verify the doublet ratios and central wavelengths of the weak nebular lines

[moustakas]

In some high signal-to-noise spectra of M31 HII regions [OIII] 7330 is not being fit well, e.g.--

Review all the central wavelengths and emissivities for the weak nebular lines

[moustakas]

It turns out that [OII] 7320 is itself a doublet (it's a magnetic dipole and electric quadrapole), as is [OII] 7330. I updated the atomic data and now I'm using the emissivity-weighted wavelengths for each of the doublets. In addition, I now fix the 7320/7330 doublet ratio to be 1.2251, not 1.9085 as before, which was very wrong because it ignored the double-doublets.

Some documentation can be found here-- https://github.com/desihub/fastspecfit/blob/m31/doc/nb/nebular-lines.ipynb

And here's the updated fit for this particular object (28675135242244), with other spectra looking just as good (although [SIII] seems off now...).

[moustakas]

Done in #87.

[SIII] 9069,9532 will likely need some updates, but I'll open a separate ticket for that doublet.