- Analysis documentation ====
These twiki pages contain the CMS synchronization status of the WW and WZ cross section analyses at 13 TeV. There we can find the datasets used in each analysis, trigger bits, lepton selections and analysis cuts.
https://twiki.cern.ch/twiki/bin/view/CMS/WW2015Variables
https://twiki.cern.ch/twiki/bin/view/CMS/WZ13TeV
https://twiki.cern.ch/twiki/bin/view/CMS/ReviewSMP16002
http://www.hep.uniovi.es/nachos/WZ/ValidationPlots_v2_2090pb/
- First time only ====
Log in to gridui.
ssh -Y gridui.ifca.es -o ServerAliveInterval=240
source /cvmfs/cms.cern.ch/cmsset_default.sh
Set a CMSSW release.
export SCRAM_ARCH=slc6_amd64_gcc530
cmsrel CMSSW_8_0_5
cd CMSSW_8_0_5/src
cmsenv
Go to the master repository (https://github.com/piedraj/AnalysisCMS) and click Fork in the top-right corner of the page. Then get the code in your working area.
git clone https://github.com/YOUR_USERNAME/AnalysisCMS
git clone https://github.com/scodella/BTagSFUtil
Create an empty Git repository or reinitialize an existing one.
git init
You should also add a remote upstream, to be able to modify both your and the master repository.
git remote add upstream https://github.com/piedraj/AnalysisCMS
Then, do a git remote in order to check if the upstream appears.
git remote -v
Do this only if you want to create a tag.
pushd AnalysisCMS
git tag -a 20160318_electron -m 'Second AnalysisCMS tag'
git push origin 20160318_electron
popd
Do this only if you want to use a tag.
pushd AnalysisCMS
git checkout tags/20160318_electron
popd
-
Always do ====
ssh -Y gridui.ifca.es -o ServerAliveInterval=240 source /cvmfs/cms.cern.ch/cmsset_default.sh cd CMSSW_8_0_5/src cmsenv cd AnalysisCMS
-
Compile and run ====
./make
It is recommended to test the code. The following example reads a latino tree and produces the corresponding histograms.
./runAnalysis /full/path/latino_WZTo3LNu.root nominal
Submit jobs to the gridui batch system. It is encouraged to first read the Basic Grid Engine Usage documentation.
rm -rf minitrees/<systematic>
rm -rf rootfiles/<systematic>
rm -rf txt/<systematic>
./submit-jobs.sh
Show the status of the submitted jobs.
qstat -u $USER
qstat -j <job-ID.ja-task-ID>
Alternatively one can login to a node and run interactively. Do this only if your jobs will take less than 2 hours.
qlogin -P l.gaes
cd CMSSW_8_0_5/src
source /cvmfs/cms.cern.ch/cmsset_default.sh
cmsenv
cd AnalysisCMS
./submit-jobs-interactive.sh
exit
- Draw ====
First hadd the files produced in the previous step.
cd test
./submit-hadd.sh
Then compile and draw your favorite distributions.
./make
./runPlotter
- Create your website ====
Follow the instructions at Permissions for your AFS folder.
ssh -Y lxplus.cern.ch
mkdir www
fs setacl www webserver:afs read
afind www -t d -e "fs setacl -dir {} -acl webserver:afs read"
cd www
wget https://raw.githubusercontent.com/piedraj/AnalysisCMS/master/test/.htaccess
Go to the CERN Web Services and click on Create a new website. Choose the "AFS folder" site type.
- Share on the web ====
Create a soft link to your figures folder.
cd figures
export FIGURES_DIR=`pwd`
cd
ln -s $FIGURES_DIR
Copy the distributions to lxplus.
ssh -Y lxplus.cern.ch
scp -r [email protected]:figures www/.
