This repository contains the code to calculate trigger efficiencies that is used by the B2G trigger contacts. It is using the orthogonal dataset method and mainly targets hadronic triggers, like our jet mass triggers (HLT_AK8PFJet400_MassSD30).
The python code used in this framework requires a few packages (and their dependencies), which can be installed using pip or conda. The needed packages are the following:
- uproot
- coffea
- awkward
- numpy
- mplhep
Additionally, the xrootd package is required if files from the GRID are to be used.
If you want to execute the code on lxplus, a pre-existing environment can be used. However, a few requirements are not yet included in any stable release (afaik), primarily the numba version 0.55.2, that coffea needs, is not available. I was able to run the code using the latest development release, so for the moment one should be able to just do
source /cvmfs/sft-nightlies.cern.ch/lcg/views/dev3/latest/x86_64-centos7-gcc11-opt/setup.sh
and be able to run the module. This README will be updated once a working stable release is available. The only other required environment setting then is to activate your GRID proxy (if files should be accessed using xrootd):
voms-proxy-init --voms cms -rfc --valid 168:00 -out $HOME/private/.proxy
export X509_USER_PROXY=$HOME/private/.proxy
A single main python file is implemented that executes the desired code: runTriggerEfficiencies.py
This file can be called as python runTriggerEfficiencies.py, and requires several arguments afterwards:
- -i or --input: a text file containing a list of input files to run on (one file per line)
- -r or --refTriggers: a text file containing the reference triggers (one trigger per line)
- -t or --testTriggers: a text file containing the triggers to be measured (one trigger per line)
- -e or --era: the data-taking era (a string, for example "22RunD"). Might be auto-detected later
Two additional, optional flags can be set:
- --doJECs: apply jet energy corrections according to the era
- --useGoldenJSON: apply a goldenJSON selection
A complete example command thus could be
python runTriggerEfficiencies.py -i example/local/inputfiles.txt -r example/local/refTriggers.txt -t example/local/testTriggers.txt -e 22RunD --doJECs --useGoldenJSON
To run the trigger efficiencies on more files, the HTCondor batch system is used. Example scripts are given in the repository, and an exemplary workflow is shown here. Note that this is written to work on CERN lxplus, and might require changes to run on other systems.
First, prepare the required input files: two text files with reference and test triggers (as explained above), and the input file list. If the input sample is on DAS, this can be done automatically, for example using
./scripts/createFileList.sh "/Muon/Run2022F-PromptNanoAODv10_v1-v2/NANOAOD" example/condor/filelist.txt
which will produce a single file with all input files corresponding to the passed DAS query. Note that no redirector is added yet, this can be done using
python scripts/add_xrootd_redirector.py example/condor/filelist.txt "root://xrootd-cms.infn.it/".
Finally, you most likely want to split this list of input files in multiple jobs. Create a directory where the split files should be located, and execute
python scripts/split_filelist.py example/condor/filelist.txt 10 example/condor/split_file_list/
where in this example, we'll create 10 condor jobs. Now, all input files are ready, we need to create a submit file for the condor jobs.
The script create_submit_file.py can do that automatically, given a input folder (in exactly the structure as the example/condor one in this repository), the era of the input files (only one, do not combine files from different eras in one condor job!) and the output file name of the submitfile (for example condor.sub). You can also add options like --useGoldenJSON or --doJECs (this might be replaced by a config file later).
Then, create a log directory to store job logs, and run your jobs using
condor_submit condor.sub
When the jobs have run through, you'll have several files named output_*.root in the directory where you executed the submit command. Combine these using
hadd output_total.root output_*.root.
To plot the results of the efficiency calculation code, the plotEfficiencies.py and plotDistributions.py scripts can be used. These expect several input parameters:
- -i or --input: the root file(s) to use for plotting
- -o or --options: the plotting option(s), see below for more details
- -t or --triggers: the trigger path(s) to create plots for
- -v or --variables: the variable(s) to plot
Each of these parameters can be any number of strings / files. The behavior of the plotting script is mainly steered by the option passed, where two options exist (of which at least one needs to be given):
- oneTrigger: each plot will contain results for a single trigger, but multiple input files. Useful to compare years / run eras.
- oneFile: each plot will contain one file, but multiple triggers. Useful to compare triggers within an era / year.
Additional options can also be passed as strings following the -o flag. At the moment, only one is implemented: mSD35 created plots with the SD mass cut at 35 applied. Further options might be added here later.
Please check in-file comments for more instructions.
Jet energy corrections are applied according to the files in the data directory, based on the files obtained from the JERC group recommendations. In case new ones need to be added, these can be added to the data/corrections directory. There, a helper script (ConvertCorrectionFiles.py) is given to obtain the correct file naming.
The golden JSON selection is made based on the file in data/goldenJSON, where new ones can be added.
The 2022 file currently located in that directory was taken from here.