Propagate XBRL column annotations to finished tables

[zaneselvans]

The FERC XBRL Taxonomy contains lots of useful metadata, which generally gets associated with columns in the DB we build from the XBRL filings. When we reshape those tables so they can be combined with the DBF data in a long / tidy format, many of those columns end up turning into rows, or categorical values that allow the selection and aggregation of certain types of values (e.g. by FERC Account). In the process, we lose easy access to the original XBRL DB column annotations, which make the final output database harder to interpret.

We need a way to preserve these annotations, and integrate them into our data dictionaries -- potentially with a structure similar to what we do right now with the code definition tables that define the meanings of various standardized codes which show up all over the place in the PUDL DB.

E.g. what does it mean for Utility Plant to be leased from others vs. in service vs. for future use?

This might also involve adopting appropriate ENUM or foreign key constraints on categorical columns