Benchmark ready

Fixes several bugs and improves documentation and usability compared to (v0.1) alpha. Early testing on public data has been completed with all models reported in the paper and results appear to align so far.