TWiki> Main Web>TWikiUsers>JordiDuarte>HiggsTriggerValidation (2012-05-02, JordiDuarte) EditAttachPDF

Higgs Trigger Validation

The cmssw release trigger validation is done within the PdmV group. Each release should be validated, just comparing with some release of reference, and reported each new release period at PdmV Trigger Validation twiki

Code Integration

The code can be found at cvs repository in HLTriggerOffline/Higgs. The tag conventions XX-YY-ZZ are as following:
  • XX: increases by one when the code breaks off with respect the old one. Therefore, the code is incompatible with the last tag. Some examples: trigger paths removed from the hltHiggsValidator_cfi.py, folders (analysis) removed from the validation,...
  • YY: increases by one when new functionalities have been incorporated in the code without breaking it. Code compatible with the last tag. Some examples: trigger paths added to the hltHiggsValidator_cfi.py, folders (analysis) added, histograms or plots added, ...
  • ZZ: increases by one mainly when a bug(s) has been fixed.

Tag RELEASES DQM integrated Description
V01-00-00 52X (missing 60Y), where X>=3 New code released
V01-00-01 52X, 60Y, where X>=3 Fixed Error -> Warning when a trigger path is not found, Removed HLT_LooseIsoPFTau35_Trk20_MET75

Tag Collector

Code description

Code Validation Process

Compilation and previous checks: using a 5_2_0 release from scratch

In order to check all the code is in good place, with all the needed libraries and so on: downloading from scratch in a clean lxplus environment: 
ssh -X lxplus.cern.ch
[duarte@lxplus440:~/scratch0/TEST_CMSSW]$ export SCRAM_ARCH=slc5_amd64_gcc462
[duarte@lxplus440:~/scratch0/TEST_CMSSW]$ cmsrel CMSSW_5_2_0
[duarte@lxplus440:~/scratch0/TEST_CMSSW]$ cd CMSSW_5_2_0/src/  && cvs co HLTriggerOffline/Higgs
[duarte@lxplus440:~/scratch0/TEST_CMSSW/CMSSW_5_2_0/src/HLTriggerOffline/Higgs]$ scramv1 b -j8
>> Local Products Rules ..... started
>> Local Products Rules ..... done
>> Entering Package HLTriggerOffline/Higgs
>> Creating project symlinks
  src/HLTriggerOffline/Higgs/python -> python/HLTriggerOffline/Higgs
Entering library rule at HLTriggerOffline/Higgs
>> Compiling edm plugin /afs/cern.ch/user/d/duarte/scratch0/TEST_CMSSW/CMSSW_5_2_0/src/HLTriggerOffline/Higgs/src/EVTColContainer.cc 
>> Compiling edm plugin /afs/cern.ch/user/d/duarte/scratch0/TEST_CMSSW/CMSSW_5_2_0/src/HLTriggerOffline/Higgs/src/HLTHiggsPlotter.cc 
>> Compiling edm plugin /afs/cern.ch/user/d/duarte/scratch0/TEST_CMSSW/CMSSW_5_2_0/src/HLTriggerOffline/Higgs/src/HLTHiggsSubAnalysis.cc 
>> Compiling edm plugin /afs/cern.ch/user/d/duarte/scratch0/TEST_CMSSW/CMSSW_5_2_0/src/HLTriggerOffline/Higgs/src/HLTHiggsValidator.cc 
>> Compiling edm plugin /afs/cern.ch/user/d/duarte/scratch0/TEST_CMSSW/CMSSW_5_2_0/src/HLTriggerOffline/Higgs/src/MatchStruct.cc 
>> Compiling edm plugin /afs/cern.ch/user/d/duarte/scratch0/TEST_CMSSW/CMSSW_5_2_0/src/HLTriggerOffline/Higgs/src/SealModules.cc 
>> Building edm plugin tmp/slc5_amd64_gcc462/src/HLTriggerOffline/Higgs/src/HLTriggerOfflineHiggs/libHLTriggerOfflineHiggs.so
Leaving library rule at HLTriggerOffline/Higgs
@@@@ Running edmWriteConfigs for HLTriggerOfflineHiggs
--- Registered EDM Plugin: HLTriggerOfflineHiggs
>> Leaving Package HLTriggerOffline/Higgs
>> Package HLTriggerOffline/Higgs built
>> Local Products Rules ..... started
>> Local Products Rules ..... done
gmake[1]: Entering directory `/afs/cern.ch/user/d/duarte/scratch0/TEST_CMSSW/CMSSW_5_2_0'
@@@@ Refreshing Plugins:edmPluginRefresh
>> Creating project symlinks
>> Done python_symlink
>> Compiling python modules cfipython/slc5_amd64_gcc462
>> Compiling python modules src/HLTriggerOffline/Higgs/python
>> Compiling python modules python
>> Pluging of all type refreshed.
>> All python modules compiled
gmake[1]: Leaving directory `/afs/cern.ch/user/d/duarte/scratch0/TEST_CMSSW/CMSSW_5_2_0'
So the code compiles fine and is ready to be tested Yes / Done

