Based on the results of DM-31281, create scripts for each ap_verify dataset that can regenerate the refcats to be used with that dataset's science processing. Consult with John Parejko on the most appropriate way to do this (e.g., generating new refcats like the existing ap_verify_hits2015/scripts/gaia_HiTS_2015.py does, or simply grabbing a subset of the /repo/main refcats). I am assuming that the solution to DM-31681 will not need special refcat support, as template generation will take place within a development repository.

This issue covers making scripts for ap_verify_ci_hits2015 and ap_verify_hits2015, but it may also cover ap_verify_ci_cosmos_pdr2 if no dataset-specific knowledge is needed to do so. Otherwise, support for ap_verify_ci_cosmos_pdr2 can wait until DM-29857 finalizes the complete workflow.