Skip to content
BranchesTags

Repository files navigation

JetMETAnalysisMCtruth

Author / Maintainer: Ilias Zisopoulos, GitHub Profile, ilias.zisopoulos@cern.ch, ilzisopoulos@gmail.com

Introduction

This package provides tools to derive, validate, and visualize L1FastJet and L2Relative MC-truth Jet Energy Corrections (JEC).

Originally based on cms-jet/JetMETAnalysis, this package was overhauled and rewritten by Ilias Zisopoulos to extend compatibility across newer CMSSW release cycles, transition from legacy AOD to native MiniAOD processing, and incorporate dedicated validation workflows. It is now actively maintained and utilized within the CMS Jet Energy Resolution and Corrections (JERC) group under cms-analysis/jme/jerc-derivation/JetMETAnalysisMCtruth.

Documentation

  1. CMS-AN-2023/061, "MC truth jet energy corrections using the 2022, 2023 and 2024 Run-3 simulations":
    https://cms.cern.ch/iCMS/user/noteinfo?cmsnoteid=CMS%20AN-2023/061

  2. CMS-AN-2021/148, "MC truth jet energy corrections using the Legacy 2016 simulations":
    https://cms.cern.ch/iCMS/user/noteinfo?cmsnoteid=CMS%20AN-2021/148

  3. CMS-AN-2020/151, "MC truth jet energy corrections using the Legacy 2017 and 2018 simulations":
    https://cms.cern.ch/iCMS/user/noteinfo?cmsnoteid=CMS%20AN-2020/151

  4. CMS-AN-2020/049, "2018 Relative and Absolute MC Truth Jet Energy Corrections":
    https://cms.cern.ch/iCMS/user/noteinfo?cmsnoteid=CMS%20AN-2020/049

  5. CMS-AN-2019/230, "2016 Relative and Absolute MC Truth Jet Energy Corrections":
    https://cms.cern.ch/iCMS/user/noteinfo?cmsnoteid=CMS%20AN-2019/230

Log book of previous changes

In some cases, when moving to a newer CMSSW version, compilation errors when doing scram b -j 8 arise. In this section the necessary changes to the framework when migrating from a CMSSW version to a newer one will be presented, for book keeping purposes. None of these need to be repeated by the user. The current version of the framework works for the most recent CMSSW_14_2_X version.

Changes from CMSSW_10_6_X to CMSSW_12_6_X

  1. Open the JetMETAnalysis/JetAnalyzers/BuildFile.xml code and in line 20 replace SimDataFormats/JetMatching with DataFormats/JetMatching
  2. Open the codes JetMETAnalysis/JetAnalyzers/interface/JetResponseAnalyzer.hh and JetMETAnalysis/JetAnalyzers/interface/JetResponseAnalyzerProducer.hh and replace SimDataFormats/JetMatching with DataFormats/JetMatching in lines 45 and 43 respectively.
  3. cp /cvmfs/cms.cern.ch/slc7_amd64_gcc900/external/gcc/9.3.0/include/c++/9.3.0/bits/stl_tree.h JetMETAnalysis/JetUtilities/interface/
  4. Open the JetMETAnalysis/JetUtilities/interface/stl_tree.h code and comment out lines 778-781 which are responsible for giving the error static assertion failed: comparison object must be invocable as const
  5. Open the JetMETAnalysis/JetAnalyzers/bin/jet_match_x.cc and JetMETAnalysis/JetAnalyzers/bin/jet_synchtest_x.cc codes and add before any other include the following: #include "JetMETAnalysis/JetUtilities/interface/stl_tree.h"
  6. Then, open the ../tmp/slc7_amd64_gcc900/src/JetMETAnalysis/JetUtilities/src/JetMETAnalysisJetUtilities/a/JetMETAnalysisJetUtilities_xr.cc code and do the same -> add before any other include the following: #include "JetMETAnalysis/JetUtilities/interface/stl_tree.h"
  7. Move the SynchFittingProcedure.hh code from JetMETAnalysis/JetUtilities/src/ to the JetMETAnalysis/JetUtilities/interface/ folder and then open the JetMETAnalysis/JetAnalyzers/bin/jet_synchplot_x.cc code and in line 35 replace src/ with inteface/ (to provide the new correct path).
  8. Modify python files to work with python3: 4 spaces instead of a tab, algsizetype.items() instead of algsizetype.iteritems(), add parentheses in print commands, list(genJetsDict.keys()).index(alg_size_type) instead of genJetsDict.keys().index(alg_size_type)

Changes from CMSSW_12_6_X to CMSSW_13_0_X and beyond (up to at least CMSSW_14_2_X)

  1. Change to edm::one::EDAnalyzer<> and edm::one::EDProducer<> from edm::EDAnalyzer and edm::EDProducer respectively. Modify the includes as well.

  2. Copy the file /cvmfs/cms.cern.ch/slc7_amd64_gcc11/external/gcc/11.2.1-f9b9dfdd886f71cd63f5538223d8f161/include/c++/11.2.1/bits/stl_tree.h to the JetMETAnalysisMCtruth/JetUtilities/interface directory and comment out lines 768-771

  3. In JetUtilities/src/JetInfo.cc line 361 change assert(words>0) to assert(words != nullptr)

  4. From CMSSW_12_6_X copy the codes JetMETCorrections/Objects/interface/JetCorrector.h and cmssw/JetMETCorrections/Objects/src/JetCorrector.cc and paste them to JetUtilities/interface/ and JetUtilities/src/ respectively. In JetCorrector.cc comment out lines 48-53, and in JetAnalyzers/src/JetResponseAnalyzer.cc, JetAnalyzers/src/JetResponseAnalyzerProducer.cc write jetCorrector_ = 0

  5. From CMSSW_12_6_X copy the codes JetMETCorrections/Configuration/python/JetCorrectionServicesAllAlgos_cff.py and JetMETCorrections/Configuration/python/JetCorrectionServices_cff.py and paste them inside JetAnalyzers/python/

HOW TO: Setup the framework

Setup the code in the AFS area and not the EOS user area, because HTCondor is used, that is not compatible with EOS.

mkdir MCtruthJEC/ 
cd MCtruthJEC/ 
cmsrel CMSSW_14_2_2 
cd CMSSW_14_2_2/src 
cmsenv 
git clone https://gitlab.cern.ch/cms-analysis/jme/jerc-derivation/JetMETAnalysisMCtruth.git

Then compile:

scram b -j 8

In the first compilation you will get a compilation error about is_invocable_v<const _Compare&, const _Key&, const _Key&>.

