`NA - Series` type shows up as `NAType` as opposed to `Series`

[MarcoGorelli]

Describe the bug

To Reproduce

1. Provide a minimal runnable pandas example that is not properly checked by the stubs.
2. Indicate which type checker you are using (mypy or pyright).
3. Show the error message received from that type checker while checking your example.

from __future__ import annotations
import pandas as pd
from typing import reveal_type

s = pd.Series([1,2,3])

reveal_type(pd.NA - s)
print(type(pd.NA - s))

outputs

$ python t.py 
Runtime type is 'Series'
<class 'pandas.core.series.Series'>

$ pyright t.py 

/home/marcogorelli/pandas-stubs-dev/t.py
  /home/marcogorelli/pandas-stubs-dev/t.py:7:13 - information: Type of "pd.NA - s" is "NAType"
0 errors, 0 warnings, 1 information 

Note how the runtime type (Series) doesn't match the type-checked one NAType

Please complete the following information:

• OS: [e.g. Windows, Linux, MacOS] linux (WSL2)
• OS Version [e.g. 22]
• python version 3.12.8
• version of type checker pyright 1.1.403
• version of installed pandas-stubs 82d1a65

Additional context
Add any other context about the problem here.

[Dr-Irv]

Looks like all the math operators in pandas-stubs\_libs\missing.pyi for class NAType need overloads to handle Series, maybe Index, maybe np_1darray and maybe lists, along with separate overloads for scalars.

[MarcoGorelli]

Related, index minus na erroneously shows Index[int], instead of Index[type[object]]

[Dr-Irv]

Related, index minus na erroneously shows Index[int], instead of Index[type[object]]

That's a tough one. Because we infer s = pd.Series([1,2,3], dtype="Int64") to be Series[int], and in that case, s - pd.NA, remains Series[int], but if you do s1 = pd.Series([1,2,3]), then s1 is also Series[int], but s1 - pd.NA will be Series[object]

So I think the current behavior is fine.

[MarcoGorelli]

sure, i've opened a separate issue about the Index overloads, because i think there is a real issue there #1347