Currently this is build using CMSSW_9_4_13
install:
source /cvmfs/cms.cern.ch/cmsset_default.csh
setenv SCRAM_ARCH slc6_amd64_gcc630
cmsrel CMSSW_9_4_13
cd CMSSW_9_4_13/src/
cmsenv
## Modified MET
git cms-merge-topic cms-met:METFixEE2017_949_v2
## Redo MET filter
git cms-addpkg RecoMET/METFilters
## if wanting to use el ID v2 (for 2017 data):
git cms-merge-topic cms-egamma:EgammaPostRecoTools
## HOT tagger part1
cd ${CMSSW_BASE}/src
git cms-merge-topic -u pastika:AddAxis1_946p1
git clone [email protected]:susy2015/TopTagger.git
### BestCalc: copy lwtnn so that BestCalc.cc will compile. ( This is bad practice, should always try to get official CMSSW recipes whenever possible)
cd ${CMSSW_BASE}/src
cp -r ~jmanagan/nobackup/CMSSW_9_4_12/src/lwtnn . ## use scp after a Fermilab kinit to copy onto non-LPC clusters
## FWLJMET code
git clone -b 9_4_X_2017data [email protected]:cms-ljmet/FWLJMET.git
## JetSubCalc currently uses uses PUPPI mass corrections:
cd ${CMSSW_BASE}/src/FWLJMET/LJMet/data/
git clone https://github.com/thaarres/PuppiSoftdropMassCorr
cd -
#Compile
scram b
## HOT tagger part2
cd ${CMSSW_BASE}/src
cmsenv
mkdir -p ${CMSSW_BASE}/src/TopTagger/TopTagger/data
getTaggerCfg.sh -o -n -t DeepResolved_DeepCSV_GR_noDisc_Release_v1.0.0 -d $CMSSW_BASE/src/TopTagger/TopTagger/data
Some info:
- LJMet/plugins/LJMet.cc : the EDAnalyzer that wraps LJMet classes
- Modified MET: https://twiki.cern.ch/twiki/bin/viewauth/CMS/MissingETUncertaintyPrescription#Instructions_for_9_4_X_X_9_or_10
- Redo MET filter : https://twiki.cern.ch/twiki/bin/viewauth/CMS/MissingETOptionalFiltersRun2#How_to_run_ecal_BadCalibReducedM
- el ID V2 : https://twiki.cern.ch/twiki/bin/view/CMS/EgammaMiniAODV2
- HOT tagger : https://github.com/susy2015/TopTagger/tree/master/TopTagger#instructions-for-saving-tagger-results-to-nanoaod-with-cmssw_9_4_11
- BoostedEventShapeTagger(BEST) :
run LJMet:
cmsRun LJMet/runFWLJMet_multiLep.py (or runFWLJMet_singleLep.py)