Open the file CMSSW_14_2_2/tmp/el9_amd64_gcc12/src/JetMETAnalysisMCtruth/JetUtilities/src/JetMETAnalysisMCtruthJetUtilities/lcgdict/JetMETAnalysisMCtruthJetUtilities_xr.cc and before any other include (among lines 6 and 7) add the following line:

#include "JetMETAnalysisMCtruth/JetUtilities/interface/stl_tree.h"

Re-compile and there should be no errors. The aforementioned JetMETAnalysisMCtruthJetUtilities_xr.cc script is autogenerated, so if this error comes back again in any given time, repeat the step above.

Very important note: Every time there you change any .cc or .hh or .h code inside $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetAnalyzers or $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetUtilities you should then re-compile, doing scram b -j 8 for the changes to take effect.

HOW TO: Produce ntuples from MINIAOD

In this section instructions are provided on how to produce a JRA ntuple which contains a tree with event and matched rec-gen jet variables, needed for the JEC derivation.

Important codes for producing ntuples

  1. JetMETAnalysisMCtruth/JetAnalyzers/test/run_JRA_cfg_MCtruth.py
    Lines 25-26: Specify the jet collections to be saved in the JRA trees
    Line 51: Insert global tag of sample to be processed
    Line 69: Specify how many events to be processed, -1 stands for all events in the sample
    Line 79: Specify which MiniAOD root file to be processed for a local test

The run_JRA_cfg_MCtruth.py script then uses the addAlgorithm.py one:

  1. JetMETAnalysisMCtruth/JetAnalyzers/python/addAlgorithm.py
    Line 382: Specify the raw jet pT cut with which the rec-gen matching will be performed, and jets will be saved in the JRA trees (default = 0 GeV)
    Lines 391-407: While reconstructing PUPPI jets the code uses the following commands to consider the stored PUPPI weights in the dataset:
process.puppi.useExistingWeights = True
process.puppiNoLep.useExistingWeights = True

  1. JetMETAnalysisMCtruth/JetAnalyzers/python/customizePuppiTune_cff_V15.py
    This is a configuration file for applying the V15 PUPPI tune recipe. If one did not want to use the default PUPPI weights in the dataset but wanted to re-calculate the V15 weights on the fly, they should load and call in addAlgorithm.py this file, while turning the options above to False. NOT needed anymore, as the V15 tune is outdated.

  2. JetMETAnalysisMCtruth/JetAnalyzers/python/Defaults_cff.py
    Line 33: Rec-gen jet pairings are saved in the ntuple, along with their deltaR. Change the maximum deltaR value that is saved (default = 999, i.e. write everything). The deltaR < 0.2 (0.4) criterion will be used later in another step, so here we save all of them.

  3. JetMETAnalysisMCtruth/JetAnalyzers/src/JetResponseAnalyzer.cc, JetMETAnalysisMCtruth/JetAnalyzers/interface/JetResponseAnalyzer.hh, JetMETAnalysisMCtruth/JetUtilities/src/JRAEvent.cc, JetMETAnalysisMCtruth/JetUtilities/interface/JRAEvent.h
    These codes produce the trees. They do not need any change at the moment. If a new variable needs to be added in the tree of the JRA ntuples these are the codes that need to be modified.

How to produce JRA ntuples

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetAnalyzers/test/

Before submitting jobs to crab run a local test first:

In the code JetMETAnalysisMCtruth/JetAnalyzers/test/run_JRA_cfg_MCtruth.py specify a MiniAOD root file and a small number of events. Then do:

cmsenv 
voms-proxy-init -voms cms 
cmsRun run_JRA_cfg_MCtruth.py

This test will produce a file named JRA.root in the directory you are in, containing the small number of events specified. If there are no errors, the JRA.root is produced, and the trees are filled properly, then CRAB jobs can be submitted in order to process the full MC sample.

How to submit jobs to CRAB:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetAnalyzers/test/

In the code JetMETAnalysisMCtruth/JetAnalyzers/test/run_JRA_cfg_MCtruth.py put as number of events to process -1

In the code JetMETAnalysisMCtruth/JetAnalyzers/test/custom_crab_JEC.py:

Lines 4-5: Submitting jobs will create a folder workArea/requestName/ inside $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetAnalyzers/test/
Line 14: Specify the MiniAOD MC sample name (from DAS) to process
Line 17: Specify how many CRAB jobs to have per MiniAOD root file (default = 1 for faster production). The larger this number, the fewer the total CRAB jobs, and thus the longer it will take for each job to run
Line 20: Specify name of folder that will be created in the output EOS directory. Here we usually use PremixedPU or FlatPU0to120 or EpsilonPU
Line 21: Specify output directory where the JRA ntuples will be saved. Note that /eos/cms/ should not be written before /store/group/phys_jetmet/

Attention: You first need to have write access to the phys_jetmet EOS area. Ask the JME conveners for that.

Once you are done, submit the CRAB jobs:

voms-proxy-init -voms cms 
crab submit -c custom_crab_JEC.py

To check the status of jobs in CRAB while inside JetMETAnalysisMCtruth/JetAnalyzers/test/:

crab status -d workArea/requestName/

To resubmit jobs if some have failed:

crab resubmit -d workArea/requestName/

When all jobs are in finished status, the output JRA root files, based on the above custom_crab_JEC.py, will be located in a directory with this format:
/eos/cms/store/group/phys_jetmet/ilias/test/QCD_Pt-15to7000_TuneCP5_Flat2018_13TeV_pythia8/outputDatasetTag/yymmdd_hhmmss/0000/

These JRA root files are the input ntuples for the MC-truth jet energy corrections.

Please note: All Run-3 MC ntuples so far have been produced by Ilias (me), and are located in:

/eos/cms/store/group/phys_jetmet/ilias/Run3MCtruthSamples/NoRawPtCut/

HOW TO: Create histograms for PU reweighting

The only event weights applied in this analysis are the ones related to the PU reweighting. This is only relevant for the L2Relative derivation where the PremixedPU MC is used. We need to produce two root files with the mu (true number of pileup interactions per crossing) distribution; one for data and one for MC. Note that they should have the same binning (we usually use 120 bins from 0 to 120).

