DarkPhotonAnalysis

Prompt Low Mass

Getting the Skimmed Ntuples

To run:

make 
./create_reduced_tree txt_file outfilepath

txt_file is a text file containing the LFN of the Ntuples. In the directory ntuples/split are text files containing ~30 Ntuples each. So for example, we can do:

./create_reduced_tree ntuples/split/xaa scout_skimmed_1.root

to get a skimmed Ntuple scout_skimmed_1.root from the Ntuples in file xaa.

The files submitBatchJobs.sh and skimJobs.sh are used to submit batch jobs to do the skimming. submitBatchJobs.sh takes the path to the directory containing the text files as a command-line argument. skimJobs.sh is the bash script submitted to batch.

You can run it with:

. submitBatchJobs.sh ntuples/split

Fitting

The executable to do the fitting is ./lowMass_prompt. There are two ways to run it:

1. Running on a single skimmed Ntuple

   make
   ./lowMass_prompt -inputFiles=inputfile
   

   If -inputFiles is not specified, the default skimmed Ntuple used is /mnt/hadoop/store/user/idutta/DarkPhoton/Samples/xcg2Dec2018/2Dec2018xcg_job0_scout_skimmed.root.

2. Running on multiple skimmed Ntuples

   make
   ./lowMass_prompt -tchain -inputFiles=scout_skimmed.txt
   

   In this case scout_skimmed.txt is a text file containing the LFN of the skimmed Ntuples. The -tchain flag specifies that a TChain is used to create a TTree from multiple ROOT files.

Additional Optional Flags

-help Displays a help message with information about the flags.

-fitOutFile Used to specify the path to the txt file to write fit results. Default path is fitSB_JPsi_output.txt.
Example:

./lowMass_prompt -fitOutFile=fit.txt

-imgtag Used to specify tag for identifying output image file and ROOT file storing RooWorkspace. No tags are used as default.
Example:

./lowMass_prompt -imgtag=allEvents

-totalEntries Specifies program to use the number of entries in the range 0 - 10 GeV from the input file for setting initial values of nsig and nbkg. Default uses the number of entries in the JPsi range (2.0 - 3.5 GeV).
Example:

./lowMass_prompt -totalEntries

Skimmed NTuples

/mnt/hadoop/store/user/ufay/lowMassPromptDP/subTree_500000.root
/mnt/hadoop/store/user/ufay/lowMassPromptDP/subTree_1000000.root

These are smaller NTuples containing 500 000 and 1 000 000 total entries respectively in the whole range 0 - 10 GeV. They can be used for debugging.

Displaced Low Mass Limits

To run:

make getLimits
./getLimits --inputFileSig=<input signal ROOT file path> --inputFileBkg=<input bkg ROOT file path> \
  --treeName=tree --outputFile=<output ROOT file name> --datacard=<datacard file path> \
  --mzd=<mzd [GeV]> --tau0=<tau0 [mm]> --binNumber=<binNumber> --vtxCut=<vertex cut [cm]> \
  --sigFit=<signal fit PDF> --bkgFit=<bkg fit PDF> > output.txt

Example:

make getLimits
./getLimits --inputFileSig=../trimscoutV2/all-trimscoutV2-darkphoton_mzd_20_tau0_100mm_full.root \
  --inputFileBkg=../trimscoutV2/all-trimscoutV2-darkphoton_mzd_20_tau0_1000mm_full.root \
  --treeName=tree --outputFile=combine10.root --datacard=datacards/combineDatacard10.txt \
  --mzd=20 --tau0=100 --binNumber=1 --vtxCut=10 \
  --sigFit=bw --bkgFit=expo > output.txt

This creates two new directories output and datacards in the current directory. output contains plots of the background and signal fits (bkgFit_bin1_cut_10cm.png and sigFit_bin1_cut_10cm.png), and the toy data thrown from the fitted PDF (dataToy_bin1_cut_10cm.png). datacards contains the combine datacard (combineDatacard10.txt) and the ROOT file with the RooWorkspace for the combine input (combine10.root).

The datacard can then be put into combine with

combine -M Asymptotic datacards/combineDatacard10.txt --minimizerStrategy=1 --X-rtd ADDNLL_RECURSIVE=0 > limit.txt

To get the correct normalization for the signal shape, the number of signal events without any cut must be changed for each lifetime. This is done by changing the N_sigNoCut variable here: https://github.com/shufay/DarkPhotonAnalysis/blob/master/src/fitDarkphoton.cc#L752.

The signal yield without any cut should also be changed for the different background fits used. This is done by changing the sigYieldNoCut variable here: https://github.com/shufay/DarkPhotonAnalysis/blob/master/src/fitDarkphoton.cc#L751.

The implemented PDFs for fitting could be specified with the following:

• Exponential — expo
• Double Exponential — doubleExpo
• Power — pow
• Double Power — doublePow
• 2nd Order Bernstein Polynomial — bernPoly2
• 3rd Order Bernstein Polynomial — bernPoly3
• 2nd Order Chebychev Polynomial — chebPoly2

Input Files

Signal

tau₀ = 10e-10 mm

/afs/cern.ch/work/u/ufay/public/CMSSW_9_4_0_patch1/src/trimscoutV2/all-trimscoutV2-darkphoton_mzd_20_tau0_10e-10mm_full.root

tau₀ = 1 mm

/afs/cern.ch/work/u/ufay/public/CMSSW_9_4_0_patch1/src/trimscoutV2/all-trimscoutV2-darkphoton_mzd_20_tau0_1mm_full.root

tau₀ = 10 mm

/afs/cern.ch/work/u/ufay/public/CMSSW_9_4_0_patch1/src/trimscoutV2/all-trimscoutV2-darkphoton_mzd_20_tau0_10mm_full.root

tau₀ = 50 mm

/afs/cern.ch/work/u/ufay/public/CMSSW_9_4_0_patch1/src/trimscoutV2/all-trimscoutV2-darkphoton_mzd_20_tau0_50mm_full.root

tau₀ = 100 mm

/afs/cern.ch/work/u/ufay/public/CMSSW_9_4_0_patch1/src/trimscoutV2/all-trimscoutV2-darkphoton_mzd_20_tau0_100mm_full.root

tau₀ = 1000 mm

/afs/cern.ch/work/u/ufay/public/CMSSW_9_4_0_patch1/src/trimscoutV2/all-trimscoutV2-darkphoton_mzd_20_tau0_1000mm_full.root

Background

/afs/cern.ch/work/u/ufay/public/CMSSW_9_4_0_patch1/src/trimscoutV2/all-trimscoutV2_2017C_primaryVtx_sub_full.root