And they should appear here,
https://amanjong.web.cern.ch/amanjong/
https://bchazinq.web.cern.ch/bchazinq/
https://cprieels.web.cern.ch/cprieels/
https://jgarciaf.web.cern.ch/jgarciaf/
https://ntrevisa.web.cern.ch/ntrevisa/
https://piedra.web.cern.ch/piedra/
- It is commit time ====
This framework is intended to be used and developed by several people. If you want your changes to be shared by others, you should first get the latest modifications from the upstream repository, if any.
git pull https://github.com/piedraj/AnalysisCMS
git fetch upstream
Now you can commit any change (both from upstream/master and master) in your repository.
git status
git add <filepattern>
git commit -m 'Modified'
git push
If there isn't any conflict left you should be able to merge with the upstream/master repository.
git merge upstream/master
Finally, go to your repository and click Pull Request.
https://github.com/YOUR_USERNAME/AnalysisCMS
- Edit a CMS Analysis Note ====
The following instructions have been extracted from the CMS TWiki. The note number used below corresponds to the IFCA/Oviedo WZ analysis at 13 TeV.
ssh -Y lxplus.cern.ch -o ServerAliveInterval=240
svn co -N svn+ssh://svn.cern.ch/reps/tdr2 svn
cd svn
svn update utils
svn update -N notes
svn update notes/AN-16-010
eval `notes/tdr runtime -sh`
cd notes/AN-16-010/trunk
tdr --style=an b AN-16-010
Get the latest modifications from the note repository. Always do this first.
svn update
If you want to add new files.
svn add file1 file2 ... fileN
If you want to know which files you have added and/or modified.
svn status
Commit your changes.
svn commit -m 'Modified'
A detailed example of a CMS Note written in LaTeX using the cms-tdr document class can be found in the svn-instructions.pdf file.
- EOS ====
To access files at eos/cms.
eosmount eoscms
eosumount eoscms
rmdir eoscms
To access files at eos/user.
/afs/cern.ch/project/eos/installation/0.3.84-aquamarine.user/bin/eos.select -b fuse mount eosuser
/afs/cern.ch/project/eos/installation/0.3.84-aquamarine.user/bin/eos.select -b fuse umount eosuser
rmdir eosuser
- Copy latino trees from cernbox to gridui ====
First of all add your lxplus pub key in ipa.
cat $HOME/.ssh/id_rsa.pub
Then log in to lxplus, mount eos and choose the input folder.
ssh -Y lxplus.cern.ch -o ServerAliveInterval=240
bash -l
rsync --chmod=Du=rwx,Dg=rwx,Fu=rw,Fg=rw -azH eos/user/j/jlauwers/HWW2015/21Jun2016_Run2016B_PromptReco/l2loose__hadd__EpTCorr__l2tight [email protected]:
Check that the input folder has be copied at the following gridui path.
/gpfs/csic_projects/tier3data/LatinosSkims/RunII/cernbox/
- brilcalc ====
Log in to lxplus.
ssh -Y [email protected] -o ServerAliveInterval=240
bash -l
Go to the Prerequisite section of the BRIL Work Suite and export the PATH that corresponds to the centrally installed virtual environment on lxplus.
export PATH=$HOME/.local/bin:/afs/cern.ch/cms/lumi/brilconda-1.0.3/bin:$PATH
Do this the first time. Do it also if you want to update the brilcalc version.
pip uninstall brilws
pip install --install-option="--prefix=$HOME/.local" brilws
Check your brilcalc version.
brilcalc --version
2.0.5
Get the 2016 luminosity. Based on the PdmV TWiki one should use the following.
brilcalc lumi -b "STABLE BEAMS" \
--normtag /afs/cern.ch/user/l/lumipro/public/normtag_file/normtag_DATACERT.json \
-u /fb \
-i /afs/cern.ch/cms/CAF/CMSCOMM/COMM_DQM/certification/Collisions16/13TeV/Cert_271036-275125_13TeV_PromptReco_Collisions16_JSON.txt \
--hltpath "HLT_Mu8_TrkIsoVVL_v*"
Get the 2015 luminosity.
brilcalc lumi --normtag /afs/cern.ch/user/l/lumipro/public/normtag_file/moriond16_normtag.json -u /fb -i /afs/cern.ch/cms/CAF/CMSCOMM/COMM_DQM/certification/Collisions15/13TeV/Reprocessing/Cert_13TeV_16Dec2015ReReco_Collisions15_25ns_JSON_v2.txt
+-------+------+-------+-------+-------------------+------------------+
| nfill | nrun | nls | ncms | totdelivered(/fb) | totrecorded(/fb) |
+-------+------+-------+-------+-------------------+------------------+
| 47 | 115 | 33208 | 33208 | 2.398 | 2.318 |
+-------+------+-------+-------+-------------------+------------------+