Skim code setup for producing Ntuples for ttbar charge asymmetry analysis.
The framework is setup for running at the LPC, and saving ntuples to the LPC Tier3.
Check out CMSSW release:
cmsrel CMSSW_9_4_4
cd CMSSW_9_4_4/src
cmsenv
git cms-init
Next, checkout BESTProducer and other necessary packages.
There are two distinct sets of instructions depending on which dataset you will be processing. These only differ for the EGamma dependencies:
### 2016 ###
# VID (Electron ID)
git cms-addpkg RecoEgamma/ElectronIdentification
git cms-addpkg PhysicsTools/SelectorUtils
### 2017 ###
# Updated Egamma IDs (creates RecoEgamma and PhysicsTools directories)
# https://twiki.cern.ch/twiki/bin/view/CMS/CutBasedElectronIdentificationRun2
git cms-merge-topic lsoffi:CMSSW_9_4_0_pre3_TnP
### COMMON ###
# BEST (top vs qcd tagging)
git clone https://github.com/cms-ttbarAC/BESTAnalysis.git -b 94X
# LWTNN (running BEST in c++)
mkdir lwtnn
git clone https://github.com/demarley/lwtnn.git -b CMSSW_8_0_X-compatible lwtnn/lwtnn
# Analysis code -- now linked with this repository
mkdir Analysis
git clone https://github.com/cms-ttbarAC/CyMiniAna.git Analysis/CyMiniAna
git clone https://github.com/cms-ttbarAC/ttbarAC_skim
scram b -j8
Instructions for producing flat Ntuples with BEST tagger inputs.
cd ttbarAC_skim/ttbarAC_skim/test
voms-proxy-init -voms cms
source /cvmfs/cms.cern.ch/crab3/crab.csh
To submit crab jobs, the script runSkim.py is processed using the cmsRun executable.
There are two options for submitting crab jobs, (1) submit many jobs at once or (2) submit jobs manually one at a time.
Write a text file that contains the datasets you would like to process, see test/crab_datasets-data.txt and test/crab_datasets-mc.txt as examples (or you can use these files).
Then, execute the script with your datsets text file as a command-line argument:
python crab-submit-multiple.py <datasets.txt> <year>
If no command-line argument is given, a default file & year will be chosen. This script will loop through the different datasets and submit crab jobs for each one.
Note: the text file needs to be written such that each sample has a 'nickname' that can be used to create the crab directory. The example text files include this structure -- replicate it if you write your own file!
Example: filename = "example.txt"
ttbar /TT_TuneCUETP8M2T4_13TeV-powheg-pythia8/RunIISummer16MiniAODv2-PUMoriond17_80X_mcRun2_asymptotic_2016_TrancheIV_v6-v1/MINIAODSIM`
where ttbar is the nickname, and /TT.../MINIAODSIM is the sample.
Then, the jobs would be submitted as:
python crab-submit-multiple.py example.txt 2016
The following method is not recommended. Instead, if you want to only submit one sample, use the method described in "Multiple Submission" and fill your text file with only 1 dataset.
Edit crab_*.py with a new dataset name to reflect the sample you are processing, e.g., crab_SMttbar.py.
Make sure to the output directory is pointed to /store/group/lpctop/ttbarAC/ttbarAC_skim/.
If you are processing data, modify the argument config.JobType.pyCfgParams = ['isMC=True'] to config.JobType.pyCfgParams = ['isMC=False'].
To submit the CRAB jobs, enter the command:
crab submit -c crab_*.py --dryrun
crab proceed
You can also submit the crab jobs directly by removing --dryrun.
For either option, monitor the CRAB jobs and resubmit any that fail.
Instructions for producing EDM Ntuples with BEST tagger inputs and then flat trees for analysis.
cd ttbarAC_skim/ttbarAC_skim/test
voms-proxy-init -voms cms
source /cvmfs/cms.cern.ch/crab3/crab.csh
To submit crab jobs, the script runSkim_data.py (runSkim.py has been modified too much) is processed for MC (data) using the cmsRun executable.
First, edit crab_*.py with a new dataset name to reflect the sample you are processing, e.g., crab_SMttbar.py.
Make sure to the output directory is pointed to /store/group/lpctop/edmNtuples/.
To submit the CRAB jobs, enter the command:
crab submit -c crab.py --dryrun
either remove --dryrun to submit the jobs, or enter crab proceed after running the above command.
Monitor the CRAB jobs and resubmit any failed jobs.
After crab jobs are finished, make a text file with the filenames (starting with /store/group/...):
xrdfsls -u /store/group/lpctop/ttbarAC/edmNtuples/.../ | grep "ana" > textfile.txt
If there are multiple EDM files in this directory, change ana to a specific keyword in the desired files.
Then, setup the condor scripts to submit jobs to the LPC condor system that process the EDM ntuples into flat ntuples.
The condor scripts to check are submit.sh, analyze.sh, and treeMaker_fwlite.py.
There shouldn't be any changes necessary in treeMaker_fwlite.py or submit.sh, but analyze.sh may need changes for the following two lines:
output_path="/store/group/lpctop/ttbarAC/flatNtuples/"
...
python treeMaker_fwlite.py --files ${file} --isMC 1
where output_path may need to be changed, and isMC needs to be 0 for data samples.
To submit the jobs, enter:
./submit.sh files.txt <process>
Where <process> is the name of the process, e.g., SMttbar.
Note:
this string gets added to the ROOT output files, so you use that same string later with
do_hadd.shto grab only the files from a certain process andhaddthem
Monitor these jobs using condor_q ${USER} or using landscape.
Once the jobs are finished, flat ntuples are available for analysis using CyMiniAna.
Contact the authors.