JobSubmitter

This package is used to submit multiple crab or condor tasks from lxplus. Moreover, this tool is also optimized to produce private Monte-Carlo (MC) CMSSW samples (from LHE or CMSSW MC Interfaces).

Installation

First, download the package from the git repository. It is also important to set the crab environment as well as CMSSW (cmsenv) where your plugins are set.

git clone https://github.com/dfigueiredo/JobSubmitter.git

Usage

The tool loads a json file which contains specific parameters for the job submission on crab or condor. There are multiple examples in this repository. The json format for the condor mode is different for the crab mode. Moreover, when running the tool, by default the mode "crab" is active. If it is not the case, "mode condor" should be active. In case you would like to create your own json file from scratch, the following scripts can create templates:

python createTemplateCondorJson.py

python createTemplateCrabJson.py 

Running

Remember to set the CMSSW version (with cmsenv) as well as the crab environment, otherwise the tool will complain. For loading the "configuration_file.json" as an example:

source /cvmfs/cms.cern.ch/common/crab-setup.sh
python SubmitterTool.py --f configuration_file.json

Options

Options:
  -h, --help            show this help message and exit
  -f FILE, --filename=FILE
                        JSON mapping file
  -p PARSING, --parsing=PARSING
                        parsing: commands which can be passed from SHELL
                        directly. [parsing: --p "submit --file samples.json"]
  -v, --verbose         make lots of noise [default]

Therefore, when running the script SubmitterTool.py:

jobs_submission> ?

Documented commands (type help <topic>):
========================================
EOF  exit  help  mode  submit

Changing the Submission Mode

By default, the tool is set to submit with crab.

python SubmitterTool.py --f configuration_condor.json
jobs_submission> mode condor

python SubmitterTool.py --f configuration_crab.json
jobs_submission> mode crab

Configuration File

All the configurations are set in a JSON file for condor or crab submission. Multiple tasks can be defined in a single file, but the tool will submit only when "enable" option is "1". The examples bellow are showing the configuration parameters for the condor or crab json file.

Condor

{
    "datasets": [
        {
            "executable": "./MissingMassNtupleAnalyzer", 
            "enable": 1, 
            "parameters": [
                "--par1", 
                "--par2" 
            ], 
            "inputfiles": [
                "../file1.txt", 
                "../file2.txt"
            ], 
            "output": "test/", 
            "inputfolder": "/eos/cms/store/group/phys_exotica/PPS-Exo/Grid/Test", 
            "id": 0
        }
    ]
}

The options tags are:

Options	Explanation	Comments
executable	(string) User code
enable	(int) 0 or 1
parameters	(vector string) executable options
inputfiles	(vector string) auxiliary input files	In case of extra files
output	(string) output folder	Condor output. Automatic created.
inputfolder	(string) multiple root files	Condor will create a job per file (*.root)
id	(int) identifier	Start from 0. Must be different

Crab

A python script is used to produce a template for the crab submission.

{
    "datasets": [
        {
            "eospath": "/store/user/dmf", 
            "enable": 0, 
            "unitsperjob": 250, 
            "localpath": "/afs/cern.ch/user/d/dmf/private/work/private/CMSPhysicsAnalysis/PrivateMCProduction/PPSMCProduction/working", 
            "lumimask": "", 
            "sample": "", 
            "output": "RunIISummer20UL17GEN.root", 
            "id": 0, 
            "name": "PYTHIA8-SD-TOP-GEN", 
            "parameters": "", 
            "mode": "mc_private_hadron_production", 
            "site": "T2_US_Wisconsin", 
            "input": "", 
            "config": "/afs/cern.ch/user/d/dmf/private/work/private/CMSPhysicsAnalysis/PrivateMCProduction/PPSMCProduction/working/SD-TOP-PYTHIA8_cfg.py"
        }
     ]
}

The options are:

Options	Explanation	Comments
eospath	(string) eos or store user area
enable	(int) 0 or 1
unitsperjob	(int) number of files per job (MC) or LS units (data)
localpath	(string) directory where the crab folder will be saved
lumimask	(string) when running with data
sample	(string) dataset	Empty for GEN creation (MC)
output	(string) crab root output file	Same name as in the CMSSW config
id	(int) identifier.	Start from 0. Must be different
name	(string) name of the task and name of dataset (when created)
parameters	(vector string) CMSSW input parameters
mode	(string) data_analysis	For data analysis
	(string) mc_analysis	For MC analysis
	(string) mc_private_analysis	For MC analysis (DBS phys03)
	(string) mc_private_hadron_production	For MC GEN production (DBS phys03)
	(string) mc_private_hadron_lhe_production	For MC LHE production (DBS phys03)
site	(string) Storage Site
input	(string) In case of extra input file
config	(string) CMSSW config file

Examples of submissions

File	Comments
samples_skimmer2018_bjets.json	to produce skimmed ttrees for BJets.
samples_skimmer2018_displacedjet.json	to produce skimmed ttrees for BJets.
samples_skimmer2018_doublemuon.json	to produce skimmed ttrees for Double Muon.
samples_skimmer2018_egamma.json	to produce skimmed ttrees for E/Gamma.
samples_skimmer2017_mc.json	to produce skimmed ttrees for MC.
samples_condor_bjets.json	to produce final bjets ntuples.
samples_condor_displacedjet.json	to produce final displaced ntuples.
samples_condor_doublemuon.json	to produce final Double Muon ntuples.
samples_condor_egamma.json	to produce final E/Gamma ntuples.
samples_condor_mc.json	to produce final MC ntuples.
samples_condor_bjets_protonfile.json	to produce final random protons histograms.
samples_create_pythia8.json	to produce CMSSW pythia8 datasets.
samples_create_toymc.json	to produce CMSSW Toy MC datasets.