Desired feature: MCMCfit only fit amplitude (not XY)

[Mikea1985]

Sometimes I'm confident in the centroiding (from SExtractor or whatever) and just want to fit the amplitude so that I can make an image with the object subtracted out, and don't want to wait for the time-intensive 3-step MCMC fit of xy, amplitude, xy+amplitude. Also, sometimes I find that the MCMC fit runs off into lalaland and doesn't actually centroid on the source (not sure what's up with that), in which case the fit isn't useful. It would be really nice if there was a way to ask MCMCfit to only fit the amplitude, maybe with a fitXY=True/False keyword.

I tried implementing this, but failed misserably.

[fraserw]

This could be implemented with small changes to MCMCfit and the realization that lnprob can fit xym, only xy, or only m. With appropriately exposed options in the MCMCfit call, this wouldn't be too difficult. Will work on it

[Mikea1985]

I think my Pull Request #36 actually fixes this old issue.