Charged Higgs Analysis Code

Table of Contents

• Installation
• Submission
• Local Run
• Configuration
• Datacards
• Development

Installation

• Install Nero package and compile the Core (not scram). Though CMSSW is not required latest full work recipe has been tested in CMSSW_10_2_13.

BASE=$PWD [ $CMSSW_BASE/src ]
cd $BASE
git clone [email protected]:MiT-HEP/NeroProducer.git -b CMSSW_92X
cd NeroProducer/Core
make -j 16

• Install the package

cd $BASE
git clone [email protected]:MiT-HEP/ChargedHiggs.git -b cmssw_94x ChargedHiggs
cd ChargedHiggs
make -j 16

• Optional: install combine

cd $CMSSW_BASE/src
cmsenv
git clone https://github.com/cms-analysis/HiggsAnalysis-CombinedLimit.git HiggsAnalysis/CombinedLimit
cd HiggsAnalysis/CombinedLimit
git checkout v8.0.1 
scram b -j 16 

Submission to Batch

• Submission to the batch system is performed with the script/submit.py:

$python script/submit.py  --help
Usage: submit.py [options]

Options:
  -h, --help            show this help message and exit
  -i INPUT, --input=INPUT
                        Input Configuration file
  -d DIR, --dir=DIR     Directory where to write the configuration. WARNING OVERWRITE THE DIRECTORY!!!
  -v DEBUG, --debug=DEBUG
                        Debug Verbosity. From 0-3. Default=0
  -n NJOBS, --njobs=NJOBS
                        Number of Job to submit
  -q QUEUE, --queue=QUEUE
                        Queue

For example:

python script/submit.py -i dat/config.dat -v 3 -n 30 -q 1nh -d mysub/sub1

• You can check the status with the command (after the job started):

python script/submit.py -d mysub/sub1 --status

• you can resubmit jobs i-j,k or the failed job with the commands:

python script/submit.py -d mysub/sub1 --resubmit [ -j fail]
python script/submit.py -d mysub/sub1 --resubmit -j i-j,k
python script/submit.py -d mysub/sub1 --resubmit -j run

Local run

• To run locally, you can call directly the looper with:

$python python/Loop.py --help
Usage: Loop.py [options]

Options:
  -h, --help         show this help message and exit
  -v, --verbose      Verbose
  -d DAT, --dat=DAT  Configuration file

For example:

python python/Loop.py -v -d dat/config.dat

Configuration

Configuration file

configuration file is a list of keys, followed by a '=' sign. The content depends on the key itself.

• 'include': load the content of filename.dat
• 'Files': the files to be run are file1,file2,/store..
• 'addfiles': add these files to the Files to be run on.
• 'Output': output file name will be...
• 'MCDB': the Monte Carlo database file is ...
• 'SFDB': the scale factor database file is...
• 'Smears = @SmearBase,JER,JES': Book the systematics smears, by name, or by constructor (@)
• 'Analysis = AnalysisBase,Analysis2 ': run the following analysis (that inherths from AnalysisBase) in that specifici order
• 'config = AnalysisBase|a=1,b=2,c(3)': configure the analysis setting the public members, or running the public member functions ...

MC Database:

• The MC database should contain the list of MCs with the info for the normalization. Can be dumped with the script:

$python python/mcDatabase.py --help
Usage: usage

Options:
  -h, --help            show this help message and exit
  -e EOS, --eos=EOS     eos directory to scout, will not read the files in the
                        pSet
  -x XSEC, --xsec=XSEC  Use external cross-section
  -l LABEL, --label=LABEL
                        MC label
  -f FILE, --file=FILE  mc_database file name
  -r, --recursive       do same for each subdir

e.g.:

python python/mcDatabase.py -r -e /store/user/amarini/Nero/v0.4/

PU Reweighting

• Generate the PU distributions (1D)

python python/puDistributions.py -r -e /store/user/amarini/Nero/v0.4/  -f aux/pileup.root

• or Run Dependent( RD )

python python/puDistributions.py -r -e /store/user/amarini/Nero/v0.4/ --run=0,1000,2000,3000 -f aux/pileupRD.root

• add to these files the pileup distributions in data, with names:

  • PU-pileup
  • PU-pileupUp
  • PU-pileupDown

• or for the RD model:

  • RD-pileup_run1_run2
  • RD-pileupUp_run1_run2
  • RD-pileupDown_run1_run2

• modify the configuration file for using these files. The pileupRun and pileupLumi keys controls the RD vs 1D model.

SF Database

The Scale factor database, contains a list of scale-factors associated to a label. This are constructed in the looper and can be used in the analysis.

Datacards

• Merge the output files (if necessary):

python script/submit.py --hadd -d mysub/sub1

• Dump datacard shape based:

python python/dumpDatacard.py -i ChHiggs.root --hist=CutFlow/CutFlow -f datacard_2.txt -s 

• or dump cut and count datacard:

python python/dumpDatacard.py -v -f datacard.txt -i ChHiggs.root --hist CutFlow/CutFlow --bin 2 --syst=JER

Development

Code structure

**********    *********    **********
* Looper * -> * Event * -> * Weight * -> MCDB, SF
**********    *********    **********

                                   *******
                        -> vector< * Jet * >
                                   *******
                        -> ...


                      *************  
           -> vector< * SmearBase * >      : set the actual systematic smear. Object or corrector should be awared.
                      *************  

                      *************  
           -> vector< * CorrectorBase * >  : change the content of the event, correcting objects like met, energy ...
                      *************  

                      ****************
           -> vector< * AnalysisBase * >   : perform analysis. Usually do not have rw access to the objects. To check.
                      ****************