Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.76%. Comparing base (648a595) to head (ccff16e).
⚠️ Report is 1 commits behind head on main.

@@           Coverage Diff           @@
##             main    #3355   +/-   ##
=======================================
  Coverage   89.76%   89.76%           
=======================================
  Files          29       29           
  Lines       31292    31292           
  Branches     5738     5738           
=======================================
  Hits        28089    28089           
  Misses       1794     1794           
  Partials     1409     1409           

Mind pointing to the part(s) of the code from which we know this is accurate?

Mind pointing to the part(s) of the code from which we know this is accurate?

Well, I think the logic is here, but I'm confused by the num_descendants thing in the C code (do individuals have "descendants" that can be counted?)

Either way, we definitely need to sort first by putting parents before children, otherwise the canonical sort will not created tables that meet the individual-table requirements.

Well, individuals have a parent attribute, and this is being used to count the number of descendants (as measured within the individual-parent relationships only): see here.

The relevant code is here, I guess - looks like it sorts first by number of (individual-parent) descendants, then by referring node ID, then by original order.