Fix crash with swapped Spectral/Stokes axes for large mosaics

[thomaswilliamsastro]

This PR fixes a crash that can occur if the spectral/Stokes axes are swapped when creating a large mosaic.

• Fixed crash if spectral/Stokes axis is swapped when making large mosaics

[codecov]

Codecov Report

❌ Patch coverage is 0% with 25 lines in your changes missing coverage. Please review.
✅ Project coverage is 6.26%. Comparing base (86f34f4) to head (8e17dea).
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@            Coverage Diff            @@
##           master    #366      +/-   ##
=========================================
- Coverage    6.27%   6.26%   -0.01%     
=========================================
  Files          38      38              
  Lines       15232   15248      +16     
  Branches     3660    3660              
=========================================
  Hits          956     956              
- Misses      14262   14278      +16     
  Partials       14      14              

Components	Coverage Δ
casa	4.90% <0.00%> (-0.02%)	⬇️
handler	6.33% <ø> (ø)
spectral-cube	7.19% <ø> (ø)
utils	5.25% <ø> (ø)

[thomaswilliamsastro]

@EOakes it'd be good to get your eyes on this, I think this is basically the same fix as you put in in some other places not too long ago. For some reason, the pipeline spits stuff out as RA/Dec/Stokes/Freq, rather than our normally assumed RA/Dec/Freq/Stokes

[EOakes]

Looks good - this approach also matches what copy_mask in casaCubeRoutines.py already does.
Would also be good to correct the hardcoded blc = [0, 0, ichan, istokes] around line 1332, which I think faces the same issue.
Other hardcoded lines (~823, 892, 1118) are iterated over symmetrically, so they'll be mislabeled but won't break things if the axes are swapped.

[thomaswilliamsastro]

Right, I've edited that now so should be fine as well

[EOakes]

Looks good to me now.