Updates for ipyvolume viewer issues

[Carifio24]

While Azmé and I were investigating the aspect ratio issue mentioned in #451, we noticed that currently, the displayed volume rendering in the ipyvolume viewer seems to be wrong (in particular, it doesn't match what's displayed in the vispy volume viewer). I believe this might also be a partial cause of #450.

I think this is an issue with axes. As an example, I looked at the reduced TAN C14 cube used in the astropy FITS cube tutorial here. If I open this in vispy, I get

but with the current ipyvolume viewer I get

which is what led me to believe that the data was being passed into the ipyvolume widget incorrectly. After rearranging the axes as is done in this PR (note that np.transpose returns a view if possible, so we shouldn't need to worry about data getting copied), we get

This also fixes #451 for me, in that using the native aspect ratio now behaves the same as it does in vispy.

It's hard to tell how the changes to the ipyvolume viewer code affect the 3D scatter viewer since as #449 points out, it isn't working right now. Only thing I can check is that with these changes, the x/y/z labels indicate that it has a right-handed z-up coordinate system, so that's good.

[codecov]

Codecov Report

Attention: Patch coverage is 96.00000% with 1 line in your changes missing coverage. Please review.

Project coverage is 86.05%. Comparing base (488287d) to head (d479fe9).
Report is 2 commits behind head on main.

Files	Patch %	Lines
glue_jupyter/common/state3d.py	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #456      +/-   ##
==========================================
+ Coverage   86.04%   86.05%   +0.01%     
==========================================
  Files          90       90              
  Lines        5223     5229       +6     
==========================================
+ Hits         4494     4500       +6     
  Misses        729      729              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[maartenbreddels]

It's hard to tell how the changes to the ipyvolume viewer code affect the 3D scatter viewer since as #449 points out, it isn't working right now.

Indeed, that was my worry as well.

[Carifio24]

Turns out that the scatter viewer does work, albeit with some issues with particular glue components - see #449 (comment). After making similar changes to the scatter viewer as well, the display is now looking the same for me using either the ipyvolume scatter or the vispy scatter viewers.

[Carifio24]

@maartenbreddels and I had a brief discussion about the scatter issues at today's glue meeting and decided that the best thing to do is to make sure that we pass floats to the ipyvolume scatter. This PR was already manipulating the scatter widget's x/y/z values, so I decided to just add that fix here as well.

It's also now easy to confirm that the axis changes here lead to a right-handed coordinate system (which is what we want).

[astrofrog]

Thanks for all the work on this!