Type of mean is not set correctly

Describe the bug The return type of mean() is set as builtins.float where in practice it’s numpy.float64.

To Reproduce typing.py

from typing_extensions import reveal_type
import pandas as pd

val = pd.Series([True, False, True]).mean()
reveal_type(val)

mypy typing.py typing.py:5: note: Revealed type is "builtins.float"

In the interactive terminal:

In [1]: import pandas as pd
In [2]: val = pd.Series([True, False, True]).mean()
In [3]: type(val)
Out[3]: numpy.float64

Please complete the following information:

• OS: Debian
• OS: 11
• python version: 3.9.7
• version of type checker: mypy
• version of installed pandas-stubs: 1.4.3.220704

Additional context This becomes a problem when using the round function on float, pd.Series([True, False, True]).mean().round() For which mypy gives this message: error: "float" has no attribute "round"

[Dr-Irv]

While the python code works, it is not supported by pandas. as it is undocumented. There is a goal in pandas to make the return type float, so we’ll leave the signature the way it is.

[rvanlaar]

Noted. I’ve changed my code to use python’s round as a function and not as a method on the float.