Allow hierarchical analysis to take in h0 posterior samples
Currently, the MassQuadrupoleDistribution
assumes that the input posterior samples are either sampled from the mass quadrupole (Q22
) or ellipticity (ELL
). However, in our normal analyses, we actually sample in h0. I had purposely wanted to use samples in Q22
/ELL
, so that any uncertainties on distance could be naturally incorporated in their inference. It would still be useful to allow h0 samples to be passed with fixed distance and rotation frequency values for each source. This would rely on the prior in h0 being uniform and true normalisations not mattering.
I'll including this within !179, so as not to get clashes.