Improve is_empty and is_finite in a few cases

[user202729]

• is_empty for cartesian_product now returns True as long as some factor can be determined to be empty, even if other factors' emptiness is unknown.
• is_finite for cartesian_product now correctly return "unknown" in case (unknown empty set × infinite set).
• as a consequence of the above, is_empty must be provided for e.g. FiniteWords("ab") for tests to pass. (Unfortunately Sage has two different functions is_empty and __bool__ that does almost the exact same thing, I guess we will have to live with it.)
• is_empty for union of two sets now returns False as long as the first set is nonempty, regardless of the second set. (also performance optimization, no need to compute the cardinality.)
• LazyFamily now has is_finite implemented in more cases.

📝 Checklist

• The title is concise and informative.
• The description explains in detail what this PR is about.
• I have created tests covering the changes.
• I have updated the documentation and checked the documentation preview.

⌛ Dependencies

[github-actions]

Documentation preview for this PR (built with commit 887d545; changes) is ready! 🎉
This preview will update shortly after each push to this PR.

[user202729]

I tried to add is_empty to Parent… but now it's overriding is_empty in Sets.CartesianProducts.ParentMethods.

Will probably patch it by adding it to FiniteWords instead.

There's the (existing) issue of #39447 which I'm not entirely sure how to fix.

[tscrim]

This needs a doctest.

[user202729]

Technically the one above is already a test, no?

FiniteWords('ab').is_finite()

(without the category being correctly set, is_finite() wouldn't exist)

[tscrim]

But it is not a test associated to that method. Also, it is good to put a TestSuite(foo).run() test in for these methods' doctests.

[tscrim]

Suggested change

	Return False, because the empty word is in the set.
	Return ``False``, because the empty word is in the set.