Simplify and fix overloads for methods with inplace parameter

[brianhelba]

The return value of these methods varies on the inplace argument.

[Dr-Irv]

Tests are failing. Please make sure to run the tests locally before resubmitting.

I understand what you are trying to do here, and it seems correct, but I also would like to know what prompted you to make these changes. Is there a test we can add that fails with the current stubs but passes with your changes?

[Dr-Irv]

I don't think you can remove this overload.

[brianhelba]

This seems safe to simplify. There are only 3 return possibilities:

Series or DataFrame or None
When drop is False (the default), a DataFrame is returned. The newly created columns will come first in the DataFrame, followed by the original Series values. When drop is True, a Series is returned. In either case, if inplace=True, no value is returned.

The return value can be fully determined by the value of drop and inplace, otherwise the signatures are identical.

---

Note, the previous 6 overloads with a split for level: Sequence[Level] and level: Level | None had the same other parameters and possible return values (DataFrame, Series[S1], or None).

I assumed this duplication was a result of some auto-generated code, but if there's something else I'm failing to understand, please let me know.

[brianhelba]

@Dr-Irv Thanks for the quick review and sorry for my delay in replying. I've just updated the PR and tests are now passing.

I understand what you are trying to do here, and it seems correct, but I also would like to know what prompted you to make these changes.

My initial reason for wanting this change was a mistake on my part (I thought that the overloads for Series.rename were not returning the correct type, but it was actually a separate bug in my project 😅). However, after looking at this codebase, I realized that the overloads for inplace seemed needlessly complicated, so I wanted to help clean it up.

There are probably a few minor improvements here as well (adding overloads or return type annotations to methods which lacked them before), but it seemed unnecessary to try and make separate PRs (one for new annotations, one for cleanup).

[brianhelba]

Is there a test we can add that fails with the current stubs but passes with your changes?

We could probably add one for DataFrame.where or DataFrame.mask (places that totally lacked an overload before), but as mentioned, my primary motivation was just cleanup. Removing lines of code is often just as helpful as adding them.