RelVal Testing procedure. DQM group instructions

For introduce the code in the DMQforMC it is needed to proceed with the instructions found in RelValTesting twiki page. The test procedure is divided in three steps. The procedure has been followed for the Full Simulation. It was run at a lxplus from scratch, so in order to reproduce it in a lxplus don't forget: 
ulimit -S -v unlimited
  1. Used the MC sample: /RelValH130GGgluonfusion/CMSSW_5_2_0-START52_V4A-v1/GEN-SIM-DIGI-RAW-HLTDEBUG (Note that if you want to generate your our datasample, you can do it using 
     cmsDriver.py SingleMuPt10.cfi -s GEN,SIM,DIGI,L1,DIGI2RAW,HLT:GRun -n 10 --eventcontent FEVTDEBUGHLT --conditions auto:mc --mc --no_exec
    The SingleMuPt10.cfi file was extracted from Configuration/Generator/python where there are other options...)
  2. Run the combined RelVal+DQMOffline histogramming (step2) 
      cmsDriver.py step2 -s RAW2DIGI,RECO,VALIDATION,DQM -n 10 --filein /store/relval/CMSSW_5_2_0/RelValH130GGgluonfusion/GEN-SIM-DIGI-RAW-HLTDEBUG/START52_V4A-v1/0000/34331C19-FE68-E111-A291-001A9281171E.root --eventcontent FEVTDEBUGHLT --conditions auto:mc --mc --no_exec
    You can search for HLTDEBUG samples in DBS (after initialized the environment with the cmssetup) 
    dbs search --query='find dataset where dataset = *RelVal*5_2_0-*GEN-SIM-DIGI-RAW-HLTDEBUG*'
  3. Run the combined RelVal+DQMOffline harvesting (step3) 
     cmsDriver.py step3 -s HARVESTING:validationHarvesting+dqmHarvesting --harvesting AtRunEnd --conditions auto:mc --filein file:step2_RAW2DIGI_RECO_VALIDATION_DQM.root --mc --no_exec

After the steps a root file 

DQM_*_CMSSW_X_Y_Z__RECO.root
has been created a filled with the histograms which have been defined. All seems to be OK Yes / Done

Trigger group procedure. TSG group instructions

TSG group instructions

Package description

The validation is done between releases by comparing the efficiencies of some selected triggers used by several Higgs analysis. These efficiencies are calculated as follows,
  • DENOMINATOR pt,eta,phi of the gen/reco objects (which pass some analysis dependent requirements)
  • NUMERATOR pt,eta,phi of the gen/reco PASSING the monitored trigger

The efficiencies calculations are done for HWW, HZZ, H, Htaunu, H2tau and Hbb, analysis within the Higgs PAG. Actually, and in the context of MC release validation, the efficiencies are going to be calculated in events which contain the objects of interest in each analysis, where have been applied some loose and minor cuts. A set of the more critical triggers for each analysis are being monitored whenever a new release is introduced. After consulting the trigger experts, the trigger paths to be monitored are: ( STILL IN PROCESS TO BE DEFINED )

