AlignDataVariableDimensionsToDatasetCoords: complete test coverage

[edoyango]

Hi,

this is a small pull request that brings the test coverage of AlignDataVariableDimensionsToDatasetCoords to 100%. It also:

• allows the apply_func method to handle dataset's who's constituent dataArray's have coordinates with names different to the dims. I accidentally ran into this problem when trying to run the datarray example on the xarray docs
• Added notimplemented error for the undo_func - to make it obvious that the op can't yet be undone.

[tennlee]

It looks like 'interrograte' and the 'pre-commit' steps are running with Python 3.14 and are experiencing issues due to that. I will manually check for those elements, and try to sort out the issue. It's probably simple to set those steps to use Python 3.13 or earlier, but it would be preferable to fix whatever the root cause is so that it can run on 3.14.

[coveralls]

Pull Request Test Coverage Report for Build 19160599959

Details

• 20 of 20 (100.0%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.06%) to 61.162%

---

Totals
Change from base Build 19160591903:	0.06%
Covered Lines:	9545
Relevant Lines:	15189

---

💛 - Coveralls

[tennlee]

I have given this a brief look - really nice job. I want to just check a couple of things before I push merge, but thanks for writing a thorough test.

[tennlee]

My only feedback is around the exception during the undo. I'm actually unsure whether the right pattern is to just return the data unmodified, or raise the exception. I think in the face of doubt, refuse the temptation to guess, so I'm happy to accept the PR. But the project could do with some clearer guidance about how "undo" is intended to work, and that should be documented in the base class of the operator so that it's clear to future developers. However, for now, let's move forward with this, and we can revise or address the undo process subsequently. Thanks again.