"; expected "str | bytes | date | datetime | timedelta | <7 more items> | complex"
   

Please complete the following information:

• OS: Windows
• OS Version: 11
• python version: 3.11.11
• version of type checker: 1.15.0
• version of installed pandas-stubs: 2.2.3.250308

[loicdiridollou]

Thanks for the report!
This is indeed a known issue because Scalar type does not contain any of the numpy variations (np.int, np.float, ...).
It is currently up to consideration with @Dr-Irv because including those would have a large impact over the codebase but this may be the right way to go.

[Dr-Irv]

I don't have an issue with adding np.float and np.int to the Scalar type in _typing.pyi, assuming it doesn't break anything else.