Some changes had to be made to the LMSimpleShape plugin. These were of two types:

(1) I changed the code to use the LMShape Runner at Erin's suggestion, and included a "Guesser" instead of creating a guess by myself.

(2) I changed the output of LMSimpleShape to include several output values from the final fit, including x,y,flux,e1, and e2.

I formerly believed that these could all be derived using a MultipleShapeletFunction created from the Gaussian mixture (recorded as a set of Gaussian ShapeletFunctions). But our MSF has limitations which meant that I had to use the GMix values produced by Erin's code.

Looks good; my only comments (on the PR) are cosmetic, but still need to be addressed before merging (including the comment on the PR main page about the commit message).

Modification made per review
Ran Jenkins build successfully

I'm a bit confused that this ticket is closed but the associated pull request is still open.

Looks like a rebased version was merged without being pushed to Github.