returns random samples from the posterior, given a prior distribution m and n observed datapoints.
post_sample(m, data, nsamp)
an nsamp by n matrix
prior distribution (eg of type normalmix)
a list with components x and s, each vectors of length n, to be interpreted as a normally-distributed observations and corresponding standard errors
number of random samples to return for each observation
exported, but mostly users will want to use `get_post_sample`