Improve results of stubtest

[loicdiridollou]

Clean up some old pandas 2.2 code that was removed, improved the results of the stubtest.

• Closes #xxxx (Replace xxxx with the Github issue number)
• Tests added: Please use assert_type() to assert the type of any return value

[twoertwein]

Looks good to me, but I'm a bit unsure about the deprecated arguments.

[loicdiridollou]

Thanks @twoertwein for the review, I was debating myself about the deprecated args, they appear in the stubtest since the code is still there.
Prob worth opening a discussion with @Dr-Irv how we want to handle it, a part of me want to drop them since we should not use them but the stubtest complains. Not sure there is a perfect answer here, the stubtest is super useful as it uncovers many bits of code that were deleted from the pandas code but it of course can't handle deprecation like we would like.

[Dr-Irv]

Probably should split into 2 overloads, based on the docs. One is where periods is int and there is no suffix argument, and the other where periods is Sequence[int] and then suffix is allowed.

I don't understand what you mean by "has no effect at the moment"

[loicdiridollou]

If you read the source code it is not used at all.
https://github.com/pandas-dev/pandas/blob/4f952b79f91f4f9d4d86b46d34e78dc657bfd4a2/pandas/core/generic.py#L10122-L10130

Maybe I missed something in the code but not sure.

[Dr-Irv]

I created pandas-dev/pandas#61955

so let's delete it from the stubs

[Dr-Irv]

Thanks @twoertwein for the review, I was debating myself about the deprecated args, they appear in the stubtest since the code is still there. Prob worth opening a discussion with @Dr-Irv how we want to handle it, a part of me want to drop them since we should not use them but the stubtest complains. Not sure there is a perfect answer here, the stubtest is super useful as it uncovers many bits of code that were deleted from the pandas code but it of course can't handle deprecation like we would like.

I prefer to not have deprecated arguments in the stubs, because we want people to stop using them.

Maybe there is a way to tell stubtest to ignore certain differences?

[loicdiridollou]

When it comes to stubtest I don't think there is a way for it to filter since the errors that we get is that the arguments in the stubs are misaligned with the runtime version (it is impossible for stubtest to know what is deprecated in pandas).
That is why I felt like reintroducing the arguments would help clear out the errors but as you said it is not the better way to go.

[Dr-Irv]

I created pandas-dev/pandas#61955

so let's delete it from the stubs

[Dr-Irv]

thanks @loicdiridollou