setenv SCRAM_ARCH slc6_amd64_gcc530
cmsrel CMSSW_8_0_20
cd CMSSW_8_0_20/src
git cms-init
git cms-merge-topic dmajumder:CMSSW_8_0_X_AnalysisDataFormats_BoostedObjects
git clone [email protected]:dmajumder/EventCounter.git Analysis/EventCounter
git clone -b CMSSW_8_0_X_NewB2GAnaFW [email protected]:dmajumder/VLQAna.git Analysis/VLQAna
#To just build this stuff
scram b -j20
cd Analysis/VLQAna/test
cmsRun vlqAna_cfg.py
#To generate VLQ trees after pre-selection
cmsRun vlqAna_cfg.py storePreselEvts=True
#To process lots of CRAB jobs producing the above trees:
python tpthPreselTrees_submit_all.py -f CRAB/tpth_tosubmit.txt
#Make sure to uncomment the dataset names in the file 'CRAB/tpth_tosubmit.txt ' you want to process. #Be mindful of the switches for processing data and MC. #To process data, change in tpthPreselTrees_submit_all.py line 52:
config.JobType.pyCfgParams = ['isData=True','doPUReweightingOfficial=False','jecShift=0','jerShift=0', 'doBTagSFUnc=False']
#To run crab jobs for skims, e.g for CR in dielectron channel
Analysis/VLQAna/test/Skim/CRAB/Skims_CR_Zelel
#Modify the exisiting crab file, by changing the relvant parameters, like config.General.requestName, config.Data.inputDataset, config.Data.outLFNDirBase, config.Site.storageSite etc.
crab submit -c crab_XX.py
#This will create an output condor directory, through which you can check the status of the jobs.
crab status -d MY_CONDOR_DIR
#Please see more details https://twiki.cern.ch/twiki/bin/view/CMSPublic/WorkBookCRAB3Tutorial.