GarethCabournDavies
commented
2 years ago I think you're right on the hierarchical removal code. However I think it does do the removal, just doesn't write it to file :facepalm:.
The check/break for whether the max hierarchical level has been reached is on this line: https://github.com/gwastro/pycbc/blob/master/bin/all_sky_search/pycbc_add_statmap#L399
However the HR foreground and background haven't been written to the file at this point.
I think this should be solved by moving the max_hierarchical_removal check to the same place as the check/break for if the hierarchical removal ifar limit has been reached: https://github.com/gwastro/pycbc/blob/master/bin/all_sky_search/pycbc_add_statmap#L471
There is a problem in pycbc_ifar_catalog if the number of signals in the data exceeds the "max_heirarchical_removals".
This basically can be seen by following a few lines. First here:
https://github.com/gwastro/pycbc/blob/master/bin/plotting/pycbc_ifar_catalog#L83
Where the number of removals is set. Lets say that this is 10.
Then here the code tries to access the "h10" entry:
https://github.com/gwastro/pycbc/blob/master/bin/plotting/pycbc_ifar_catalog#L122
However, with these settings the file will contain h0, h1 ... h9. It will not contain h10.
Now that I write this, it might be an off-by-one error in the heirarchical removal code itself if h9 means "I've done 9 removals". Perhaps it will always do one fewer removal than what was requested?