Histogram for data

To produce the root file for data:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/Histos_PU/ 
cmsenv 
pileupCalc.py -i Cert_Collisions2024_erasBCDEFGHI.json --inputLumiJSON pileup_latest_2024.txt --calcMode true --minBiasXsec 69200 --maxPileupBin 120 --numPileupBins 120  MyDataPUHisto_2024CDEFGHI_120Bins_69200.root

where Cert_Collisions2024_erasBCDEFGHI.json is the Golden JSON file for the corresponding era you want to process and pileup_latest_2024.txt the pileup JSON for the corresponding year (ask around if you cannot find where they have it). The minimum bias cross section of 69.2 mb has been used so far for the 2022, 2023 and 2024 MCs but studies have shown that for 13.6 TeV data a more representative value is 75.3 mb. The value of 75.3 mb will be used for 2025 MCs onwards (and when 2022, 2023 and 2024 MCs are reproduced). In principle, the MC should have been generated with a mu distribution close to the one of data, such that the weights of the PU reweighting are small and effective statistics is not lost.

If not sure, please ask around for the recommendation before producing this and proceeding to the next steps.

The output file is MyDataPUHisto_2024CDEFGHI_120Bins_69200.root that contains a histogram of the mu distribution named pileup.

Histogram for MC

To produce the corresponding root file for MC:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/ 
python3 create_mu_distribution.py --era RunIII2024Summer24

The output file is MyMCPUHisto_RunIII2024Summer24_PremixedPU.root inside Histos_PU/ that contains a histogram of the mu distribution named pileup.

Plotting and validation of PU reweighting

  1. Script that plots the mu distribution of various eras in data:
cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/ 
python3 plot_mu_distributions_data.py

Input: The MyDataPUHisto_*.root files for various eras in data
Output: Plot in PDF format of the mu distribution for various eras in data

  1. Script that plots the mu distribution of data, and the corresponding mu distribution of the MC before and after the application of the PU reweighting. If the distributions between data and MC after the reweighting match, this validates that that the PU reweighting was performed properly.
cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/ 
python3 plot_mu_distributions_data_vs_mc.py --data 2024CDEFGHI --mc RunIII2024Summer24

Input: The MyDataPUHisto_*.root file for the era in data, the MyMCPUHisto_*_PremixedPU.root file for the MC before the PU reweighting, the hadded file of Step4Output from step4 for the MC after the PU reweighting
Output: Plot in PDF format with the mu distribution of data, MC before and MC after

HOW TO: Jet veto maps

The jet veto map is a root file that contains TH2D histograms which define the jet eta-phi zones that should be excluded from your selection.

Take the most recent jet veto map for the corresponding era/MC campaign from the JERCProtolab (https://gitlab.cern.ch/cms-jetmet/JERCProtoLab) and put it inside Histos_JetVetoMaps/

Script that plots the jet veto map:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/ 
python3 plot_jet_veto_maps.py --era 2023D

Input: The JetVetoMap_*.root inside Histos_JetVetoMaps/
Output: Plot in PDF format with the veto maps for each issue (hot, cold, eep, bpix, fpix, etc) in each era

HOW TO: Gen pT spectrum

This is not necessary for the derivation of the MC truth JECs, but it might be useful to examine the gen pT spectrum. We are using flatQCD MC datasets and we do not apply pT reweighting, so this spectrum should be flat.

Script to create the histogram of the gen pT spectrum:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/
python3 create_ptgen_distribution.py --era RunIII2024Summer24

Input: The JRA ntuples of the MC dataset
Output: A root file named GetPT_RunIII2024Summer24_PremixedPU.root inside Histos_Pt/

Then, to plot this spectrum for various MC datasets:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/
python3 plot_ptgen_distributions.py

Input: The root files named GetPT_*_PremixedPU.root for various MC datasets (defined in line 62) inside Histos_Pt/
Output: A plot in PDF format with the gen pT spectra of these MC datasets

HOW TO: derive MC truth JECs, perform sanity checks and validations

Here you will learn how to produce pileup offset and jet response histograms and plots, derive the L1FastJet (optional for PUPPI) and L2Relative corrections, and plot them.

Getting started

Based on the jet collection you want to process, your work area will be one of the following directories:

$CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/ 
$CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK8PUPPI/
$CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4CHS/

Given that AK4 PUPPI jets are the main jet collection in Run-3, this README will refer to this jet collection only. However, nothing changes technically if one wants to process one of the other two collections.

Please note that the L1FastJet corrections are not part of the mainstream calibration for AK4 and AK8 PUPPI jets, and are not needed. However, in this README, the instructions on how to derive such corrections will be provided anyway. These instructions are identical for AK4 CHS jets, which do need pileup offset corrections.

Only for the first time do the following:

  1. Open the CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Setup_CMSSW.sh file and put the correct SCRAM_ARCH in the first line (e.g. SCRAM_ARCH=el9_amd64_gcc12). Then in the second line put the full path leading to condor_AK4PUPPI/ (e.g. cd /afs/cern.ch/work/<u>/<username>/public/MCtruthJEC/CMSSW_14_2_2/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/)

  2. Open the CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/makefile, copy from it the following part and paste in the terminal, exactly as it is written:

	g++ ListRunLumi.cpp -o RunListRunLumi \
		`root-config --cflags --libs`

Brief overview

The derivation of MC truth JECs consists of four steps, each of which consist of a submission part and a harvest part. These will be explained in detail in the next sections. Here we present a brief overview of what each step does:

Step1: Matches events between the FlatPU and EpsilonPU datasets and produces lists with the matched events.

Step2: Matches jets between these events, calculates the offset and produces the L1FastJet text file by fitting the <offset>/Aj as a function of <pT> and <rho>. Also produces the necessary plots of the average pileup offset.

Step3: Applies the L1FastJet text file (if one was derived) and produces histograms of the response distributions vs pT and eta. Then fits the inverse of the median response as a function of pT in fine bins of eta and produces the L2Relative text file.

Step4: Produces 2D histograms of the response vs pT and eta so as to examine the median response before and after the application of corrections (closure). This step is not necessarily the last one chronologically: if a new MC dataset is produced and one wants to first examine the behavior of the raw jet response, then this step is needed, with the option that no JECs are applied.

