Skip to content

New nightly trigger tests for ITk using TIDA

Elliot Watton requested to merge ewatton/athena:ewatton_ITkFTFTests_TIDA into main

This MR adds three new tests to the TrigInDetValidation package running the trigger with ITk. This means we would now be able to monitor the triggers performance on an ITk dataset using TIDA with each nightly build. The three tests are:

  • An electron test (single 100 GeV electron events)
  • A muon test (single 100 GeV muon events)
  • A ttbar test (semileptonic, 200 pileup)

This MR includes quite a few code changes and additions, so I will briefly go over them:

  • Three files, one for each test, is added to the TrigInDetValidation/test folder. These follow the same format as the Run3 tests currently in the folder. They include a variable called "useCA_Reco", which is defined to be True. This is required because in TrigInDetArtSteps.py the configuration of the Reco_tf command that the job uses needs to be different for these Run4 jobs. In particular, for the current Run3 setup, the RDOtoRDOTrigger step of Reco_tf does not use CA by default (--CA "default:True" "RDOtoRDOTrigger:False"). The use of this "useCA_Reco" (it's a working variable name -> can be changed to something more appropriate) defines the value of the pre-exisitng _isCA member of the ExecStep class, such that the CA is used for the RDOtoRDOTrigger step. Additionally, the setting of _isCA is used in other conditions to further configure the Reco_tf command for these new Run4 tests. _isCA is False by default (to my knowledge, please correct me if I am wrong!)

  • Connected to the point above, TrigInDetArtSteps.py needed modifications such that the Reco_tf command steps away from the Run3 version to the one needed for Run4.

  • TrigInDetValidation_Base.py had to be modified to correctly handle the "useCA_Reco" variable such that Run3 should not be affected, but it is still set to True for the Run4 tests.

  • New TIDA files added, and the cuts file modified, to allow a different eta cut be used for Run4. However, this is only a step in the right direction as there are more changes required including some hardcoding of eta ranges up to |eta|=2.5 or 3.0 in the TIDA source code when making the summary plots at the end with TIDAcomparitor. @sutt, what do you think is the best course of action here? We could try to change TIDA in to be more dynamic to allow for the different eta cut for Run4, but I understand that might not be so simple.

  • Added the Run4 single electron file RDO recently put on CVMFS into TrigValInputs.json for the electron test.

Tagging @baines @smh @demelian @jmasik @sutt @hartj

Edited by Elliot Watton

Merge request reports