search

pynbody / tangos

The Agile Numerical Galaxy Organisation System
BSD 3-Clause "New" or "Revised" License
18 stars 13 forks source link

consistent_trees_ids missing for phantoms #217

Open anchwr opened 1 year ago

anchwr commented 1 year ago

Consistent_trees_ids are no longer appearing for phantoms, although they still appear for halos when the --with-ids option is used with import-consistent-trees. I'm still trying to track down the commit (likely one of mine) that did this.

anchwr commented 1 year ago

It looks like ln 116 of consistent_trees_importer.py is returning None for phantoms. I'm wondering if this is because the phantom finder_ids have also gone from being extremely large numbers to being the same as their halo_numbers. Was this intentional? I know there were some changes made a while back to make consistent_trees_importer.py compatible with unsigned int finder_ids.

anchwr commented 1 year ago

It looks like the extremely large numbers that were being used as finder_id values were close to the rockstar 'magic' parameter (i.e., 18077126535843729616). I'm struggling to find documentation of the origin of this parameter or any sign of it being used for anything beyond basic dataset type validation in yt (or of it ever being directly accessed by tangos). Adopting something like this for phantom halo finder_id values again could be useful, though. Would it be reasonable to start these at maybe one order of magnitude higher than the maximum halo finder_id value at a given timestep?