This is aimed at removing the 'PS:Online' group label, potentially replacing it with other labels as needed.
Data-only triggers simply go in the PhysicsP1_pp_run3_v1.py menu, but this is no longer a single list of chains. Instead we define different functions:
addCommonP1Signatureswhich can then be used to add the P1 chains to the physics lists in different menus. For now I have just transplanted anything that had 'PS:Online' from the various menus (mostly in Calib, EnhancedBias and Streaming slices) into their respective
addXP1Signaturesfunction, with one exception:
HLT_noalg_L1RD0_FILLEDmove to the common function, as they were in all menus.
The current organisation has sizeable overlap between the P1 streamers in different cases, but this is needed in the split as defined above. If we wanted to improve on the segmentation we could imagine having another
addCollisionP1Signatures function or
addProtonProtonP1signatures to hold common P1 triggers that should not be in cosmics or HI, but at some point this is probably more confusing than having the duplication.
The only streamers retained in the Physics menu now are various jet and MET streamers wanted in MC for monitoring etc, but if there is a motivation we can restore other streamers to Physics or other menus that will be simulated. Clarified with Tim that this doesn't have any relevant cost, so it is more the triggers with nontrivial HLT that we would need to be careful with.
I didn't yet implement any new group labels, but we had at least one case to discuss, i.e. switching
HLT_timeburner_L1All off in reprocessings. We would need to plan where we want the removal to happen, e.g. using the prescale sets vs the RB -- labels could still be useful for both.
One other observation is outside of the scope of the MR, but 'HLT_noalg_CostMonDS_L1All' has to be included in all menus explicitly and it might be nice to centralise that addition somehow.