General tips (for HTCondor and others)

  1. Throughout all four steps of the MC truth JECs you will be submitting jobs to HTCondor. You can find more information for HTCondor here: https://batchdocs.web.cern.ch/index.html

  2. Once you submit jobs to HTCondor you can check their status by doing:

condor_q
  1. You can also check the priority of your jobs with:
condor_userprio
  1. If many people are using the particular bigbird scheduler you are in, you can change bigbird and move to a less crowded one with the following commands:
tcsh
setenv _condor_SCHEDD_HOST bigbird26.cern.ch

  1. When a job is finished it will disappear from condor_q, and a corresponding root file should appear in EOS. Additionally, in the directory $CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Log/ 3 files will appear for each job: err, log, out. Check the err and out ones to see if there was an error in your codes that made the jobs crash. If you have a bug somewhere then the output root files in EOS will not be created or they will be created empty, so you can also check them by going to the EOS directory and doing ls -lh to see the size of the files and if they have closed properly. If the code does not have any bugs then these files should be a few MB (or at least a few hundred KB) each.

  2. When running some steps you will get the following error messages, which you can safely ignore, as we are not doing flavor corrections and do not care that our trees do not have these related branches:

Error in <TTree::SetBranchAddress>: unknown branch -> refpdgid_algorithmicDef
Error in <TTree::SetBranchAddress>: unknown branch -> refpdgid_physicsDef

  1. If there are no bugs and the root files have been created correctly then it is usual (especially in step2) that condor did not run all jobs (due to technical issues related to condor, wall time etc). Therefore you should always check how many output root files were created in EOS by doing ls | wc -l: they should be the same number as the jobs you submitted. If they are fewer then you can resubmit the jobs (./SubmitStep*.sh as you did in the first time) until all root files are processed.

  2. ATTENTION: In the bash scripts (SubmitStep.sh etc) the .cc executables are run with several input arguments, for example:

jet_response_analyzer_x jra.config \
   -input Input.root \
   -nbinsabsrsp 0 \
   -nbinsetarsp 0 \
   -nbinsphirsp 0 \
   -nbinsrelrsp 60 \
   -doflavor false \
   -flavorDefinition phys \
   -MCPUReWeighting MyMCPUHisto_Run3Summer23_PremixedPU.root \
   -MCPUHistoName pileup \
   -DataPUReWeighting MyDataPUHisto_2023_erasC_100Bins.root \
   -DataPUHistoName pileup \
   -output jra.root \
   -useweight false \
   -nrefmax 3 \
   -algs ak4puppi \
   -drmax 0.2 \
   -relrspmin 0.0 \
   -relrspmax 3.0 \
   -jtptmin 0 \
   -doDZcut true \
   -doNMcut true \
   -doVetoMap true \
   -JetVetoMapRootName JetVetoMap_2023C.root \
   -JetVetoMapHistName jetvetomap_all

If you need to remove a specific argument (for example, to disable PU reweighting), delete the corresponding lines entirely. Do not comment them out, as this will break the command structure and cause the subsequent arguments to be misinterpreted (and you may not get an error for this!!). Finally, always check the Condor output files to confirm that the job was executed with the intended set of arguments.

Pileup offset corrections (L1FastJet) and pileup offset monitoring

Go to the corresponding work area, make a directory with the name of the MC dataset and version of corrections you want to derive (e.g. RunIII2024Summer24_V1_PhiIndependent/) and then create an L1_output directory:

cd CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/
cd Files/
mkdir RunIII2024Summer24_V1_PhiIndependent/
cd RunIII2024Summer24_V1_PhiIndependent/
mkdir L1_output/

Copy the jet veto map inside the directory above:

cd CMSSW_BASE/src/JetMETAnalysisMCtruth/Histos_JetVetoMaps/
cp JetVetoMaps_2024CDEFGHI.root $CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Files/RunIII2024Summer24_V1_PhiIndependent/L1_output/

Modify the Setup_FileLocation.sh script with the paths of the input JRA ntuples, and the output root files in EOS:

  1. In NoPUFiles you should have the path to the EpsilonPU JRA ntuples, while in WithPUFiles you should have the path to the FlatPU JRA ntuples.

  2. Modify the Step1Output and Step2Output paths accordingly: this is where the output root files of these steps will appear.

Step1 -- event matching between EpsilonPU and FlatPU MC datasets

Run the following script which submits HTCondor jobs:

./SubmitStep1.sh

Input: The EpsilonPU and FlatPU JRA ntuples
Output: Txt files in Step1Output that contain lists of matched events

When all jobs are finished and the err files inside Log/ are empty, proceed with the harvesting step:

./HarvestStep1.sh

Input: The txt files in Step1Output
Output: A file named MatchedFiles inside CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Files

Rename and copy the output file with the name of the MC dataset, so that it won't get overwritten in the future e.g.:

cd CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Files/
cp MatchedFiles MatchedFiles_RunIII2024Summer24

Step2 -- derivation of L1FastJet corrections

First do:

./RunPrepareStep2Submission 1 > SubmitStep2.sh

Input: The $CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Files/MatchedFiles file
Output: Rewrites the SubmitStep2.sh script.

Open the SubmitStep2.sh script to see how it has changed. It should list the paths with the locations of the JRA_*.root files.

In line 3 you should write the full path of where the Setup_FileLocation.sh is located, e.g. source /afs/cern.ch/work/<u>/<username>/public/MCtruthJEC/CMSSW_14_2_2/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Setup_FileLocation.sh

Then, replace

echo "+JobFlavour = microcentury" >> $SubmissionFile

with:

echo "+JobFlavour = testmatch" >> $SubmissionFile
echo "+request_cpus=3">>$SubmissionFile
echo "requirements = (TARGET.OpSysAndVer =?= \"AlmaLinux9\")" >> $SubmissionFile

The lines above will help the HTCondor jobs to run faster.

