Tricks for making OFTI more amenable to RVs

[sblunt]

Lea and I discussed a few ways we might speed up OFTI for datasets including RVs. A summary:

• Remove gamma as a free parameter by marginalizing analytically (http://cadence.caltech.edu/~bfulton/share/Marginalizing_the_likelihood.pdf)
• Similarly marginalize over K, a la the Joker (https://iopscience.iop.org/article/10.3847/1538-4357/aa5e50/pdf)

OR

• Scale K to match min/max for particular generated model
• Translate gamma
• Potentially flip omega by 180deg to match data