Dijet Scouting Run3 nTuple-Maker

Important

This document covers the nTuple production workflow.
The analysis workflow is documented in README_Analysis.md.

Warning

This framework is only tested for Fermilab LPC machines.
It might need some adjustments for Lxplus.

---

---

Setup Environment:

source /cvmfs/cms.cern.ch/cmsset_default.sh

cmsrel CMSSW_15_0_6
cd CMSSW_15_0_6/src
cmsenv

## Clone COMBINE tool
git -c advice.detachedHead=false clone --depth 1 --branch v10.5.1 https://github.com/cms-analysis/HiggsAnalysis-CombinedLimit.git HiggsAnalysis/CombinedLimit

## Clone Combine Harvester
git clone https://github.com/cms-analysis/CombineHarvester.git
cd CombineHarvester
git checkout v3.0.0-pre1
cd ..

## Dijet Framework
git clone https://github.com/asimsek/DijetScoutingRun3Analyzer.git

## CMSSW_15_X: fix for CombineHarvester BuildFiles
sed -i 's/name="python"/name="python3"/g' \
  CombineHarvester/CombinePdfs/bin/BuildFile.xml \
  CombineHarvester/CombineTools/bin/BuildFile.xml

## Build once from CMSSW src
scram b clean; scram b -j"$(nproc --ignore=2)"

---

Pull the necessary/custom files for Higgs Combine Tool from a different repo:

From your new COMBINE area:

cd $CMSSW_BASE/src/HiggsAnalysis/CombinedLimit

# Add an extra repo for the necessary files (will remove at the end)
git remote add cmsdijetfunctions https://github.com/asimsek/CMSDijetFunctions.git 2>/dev/null || true

# Validate:
git remote -v

# Set the branch
git fetch --depth 1 cmsdijetfunctions main

Now define the explicit list of files you want to copy and check them out:

# Create a minimal list for your need:
FILES=(
  'src/RooDijet*.cc'
  'src/RooModExp*.cc'
  'src/RooAtlas*.cc'
  'src/RooModDijet*.cc'
  'interface/RooDijet*.h'
  'interface/RooModExp*.h'
  'interface/RooAtlas*.h'
  'interface/RooModDijet*.h'
)

git restore --source=cmsdijetfunctions/main --worktree -- "${FILES[@]}"

Remove the newly added remote (cmsdijetfunctions) repo and validate:

# Remove:
git remote remove cmsdijetfunctions 2>/dev/null || true

# Validate:
git remote -v

---

Register the new classes:

Add the following to the src/classes_def.xml file (just before </lcgdict> line):

  <class name="RooDijet3ParamBinPdf" />
  <class name="RooDijetBinPdf" />
  <class name="RooDijet5ParamBinPdf" />
  <class name="RooDijet6ParamBinPdf" />
  <class name="RooDijet5ParamPolyExtBinPdf" />
  <class name="RooDijet6ParamPolyExtBinPdf" />
  <class name="RooDijet7ParamPolyExtBinPdf" />
  <class name="RooModExp3ParamBinPdf" />
  <class name="RooModExp4ParamBinPdf" />
  <class name="RooModExpBinPdf" />
  <class name="RooModExp6ParamBinPdf" />
  <class name="RooAtlas3ParamBinPdf" />
  <class name="RooAtlas4ParamBinPdf" />
  <class name="RooAtlasBinPdf" />
  <class name="RooAtlas6ParamBinPdf" />
  <class name="RooModDijet3ParamBinPdf" />
  <class name="RooModDijet4ParamBinPdf" />
  <class name="RooModDijet5ParamBinPdf" />
  <class name="RooModDijet6ParamBinPdf" />

Add the following to the src/classes.h file (end of the file is fine):

#include "HiggsAnalysis/CombinedLimit/interface/RooDijet3ParamBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooDijetBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooDijet5ParamBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooDijet6ParamBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooDijet5ParamPolyExtBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooDijet6ParamPolyExtBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooDijet7ParamPolyExtBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooModExp3ParamBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooModExp4ParamBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooModExpBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooModExp6ParamBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooAtlas3ParamBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooAtlas4ParamBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooAtlasBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooAtlas6ParamBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooModDijet3ParamBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooModDijet4ParamBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooModDijet5ParamBinPdf.h"
#include "HiggsAnalysis/CombinedLimit/interface/RooModDijet6ParamBinPdf.h"

