Check tuples of abstract types

[sobolevn]

The PR is quite simple (and I would like to keep it this way):

• Before we were only checking type[] type
• Now we also check tuple[type[], ...] type

There might be other types that we want to add in the future here: TypedDictType, etc?
But, let's do it one by one, because smaller PRs are easier to merge :)

Closes #15264

[ikonst]

First impression: shouldn't there be something else that recurses into a tuple already to validate types of items, so we could avoid teaching has_abstract_type_part about TupleType (and then about TypedDictType etc.) ?

[sobolevn]

I don't see a clean way to do that.

[JukkaL]

I don't see a clean way to do that.

What about just leaving a TODO comment about this, in case somebody figures out a good way to do that?

[JukkaL]

Even if this doesn't fully solve the issue, this seems like a reasonably incremental improvement, though it would be good to add a comment explaining that this is incomplete (and perhaps why), as I mentioned in my other comment.

[sobolevn]

@JukkaL thanks for the review! I've left a TODO note for future work :)
I will merge it once CI passes.

[github-actions]

According to mypy_primer, this change doesn't affect type check results on a corpus of open source code. ✅