The affected properties were Xraylum , Xrayphlum , Xraylum_no_agn and Xrayphlum_no_agn.

All looks good to me. The fix matches what is done for other similar quantities, like the luminosities.

As a quick check, could @wullm or @jemme07 try on a snapshot of the recent runs?

I am runnig the z=0 Jets box atm.

We also need to merge the chages in this other branch: https://github.com/jchelly/SOAP/compare/master...SWIFTSIM:SOAP:flamingo-v2

@jchelly @MatthieuSchaller. Runnig the script halo_properties_L1000N1800.sh, which calls compute_halo_properties.py failed for the Jets run with the following message:

*** EXCEPTION ***
'NoneType' object has no attribute 'shape' on rank 126

Traceback (most recent call last):
  File "/snap8/scratch/dp004/fkgm22/SOAP//compute_halo_properties.py", line 477, in <module>
    compute_halo_properties()
  File "/snap8/scratch/dp004/fkgm22/SOAP//compute_halo_properties.py", line 414, in compute_halo_properties
    metadata = task_queue.execute_tasks(
  File "/snap8/scratch/dp004/fkgm22/SOAP/task_queue.py", line 195, in execute_tasks
    result.append(task(*args))
  File "/snap8/scratch/dp004/fkgm22/SOAP/chunk_tasks.py", line 265, in __call__
    total_time, task_time, nr_left, nr_done = process_halos(comm, cellgrid.snap_unit_registry, data, mesh,
  File "/snap8/scratch/dp004/fkgm22/SOAP/halo_tasks.py", line 273, in process_halos
    halo_result = process_single_halo(
  File "/snap8/scratch/dp004/fkgm22/SOAP/halo_tasks.py", line 118, in process_single_halo
    halo_prop.calculate(
  File "/snap8/scratch/dp004/fkgm22/SOAP/subhalo_properties.py", line 942, in calculate
    assert subhalo[name].shape == val.shape, f"Attempting to store {name} with wrong dimensions"
AttributeError: 'NoneType' object has no attribute 'shape'

That has been caused by the check I put in to prevent further bugs like this one. The calculation can fail if it turns out too small a region was read in, in which case it returns None and it shouldn't try to verify the dimensions of the result. I thought I dealt with that case, but apparently not. I'll take another look at it.

I think I just missed a "if val is not None" in the subhalo properties case.

Thanks John. Jaime, could you test it once more?

Ah wait, looks like no changes were actually made in the branch. Just an analysis of what went wrong.

@jemme07 could you open a separate pull request with the changes you mentioned above?

Yes, sorry, I looked at it a bit more and wasn't 100% sure where those None values were coming from so I haven't pushed a fix yet. I'll take another look at it now.

I think the Nones actually come from halos which have no particles of some type. That doesn't change the fix though.

@jemme07 could you open a separate pull request with the changes you mentioned above?

I alread have. It is here: https://github.com/jchelly/SOAP/pull/19

It is all working now. I tested it with the HYDRO_STRONG_JETS model. The z=0 SOAP catalogue is temporarily stored here:
/cosma8/data/dp004/fkgm22/FLAMINGO_SOAP/ScienceRuns/L1000N1800/HYDRO_STRONG_JETS/SOAP/

Once I have write permissions in the FLAMINGO data directory I'll copy it over.

@jemme07 could you open a separate pull request with the changes you mentioned above?

I alread have. It is here: jchelly#19

I don't understand. If this is missing in the main version then we should merge here not in John's version. We should really try to keep a single maintained official version otherwise we gonna have confusion about what to run.

Thanks for checking it all works.

I'll accept.