---

Rebuild

cd $CMSSW_BASE/src/HiggsAnalysis/CombinedLimit

scramv1 b clean; scramv1 b -j$(nproc --ignore=2)

---

---

Important Notes before the nTuple production:

Important

Do NOT forget to validate/define the JEC paths to the data/cfg/data_jec_list.txt for data and data/cfg/mc_jec_list.txt for MC.
MC should be defined as a plain year (e.g., 2024, 2025, etc.); However, data should be defined era by era since each era usually has a different JEC configuration.

Important

GoldenJSON files are defined in the input cfg file (e.g.; inputFiles_PFScouting_NanoAOD/PFScouting_2024H_cfg.txt).
Validate/Pull the latest GoldenJSON file for your dataset from official repo: CAF-Certification
Collisions24 is for 2024 data and Collisions25 is for 2025 data.
Please also note that the GoldenJSON files are updated regularly; However, the DQM group can also update these files outside the data taking year! Validating the latest GoldenJSON file usage will be beneficial before any nTuple production activities.

---

---

Local Test - NanoAOD nTuple Maker:

Caution

Run the following commands inside the main directory: cd $CMSSW_BASE/src/DijetScoutingRun3Analyzer

For Scouting Data:

## Create a temporary list:
cat > /tmp/nano_test_2024H.txt << 'EOF'
root://cms-xrd-global.cern.ch//store/data/Run2024H/ScoutingPFRun3/NANOAOD/ScoutNano-v1/2520000/03e073d9-ab61-450d-aa10-fa050e94a16b.root
EOF

## Framework uses `analysisClass.C` file to process data.
## Therefore, create a symbolic link (`symlink`/`soft link`) to use any of your "analysisClass" as `analysisClass.C`:
ln -sf analysisClass_mainDijetPFScoutingSelection_Run3_NanoAOD_Recluster.C src/analysisClass.C

## Generate the ROOT reader class (`rootNtupleClass.*`) so branch declarations and addresses match the NanoAOD file/tree you are about to process:
yes | ./scripts/make_rootNtupleClass.sh -f root://cms-xrd-global.cern.ch//store/data/Run2024H/ScoutingPFRun3/NANOAOD/ScoutNano-v1/2520000/03e073d9-ab61-450d-aa10-fa050e94a16b.root -t Events

## Compile your analyzer:
make clean && make -j8

## Start nTuple production:
./main /tmp/nano_test_2024H.txt config/cutFile_mainDijetPFScoutingSelection_Run3.txt Events test_NanoAOD_2024H_n0 test_NanoAOD_2024H_n0

## Optional: Validate the output:
root -l -q -e 'TFile f("test_NanoAOD_2024H_n0_reduced_skim.root"); TTree* t=(TTree*)f.Get("rootTupleTree/tree"); if(t) t->Print(); else f.ls();'

For Monitoring Data:

## Create a temporary list:
cat > /tmp/nano_test_monitoring_2024H.txt << 'EOF'
root://cms-xrd-global.cern.ch//store/data/Run2024I/ScoutingPFMonitor/NANOAOD/PromptReco-v1/000/386/508/00000/8d57403b-0253-4f30-9624-494d9c843bb5.root
EOF

## Framework uses `analysisClass.C` file to process data.
## Therefore, create a symbolic link (`symlink`/`soft link`) to use any of your "analysisClass" as `analysisClass.C`:
ln -sf analysisClass_mainDijetPFScoutingSelection_Run3_NanoAOD_Recluster.C src/analysisClass.C

## Generate the ROOT reader class (`rootNtupleClass.*`) so branch declarations and addresses match the NanoAOD file/tree you are about to process:
yes | ./scripts/make_rootNtupleClass.sh -f root://cms-xrd-global.cern.ch//store/data/Run2024I/ScoutingPFMonitor/NANOAOD/PromptReco-v1/000/386/508/00000/8d57403b-0253-4f30-9624-494d9c843bb5.root -t Events

