This fixes a bug with isolated nodes that contain only self-loops.

This solves a problem when NestedBlockState.mcmc_sweep() is invoked with
nonstandard values of the parameter 'c'.

This adds support for exact likelihood calculation, correct posterior
sampling of models with nonempty groups, much improved "distributed"
degree-corrected prior, as well as continuous edge covariates.

This is a significant refactoring of the statistical inference code,
with many cleanups and improvements, and is aimed towards future
extensibility.

The old community module is available as 'community_old' and will be
completely removed in the near future.