How are we using etl_group?

[bendnorman]

It seems like the meaning of the etl_group attribute of resources objects might be outdated? I think it would be nice to align our notion of etl_group with our Asset Groups. Asset groups right now are determined by the resource/assets's datasource and processing layer: {layer}_{datasource}.

Why did we originally create etl_group? It's being used in various places as a way to filter out groups of assets we might not want metadata for.

[zaneselvans]

It was largely a convenience -- just a way to address various groups of tables that needed to be iterated over in the ETL rather than compiling lists of them. If there's a better way to group assets now, then we can use that instead, but I doubt that our current asset groups map cleanly to the structural similarities that were reflected by etl_group and IIRC actually accessing the list of assets in an asset group was kind of a pain in the ass.