My main reason for dropping these table writes into the make-the-plots tasks was to capture the same tables being used to generate the plots, with as little development as possible---certainly not to be economical. If we want to make this its own Task, it seems like that could (should?) be completely divorced from the current plotting scripts. This relates to what we talked about in the meeting yesterday about maybe with the interactive plots we shouldn't do even the pre-culling that the pipe_analysis scripts are doing (which I'm currently capturing with my table writes).
I do agree in principle that the table-preparation should eventually be an independent step (in which case, the catalog preparation within the plotting tasks should probably also be broken out into its own task), but before we get there we want to try to view the whole PDR1 with the interactive plots (
DM-12043), and I think the quickest way to get there is with the writes just snuck in there.
I'm happy to make the import of writeParquet explicit, but as you point out, that's just a drop in the bucket, so I figured it probably wasn't worth it.