minor refactor of mcmc internals

[njtierney]

In attempting to do #765 via #766 I've found some parts of the code a bit harder to wrangle and reason with, and I've gotten stuck in a few places.

Specifically:

• sampler methods don't have a "sampler_name" slot - it can be accessed by "class", but it doesn't feel idiomatic to R6
• progress bar creation clutters the intent - e.g., in run_chain,run_samplers
• improve clarity in run_samplers for when they are running warmup vs sampling
• improve abstraction in run_samplers so we can perform different warmup vs sampling steps based upon the sampler used.

I'm going to try doing a small refactor of the code to make it a bit easier to reason with, before coming back to either start #766 again or branch off of #766 .