Create FFPR script

[zhangderby]

No description provided.

[zhangderby]

First pass at FFPR has yielded some positive results and a bit of weirdness that I'm still trying to figure out.

The following results are for the simple case of decentering M2 50 microns in both x and y relative to it's nominal alignment. The resulting field-dependent PSFs are:

Feeding this into FFPR, we can see the first major improvement in how quickly the optimizer converges. This is 2-3x faster than the previous "bing bong" method:

And here we can see that the wavefront estimates are quite good as well:

Now the weirdness. FFPR is supposed to jointly estimate the field-dependent aberrations and how they evolve over the field. This is also used when estimating the wavefront error of individual PSFs. We saw above that individual wavefront errors were estimated well, but below we can see that FFPR is completely wrong when estimating field-dependent aberrations: