These are distance estimates based on Gaia eDR3 data.

We've originally ingested these from CSV files rsynced up by Coryn
(which are in data/original_upload.tar.xz).

If we ever need to restore the table, do it from the dump (which is also
what ESAC used.

Test set was data/dist_3663.csv (to match the edr3lite test set);
to re-create later, extract healpix 3663.


Historically, I've dumped Coryn's input from Mintaka (a.k.a. ARI-Gaia)
using bin/dump_for_coryn.py (see instructions there), which he then
rsynced  down.  Note for next time: make sure the rsync only covers a
data directory.  This time around, we rsynced the whole resdir, which
was nothing but pain.