FEAT: with_spectral_axis_and_flux_units

[pllim]

Follow-up of #1181 . I think it would really benefit Cubeviz parser if we can do both spectral axis and unit conversions at the same time, to avoid copying an object twice unnecessarily.

I also couldn't help myself by cleaning up some existing test code after noticing unnecessary usage of Numpy random number generator.

🐱

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 86.89%. Comparing base (2329d11) to head (7097d8a).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1184      +/-   ##
==========================================
+ Coverage   86.87%   86.89%   +0.01%     
==========================================
  Files          63       63              
  Lines        4556     4562       +6     
==========================================
+ Hits         3958     3964       +6     
  Misses        598      598              

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

[pllim]

devdeps failure is unrelated: http.client.RemoteDisconnected: Remote end closed connection without response

[keflavich]

This makes some sense to me - with single 1D spectra, copying the spectral axis (which is needed for changing the spectral unit) is just as expensive as copying the data (changing the flux unit). In spectral-cube, we have considered copies of the spectral axis 'cheap' under the assumption that the 1D axis is always << the 3D data. So I'm onboard with this - but I had to think through it to see why there's value.

[rosteen]

LGTM, thanks for the test cleanups.

[pllim]

@keflavich , there is also cost in memory.

Thanks for the reviews, everyone!