Implement single Gaussian model

[wtbarnes]

One of the simplest models for the DEM distribution is assuming a single Gaussian and then doing a chi-squared between forward modeled intensities and the observed intensities to get the nominal value, the mean, and and standard deviation of the Gaussian.

I have a proof of concept for this in this notebook, but it is very slow. I struggled with doing the fit using astropy models because the minimization is on the intensities rather than the model itself. My hand-rolled chi-squared minimization is basically useless for anything except a single spectra.

I'd be interested in seeing if someone else could work out the best way to do this. More generally, it would be interesting to provide a way to do this for any arbitrary astropy model. This would mean the user could specify any arbitrary compound model (e.g. double-gaussian, multi-gaussian + some other function) without having to implement a full model themselves.

[PaulJWright]

Hello Will!

I implemented this notebook in a sandbox version of sunkit-dem I am playing with. While I agree, it is slow, it is definitely a start, and perhaps that is what we really need in the package?

I am curious about the last paragraph though, and will think more about it.