##Instructions of calculating csv Scale Fators

###Step 0: get a GitHub account. If you already have one, go to step 1.

###Step 1: Check out the ttH BEAN code following instructions here: https://github.com/cms-ttH/BEAN

###Step 2: Check out treeMaking code Following instructions here: https://github.com/cms-ttH/ttHMultileptonAnalysis

###Step 3: Making trees

1. tree making script: ttHMultileptonAnalysis/TemplateMakers/bin/csvSF.C
2. file lists for the skims of all samples: ttHMultileptonAnalysis/listsForSkims2012_53x_v2_hadoop/*list

// you need to change the path of the skim files accordingly in *list

$ To run a test job

cd ttHMultileptonAnalysis/TemplateMakers/test/

csvSF ssCondor.py wz test NA 1 10

// csvSF ssCondor.py sample lable JES_syst(NA, JESUp, or JESDown) jobNumber totalJobs

// This runs a test job on sample zz, year 2012_53x, job_label test, JES_syst nominal, job number 1 out of 10 total jobs. The command runs locally, the output goes in batchBEAN/zz_test/

$ To run the full set of jobs, do the following (runs on batch queue!):

./submit_condor_jobs.py csvSF label -j JES_syst

// this submits one group of jobs for each list file in csvSF_lists.txt. The number of jobs in the group is one for each line in the list file.

// The output goes in batchBEAN/${sample}_${label}/

$ hadd output root files

• datasets

hadd -f TwoMuon.root DoubleMu_Run2012**{*label}/*root

hadd -f TwoEle.root DoubleElectron_Run2012**{*label}/*root

hadd -f MuonEle.root MuEG_Run2012**{*label}/*root

• MC

hadd -f allMC.root ttbar*{label}/root singlet{label}/root wjets{label}/root zjets{label}/root ww{label}/root wz{label}/root zz{label}/*root

###Step 4: Calculate csvSFs

1. Check out csv SF calculation code:

cd $CMSSW_BASE/src

git clone https://github.com/cms-ttH/csvReweighting.git

cd csvReweighting/

mkdir dataMCRootFiles

(mv the root files from hadd above to dataMCRootFiles)

2. version0 SF calculation

• light flavor scale factor

root -l lightflavorCSVSF.C

• heavy flavor scale factor

root -l heavyflavorCSVSF.C

the scale factors version 0 will be stored at csv_rwt_hf_v0.root and csv_rwt_lf_v0.root

3. Do multiple Iterations (tricky part)

• 3a: replace SF files (x --> iteration version number)

  mv csv_rwt_lf_v{x}.root BEAN/BEANmaker/data/csv_rwt_lf_IT.root

  mv csv_rwt_hf_v{x}.root BEAN/BEANmaker/data/csv_rwt_hf_IT.root

• 3b: remake the MC trees following Step 3, new csv weigths (csvWgtlf, csvWgthf) based on SFs version{x} will be saved

//don't need to rerun data trees since there will be no change

• 3c: recalculate SFs version{x+1} following Steps 4 using SFs version{x},

//-- change the input in lightflavorCSVSF.C/heavyflavorCSVSF.C accordingly

• input 1 --> the version number

• input 2 --> the input MC root file

// the scale factors version x+1 will be stored at csv_rwt_hf_v{x+1}.root and csv_rwt_lf_v{x+1}.root

###Step5: SF fitting and extrapolation

1. Create HistoFiles directory and put all inputs there

mkdir HistoFiles

cp $CMSSW_BASE/src/BEAN/BEANmaker/data/mc_btag_efficiency_8TeV_53x.root HistoFiles/

cp csv_rwt_f_IT_v2.root HistoFiles/

2. Run scripts to create Statistics scale factors and wait for that to finish

root -b -q fit_csvSF_HF_statErr.C+ > & ! out_fit_csvSF_HF_statErr.log &

root -b -q fit_csvSF_LF_statErr.C+ > & ! out_fit_csvSF_LF_statErr.log &

3. Run scripts to do the fitting and making image files

mkdir Images

root -b -q fit_csvSF_HF.C+

root -b -q fit_csvSF_LF.C+

4. Remove unnecessary histograms from files to reduce size, bloat

root -b -q drop_histograms.C+'("HistoFiles/csv_rwt_hf_final_IT_v2.root","csv_rwt_hf_final_IT_v2.root")'

root -b -q drop_histograms.C+'("HistoFiles/csv_rwt_lf_final_IT_v2.root","csv_rwt_lf_final_IT_v2.root")'