BstarToTW_CMSDAS2023

Getting started (in bash shell)

Setup CMSSW environment:

ssh -Y USERNAME@lxplus.cern.ch
export SCRAM_ARCH=slc7_amd64_gcc820 
cd ~/public/
mkdir CMSDAS2023 && cd CMSDAS2023
cmsrel CMSSW_11_1_4
cd CMSSW_11_1_4/src
cmsenv

In the CMSSW_11_1_4/src/ directory, clone this exercise repo:

git clone https://github.com/ammitra/BstarToTW_CMSDAS2023.git

OR fork the code onto your personal project space and set the upstream:

git clone https://github.com/<GitHubUsername>/BstarToTW_CMSDAS2023.git
cd BstarToTW_CMSDAS2023
git remote add upstream https://github.com/ammitra/BstarToTW_CMSDAS2023.git
git remote -v

In the CMSSW_11_1_4/src/ directory, create a python virtual environment and install TIMBER within it:

cd ~/public/CMSDAS2023/CMSSW_11_1_4/src/
git clone https://github.com/ammitra/TIMBER.git
python -m virtualenv timber-env
source timber-env/bin/activate
cd TIMBER
source setup.sh
cd ..

You can test that the TIMBER installation is working by running the following in your shell:

python -c 'import TIMBER.Analyzer'

If all went well, the command should be executed with no output.

At this point you should have a directory structure that looks like this:

└── ~/public/CMSDAS2023/CMSSW_11_1_4/src/
    ├── TIMBER/
    ├── timber-env/
    └── BstarToTW_CMSDAS2023/

Starting up once environment is set:

Once you have an environment:

cd CMSSW_11_1_4/src/
cmsenv
source timber-env/bin/activate

You will need to perform this step every time you log on to the LPC cluster.

If you need to update TIMBER

cd TIMBER/
git fetch --all
git checkout master
python setup.py develop
cd ../

If you need to update BstarToTW_CMSDAS2023

cd BstarToTW_CMSDAS2023
git fetch --all
git pull origin master
cd ../

Submitting Condor jobs

To submit Condor jobs on LXPLUS, first create a valid proxy:

voms-proxy-init --rfc --voms cms -valid 192:00

The x509 proxy will be created and automatically outputted to the default location /tmp/x509up_u$(id -u)

Next, copy your proxy to your AFS home directory:

cp /tmp/x509up_u$(id -u) ~/

You'll now have to modify the JDL (job description language) template provided for you. Open the file condor/jdl_template and replace the following two lines with your proxy filename and proxy path. The filename is of the form x509up_uXXXXXX, where you can find your user id from the file name, or from echo $(id -u). The proxy path is wherever your AFS home is. Replace the /a/ with the first letter of your username, and replace the username in the template with your own:

Proxy_filename = x509up_u133882
Proxy_path = /afs/cern.ch/user/a/ammitra/$(Proxy_filename)

Finally, before submitting the Condor job, you just need to change the directory that the output ROOT file from the job will be sent to. Change the last line in the condor/run_*.sh scripts to your username to send the job results to your CERN EOS space:

cp -f rootfiles/*.root /eos/home-a/ammitra/

should become:

cp -f rootfiles/*.root /eos/home-u/username/

where u is the first letter of your username and username is your LXPLUS username.

You can now run either your selection, N-1 script, or the script for generating 2D template histograms for 2DAlphabet:

Selection:

python $CMSSW_BASE/src/BstarToTW_CMSDAS2023/CondorHelper.py -r condor/run_selection.sh -a condor/2016_args.txt -i "bstar.cc bstar_config.json helpers.py"

N - 1:

python $CMSSW_BASE/src/BstarToTW_CMSDAS2023/CondorHelper.py -r condor/run_Nminus1.sh -a condor/2016_args.txt -i "bstar.cc bstar_config.json helpers.py"

2D template histos:

python $CMSSW_BASE/src/BstarToTW_CMSDAS2023/CondorHelper.py -r condor/run_bstar.sh -a condor/2016_args_2DTemplates.txt -i "bstar.cc bstar_config.json helpers.py"

The argument files for the various years for selection and N - 1 are:

condor/2016_args.txt
condor/2017_args.txt
condor/2018_args.txt

The argument files for the various years for 2D template histos are:

condor/2016_args_2DTemplates.txt
condor/2017_args_2DTemplates.txt
condor/2018_args_2DTemplates.txt

To check the status of your submitted jobs:

condor_q $USER

To check the progress of your submitted jobs:

condor_tail -name bigbird28@cern.ch <job ID>

To remove a job:

condor_rm -name bigbird28@cern.ch <job ID>