Updates in preparation for next release

[weaverba137]

This PR:

• Closes ModuleNotFoundError: No module named 'imp' #89
• Closes matplotlib deprecated cm.get_cmap #88
• Closes Padding method not propagated for filter group #64
• Closes speclite.filters.get_ab_maggies alters input wavelength array? #34
• Closes pad_spectrum does not fully obey axis argument #25
• Restores test coverage to 100%.
• Fixes the speclite_benchmark command.

[weaverba137]

Additional commentary: I tried to make as few changes as possible in this PR, and I think this is a release candidate for v0.20. However, I recommend that subsequent to this PR and tagging v0.20 that a major overhaul of the package infrastructure be performed, e.g. removing astropy_helpers, minimizing the use of setup.py. This isn't hard, since there are plenty of examples already in the Astropy and DESI ecosystems.

[weaverba137]

Bonus: I have been able to fix #25. For both #25 and #34, the issue was that FilterSequence.pad_spectrum was not passing the axis or method keywords to FilterResponse.pad_spectrum. That's all!

[dkirkby]

Looks great, thanks!

[weaverba137]

Thank you @dkirkby. I'll leave this open longer in case anyone wants to take a look: @segasai, @moustakas, @araichoor.

[segasai]

I can see that you added the previously failing example to the tests, so it looks good to me in terms of fixing #34
Thanks!

[weaverba137]

Thank you everyone, merging and tagging momentarily.

[moustakas]

Could v0.20 be installed at NERSC, please?

module spider speclite

--------------------------------------------------------------------------------------------------------------------
  speclite:
--------------------------------------------------------------------------------------------------------------------
     Versions:
        speclite/main
        speclite/master
        speclite/v0.13
        speclite/v0.15
        speclite/v0.16
        speclite/v0.17
        speclite/v0.18
        speclite/v0.19

[weaverba137]

Done, but only for desiconda/20240425-2.2.0.