Fixed bugs with WCS slicing, - #18417

Thank you for your contribution to Astropy! 🌌 This checklist is meant to remind the package maintainers who will review this pull request of some common things to look for.

• Do the proposed changes actually accomplish desired goals?
• Do the proposed changes follow the Astropy coding guidelines?
• Are tests added/updated as required? If so, do they follow the Astropy testing guidelines?
• Are docs added/updated as required? If so, do they follow the Astropy documentation guidelines?
• Is rebase and/or squash necessary? If so, please provide the author with appropriate instructions. Also see instructions for rebase and squash.
• Did the CI pass? If no, are the failures related? If you need to run daily and weekly cron jobs as part of the PR, please apply the "Extra CI" label. Codestyle issues can be fixed by the bot.
• Is a change log needed? If yes, did the change log check pass? If no, add the "no-changelog-entry-needed" label. If this is a manual backport, use the "skip-changelog-checks" label unless special changelog handling is necessary.
• Is this a big PR that makes a "What's new?" entry worthwhile and if so, is (1) a "what's new" entry included in this PR and (2) the "whatsnew-needed" label applied?
• At the time of adding the milestone, if the milestone set requires a backport to release branch(es), apply the appropriate "backport-X.Y.x" label(s) before merge.

Specifically slicing WCS objects wi…th more than 2 dimensions but with _naxis not set would fail, and that slicing a WCS with an ellipsis would not return a WCS object but a SlicedFITSWCS instead

Pardon my English, but it is not clear to me whether the above is true before this PR or will it be true after this PR will be merged.

Sorry for the confusing wording. This PR makes is so we can slice 3+D WCS (it failed before) and makes it so we can do wcs[...] and get back a WCS object (we got back SlicedFITSWCS back before)

While it looks like your PR may fix the specific failure, it seems to me it basically is a patch hiding the underlying issue. Based on what I see so far in the code, a lot of things depend on _naxis being set correctly but I feel that code that sets/modifies _naxis is bad. For example, doing this on the 3D wcs object from #18410 results in _naxis = [0, 0] for a 3D object:

>>> wcs.pixel_shape = None
>>> wcs._naxis
[0, 0]

This leads to the following puzzling printout of a 3D WCS (even with your changes):

WCS Keywords

Number of WCS axes: 3
CTYPE : '' '' ''
CRVAL : 0.0 0.0 0.0
CRPIX : 0.0 0.0 0.0
PC1_1 PC1_2 PC1_3  : 1.0 0.0 0.0
PC2_1 PC2_2 PC2_3  : 0.0 1.0 0.0
PC3_1 PC3_2 PC3_3  : 0.0 0.0 1.0
CDELT : 1.0 1.0 1.0
NAXIS : 0  0

Notice NAXIS has only two values even though at the top it reports "Number of WCS axes: 3".

I think a better approach would be to replace _naxis = [0, 0] throughout the code with _naxis = len(self._wcs.crpix) * [0]. I am not sure yet about this approach but it is the first thing I thought about. Of course, _naxis should be updated after WCSBase.__init__() in wcs.py if _get_naxis() was called with _get_naxis(header=None).

Oh, another bad example:

In [13]: w = WCS(naxis=1)

In [14]: w
Out[14]:
WCS Keywords

Number of WCS axes: 1
CTYPE : ''
CRVAL : 0.0
CRPIX : 0.0
PC1_1  : 1.0
CDELT : 1.0
NAXIS : 0  0

In [15]: w._naxis
Out[15]: [0, 0]

Why is _naxis being forced to be of length 2 in

See #18405, which fixes/hides some of the above