nespinoza
commented
8 months ago Hi! We could, yes --- but I'm not sure philosophically we should.
The reason is that the posterior parameters are supposed to be the parameters that you are actually constraining from your fit (ie, the ones you set priors to). Adding other derived quantities is left to the user in general a bit on purpose.
If there's hard push on this I think we could add it though. I would have to check this doesn't break something else down the road...
N.
I'm trying to fit a transit using r1, r2, and rho parameters. The posteriors.dat has an estimation of the semimajor axis (a) and the impact parameter (b), as well as inclination (i). I can see at some point that writepp( ) estimates these parameters and gets quantiles from samples. I was wondering if b, i and a could be retrieved from results.posteriors['posterior_samples']? Right now I couldn't find any other option than to copy the evaluation steps and insert my P and rho samples to get samples of i, b and a.