Notes:
This file contains some example python code to find a beta distribution when given its confidence interval, e.g., it's 90%, 95% confidence interval.
I am just presenting a very barebones implementation, and I'm not producing the glue code to integrate this with the rest of Squigglepy. In particular, if I move the file one folder up, I encounter this error: AttributeError: 'module' object has no attribute 'Number'
It does so by means of calling my personal server, which is running an instance of this javascript package. I am not making guarantees about it, but I don't intend to take it down in the near future either.
The other possible approach would be to implement Nelder Mead, or other such method within Python. I don't think it would be fundamentally difficult, but it would take half a day to a few days worth of work, which feels wasteful. For that approach, see the documentation for the fit-beta repository