The SubmitStep2.sh script uses the Step2PUMatching.sh one so open it and modify it appropriately. There, you need to specify:

  1. -ApplyJEC false -> Do not apply any L1FastJet JECs for now (we haven't derived them yet!)
  2. -npvRhoNpuBinWidth 20 -> Bin width of the mu bins (relevant for the pileup offset plots)
  3. -NBinsNpvRhoNpu 6 -> Number of mu bins (relevant for the pileup offset plots)
  4. -useweight false -> We do not apply pT reweighting
  5. -nrefmax 3 -> Consider only the 3 leading gen jets
  6. -doDZcut false -> We do not usually apply this cut for the pileup offset (but can turn this on if needed)
  7. -doVetoMap true \ -JetVetoMapRootName JetVetoMap_20234CDEFGHI.root \ -JetVetoMapHistName jetvetomap_all -> Apply jet veto map and specify root file name and TH2D hist name

In turn, the Step2PUMatching.sh code uses the $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetAnalyzers/bin/jet_match_x.cc script. This does not need any modifications but here are some relevant snippets:

  1. in line 1386 the offset is calculated
  2. in lines 1436-1438 the TProfile3D objects are filled for <offset/Aj>, <rho>, <pTrec>
  3. many more histograms are filled further below (the relevant ones for plotting the pileup offset are the p_offresVsrefpt_XX_tnpuYY_YY)

Once you make sure everything is in place, submit the HTCondor jobs:

./SubmitStep2.sh

Input: None, as the script has already been modifed appropriately.
Output: Root files in Step2Output

When all jobs are finished properly and all (or most) root files are produced in EOS, proceed with the harvesting step. Modify the HarvestStep2.sh script:

  1. This scripts hadds the root files in Step2Output.
  2. If you don't want to derive L1FastJet corrections then keep everything else commented out and go immediately to the step where you plot the pileup offset (here)!!!
  3. If you want to derive L1FastJet corrections, un-comment out the part that uses the $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetAnalyzers/bin/jet_synchfit_x.cc script and modify accordingly.

In the jet_synchfit_x.cc script:

  1. In lines 425-450 you determine what fit function should be used (Complex, Simple, SemiSimple): default is SemiSimple
  2. In lines 640-648 you determine the range of the 2D fit.

Once ready, do:

./HarvestStep2.sh

Input: The root files in Step2Output, which are hadded
Output if jet_synchfit_x.cc is commented out: None besides the hadded root file in Step2Output since you don't want to derive L1FastJet corrections
Output if jet_synchfit_x.cc is used: The L1FastJet JEC text file will be created inside $CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Files/RunIII2024Summer24_V1_PhiIndependent/L1_output/

Pileup offset monitoring and plotting

The hadded root file (output_ak4puppi.root) in Step2Output contains 2D histograms of the pileup offset. There are a ton of 2D histograms, but the ones we want are the p_offresVsrefpt_XX_tnpuYY_YY where XX = BB, EI, 1EO, 2EO, FF (the 5 detector regions in abs eta) and YY are the mu bins, which were determined in Step2PUMatching.sh.

In the pileup offset plots we show the following bins in mu: [0-20], [20-40], [40-60], [60-80], [80-100] and mu approximately zero. To produce the necessary histograms for the latter bin, rerun step2 (specifying a different Step2Output to avoid overwriting) but this time do -npvRhoNpuBinWidth 1 \ -NBinsNpvRhoNpu 6 \ . The first of these bins will be [0-1] which is mu approximately 0.

In order to produce 1D histograms of the average offset divided by gen pT do:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/
jet_synchplot_x -inputDir ./ -algo1 ak4puppi -algo2 ak4puppi -outDir ./ -outputFormat .png .pdf -fixedRange false -tdr true -npvRhoNpuBinWidth 20 -NBinsNpvRhoNpu 6

Input: The hadded root file in Step2Output (output_ak4puppi.root)
Output: A root file named canvases_synchplot_ak4puppi.root inside outDir

You need to use the same -npvRhoNpuBinWidth 20 -NBinsNpvRhoNpu 6 values as the ones used to create the input output_ak4puppi.root file.

The script above uses the $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetUtilities/interface/SynchFittingProcedure.hh script. In it, in lines 463-465 the average offset is divided by get pT. If you do not want to divide by pT, then comment out these lines and use the lines 459-460 instead (make sure to recompile if you change this).

OK, now you have two canvases_synchplot_ak4puppi.root files; one with 6 bins in mu with a bin width of 20, and one with 6 bins with a bin width of 1. Use the following script to stitch them in a single root file that contains the mu bins [~0, 0-20, 20-40, 40-60, ...]:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/
python3 create_input_histos_for_pu_offset.py --jetCone 4 --jetAlgo puppi --era RunIII2024Summer24 --version V1_PhiIndependent --JEC 0 --DivByPt 1

Finally, plot the average offset (divided by pT or not) as a function of pT and eta:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/ 
python3 plot_pu_offset_vs_ptgen_eta_mu.py --jetCone 4 --jetAlgo puppi --era RunIII2024Summer24 --JEC 0 --DivByPt 1

Input: The root file created from stitching up the canvases_synchplot_ak4puppi.root files
Output: Plots in PDF format with the raw pileup offset

!!! Please note !!!
You have now plotted the raw pileup offset because you have used the output_ak4puppi.root file that was produced without any JEC applied. If you've derived L1FastJet corrections then you will have to rerun the entire step2 chain, but this time in Step2PUMatching.sh instead of:

-ApplyJEC false \

write:

-ApplyJEC true \
-JECpar RunIII2024Summer24_V1_MC_L1FastJet_AK4PUPPI.txt \

That way, the new output_ak4puppi.root file in Step2Output (careful not to overwrite, change this path) will contain histograms with the L1 JECs applied. Repeat all steps to produce the pileup offset plots again.

!!! Important detail to remember !!!
Most FlatPU samples are generated from 0 to 120 in PU (mu). That's why we usually make 6 bins with a bin width of 20. If you were to create 3 bins with a bin width of 20 then these bins would not be [0-20], [20-40], [40-60]. They would be [0-20], [20-40], [40-inf] (https://gitlab.cern.ch/cms-analysis/jme/jerc-derivation/JetMETAnalysisMCtruth/-/blob/master/JetUtilities/src/JetInfo.cc?ref_type=heads#L369). So always make sure that you know what mu bins you actually plot.

