Adding functions related to Bayesian hypothesis testing

[sethaxen]

Python ArviZ includes the following functionality related to Bayesian hypothesis testing:

• region of practical equivalence (ROPE, e.g. plot_posterior)
• Bayes factors (plot_bf)
• Bayesian p-values in section 6 of BDA3 (plot_bpv)

While these are not part of the stats API in Python ArviZ, we should add them here, with appropriate suggestions/qualifications for their use.

Additionally, in TuringLang/MCMCDiagnosticTools.jl#90 @DominiqueMakowski suggested including the probability of direction, which is related to a p-value on significance of an effect.

See also: bayestestR