A suite of analysis and diagnostics tools in R and python for working with
Markov chain Monte Carlo generally and Hamiltonian Monte Carlo specifically.
The suite includes functions for interfacing with RStan, PyStan2, and
PyStan3 and notebooks demonstrating their use.
These tools can also be interfaced with any Hamiltonian Monte Carlo code by
implementing appropriate extract_expectands, extract_hmc_diagnostics, and
plot_inv_metric functions.
Recommendations for code optimization are welcomed and appreciated.
I thank Sean Talts and Dan Waxman for Python code improvements. Raoul Kima originally suggested separating divergent transitions by numerical trajectory length.