/Timing

Displaced Photon Analysis

Primary LanguageC++MIT LicenseMIT

Displaced Photon Analysis Framework

HUGE N.B.

OOT and GED VIDs all messed up! Make sure to apply by hand (or with new VID producers)! If doing by hand, will need to compute Smajor/Sminor!

Standard Setup

cmsrel CMSSW_9_4_13
cd CMSSW_9_4_13/src
cmsenv
git cms-init

git cms-merge-topic kmcdermo:post_reco_OOT_AND_add_OOT_VID
git cms-addpkg RecoMET/METFilters

git clone git@github.com:kmcdermo/Timing.git
pushd Timing
git remote rename origin kmcdermo
popd

scram b -j 8

Will also need to setup combine somewhere in a separate CMSSW release area (see links below). Change the relative path for the setup in scripts/common_variables.sh. Currently using Higgs combine with CMSSW_8_1_0.

N.B. The previous instructions included: git cms-merge-topic cms-met:METFixEE2017_949_v2 has been dropped as this code is merged in 9_4_12.

Made kmcdermo:post_reco_OOT_AND_add_OOT_VID with the following:

  • kmcdermo:post_reco_OOT: adds scale/smear + GED cut-based VID for OOT photons
  • combo of squashed commits from: lsoffi:CMSSW_10_1_1_OOT_PhotonID_wSMaj and kmcdermo:add_OOT_VID
  • add commits for OOT VID for GED and OOT Photons

Additional useful links

Notation on variable names in nTuples

  • rh = recHit

  • sc = SuperCluster

  • ph = photon

  • pho = photon (used in branch names)

  • is* = "bool" for is*

    • isHLT = is matched to HLT object
    • isGen = is matched to gen particle

  • genVpion = generator V-pion (from HVDS)

  • gengmsb = generator GMSB matched neutralino

  • object*

    • E = energy
    • pt = transverse energy

  • jet* (vector up to 10)

    • ID = jet cut based ID
    • NHF = neutral hadron energy fraction
    • NEMF = neutral EM energy fraction
    • CHF = charged hadron energy fraction
    • CEMF = charged EM energy fraction
    • MUF = muon energy fraction
    • NHM = neutral multiplicity
    • CHM = charged multiplicity

  • rh* = vector of rechit values

    • ID = raw detector ID
    • TOF = time of flight corrected to PV
    • isOOT = if true, recHit is flagged OOT collection
    • isGS6,isGS1 = if true, photon has Gain Swith 6, 1 (reminder, base gain is 12, then 6, then 1)
    • ped* = pedestal noise (Gain 12, 6, 1)
    • pedrms* = pedestal noise error (Gain 12, 6, 1)

  • pho* [photons 0-3]

    • passEleVeto = photon passes electron veto
    • hasPixSeed = photon has pixel seed
    • HoE = hadTowerOverEM (single HCAL tower over EM energy)
    • r9 = seed E / 3x3 E
    • sieie = sigma_ieta,ieta
    • sipip = sigma_iphi,iphi
    • sieip = sigma_ieta,iphi
    • smaj = S_major (semi-major axis)
    • smin = S_minor (semi-minor axis)
    • alpha = angle of smaj to origin
    • suisseX = swisscross, seed energy / four surrounding energy
    • gedID = GED cut-based ID
    • ootID = OOT cut-based ID
    • Iso = Isolation
    • IsoC = Isolation corrected for (rho, pt)
      • ChgHadIso = Charged Hadron Iso
      • NeuHadIso = Neutral Hadron Iso
      • PhoIso = Photon Iso
      • EcalPFClIso = ECAL PF Cluster Iso
      • HcalPFClIso = HCAL PF Cluster Iso
      • TrkIso = Tracker Iso (hollow cone dR < 0.3)
    • recHits = vector of indices inside rh* branches for each photon
    • seed = index of seed rec hit for photon
    • seed* = value of seed rec hit branch
    • weighttime* = cluster weighted time, computed at first level skim
    • is (bools)
      • isEB = seed rechit is from ECAL barrel, else assumed endcap
      • isSignal = photon is matched to gen photon from signal particle
      • isOOT = if true, photon is from OOT collection, else from GED collection
      • isTrk = if true, object is matched to a single track pT > 5 within dR < 0.2

uncertainties:

