Include additional posterior components of interest

[MansMeg]

posterior (default) posterior predictive prior prior predictive elementwise loglikelihoods

The question is how to do this efficiently. Probably by including likelihood functions that take the data and posterior gold standard and produce it.