BoostedTTH

CMSSW tools for analyzing TTH events with boosted objects

Installation

Follow These Steps: If it is your first time setting up a CMSSW release CMSSW will create a hidden .cmsgit-cache directory in your home directory that can grow quite large. Therefore it is a good idea to specify the path of this directory to be on the dust (or other high volume storage) Do for example: #export CMSSW_GIT_REFERENCE=/nfs/dust/cms/user/$USER/.cmsgit-cache

# setup environment
# SL6
#export SCRAM_ARCH="slc7_amd64_gcc700"
# CC7
export SCRAM_ARCH="slc7_amd64_gcc700"

export CMSSW_VERSION="CMSSW_10_6_27"

# create new CMSSW environment
scram project $CMSSW_VERSION
cd $CMSSW_VERSION/src
export CMSSWSRCDIR="$( pwd )"
eval `scramv1 runtime -sh` 

# producer of deterministic seeds for physics objects to be able to do synchronization

# producer to apply JER to jets
# git cms-merge-topic sebwieland:CMSSW_10_2_X_SmearedJetProducer
# producer of deterministic seeds for physics objects to be able to do synchronization
# git cms-merge-topic yrath:deterministicSeeds_102X

# for rerun pileupJetID
git cms-addpkg RecoJets/JetProducers

# adds function to easily recalculate electron/photon IDs and energy corrections
# For UL:
# https://twiki.cern.ch/twiki/bin/view/CMS/EgammaUL2016To2018#Recipe_for_running_scales_and_sm

git cms-addpkg RecoEgamma/EgammaTools  ### essentially just checkout the package from CMSSW
git clone https://github.com/cms-egamma/EgammaPostRecoTools.git
mv EgammaPostRecoTools/python/EgammaPostRecoTools.py RecoEgamma/EgammaTools/python/.
git clone -b ULSSfiles_correctScaleSysMC https://github.com/jainshilpi/EgammaAnalysis-ElectronTools.git EgammaAnalysis/ElectronTools/data
git cms-addpkg EgammaAnalysis/ElectronTools

# mitigation of EE noise to MET in 2017 data
# not needed anymore for UL
# https://twiki.cern.ch/twiki/bin/view/CMS/JetMET#Quick_links_to_current_recommend
# git cms-merge-topic cms-met:METFixEE2017_949_v2_backport_to_102X

# not needed anymore for UL
# https://twiki.cern.ch/twiki/bin/viewauth/CMS/MissingETOptionalFiltersRun2#2018_2017_data_and_MC_UL
# needed to run ecalBadCalibReducedMINIAODFilter
# git cms-addpkg RecoMET/METFilters

# needed to include tt+4b and tt+bbb categorization
git cms-addpkg TopQuarkAnalysis/TopTools
git remote add wwei git@github.com:wweiphy/cmssw.git
git fetch wwei
git checkout wwei/ttHH_with_tt4b
or you can add the changes in your own cmssw repository: https://github.com/wweiphy/cmssw/commit/e305554a08d0dd22b43135db34a931104d44108a

# needed to rerun DeepJet
# Not needed for UL
# git cms-addpkg RecoBTag/TensorFlow
# git cherry-pick 94ceae257f846998c357fcad408986cc8a039152

# install common classifier (currently work in progress)
mkdir TTH
cd TTH
git clone https://gitlab.cern.ch/wewei/CommonClassifier.git CommonClassifier -b ttHH_Run2
git clone https://gitlab.cern.ch/wewei/MEIntegratorStandalone.git MEIntegratorStandalone -b ttHH_Run2

# mkdir -p $CMSSW_BASE/lib/$SCRAM_ARCH/
cp -R MEIntegratorStandalone/libs/* $CMSSW_BASE/lib/$SCRAM_ARCH/
scram setup lhapdf

# modify gsl.xml file: <environment name="GSL_BASE" default="/cvmfs/cms.cern.ch/slc7_amd64_gcc700/external/gsl/2.2.1-omkpbe2"/>
# modify gsl.xml file: <tool name="gsl" version="2.2.1">
scram setup MEIntegratorStandalone/deps/gsl.xml

