Make an EIA-generator-level table w/ granular FERC1 data

[cmgosnell]

We are now publishing the granular connection between FERC1's plant tables and EIA plant parts (out__yearly_plants_all_ferc1_plant_parts_eia). The backbone of this table is FERC1's plant_all_ferc1 (currently denorm_plants_all_ferc1 soon to be out_ferc1__yearly_all_plants), which is FERC's heterogeneously reported clumps of plants squashed together with whatever granularity is most appropriate from the EIA plant parts table. which is to say that one record is a whole plant while the next is a generator while the next is a plant's CT prime movers. Along with this table having lots of different granularities reported in it, you also need to have some understanding of how the EIA plant parts table works in order to make it talk to any of the other EIA tables. This is a pain/complex and it would be good to not need to expose users to it.

still. wahoo. big step forward. nonetheless, probably most users would rather a table with uniformly reported plant chunks. preferably the most granular - generators. This would require proportionally allocating FERC1 data that is reported at a larger granularity down to the level of the generator.

Luckily we've basically already done all of this over in this repo while prepping data for RMI's optimus model. Specially with the PlantPartScaler. One complication is the ownership slices. For the RMI outputs, each record is a generator/ownership slice (i.e. if a generator was owned 70% by utility A and 30% by utility B, that generator would have two records w/ all its data scaled to the % ownership).

So I think the main work here would be in migrating this code over into PUDL, updating it (bc it is a little stale), determining what to do about ownership and creating an out_ db table for the outcome.