production but is called something else) and registering "legacy" datasets:
* eliminate old script create_registry_db.py
* add column to provenance table to keep track of associated production db
* make all column names lowercase (easier to deal with in psql)
* hide imports of git as much as possible.  It's not required for typical
  user, only for those creating schemas
* try to keep line length <= 80 in most cases

…ersion