# use recent version of LHAPDF header
sed -i '6i#include "LHAPDF/LHAPDF.h"' MEIntegratorStandalone/interface/Integrand.h
sed -i '32i /*' MEIntegratorStandalone/interface/Integrand.h
sed -i '44i */' MEIntegratorStandalone/interface/Integrand.h
# install reco likelihood variables (deprecated?)
source CommonClassifier/setup/install_recoLikelihood.sh

# install miniaod and boostedtth
cd $CMSSWSRCDIR
git clone -b TTHHUL2 https://github.com/wweiphy/MiniAOD.git
git clone -b ttHH_Run2 https://github.com/wweiphy/BoostedTTH.git

# Download the JER correction files
cd $CMSSWSRCDIR/BoostedTTH/BoostedAnalyzer/data
mkdir jerfiles
cd jerfiles
# 2017
wget "https://raw.githubusercontent.com/cms-jet/JRDatabase/master/textFiles/Fall17_V3_MC/Fall17_V3_MC_PtResolution_AK4PFchs.txt"
wget "https://raw.githubusercontent.com/cms-jet/JRDatabase/master/textFiles/Fall17_V3_MC/Fall17_V3_MC_SF_AK4PFchs.txt"
wget "https://raw.githubusercontent.com/cms-jet/JRDatabase/master/textFiles/Fall17_V3_MC/Fall17_V3_MC_PtResolution_AK8PFchs.txt"
wget "https://raw.githubusercontent.com/cms-jet/JRDatabase/master/textFiles/Fall17_V3_MC/Fall17_V3_MC_SF_AK8PFchs.txt"
# 2016
wget "https://raw.githubusercontent.com/cms-jet/JRDatabase/master/textFiles/Summer16_25nsV1_MC/Summer16_25nsV1_MC_PtResolution_AK4PFchs.txt"
wget "https://raw.githubusercontent.com/cms-jet/JRDatabase/master/textFiles/Summer16_25nsV1_MC/Summer16_25nsV1_MC_SF_AK4PFchs.txt"
wget "https://raw.githubusercontent.com/cms-jet/JRDatabase/master/textFiles/Summer16_25nsV1_MC/Summer16_25nsV1_MC_PtResolution_AK8PFchs.txt"
wget "https://raw.githubusercontent.com/cms-jet/JRDatabase/master/textFiles/Summer16_25nsV1_MC/Summer16_25nsV1_MC_SF_AK8PFchs.txt"
# 2018
wget "https://raw.githubusercontent.com/cms-jet/JRDatabase/master/textFiles/Autumn18_V1_MC/Autumn18_V7_MC_SF_AK4PFchs.txt"
wget "https://raw.githubusercontent.com/cms-jet/JRDatabase/master/textFiles/Autumn18_V1_MC/Autumn18_V7_MC_PtResolution_AK4PFchs.txt"
wget "https://raw.githubusercontent.com/cms-jet/JRDatabase/master/textFiles/Autumn18_V1_MC/Autumn18_V7_MC_SF_AK8PFchs.txt"
wget "https://raw.githubusercontent.com/cms-jet/JRDatabase/master/textFiles/Autumn18_V1_MC/Autumn18_V7_MC_PtResolution_AK8PFchs.txt"

cd $CMSSWSRCDIR

#compile
scram b -j 12

### only for crab use ###
cp /cvmfs/cms.cern.ch/$SCRAM_ARCH/external/gsl/2.2.1/lib/* $CMSSWSRCDIR/../lib/$SCRAM_ARCH
scram b -j 12

Overview

BoostedObjects contains the classes needed for subjet-analysis. They associate fat jets with the corresponding filtered objects.

BoostedProduces contains the tools used to run the HEPTopTagger and SubjetFilterJet algorithm on MiniAOD and add the output as the above collections.

BoostedAnalyzer can be used to analyze MiniAOD files. The plugin itself takes care of objectselections and stores the objects in InputCollections. Different event selection can be used in this step, too. The inputcollections can be analyzed with a TreeWriter that can load different Processors. Every processor writes a certain class of output variables in a flat TTree.