link: https://cms-analysis.github.io/HiggsAnalysis-CombinedLimit/
voms-proxy-init --voms cms
cmssw-el7 (on lxplus)
setenv SCRAM_ARCH slc7_amd64_gcc700 (on lpc)
cmsrel CMSSW_10_2_13
cd CMSSW_10_2_13/src
cmsenv
git clone https://github.com/cms-analysis/HiggsAnalysis-CombinedLimit.git HiggsAnalysis/CombinedLimit
cd $CMSSW_BASE/src/HiggsAnalysis/CombinedLimit
git fetch origin
git checkout v8.2.0
scramv1 b clean; scramv1 b
cd ../../../../
git clone -b ParticleNet_TopW_SFs_NanoV9 https://github.com/cms-jet/ParticleNetSF.git
cd ParticleNetSF/
cp TagAndProbeExtended.py ../CMSSW_10_2_13/src/HiggsAnalysis/CombinedLimit/python
cd ../CMSSW_10_2_13/src/
git clone -b 102x https://github.com/cms-analysis/CombineHarvester.git
cd CombineHarvester/
scram b -j 8
cd ../../../ParticleNetSF/
./runFullChain.sh [T|W] [2016|2017|2018] ["Nominal"|"MD"]
N.B.: Small tweaks are needed to be done inside the "configuration.h" and "makeFits.C" according to the object/version selected for the SF extraction.
github link: https://github.com/cms-jet/ParticleNetSF
[this is to speed up the process]:
root [0] .L make2DTemplates.C
root [1] mainfunction("tt1l","2017","0.90","1.”)
root [0] .L HeavyFlavourZCandleStudies.C
root [1] HeavyFlavourZCandleStudies("2017","tt1l","bb","0.90","1.",false,"pass")
The Datacards will be produced under particlenet_sf/fitdir [if one keeps the default naming/settings]
root [0] .L makeDatacards.C
root [1] makeDatacards("2017","tt1l","bb","0.90","1.")
root [0] .L makeFits.C
root [1] makeFits("2017","bb","0.90","1.","tt1l")
or “zqq” or “tt1L” to fit only specific samples