#Ntuple code runs over miniAOD and produces ROOT ntuples
# The Ntuple code is checked into my own git repository on gitlab
#

#THIS IS ONLY FOR WHEN WORKING AT RAL. Set up environment for sh/bash etc. Ignore the next two lines at CERN or elsewhere
export VO_CMS_SW_DIR=/cvmfs/cms.cern.ch
source $VO_CMS_SW_DIR/cmsset_default.sh

#checkout release
cmsrel CMSSW_8_0_12/
CMSSW_8_0_12/src
cmsenv
git clone https://github.com/Sam-Harper/usercode.git SHarper
cd SHarper/
git checkout HEEPID80X
cd ..

#check out NTupler package (you need to have setup ssh keys correctly at CERN)
git clone ssh://git@gitlab.cern.ch:7999/olaiya/NTupler.git

#build
scramv1 b -j 16

Producing the Ntuples
======================
#If you are not interested in producing the Ntuples because you already have them skip to the Running over the Ntuples section
#change into ntuple directory
cd NTupler/PATNTupler/plugins
#If you want to run the ntuple code directly edit nTupleProduction.py to specify the file you want to run over
cmsRun nTupleProduction.py 
#Alternatively if you want to run a crab production edit the crab config file crabConfig_nTupleProd.py to run over the dataset and specify the output directory on the T2 where you want the ntuple to be written. Then submit
source /cvmfs/cms.cern.ch/crab3/crab.csh
crab submit -c crabConfig_nTupleProd.py
#check the status
crab status crab_projects/crab_RAL_nTuple_production --long


Running over the Ntuples
========================
#To run over the ntuples go into the main directory where there is an example of running over the ntuples. Go into this directory and compile the code.

cd NTupler/PATNTupler/main
gmake

#The code main.cc gives examples of how you can access the variables in the ntuple.
#To run over a list of file execute the command
./nTupAna output.root ntuples.list
# output.root is the name of the ROOT file you want to generate. The content is specified in main.cc
# the file ntuples.list contains the list of root files to run over.
cat test.list          
../plugins/nTuple.root
# If you have a json file you want to use to run on certified data you can specify it too as follows:
./nTupAna output.root ntuples.list goodRuns_246908-260627_13TeV.txt
# The format of the json file is a little different to that provided by CMS. To generate the correct json format use the program in the main directory called convertJsonToCMSSWConfig.py as follows
./convertJsonToCMSSWConfig.py  Cert_246908-260627_13TeV_PromptReco_Collisions15_25ns_JSON.txt > goodRuns_246908-260627_13TeV.txt

#main.cc has two examples of accessing root file. In here is where you code your analysis
#Jobs can also be submitted to the batch machine if there are too many files to run over in one go
./submitCondorNtupleJobs.py --inputFiles ntuples.list  --jobName Zprime --filesPerJob 50 --json goodRuns_246908-260627_13TeV.txt
# --inputFiles requires the filename that contains the list of ntuples. --jobName is what you want to call the jobs. The default is EE.--filesPerJob is used to specify how many files per job you want. The files are taken from ntuples.list. --json wants the json file. This is optional
#./submitCondorNtupleJobs.py will prepare the condor jobs in the tmp directory. So you will have to go there to run them. You will find a file in tmp called subCondorJobs_Zprime.sh which will submit the job
cd ../tmp
./subCondorJobs_Zprime.sh