pyblocxs: It is a Python Sherpa Extension for Bayesian Low Counts X-ray Spectral Analysis. It provides MCMC methods to explore a model parameter space and summarized the full posterior or profile posterior distributions. Computed parameter uncertainties can include calibration errors. It simulates replicate data from the posterior predictive distributions. It can test for added spectral components by computing the Likelihood Ratio Statistic on replicate data and the ppp-value.
It has two samplers:
(1) Metropolis-Hastings: centered on the best fit values
(2) Metropolis-Hastings mixed with Metropolis jumping rule: centered on the current draw of parameters with the prior scale specfied as a scalar multiple of the Sherpa covar() output.