There is currently no way with gen3 to ingest a directory of files without writing a small program each time. We need the equivalent of the gen2 ingestImages.py task.

This could go into obs_base although we have to understand how it's configured, and how it would differ for DECam data where there are multiple datasets in a single file.

There are open questions concerning how we tell the command what type of instrument is being ingested and whether there should always be a separate command to prepare a butler repository for that instrument.