Trigger object type - useful for the application of trigger scale factors

[rmanzoni]

Each muon of the tau candidate now carries with it, as an attribute, the trigger object to which it is matched. This is valid for each trigger that is tested.

For example, for the main tau3mu trigger HLT_DoubleMu3_Trk_Tau3mu_v* (notice _v* is dropped)

triplet.mu1().best_trig_match['HLT_DoubleMu3_Trk_Tau3mu']
triplet.mu2().best_trig_match['HLT_DoubleMu3_Trk_Tau3mu']
triplet.mu3().best_trig_match['HLT_DoubleMu3_Trk_Tau3mu']

similarly, if another trigger, such as HLT_IsoMu24_v*, is tested, you'll also find

triplet.mu1().best_trig_match['HLT_IsoMu24']
triplet.mu2().best_trig_match['HLT_IsoMu24']
triplet.mu3().best_trig_match['HLT_IsoMu24']

If a given muon of the tripled did not match to any trigger object, for instance, because you only care about a one- or two- prong trigger, then you'll get None.

The trigger object types are defined here
http://cmslxr.fnal.gov/source/DataFormats/HLTReco/interface/TriggerTypeDefs.h

The most relevant to our case are:

TriggerMuon = +83,
TriggerTrack = +91,

This information can be accessed this way

triplet.mu1().best_trig_match['HLT_DoubleMu3_Trk_Tau3mu'].triggerObjectTypes()[0]

[rmanzoni]

@lguzzi @sarafiorendi

[rmanzoni]

Guess what, I had to change the code:

• now it is guaranteed that there are always 2 muons and a track
• in case of ambiguities, such those that arose from @lguzzi's tests, the assignment is such that the sum dR1 + dR2 + dR3 is at its minimum

Now you can access the trigger type by doing

event.tau3mu.best_trig_match[1]['HLT_DoubleMu3_Trk_Tau3mu'].triggerObjectTypes()[0]
event.tau3mu.best_trig_match[2]['HLT_DoubleMu3_Trk_Tau3mu'].triggerObjectTypes()[0]
event.tau3mu.best_trig_match[3]['HLT_DoubleMu3_Trk_Tau3mu'].triggerObjectTypes()[0]

From this commit on
fad6c47

[rmanzoni]

There remains a small point that can be improved: so far, for triggers with less than 3 objects at HLT, the matching is checked only for the first and possibly second muon in the triplet, not for any muons.

As long as we don't use any HLT other than HLT_DoubleMu3_Trk_Tau3mu we can live with this.