Analysis Cuts Triggers
HWW 2 $\ell$ with $p_t > 10\;GeV/c$, $abs(\eta) < 2.4$ AND ( ($\ell==\mu$ AND isGlobalMuon ) OR ($\ell==e$ AND H/E < 0.05 AND 0.6 < E/p < 2.5)) HLT_Mu17_Mu8
HLT_Mu17_TkMu8
HLT_Mu8_Ele17_CaloIdT_CaloIsoVL_TrkIdVL_TrkIsoVL
HLT_Mu17_Ele8_CaloIdT_CaloIsoVL_TrkIdVL_TrkIsoVL
HLT_Ele17_CaloIdT_CaloIsoVL_TrkIdVL_TrkIsoVL_Ele8_CaloIdT_CaloIsoVL_TrkIdVL_TrkIsoVL
HZZ 4 $\ell$ with $p_t > 10\;GeV/c$, $abs(\eta) < 2.4$ AND ( ($\ell==\mu$ AND isGlobalMuon ) OR ($\ell==e$ AND H/E < 0.05 AND 0.6 < E/p < 2.5)) HLT_Ele17_CaloIdT_CaloIsoVL_TrkIdVL_TrkIsoVL_Ele8_CaloIdT_CaloIsoVL_TrkIdVL_TrkIsoVL
HLT_Mu17_TkMu8
HLT_Ele20_CaloIdVT_CaloIsoVT_TrkIdT_TrkIsoVT_SC4_Mass50
HLT_Mu24_eta2p1
HLT_Mu5_Track3p5
Hgg 2 photons with $p_t > 30,20\;GeV/c$, $H/E<0.1$ AND { R9 > 0.85 OR { $ \sigma_{i\eta i\eta} < 0.014\;(0.0035$ if endcap) AND $\sum_{ECAL} E_T <5.0 + 0.012E_T$ AND $\sum_{HCAL} E_T <(5.0 + 0.005E_T)$ AND $\sum_{ECAL} E_T <(5.0 + 0.002E_T)$ ] } HLT_Photon26_R9Id85_OR_CaloId10_Iso50_Photon18_R9Id85_OR_CaloId10_Iso50_Mass50
HLT_Photon36_R9Id85_OR_CaloId10_Iso50_Photon22_R9Id85_OR_CaloId10_Iso50
Htaunu MISSING INFORMATION HLT_LooseIsoPFTau35_Trk20_MET75
H2Taus MISSING INFORMATION HLT_Ele20_CaloIdVT_CaloIsoTRho_TrkIdT_TrkIsoT_LooseIsoPFTau20
HLT_IsoMu18_LooseIsoPFTau20
Hbb   HLT_Mu17_Mu8
HLT_Ele17_CaloIdT_CaloIsoVL_TrkIdVL_TrkIsoVL_Ele8_CaloIdT_CaloIsoVL_TrkIdVL_TrkIsoVL
HLT_IsoMu24_eta2p1_CentralPFJet30_CentralPFJet25_PFMHT20
HLT_Ele27_WP80_CentralPFJet30_CentralPFJet25_PFMHT20
HLT_DiCentralPFJet30_CaloMET50_dPhi1_PFMHT80_HBHENoiseFiltered_dPhi1
HLT_Jet80Eta1p7_Jet70Eta1p7_L1FastJet_DiBTagIP3D
HLT_Mu12_eta2p1_DiCentralL1FastJet_40_20_DiBTagIP3D1stTrack
HLT_Mu15_eta2p1_TriCentralL1FastJet_40_20_20_DiBTagIP3D1stTrack
The trigger paths also can be found, and probably with the latest changes, on this spreadsheet

Configuration

The hltHiggsValidator_cfi.py configuration file is the place where to include the path to be monitored

Validation: defining the baseline release

After put the new code in place and define the trigger paths we want to monitor, we have to define the baseline plots which are going to be used as reference. We have to found a release we agree is ok: 5_2_2 has been validated by all the other PAG/POG, so we would assume it is ok. The samples needed to be able of validate all the channels are not fully defined, trying to do it here:
Analysis GEN/RECO objects involved RelVal Sample Availability
HWW electrons, muons -------- H165WW2L config python
HZZ electrons, muons -------- H200ZZ4L config python
Hgg photons /RelValH130GGgluonfusion GEN-SIM-RECO, GEN-SIM-DIGI-RAW-HLTDEBUG
Htaunu taus, MET (neutrinos) ?? GEN-SIM-RECO, GEN-SIM-DIGI-RAW-HLTDEBUG
H2tau taus /RelValHiggs200ChargedTaus GEN-SIM-RECO, GEN-SIM-DIGI-RAW-HLTDEBUG
Hbb jets, leptons, MET,... /RelValTTbar GEN-SIM-RECO, GEN-SIM-DIGI-RAW-HLTDEBUG

samples

-- JordiDuarte - 19-Mar-2012

Edit | Attach | Watch | Print version | History: r12 < r11 < r10 < r9 < r8 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r12 - 2012-05-02 - JordiDuarte
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Main All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright &© 2008-2024 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback