feat(series): add and truediv for bools

[cmp0xff]

This is inspired by #1312 (comment) and a follow-up for #1311.

• Tests added: Please use assert_type() to assert the type of any return value

[Dr-Irv]

Just have to add a few comments so we understand why the ignores are there in the tests.

I caught something when eyeballing the difference between __truediv__() and truediv() .

It's odd that truediv() works for mypy but __truediv__() does not.

[Dr-Irv]

I see that you were able to remove the 2 ignores. Request a review when you've handled my comments from the previous review (2 of which you can now dismiss) and from this review.

[cmp0xff]

It's odd that truediv() works for mypy but truediv() does not.

Our test coverage used to be poorer than it is now. truediv was not tested to the same standard as __truediv__.

[Dr-Irv]

I checked and __truediv__() and truediv() have the same patterns.

Then checked __rtruediv__() and rtruediv() and found discrepancies.

Probably should do the same for add() and radd() at some point, but that can be another PR

[Dr-Irv]

rtruediv() has Series[int]andSeries[float]forother` - please reconcile

[Dr-Irv]

rtruediv() has Series[_T_COMPLEX] for other - please reconcile

[Dr-Irv]

Thanks for also checking on __radd__()/radd() . I've checked them all now, and found one that I commented on, and one that is missing but I couldn't write a comment. This one is missing from add():

    @overload
    def __add__(self, other: Series[Never]) -> Series: ...

Otherwise, it's all good

[Dr-Irv]

Thanks @cmp0xff for the nice contribution!

Might want to add some tests for things we don't want to allow, e.g.,

s = pd.Series(["abc", "def"])  # inferred as Series[str]
s / "456"

The latter division will fail, and we are detecting it via the type checks, but having tests for that would be a nice addition, for another PR. Having said that, eagerly awaiting your PRs for subtraction and multiplication!