Remove absorption_spectrum_fit code

[chummels]

It isn't clear that the absorption_spectrum_fit code actually works, and I'm not aware of anyone using it in several years. It has minimal testing coverage and just seems like the codebase would be simpler without it. I advocate for pulling it out of Trident and moving it to the yt attic repository.
Please speak up if you actually use this code. https://github.com/yt-project/yt_attic

[brittonsmith]

Might check on the mailing list?

[Anikbh11]

I've used this part only a couple times to see how it works. But I'm for sure going to use this feature once I have my light ray ion balance and absorption spectra sorted (hopefully by the end of next week). Primarily to fit the synthetic spectra I'll generate for these light rays. As far as I know, there are a couple of other chi-squared Voigt fitters that are widely used (VPFIT, AutoVP) which either are old or need manual work. Which is why I thought this feature within trident was very cool. If you say no one has tested it, I'll have to test it when I use it very soon anyway.

I've also written a Bayesian mcmc line fitting package which I was going to compare with this just as an exercise.