  • *Abs = (Absolute) Uncertainty = energy +/- *Abs
  • *Rel = (Relative) Uncertainty = energy * (1 +/- *Rel)
  • *SF = (Scale Factor) = energy * SF

Core chain of analysis

Produce ntuples with multicrab:

  • test/multicrab_dispho_Data.py
  • test/multicrab_dispho_Bkgd.py
  • test/multicrab_dispho_GMSB.py
  • test/multicrab_dispho_HVDS.py

N.B. XSecs taken from GenXSecAnalyzerTool (see fork of genproductions) --> Currently sitting here: /afs/cern.ch/work/k/kmcdermo/private/dispho/XSEC/CMSSW_9_4_10/src/genproductions/test/calculateXSectionAndFilterEfficiency

Each script calls config: test/dispho.py. Make sure to configure correctly! Plugins used:

  • NTuplizer : plugins/Dispho.cc/hh and plugins/DisPhoTypes.hh
  • HelperUtils: plugins/CommonUtils.cc/hh

Output of ntuples end up in EOS group space: currently configured for /eos/cms/group/phys_exotica/displacedPhotons/nTuples/2017. Now begin macro chain of analysis:

  • cd to dispho_work/macros.

  • Produce skims with new time branches, event weights, basic core skim + met flags: scripts/skimAndMerge/*.sh

    • Also makes PU weights...
    • Can launch skims in parallel with loops in background

  • Produce data+bkgd mc skims into a single file: scripts/runFastSkimmer.sh

  • Produce signal skims into single file: scripts/runSignalSkimmer.sh

  • Make sure to do both for GJets CR, QCD CR, and SR selection

  • Then run the scripts for making the variable weight trees in CR to SR: scripts/makeWgtsAndPlots.sh

  • Then check 1D final plots: scripts/makePlotsForSR.sh

  • Finally, run 2D plots for fis: scripts/makeLimits.sh

Plots outputted to lxplus eos user space: /eos/user/k/kmcdermo/www/dispho/plots

Estimating Uncertainties

"Nominal skims" are in skims/v4/unc_tests/signals_nominal.root. To produced new skims, first set ${outbase} in scripts/skimAndMerge/common_variables.sh and change the skim config input to the unc being tests. Then run the first Skimmer.

Next, using the new ${outbase}, pass this to the SignalSkimmer, with the cuts_v2 nominal file. Then, move the resulting file into skims/v4/unc_tests, and add a new input line in scripts/common_variables.sh to where you placed the file.

Finally, run the TimeAdjuster, skipping bkgd+data. uncertaintyChain.sh will do the rest.

Extracting lumi info

Compute lumi for 2017:

brilcalc lumi --normtag /cvmfs/cms-bril.cern.ch/cms-lumi-pog/Normtags/normtag_PHYSICS.json -u /fb -i /afs/cern.ch/cms/CAF/CMSCOMM/COMM_DQM/certification/Collisions17/13TeV/ReReco/Cert_294927-306462_13TeV_EOY2017ReReco_Collisions17_JSON_v1.txt

Compute PU info for 2017 with 94X environment:

pileupCalc.py -i ~/public/input/golden2017.json --inputLumiJSON ~/private/bril/pu2017json.txt --calcMode true --minBiasXsec 69200 --maxPileupBin 200 --numPileupBins 200 MyDataPileupHistogram.root

Old files of interest stored in cms user kmcdermo space

  • Backup of older fast skims: /eos/cms/store/user/kmcdermo/nTuples/BACKUP/afs/cern.ch/work/k/kmcdermo/private/dispho/Analysis
    • production directory is older mirror of development directory
    • CMSSW_9_4_10 is copy of since deprecated skims
  • Backup of since deprecated skims (original output from CRAB deleted): /eos/cms/store/user/kmcdermo/nTuples/skims
  • Backup of since deprecated unskimmed CRAB output for OOT VID studies: /eos/cms/store/user/kmcdermo/nTuples/unskimmed

To do:

  • switch ctau to tau [ns]
  • make sure to use REAL ctau in limit plots... update scripts to read in this value
  • make exclusion in 2D vs neutralino mass
  • make exclusion in 1D vs REAL ctau (slices of neutralino mass)