Scripts for Clustering, Classifying and Regressing Diphoton Events

Runs over MiniAOD, creates clusters based on hits in ECAL. Then applies classification NN to classify clusters as monophoton, diphoton, or hadronic, and then predicts the mass of the clusters.

Includes configs for NanoAODv9 production in UL2018.

Here are some instructions for deploying at the ND CRC.

Start SLC7 Container:

/cvmfs/cms.cern.ch/common/cmssw-el7 --cleanenv -B /scratch365/ -B /opt/ -B /cvmfs/

Create release, get code, and compile:

mkdir MLPhotons
cmsrel CMSSW_10_6_19_patch2
cd CMSSW_10_6_19_patch2/src
cmsenv
git clone https://github.com/atownse2/MLPhotons.git .
scram b

Get a test file (you might need to find a new file):

xrdcp root://ndcms.crc.nd.edu//store/mc/RunIISummer20UL18MiniAODv2/BkkToGRadionToGGG_M1-1000_R0-12p5_TuneCP5_13TeV-madgraph-pythia8/MINIAODSIM/106X_upgrade2018_realistic_v16_L1v1-v3/50000/128D50BB-33ED-E742-BAC1-A4F60467D2AA.root ./test.root

Test run:

./run_signal_test.sh