Implement four spatial profiles

[Russell-Ryan]

We should consider three types of profiles for the cross-dispersion weighting, in order of priority:

1. flat/uniform profile. This provides the default case
2. Analytic form. This should be a Gaussian or Moffat, but other forms may be considered in the future. Care should be taken in constraining the fit to be realistic, such as the peak value must be positive, the centroid should not move far from the trace expected position (say +/- 1 pix), and the width parameter should be consistent with the known value (say require width <=2 pix, depending on how width is defined).
3. Empirical model. This would implement an empirically derived cross dispersion profile, this likely won't come until sometime in the mission. We can build a "model" empirical model from galsim simulations to build the algorithm. This should be consider "experimental", but expectations are this will give the highest S/N with the least spectral bias.