Minor mcmc refactor 770

[njtierney]

Resolves #770

[njtierney]

This PR does the following:

• Implements name attribute of samplers. E.g., hmc()$name returns "hmc". This should help make some future code for dispatch a bit simpler.

Some tidying up of samplers:

• Creates "sampler_class.R" and moved relevant parts of "inference_class.R" into that
• Moves R6 sampler objects into "samplers.R

Improves abstraction in run_samplers

• Created self$run_warmup and self$run_sampling functions. This means we can now write our own custom warmup or sampling methods for each sampler, if we need to (which we will need to be able to do in order to get adaptive HMC working)

Created build_samplers in mcmc

• Does the lapply(...build_sampler) step, plus the other bits around that, which are a bit distracting from the overall intention

Adds some helper functions to re-express code intent

• create_empty_matrix
• set_progress_bar_type
• inform_if_remote_machine
• inform_if_local_parallel_multiple_samplers
• sampler_parallel_reporting - manage complexity around writing to log files and greta stash

I couldn't quite get the abstraction of the progress bars down. I think there is scope for that, but in the grand scheme of things, it's actually quite a small issue, it's only about 25 lines of code. It might be a nice idea to come back to as something bite sized. I'll make an issue for it.