Jet response corrections (L2Relative)

Go to the corresponding work area, make a directory with the name of the MC dataset and version of corrections you want to derive (e.g. RunIII2024Summer24_V1_PhiIndependent/) and then create an L2L3_output directory:

cd CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/
cd Files/
mkdir RunIII2024Summer24_V1_PhiIndependent/
cd RunIII2024Summer24_V1_PhiIndependent/
mkdir L2L3_output/

Copy the jet veto map inside the directory above:

cd CMSSW_BASE/src/JetMETAnalysisMCtruth/Histos_JetVetoMaps/
cp JetVetoMaps_2024CDEFGHI.root $CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Files/RunIII2024Summer24_V1_PhiIndependent/L2L3_output/

Copy the histograms for PU reweighting inside the directory above:

cd CMSSW_BASE/src/JetMETAnalysisMCtruth/Histos_PU/
cp MyDataPUHisto_2024CDEFGHI_120Bins_69200.root $CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Files/RunIII2024Summer24_V1_PhiIndependent/L2L3_output/
cp MyMCPUHisto_RunIII2024Summer24_PremixedPU.root $CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Files/RunIII2024Summer24_V1_PhiIndependent/L2L3_output/

Modify the Setup_FileLocation.sh script with the paths of the input JRA ntuples, and the output root files in EOS:

  1. In WithPUFiles you should have the path to the PremixedPU JRA ntuples (NoPUFiles not needed).
  2. Modify the Step3Output and Step4Output paths accordingly: this is where the output root files of these steps will appear.

Step3 -- derivation of L2Relative corrections

The submission step produces histograms of the jet response and the rec pT in fine bins of gen pT and eta, while the harvest step performs the fits and creates the L2Relative text file.

Modify the Step3ApplyL1.sh script accordingly. There is the following snippet inside this code that applies the L1FastJet corrections before producing the histograms:

jet_apply_jec_x \
   -input Input.root \
   -output JRA_jecl1.root \
   -jecpath ./ \
   -era RunIII2024Summer24_V1_MC \
   -levels 1 \
   -algs ak4puppi \
   -L1FastJet true \
   -saveitree false

In the case of PUPPI jets where we do not produce L1FastJet corrections, this is commented out and no JECs are applied.

The Step3ApplyL1.sh code calls the $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetAnalyzers/bin/jet_response_analyzer_x.cc script and configures it. Below is an explanation of each option:

  1. -MCPUReWeighting MyMCPUHisto_RunIII2024Summer24_PremixedPU.root \ -MCPUHistoName pileup \ -DataPUReWeighting MyDataPUHisto_2024CDEFGHI_120Bins_69200.root \ -DataPUHistoName pileup \ -> PU reweighting root files and histograms
  2. -useweight false -> We do not apply pT reweighting
  3. -nrefmax 3 -> Consider only the 3 leading gen jets
  4. -drmax 0.2 -> Consider only the matched rec-gen jets, namely those with DR < 0.2
  5. -nbinsrelrsp 60 \ -relrspmin 0.0 \ -relrspmax 3.0 \ -> Jet response is binned in 60 bins from 0 to 3
  6. -jtptmin 0 -> No raw jet pT cut
  7. -doDZcut true -> Apply the DZ cut
  8. -doNMcut true -> Apply the neutral multiplicity > 1 cut to PUPPI jets with abs eta > 3
  9. -doVetoMap true \ -JetVetoMapRootName JetVetoMaps_2024CDEFGHI.root \ -JetVetoMapHistName jetvetomap_all -> Apply the jet veto map and specify the root file name and TH2D hist name. The full veto map (with BPix and FPix areas) should be used

The main loop with all the important stuff is in lines 1468 onwards in $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetAnalyzers/bin/jet_response_analyzer_x.cc.

When ready, submit HTCondor jobs:

./SubmitStep3.sh

Input: The JRA ntuples of the PremixedPU MC dataset
Output: Root files in Step3Output

When all jobs are finished, edit the -outputDir and -era of the HarvestStep3.sh code, which calls the code $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetUtilities/src/L2Creator.cc.

./HarvestStep3.sh

Input: The root files in Step3Output which are hadded
Output: The L2Relative JEC text file and a root file named l2.root will be created inside the output directory (CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Files/RunIII2024Summer24_V1_PhiIndependent/L2L3_output/)

You will have to run this command many times because it performs 82 fits (for each eta bin, ieta) and many of them will not converge. Open the $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetUtilities/src/L2Creator.cc script. In lines 415-542 I have a few sets of alternative initial parameters. Do the following:

  1. Run ./HarvestStep3.sh and based on the printouts of the fit probabilities determine which ieta failed (those with fit prob < 0.05).
  2. Put all those ieta that failed in the first alternative set of initial parameters (line 415), recompile (scram b -j 8) and run ./HarvestStep3.sh again.
  3. Some of those ieta will fail again, so put those in the second set of parameters (line 437), recompile and re-run ./HarvestStep3.sh
  4. Repeat until no fit fails.
  5. In lines 336-343 you can change the starting point of each fit (i.e. each ieta).
  6. In lines 300-303 we have used a minimum uncertainty of 0.1% for the points of the spectrum that we fit. You can also tweak that for particular ieta fits that keep failing.

Once you make sure that all fits have a fair fit probability (usually above 5%), check the L2Relative file which contains the post-fit parameters. There should not be any insanely large values like e+150 or inf values.

Then make sure that there is no asymptotic discontinuity in the fits. Use the script below that also plots these 82 fits:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/
python3 plot_inverse_median_response_vs_ptrec.py --jetCone 4 --jetAlgo puppi --era RunIII2024Summer24 --version V1_PhiIndependent

Input: The $CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Files/RunIII2024Summer24_V1_PhiIndependent/L2L3_output/l2.root file
Output: Plots in PDF format of each fit. Also, if a warning message about asymptotic discontinuity is printed, then fix that fit because it is problematic.

You can merge all 82 PDF files into a single PDF file:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/condor_AK4PUPPI/Files/RunIII2024Summer24_V1_PhiIndependent/L2L3_output/plots/
pdfunite $(ls *.pdf | awk -F'RecEta' '/RecEta/ {
    split($2, r, ".pdf")
    split(r[1], a, "to")
    eta1 = a[1]+0; if (eta1 < 0) eta1 = -eta1
    eta2 = a[2]+0; if (eta2 < 0) eta2 = -eta2
    avg = (eta1 + eta2) / 2
    printf "%.6f %s\n", avg, $0
}' | sort -n | cut -d' ' -f2-) Fits_InverseOfResponseVsRecPt_AK4PUPPI_RunIII2024Summer24_V1_PhiIndependent.pdf

You can also plot these fits for multiple eras, with the following script:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/
python3 plot_inverse_median_response_vs_ptrec_multiple_sets.py --jetCone 4 --jetAlgo puppi

Input: The l2.root file for various MC datasets (defined in line 116)
Output: Plots in PDF format for each eta bin with the fits superimposed for all these MC datasets

Step4 -- plotting of jet response

Edit the Step4Closure.sh script that uses the $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetAnalyzers/bin/jet_correction_analyzer_x.cc script. These produce 2D plots of jet response vs pT and eta. There is the option to either apply JEC or not so that both the raw and corrected jet response can be examined.

