Fix Dataset.isin and DataArray.isin when passed a set

[NoiceHax]

Description

isin passed its argument straight to numpy, and numpy turns a set into a 0-dimensional object array. That array never compares equal to any element, so xr.DataArray([1, 2, 3, 4]).isin({1, 2}) came back False everywhere instead of matching 1 and 2. The same happened with a frozenset or with dict.keys(). pandas returns the answer you would expect here, so the difference is easy to trip over.

The fix adds one branch to the existing isinstance chain in DataWithCoords.isin that converts a collections.abc.Set to a list before the numpy call. That covers sets, frozensets and dict views. Passing a Dataset still raises TypeError as before.

Tests: set and frozenset added to the parametrize on test_isin in test_dataset.py, set and dict.keys() added to test_isin in test_dataarray.py. Both fail before the change.

The issue thread also floats raising for scalars and strings the way pandas does. That is a larger behavior change and is not part of this PR.

Checklist

• Closes DataArray.isin does not accept sets intuitively #10022
• Tests added
• User visible changes (including notable bug fixes) are documented in whats-new.rst
• New functions/methods are listed in api.rst

AI Disclosure

• This PR contains AI-generated content.
  • I have tested any AI-generated content in my PR.
  • I take responsibility for any AI-generated content in my PR.
    Tools: Claude Code. I asked it to special-case sets in isin as suggested in the issue, then reviewed the change and ran the isin tests and pre-commit locally.