## Compile your analyzer:
make clean && make -j8

## Start nTuple production:
./main /tmp/nano_test_monitoring_2024H.txt config/cutFile_mainDijetPFScoutingSelection_Run3.txt Events test_monitoring_NanoAOD_2024H_n0 test_monitoring_NanoAOD_2024H_n0

## Optional: Validate the output:
root -l -q -e 'TFile f("test_monitoring_NanoAOD_2024H_n0_reduced_skim.root"); TTree* t=(TTree*)f.Get("rootTupleTree/tree"); if(t) t->Print(); else f.ls();'

For QCD MC:

## Create a temporary list:
cat > /tmp/nano_test_QCD.txt << 'EOF'
root://cms-xrd-global.cern.ch//store/mc/RunIII2024Summer24NanoAOD/QCD_Bin-PT-1000to1500_TuneCP5_13p6TeV_pythia8/NANOAODSIM/140X_mcRun3_2024_realistic_v26-v2/120000/00ba6769-d1a4-4faa-be8f-2157b2a039d7.root
EOF

## Framework uses `analysisClass.C` file to process data.
## Therefore, create a symbolic link (`symlink`/`soft link`) to use any of your "analysisClass" as `analysisClass.C`:
ln -sf analysisClass_mainDijetPFScoutingSelection_Run3_NanoAOD_Recluster.C src/analysisClass.C

## Generate the ROOT reader class (`rootNtupleClass.*`) so branch declarations and addresses match the NanoAOD file/tree you are about to process.
## IMPORTANT: Don't forget to update the root file. You may need to use `root://cms-xrd-global.cern.ch/` or `root://cmsxrootd.fnal.gov/` to access dataset root files.
yes | ./scripts/make_rootNtupleClass.sh -f root://cms-xrd-global.cern.ch//store/mc/RunIII2024Summer24NanoAOD/QCD_Bin-PT-1000to1500_TuneCP5_13p6TeV_pythia8/NANOAODSIM/140X_mcRun3_2024_realistic_v26-v2/120000/00ba6769-d1a4-4faa-be8f-2157b2a039d7.root -t Events

## Compile your analyzer:
make clean && make -j8

## Start nTuple production:
./main /tmp/nano_test_QCD.txt config/cutFile_mainDijetPFScoutingSelection_Run3.txt Events test_NanoAOD_QCD_n0 test_NanoAOD_QCD_n0

## Optional: Validate the output:
ls -lh test_NanoAOD_QCD_n0_reduced_skim.root
root -l -q 'TFile f("test_NanoAOD_QCD_n0_reduced_skim.root"); TTree* t=(TTree*)f.Get("rootTupleTree/tree"); if(t) t->Print();'

---

nTuple production on HTCondor:

Note

HTCondor is used to parallelize nTuple production by splitting large datasets into many independent jobs, which significantly reduces total processing time and improves large-scale production throughput.

Create & Send jobs to Condor:

Caution

Run the HTCondor scripts/commands inside the dijetCondor directory:
cd $CMSSW_BASE/src/DijetScoutingRun3Analyzer/dijetCondor

Data (PFScouting):

python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2024/PFScouting_2024C_cfg.txt --force-new-list --request-memory-mb 4096
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2024/PFScouting_2024D_cfg.txt --force-new-list --request-memory-mb 4096 
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2024/PFScouting_2024E_cfg.txt --force-new-list --request-memory-mb 4096
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2024/PFScouting_2024F_cfg.txt --force-new-list --request-memory-mb 4096
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2024/PFScouting_2024G_cfg.txt --force-new-list --request-memory-mb 4096
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2024/PFScouting_2024H_cfg.txt --force-new-list --request-memory-mb 4096
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2024/PFScouting_2024I_cfg.txt --force-new-list --request-memory-mb 4096