The configuration options of the jet_correction_analyzer_x.cc script are similar to the ones of the jet_response_analyzer_x.cc script in step3. The main loop with all the important stuff is in lines 590 onwards.

Particularly for the application of JEC:

  1. If you do not want any JECs applied then only have -era RunIII2024Summer24_V1_MC \ without the -levels option
  2. If you want to apply the L2Relative text file then have -era RunIII2024Summer24_V1_MC \ -levels 2 \
  3. If you want to apply both the L1FastJet and L2Relative text files then have -era RunIII2024Summer24_V1_MC \ -levels 1 2 \

Once ready, submit the HTCondor jobs:

./SubmitStep4.sh

Input: The JRA ntuples of the PremixedPU MC dataset
Output: Root files in Step4Output

Once all jobs are finished, edit the -outputDir in the HarvestStep4.sh script which calls the $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetUtilities/src/ClosureMaker.cc code, and run:

./HarvestStep4.sh

Input: The root files in Step4Output which are hadded
Output: The ClosureVsRefPt.root root file inside -outputDir

The ClosureVsRefPt.root file contains histograms of the jet response vs pt and eta.

In order to plot the median response as a function of gen pT for the 5 different detector regions:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/
python3 plot_median_response_vs_ptgen_eta_overview.py --jetCone 4 --jetAlgo puppi --era RunIII2024Summer24 --version V1_PhiIndependent --JEC 1 --ymin 0.92 --ymax 1.08

Input: The ClosureVsRefPt.root file which is renamed to the format of (L1)L2L3ClosureVsPt_AK4PUPPI_RunIII2024Summer24.root if JECs are applied (--JEC 1) or RawResponseVsPt_AK4PUPPI_RunIII2024Summer24.root if JECs are not applied (--JEC 0).
Output: A plot in PDF format of the median jet response vs gen pT for various abs eta bins

The median jet response can also be plotted for each abs eta bin and compared between two different sets of samples, alongside their ratio:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/
python3 plot_median_response_vs_ptgen_with_ratio.py --jetCone 4 --jetAlgo puppi --MC1 Winter24 --MC2 Summer24 --JEC 0

Input: The ClosureVsRefPt.root files for the two MC campaigns, renamed as before
Output: Five plots in PDF format for each abs eta bin with the median jet response vs gen pT and the ratio for the two MCs

The response distributions, from which the median is extracted, can also be plotted using the script below:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/
python3 plot_response_distributions.py --jetCone 4 --jetAlgo puppi --JEC 0 --JetPtMin 30 --JetPtMax 35

Input: The hadded root file in Step4Output
Output: Five plots in PDF format for each abs eta bin with the response distribution for various MCs (defined in line 88) and 30 < gen pT < 35 GeV

Dedicated JECs for the BPix and FPix regions

The steps for deriving L2Relative corrections explained previously refer to the default JECs that are derived with the BPix and FPix regions excluded (i.e. included in the veto map).

After deriving these V1_PhiIndependent corrections, you need to derive dedicated ones for the BPix and FPix regions (if these issues are simulated in the MC dataset).

Rerun step3 but this time do not apply any veto map (-doVetoMap false in Step3ApplyL1.sh) and at the same time only consider jets inside the BPix region, by inserting the following inside the jet loop (right after line 1516):

//BPix area      
if( !(JRAEvt->jteta->at(iref)>=-1.479 && JRAEvt->jteta->at(iref)<=0.087 && JRAEvt->jtphi->at(iref)>=-1.2217305 && JRAEvt->jtphi->at(iref)<=-0.78539816) ) continue;

Then, proceed with running ./HarvestStep3.sh in order to create a dedicated L2Relative JEC txt file for the BPix region (eta bins between -1.479 and 0.087).

Then, do exactly the same for the FPix region, where this time use:

//FPix area
if( !( (JRAEvt->jteta->at(iref)>=-2.043 && JRAEvt->jteta->at(iref)<=-1.566 && JRAEvt->jtphi->at(iref)>=2.443461 && JRAEvt->jtphi->at(iref)<=2.7925268) || (JRAEvt->jteta->at(iref)>=-2.043 && JRAEvt->jteta->at(iref)<=-1.83 && JRAEvt->jtphi->at(iref)>=2.7925268 && JRAEvt->jtphi->at(iref)<=3.0543262) ) ) continue;

and derive a dedicated L2Relative JEC txt file for the FPix region (eta bins between -2.043 and -1.566).

Lastly, you will have to merge the default L2Relative text file with the two dedicated ones for BPix and FPix. Currently, there is no script or automated way to do this, but it is rather simple, as explained below:

  1. Modify the first line of the text file:
Replace 

{1 JetEta 1 JetPt max(0.0001,((x<[10])*([9]))+((x>=[10])*([0]+([1]/(pow(log10(x),2)+[2]))+([3]*exp(-([4]*((log10(x)-[5])*(log10(x)-[5])))))+([6]*exp(-([7]*((log10(x)-[8])*(log10(x)-[8])))))))) Correction L2Relative}

with

{2 JetEta JetPhi 1 JetPt max(0.0001,((x<[10])*([9]))+((x>=[10])*([0]+([1]/(pow(log10(x),2)+[2]))+([3]*exp(-([4]*((log10(x)-[5])*(log10(x)-[5])))))+([6]*exp(-([7]*((log10(x)-[8])*(log10(x)-[8])))))))) Correction L2Relative}
  1. Modify the eta bins that are not part of the BPix or FPix region, such that they are explicitly valid for phi between -3.1416 and +3.1416. Example below:
Replace

