fix: prevent sets from being reported as InPlaceSortAnalysis warnings

[mohasarc]

Hi @AryazE,

I’ve noticed that InPlaceSortAnalysis incorrectly raises warnings for Set objects, even though sets can't be sorted and don’t have an in-place .sort() method.

For example:

a = set(range(0, 1001))
sorted(a)  # <--- This would currently raise a warning, but it shouldn't.

I’ve implemented a simple fix for this. Please have a look and let me know if any further changes are needed.

[AryazE]

Thank you for the PR.
Yes, this seems like a problem.
However, I think the fix you suggested restricts the checker too much. I think any collection or object that has .sort and has a length (too check for long sequences) should be caught.
I have made a suggestion for the change.
Could you also extend the tests with such cases?

[AryazE]

Suggested change

	if type(pos_args[0]) is list and len(pos_args[0]) > self.threshold:
	if hasattr(pos_args[0], "sort") and hasattr(pos_args[0], "__len__") and len(pos_args[0]) > self.threshold:

[mohasarc]

I think your suggestion makes perfect sense, though we also need to update line 48 to perform the same check:

should be:

if len(self.stored_lists) > 0 and hasattr(val, "sort") and hasattr(val, "__len__"):

I'll perform both changes now

[mohasarc]

I made the changes. I also made a small refactor that I thought would be nice to reduce repetition.

[AryazE]

Thanks for the update.
I have renamed the function and added some tests.

[mohasarc]

It look great, thanks!