Started this because it was both the tallest pole in the QG generation profiles and a schema change. But now I think I'm going to go do DM-24938 first, as that'll inform how I want to structure the tables to make sure the lookups in QG generation are fast. There is a branch here, but it just adds an enum value and documentation for it.

I think I see a path towards merging this before implementing much of the query/dimensions changes prototyped on DM-26336. That will leave some clearly temporary code in pipe_base (basically, we'll need to special-case both calibration lookups and refcat lookups in QG generation), and it will defer the big QG gen optimization until later as well. I suspect this ticket would yield at most modest improvements to QG gen performance, but it would still include the big schema and calibration-certification changes, and that means it's still worth landing as soon as possible, given that most of it is already done.

Calling this done as its child tickets are done.  The QG generation optimization work once planned for this ticket will happen on DM-26692.