iscc.inference.Prior

Prior(specs)

A product prior over named parameters.

Each parameter maps to a sampler, given as one of: * (low, high) tuple -> Uniform(low, high); * a frozen scipy.stats distribution (anything with .rvs / .pdf); * a callable rng -> sample (no density; logpdf falls back to a flat 0).

Methods:

Name	Description
sample	Draw n parameter sets -> (n, n_params) array, columns ordered as names.

sample

sample(rng, n)

Draw n parameter sets -> (n, n_params) array, columns ordered as names.