Transform wind and solar aggregations into core tables

[zaneselvans]

Tasks

Beta Give feedback

1. Split composite aggregation keys into plant & generator IDs
2. Rename columns using PUDL naming conventions
3. Set column dtypes
4. Create field and resource definitions.
5. Add FK constraints linking aggregations to EIA plant & generator tables
6. Debug plant_id_eia and generator_id FK constraint violations
7. Fix generator IDs with leading zeroes at the beginning of the transform (unnecessary)
8. Replace plant_id_eia 64272 with 64481 at the beginning of the transform (unnecessary)
9. Use an @asset_check() to ensure that all aggregation keys appear in aggregation table
10. Ask Ana / Elaine whether Hybrid_Solar_64481 should include both generators AEC1 and AEC2.
11. Check hour starting vs. ending conventions in our hourly tables.

[zaneselvans]

To nobody's great surprise, the FK constraints between the aggregation tables and the generator entity tables are being violated. Need to investigate the details of that and see why, or if some of the generator IDs have been munged badly upstream.

There were 5 bad generator IDs.

4 of them had leading zeroes and so didn't match the generator IDs in PUDL (which we've purged of leading zeroes, since they are not reported consistently across years, and removing the zeroes results in consistent IDs)

The last bad ID was plant_id_eia==64272 and generator_id=="AEC1". That plant_id_eia does not exist in the PUDL database at all. However, plant_id_eia==64481 is the only plant with a generator_id=="AEC1" and it's solar, and has the capacity expected from the missing generator, and that plant also has an AEC2 generator which shows up with the right plant_id_eia in the GridPath data, so I think we should probably just change the plant ID.