## 2025C
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2025/PFScouting_2025C_v1_cfg.txt --force-new-list --request-memory-mb 4096
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2025/PFScouting_2025C_v2_cfg.txt --force-new-list --request-memory-mb 4096
## 2025D
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2025/PFScouting_2025D_v1_cfg.txt --force-new-list --request-memory-mb 4096
## 2025E
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2025/PFScouting0_2025E_v1_cfg.txt --force-new-list --request-memory-mb 4096
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2025/PFScouting1_2025E_v1_cfg.txt --force-new-list --request-memory-mb 4096
## 2025F
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2025/PFScouting0_2025F_v1_cfg.txt --force-new-list --request-memory-mb 4096
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2025/PFScouting0_2025F_v2_cfg.txt --force-new-list --request-memory-mb 4096
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2025/PFScouting1_2025F_v1_cfg.txt --force-new-list --request-memory-mb 4096
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2025/PFScouting1_2025F_v2_cfg.txt --force-new-list --request-memory-mb 4096
## 2025G
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2025/PFScouting0_2025G_v1_cfg.txt --force-new-list --request-memory-mb 4096
python3 condor_submit_nanoAOD.py -c inputFiles_PFScouting_NanoAOD/2025/PFScouting1_2025G_v1_cfg.txt --force-new-list --request-memory-mb 4096

Tip

--no-submit: dryRun (create files without sending jobs to condor). --cms-connect: Additional HTCondor arguments for CMS Connect machines.

Data (PFMonitoring - for Trigger Efficiency):

python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2024/PFMonitoring_2024C_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2024/PFMonitoring_2024D_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2024/PFMonitoring_2024E_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2024/PFMonitoring_2024F_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2024/PFMonitoring_2024G_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2024/PFMonitoring_2024H_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2024/PFMonitoring_2024I_cfg.txt --force-new-list

## 2025C
python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2025/PFMonitoring_2025C_v1_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2025/PFMonitoring_2025C_v2_cfg.txt --force-new-list
## 2025D
python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2025/PFMonitoring_2025D_v1_cfg.txt --force-new-list
## 2025E
python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2025/PFMonitoring_2025E_v1_cfg.txt --force-new-list
## 2025F
python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2025/PFMonitoring_2025F_v1_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2025/PFMonitoring_2025F_v2_cfg.txt --force-new-list
## 2025G
python3 condor_submit_nanoAOD.py -c inputFiles_PFMonitoring_NanoAOD/2025/PFMonitoring_2025G_v1_cfg.txt --force-new-list

Tip

--no-submit: dryRun (create files without sending jobs to condor). --cms-connect: Additional HTCondor arguments for CMS Connect machines.

QCD MC:

python3 condor_submit_nanoAOD.py -c inputFiles_QCD_NanoAOD/QCDMC_2024_PT50to80_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_QCD_NanoAOD/QCDMC_2024_PT80to120_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_QCD_NanoAOD/QCDMC_2024_PT120to170_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_QCD_NanoAOD/QCDMC_2024_PT170to300_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_QCD_NanoAOD/QCDMC_2024_PT300to470_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_QCD_NanoAOD/QCDMC_2024_PT470to600_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_QCD_NanoAOD/QCDMC_2024_PT600to800_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_QCD_NanoAOD/QCDMC_2024_PT800to1000_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_QCD_NanoAOD/QCDMC_2024_PT1000to1500_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_QCD_NanoAOD/QCDMC_2024_PT1500to2000_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_QCD_NanoAOD/QCDMC_2024_PT2000to2500_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_QCD_NanoAOD/QCDMC_2024_PT2500to3000_cfg.txt --force-new-list
python3 condor_submit_nanoAOD.py -c inputFiles_QCD_NanoAOD/QCDMC_2024_PT3000toInf_cfg.txt --force-new-list

Warning

A dummy (or real) GoldenJSON file definition in the MC config files is necessary to avoid any possible issues.
Given GoldenJSON won't be used for MC samples.

