feature/mx-1711 prepare artificial data extractor for roundtrip workflow

[cutoffthetop]

PR Context

• main motivation was to restore determinism, ie running artificial extractor twice with same seed should get you the exact same data. this was ruined by the weight per types, i reckoned. at least after removing that i did get consistent data. i also removed and fixed some unused config options.

Changes

• increase minimum valid artificial data count to two times the number of entity types

Removed

• remove matched setting for the artificial extractor, since that was not implemented
• stop configuring entity-type weights for artificial data, since that broke determinism
• removed unused -c alias for the count setting of the artificial extractor

[erichesse]

Alright looks good!