streamline gp_baseline workflow

[ncatlett]

Currently difficult to extend to new namespaces - requires modifications to gp_baseline, parsers, parsed, datasets, config, and namespaces/equivalence scripts

[ncatlett]

Methods for datasets now better aligned, so outputs match and default methods can be inherited. namespaces.py removed and replaced with write_ns_values method in datasets.
Still needs significant work, particularly in gp_baseline around loading/re-loading of data for starting workflow in different phases.