---

Condor output (root) check & re-submit:

cd $CMSSW_BASE/src/DijetScoutingRun3Analyzer/dijetCondor

# Build once
c++ -O3 -march=native -DNDEBUG -o check_condor_outputs check_condor_outputs.cpp $(root-config --cflags --libs)

# Data (Scouting)
./check_condor_outputs cjobs_ScoutingPFRun3_Run2024G_ScoutNano_v1_NANOAOD_05March2026_13 /eos/uscms/store/group/lpcjj/Run3PFScouting/nanoAODnTuples/2024/ScoutingPFRun3/ScoutingPFRun3_Run2024G_ScoutNano_v1 --config inputFiles_PFScouting_NanoAOD/2024/PFScouting_2024G_cfg.txt

# Data (Monitoring)
./check_condor_outputs cjobs_ScoutingPFMonitor_Run2024H_PromptReco_v1_NANOAOD_03March2026_02 /eos/uscms/store/group/lpcjj/Run3PFScouting/nanoAODnTuples/2024/ScoutingPFMonitor/ScoutingPFMonitor_Run2024H_PromptReco_v1 --config inputFiles_PFMonitoring_NanoAOD/2024/PFMonitoring_2024H_cfg.txt

# QCD MC
./check_condor_outputs cjobs_QCD_Bin-PT-80to120_TuneCP5_13p6TeV_pythia8_NANOAODSIM_03March2026_04 /eos/uscms/store/group/lpcjj/Run3PFScouting/nanoAODnTuples/2024/QCD_Bin-PT-80to120_TuneCP5_13p6TeV_pythia8 --config inputFiles_QCD_NanoAOD/QCDMC_2024_PT80to120_cfg.txt --check-subdirs

Tip

Use --check-subdirs to include sub-folders of the given directory for the root file search.
Use --job-start and --job-end in order to give a range that requested to be checked.
Use --noPerFile: skip per-job matching and do only total event comparison.
Use --threads N: override the C++ worker count. If omitted, it uses hardware concurrency.
Use --resubmit --request-memory-mb 4096 to resubmit jobs for the missing files and request more memory for them.
Only target EOS folder can be given (without cjobs_ folder) to count the event number and compare with the dataset - even single root file can be given for this purpose.

Processed Lumi Calculation:

Warning

The following brilcalc setup instruction is prepared for Fermilab LPC.
Please follow the official brilcalc instructions for lxplus.

Setup brilcalc (for LPC):

### Find which brilconda python exists on your node (pick one that exists)
ls -d /cvmfs/cms-bril.cern.ch/brilconda* 2>/dev/null

### Prefer a python3-based one if present (example: brilconda310)
BRILPY=/cvmfs/cms-bril.cern.ch/brilconda310/bin/python3

### Install brilcalc
$BRILPY -m pip install --user --upgrade brilws

export PATH=$HOME/.local/bin:$PATH

## Validate
which brilcalc
brilcalc --version

Produce crab-like processed_lumi.json file:

python3 condor_report.py -c inputFiles_PFScouting_NanoAOD/PFScouting_2024H_cfg.txt -o processed_lumis_2024H.json \
    --brilcalc --unit /fb --normtag /cvmfs/cms-bril.cern.ch/cms-lumi-pog/Normtags/normtag_PHYSICS.json \
    --no-golden \
    --brilcalc-extra -c "$CMS_PATH/SITECONF/local/JobConfig/site-local-config.xml" 

python3 condor_report.py -c inputFiles_PFScouting_NanoAOD/PFScouting_2024H_cfg.txt -o processed_lumis_2024H.json \
    --brilcalc --unit /fb --normtag /cvmfs/cms-bril.cern.ch/cms-lumi-pog/Normtags/normtag_PHYSICS.json \
     --golden-json \
     --brilcalc-extra -c "$CMS_PATH/SITECONF/local/JobConfig/site-local-config.xml"

Tip

--brilcalc argument can be used to perform brilcalc on the newly generated json file.
--input-json <fileName>.json argument can be used to process an existing json file with brilcalc.
--golden-json can be used to filter output json according to the goldenJSON file (oposite is --no-golden).

brilcalc lumi -c "$CMS_PATH/SITECONF/local/JobConfig/site-local-config.xml" -u /fb \
    -i processed_lumis_2024H.json \
    --normtag /cvmfs/cms-bril.cern.ch/cms-lumi-pog/Normtags/normtag_PHYSICS.json

Merge MC Samples:

Note

This script allows you to merge (hadd) the sample outputs.
The result is one ROOT file per sample type (e.g.; one output for 80to120, one for 120to170, etc.).
This is a necessary step for MC scaling with xSec values.

python3 merge_samples.py /eos/uscms/store/group/lpcjj/Run3PFScouting/nanoAODnTuples/2024/QCDSamples --sample-glob "QCD_Bin-PT-*"

Tip

--sample-glob argument lets you select which sample folders under the base directory will be merged.