   0.261   0.348    13    3.718792   5300.1131     0.5479962617      13.77220075      12.01276268   -0.03938225575      3.566270518      1.078901922    -0.4527825851     0.2572148366     0.8225844628      1.134272805                8

with

   0.261   0.348       -3.1416         3.1416      13    3.718792   5300.1131     0.5479962617      13.77220075      12.01276268   -0.03938225575      3.566270518      1.078901922    -0.4527825851     0.2572148366     0.8225844628      1.134272805                8
  1. Modify the eta bins that are part of the BPix or FPix region by providing the JEC for all 3 phi bins for each particular eta bin. Example below:
If the default JEC txt file (excluding BPix) has:
  -1.131  -1.044    13   3.6936659   3438.1598     0.6239476945       9.89373998      10.16377066    -0.1389986052       2.50916051     0.7450651239     -0.275431022     0.4908949689      1.185593303      1.129665482                8
and the dedicated JEC txt file (inside BPix) has:
  -1.131  -1.044    13   4.5073879   2999.6713     0.6186154427      11.49065069      14.35726889   -0.05773817491      2.463488477      2.060894849    -0.1300350931      4.025267462      1.078775466      1.258965243                8  

then in the merged, phi-dependent JEC txt file write:

  -1.131  -1.044       -3.1416     -1.2217305      13   3.6936659   3438.1598     0.6239476945       9.89373998      10.16377066    -0.1389986052       2.50916051     0.7450651239     -0.275431022     0.4908949689      1.185593303      1.129665482                8
  -1.131  -1.044    -1.2217305    -0.78539816      13   4.5073879   2999.6713     0.6186154427      11.49065069      14.35726889   -0.05773817491      2.463488477      2.060894849    -0.1300350931      4.025267462      1.078775466      1.258965243                8
  -1.131  -1.044   -0.78539816         3.1416      13   3.6936659   3438.1598     0.6239476945       9.89373998      10.16377066    -0.1389986052       2.50916051     0.7450651239     -0.275431022     0.4908949689      1.185593303      1.129665482                8

Now that you have a new, phi-dependent L2Relative JEC txt file you can run step4, as usual, providing this txt file now.

Additionally, in step4, you can isolate the BPix and FPix regions by adding the previous snippets to the jet_correction_analyzer_x.cc file:

//BPix area      
if( !(JRAEvt->jteta->at(iref)>=-1.479 && JRAEvt->jteta->at(iref)<=0.087 && JRAEvt->jtphi->at(iref)>=-1.2217305 && JRAEvt->jtphi->at(iref)<=-0.78539816) ) continue;	    

//FPix area
if( !( (JRAEvt->jteta->at(iref)>=-2.043 && JRAEvt->jteta->at(iref)<=-1.566 && JRAEvt->jtphi->at(iref)>=2.443461 && JRAEvt->jtphi->at(iref)<=2.7925268) || (JRAEvt->jteta->at(iref)>=-2.043 && JRAEvt->jteta->at(iref)<=-1.83 && JRAEvt->jtphi->at(iref)>=2.7925268 && JRAEvt->jtphi->at(iref)<=3.0543262) ) ) continue;

and apply either the V1_PhiIndependent or the V1_PhiDependent JEC txt file to see the jet response in those regions.

A relevant script that plots the fits (inverse of median jet response vs pT) for the phi bins inside and outside the BPix or FPix regions can be used:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/
python3 plot_inverse_median_response_vs_ptrec_bpix_fpix.py --jetCone 4 --jetAlgo puppi --era RunIII2024Summer24 --version V1 --issue BPix

Input: The l2.root file for the default and the dedicated JECs
Output: Plots in PDF format with the fits to the inverse of median response for the relevant eta bins. Superimposed in each plot is the fit (correction factor) inside and outside the BPix or FPix region

Furthermore, you can plot the median jet response inside the BPix or FPix areas, using the following script:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/
python3 plot_median_response_vs_ptptcl_bpix_fpix.py --jetCone 4 --jetAlgo puppi --MC RunIII2024Summer24 --version V1 --DoBPix 1 --DoFPix 1 --JECvsPhi 1

Input: The hadded root file in Step4Output inside the BPix and/or the FPix region. Option -JECvsPhi determines what Step4Output to consider; the one derived when applying the V1_PhiIndependent or V1_PhiDependent JECs
Output: Plot in PDF format with the median jet response vs gen pT for jets inside the BPix and/or FPix region

Correction factors

The two JEC text files L1FastJet and L2Relative can be used as input to plot the correction factors as a function of eta (and as a function of pT but these are already plotted as fits to the inverse of the median response).

First do:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/
jet_draw_corrections_x -algs ak4puppi -path ./ -outputDir ./ -useL2Cor true -era RunIII2024Summer24_V1_MC

Input: If -useL2Cor true then the L2Relative text file. If -useL1FasCor true then the L1FastJet text file. If -useL1FasCor true -useL2Cor true then the L1FastJet * L2Relative text files.
Output: The Corrections_Overview_ak4puppi.root file inside the outputDir folder

Inside the $CMSSW_BASE/src/JetMETAnalysisMCtruth/JetAnalyzers/bin/jet_draw_corrections_x.cc script, in lines 307-331 we specify for what raw jet pT values we want to plot the correction factors vs eta. Currently these are: ptraw = [10, 15, 20, 30, 50, 100, 300, 500, 1000, 3000] GeV.

In order to plot the correction factors vs eta for these raw pT values use the following script:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/
python3 plot_correction_vs_eta.py --jetCone 4 --jetAlgo puppi

Input: The Corrections_Overview_ak4puppi.root files for each MC dataset (defined in line 83), renamed in the format L2L3Correction_AK4PUPPI_RunIII2024Summer24.root.
Output: Plots in PDF format of the correction factor vs eta for each raw pT value

You can also plot the correction factors of two MC dataset alongside their ratio, using the following script:

cd $CMSSW_BASE/src/JetMETAnalysisMCtruth/scripts/
python3 plot_correction_vs_eta_with_ratio.py --jetCone 4 --jetAlgo puppi --MC1 Winter24 --MC2 RunIII2024Summer24

Input: The Corrections_Overview_ak4puppi.root files for the two MC datasets, renamed in the format L2L3Correction_AK4PUPPI_RunIII2024Summer24.root.
Output: Plots in PDF format of the correction factor vs eta for each raw pT value

About

No description, website, or topics provided.

Resources

Readme
Activity

Stars

0 stars

Watchers

1 watching

Forks

1 fork
Report repository

Releases

Packages

Contributors

Languages