Razor ntuplizer for running over LHC Run 2 miniAOD compatible with CMSSW_9_4_9.
N.B. For 2018 data, please use CMSSW_10_2_X. Note that CMSSW_10_6_X will break and you will need to use the Tuplize_2018UL branch. Full recipe by PPD can be found here https://twiki.cern.ch/twiki/bin/viewauth/CMS/PdmVAnalysisSummaryTable. Also, extra packages need to be installed for 2018 data to obtain the photon energy corrections, according to https://twiki.cern.ch/twiki/bin/view/CMS/EgammaMiniAODV2#2018_MiniAOD
If you are running on lxplus, you will likely need to use Singularity because 10_2_X only compiles in SLC6, which is not supported on lxplus.
To start singularity:
export SINGULARITY_CACHEDIR=/tmp/$(whoami)/singularity;singularity shell -B /cvmfs -B /afs -B /eos /cvmfs/singularity.opensciencegrid.org/bbockelm/cms:rhel6
source /cvmfs/cms.cern.ch/cmsset_default.sh
To install extra packages for photon energy corrections:
export SCRAM_ARCH=slc6_amd64_gcc700
cmsrel CMSSW_10_2_10
cd CMSSW_10_2_10/src
cmsenv
git cms-init
git cms-merge-topic cms-egamma:EgammaPostRecoTools #just adds in an extra file to have a setup function to make things easier
git cms-merge-topic cms-egamma:PhotonIDValueMapSpeedup1029 #optional but speeds up the photon ID value module so things fun faster
git cms-merge-topic cms-egamma:slava77-btvDictFix_10210 #fixes the Run2018D dictionary issue, see https://github.com/cms-sw/cmssw/issues/26182, may not be necessary for later releases, try it first and see if it works
#now to add the scale and smearing for 2018 (eventually this will not be necessary in later releases but is harmless to do regardless)
git cms-addpkg EgammaAnalysis/ElectronTools
rm EgammaAnalysis/ElectronTools/data -rf
git clone [email protected]:cms-data/EgammaAnalysis-ElectronTools.git EgammaAnalysis/ElectronTools/data
#now build everything
scram b -j 8
Now install and compile our tuplizer:
git clone [email protected]:cms-lpc-llp/DelayedPhotonTuplizer.git Tuplizer/DelayedPhotonTuplizer
scram b
cmsRun python/razorTuplizer_MC_Fall17_EcalRechits_reMiniAOD_OOT.py
Before running, check python/razorTuplizer.py to make sure that the correct global tag is defined. (process.GlobalTag.globaltag = ...)
To run a test job with CRAB3:
source /cvmfs/cms.cern.ch/crab3/crab.sh
crab submit -c crabConfigRazorTuplizer.py
To submit mass production, use bash scripts in crab_scripts directory. Make 2018 copies based on submit_data.sh, submit_DiPhoton.sh, submit_GJet16_DoubleEMEnriched.sh, and submit_QCD16_DoubleEMEnriched.sh (we don't have signals for 2018 yet). Before running each script, please check every line to make sure it's correct and updated. For data lumi mask, use this https://cms-service-dqm.web.cern.ch/cms-service-dqm/CAF/certification/Collisions18/13TeV/PromptReco/Cert_314472-325175_13TeV_PromptReco_Collisions18_JSON.txt (please cross check with PPD recommendation for update https://twiki.cern.ch/twiki/bin/view/CMS/PdmV2018Analysis#DATA).
After crab submission, job status can be checked with python check_crab.py -s [KEY_WORD] where KEY_WORD is any word matched with the crab directory of the sample you want to check. For example python check_crab.py -s QCD. Optional arguments:
-r: to resubmit failed jobs even when the task is not finished,-k: to kill unfinished task.