---

---

Condor Job Diagnosis (Fermilab LPC):

Note

If you have some jobs that are in the "Hold/Held" state, use the following commands CAREFULLY to understand the problem!
And then either release the job(s) for re-processing, or remove the it/them completely!

## Learn the hold reason of the jobs:
condor_q -hold -af HoldReason

## Release the job for re-processing:
condor_release -const 'jobstatus==5' -name lpcschedd6

## WARNING!!
## This removes all the jobs with the same `.sh` file name PERMENANTLY!!:
## ONLY USE THIS IF YOU NEED TO REMOVE ALL THE SAME TYPE OF JOBS!!
## Don't forget to change the lpcschedd number and also the file name prefix (until & including `_n`).
condor_rm -name lpcschedd5 -constraint "Owner==\"$USER\" && regexp(\".*QCD_Bin-PT-80to120_TuneCP5_13p6TeV_pythia8_NANOAODSIM_n[0-9]+\\.sh$\", Cmd)"


## Shows a detailed match diagnosis for the job, explaining why it is not running
## e.g., requirement mismatch, memory, site constraints.
condor_q -better-analyze <jobID>

Dataset Event Count Cross-Check:

Caution

Run this command inside the dijetCondor directory:
cd $CMSSW_BASE/src/DijetScoutingRun3Analyzer/dijetCondor

# Data (Scouting)
python3 check_dataset_entries.py /ScoutingPFRun3/Run2024H-ScoutNano-v1/NANOAOD --workers 8 --backend root

# Data (Monitoring)
python3 check_dataset_entries.py /ScoutingPFMonitor/Run2024H-PromptReco-v1/NANOAOD --workers 8 --backend root

# QCD MC
python3 check_dataset_entries.py /QCD_Bin-PT-80to120_TuneCP5_13p6TeV_pythia8/RunIII2024Summer24NanoAOD-140X_mcRun3_2024_realistic_v26-v2/NANOAODSIM --workers 8 --backend root

Tip

This script opens all ROOT files in the given DAS dataset, counts the total number of entries in the Events tree, and compares the summed value with DAS nevents.
--backend root is recommended on LPC/CMSSW setups.
--workers controls the parallel file checks, and --tree can be used if the input tree name is different from Events.

---

---

CMS Connect Usage:

Important

Need a Fermilab (LPC) computing account in order to request and use this sytem.
If you're working on lxplus, skip this part!

LPC Condor resources are limited and might take very long due to the user-priority and high-demand. CMS Connect provides a Tier3-like environment for condor jobs and allow to use all resources available in the CMS Global Pool.

Registration & Instructions:

Register to ci-connect

Caution

ci-connect is the new platform and cilogon is deprecated. Follow the Twiki page instruction for SSH Key steps but use ci-connect page instead of cilogon. Allow system a few hours after uploading the SSH Key to your ci-connect profile.

• CMS Connect Setup
• CMS Connect Examples

Warning

The following steps can only be used once the system recognizes your new ssh-key.

Login to CMS Connect:

ssh <username>@login.uscms.org

Upload your grid cerificate from your LPC Area:**

Note

You might need to delete the old ones in ~/.globus folder (if exist)!

copy_certificates

Create a new proxy:

voms-proxy-init -voms cms

Warning

Important: Set a default project for your account with the following commands to be able to submit any Condor jobs:

mkdir ~/.ciconnect
echo "cms.org.cms" > ~/.ciconnect/defaultproject

---

---

Useful Links

• Run3 Luminosity and uncertainty recommendations
• Golden JSON
• BrilCalc - Lumi Calculation
• Run3 JERC
• Run3 PdmV
• Rucio CLI Twiki
• Rucio WebUI
• LPC EOS Guidelines