AOD-compatible TreeMaker

This is a customized fork of TreeMaker compatible with sideloading AOD files.

Workflow

Custom scripts are available in Production/aod_support/:

1. Run create_AOD_filelists_from_datasets.py to create your own list of AOD file names you want to process
2. Place file lists in Production/python/campaign/
3. Create dictionaries in Production/test/condorSub/ with create_dictionaries_for_AOD_filelists.py
4. Create a file name catalogue which maps AOD files to miniAOD files with create_miniAOD_filename_catalogue.py

Supporting scripts

In Production/aod_support/, you can find

• scripts to create AOD file lists for custom dataset paths
• mark already processed ntuples as done
• run unit tests with AOD files

TreeMaker

• Instructions
• Unit Tests (Interactive Runs)
• Submit Production to Condor
• Calculate Integrated Luminosity
• Calculate Pileup Corrections
• Info for New Samples
  • Samples with Negative Weight Events
• Options

Instructions

The following installation instructions assume the user wants to process 2016, 2017, or 2018 miniAOD.

For 2016 and 2017:

wget https://raw.githubusercontent.com/LongLivedSusy/TreeMaker/Run2_2017_aodv2/setup.sh
chmod +x setup.sh
./setup.sh
cd CMSSW_9_4_11/src/
cmsenv
cd TreeMaker/Production/test

For 2018, replace the middle lines:

./setup.sh -c CMSSW_10_2_7
cd CMSSW_10_2_7/src/

The script setup.sh has options to allow installing a different fork or branch of TreeMaker (though some branches may have different setup scripts, so check carefully which one you download):

• -f [fork]: which fork to download ([email protected]:fork/TreeMaker.git, default = TreeMaker)
• -b [branch]: which branch to download (-b branch, default = Run2_2017)
• -c [version]: which CMSSW version to use (default = CMSSW_9_4_11)"
• -a [protocol]: which protocol to use for git clone (default = ssh, alternative = https)
• -j [cores]: run CMSSW compilation on # cores (default = 8)
• -h: display help message and exit

Several predefined scenarios are available for ease of production. These scenarios define various sample-dependent parameters, including:
global tag, collection tag name, generator info, fastsim, signal, JSON file, JEC file, residual JECs, era.
The available scenarios are:

1. Summer16MiniAODv3: for Summer16 miniAODv3 MC
2. Summer16MiniAODv3sig: for Summer16 miniAODv3 MC (signal)
3. Summer16MiniAODv3Fastsig: for Summer16 miniAODv3 FastSim MC (signal)
4. 2016MiniAODv3: for 2016 miniAODv3 data (17Jul2018), periods B-H
5. Fall17: for Fall17 miniAOD MC
6. Fall17sig: for Fall17 miniAOD MC (signal)
7. Fall17Fastsig: for Fall17 miniAOD FastSim MC (signal)
8. 2017ReReco31Mar: for 2017 ReReco data (31Mar), periods B-F
9. Spring16Fastsig: for Spring16 miniAOD FastSim MC (80X) (signal scans)
10. Spring16Pmssm: for Spring16 miniAOD 25ns PMSSM MC scan (80X) (signal)
11. Summer16: for Summer16 miniAOD 25ns MC (80X)
12. Summer16sig: for Summer16 miniAOD 25ns MC (80X) (signal)
13. Autumn18: for Autumn18 miniAOD 25ns MC (102X)
14. Autumn18sig: for Autumn18 miniAOD 25ns MC (102X) (signal)
15. 2018B26Sep: for 2018 prompt data, partial period B
16. 2018B26SepHEM: for 2018 prompt data, partial period B, with the HEM 15/16 issue
17. 2018PromptReco: for 2018 prompt data, run period D
18. 2018ReReco17Sep: for 17Sep2018 rereco data, periods A, B, and C

Unit Tests (Interactive Runs)

Several predefined run commands (at least one for each scenario) are defined in a script called unitTest.py. It has several parameters:

• test: number of the test to run (default=-1, displays all tests)
• name: name of the output ROOT and log files for the test (default="", each test has its own default name)
• run: run the selected test (default=False)
• numevents: how many events to run (default=100)

A few examples of how to run the script:

1. To see all tests:

python unitTest.py

2. To run test 2:

python unitTest.py test=2 run=True

Note that all of the background estimation processes (and some processes necessary to estimate systematic uncertainties) are turned ON by default in runMakeTreeFromMiniAOD_cfg.py.

Submit Production to Condor

Condor submission is supported for the LPC batch system or for the global pool via CMS Connect. Job submission and management is based on the CondorProduction package. Refer to the package documentation for basic details.

The test/condorSub directory contains all of the relevant scripts. If you make a copy of this directory and run the submitJobs.py script, it will submit one job per file to Condor for all of the relevant samples. Example:

./lnbatch.sh myProduction
cd myProduction
python submitJobs.py -p -o root://cmseos.fnal.gov//store/user/YOURUSERNAME/myProduction -s

submitJobs.py can also:

• count the expected number of jobs to submit (for planning purposes),
• check for jobs which were completely removed from the queue and make a resubmission list.

The class jobSubmitterTM.py extends the class jobSubmitter from CondorProduction. It adds a few extra arguments:

Python:

• -d, --dicts [list]: comma-separated list of input dicts; each prefixed by dict_ and contains scenario + list of samples (default taken from .prodconfig)
• -o, --output [dir]: path to output directory in which root files will be stored (required)
• -J, --json [jsonfile]: manually specified json file to check data (override scenario)
• -N, --nFiles [num]: number of files to process per job
• -A, --args [list]: additional common args to use for all jobs (passed to runMakeTreeFromMiniAOD_cfg.py)
• -v, --verbose: enable verbose output (default = False)
• -x, --redir: input file redirector

Shell (in step2.sh):

• -o [dir]: output directory
• -j [jobname]: job name
• -p [process]: process number
• -x [redir]: xrootd redirector

When the python file list for a given sample (usually data) is updated, it may be desirable to submit jobs only for the new files. The input dictionary format for jobSubmitterTM.py optionally allows a (non-zero) starting number to be placed after the sample name. To get the number of the first new job, just use len(readFiles) from the python file list before updating it.

When submitting jobs for prompt data, each data file will be checked to see if the run it contains is certified in the corresponding JSON file. The JSON file is taken by default from the scenario; an alternative can be specified with the --json option, e.g. if the JSON is updated and you want to submit jobs only for the newly certified runs. (Use compareJSON.py to subtract one JSON list from another, following this twiki.)

Calculate Integrated Luminosity

Scripts are available to calculate the integrated luminosity from data ntuples (produced with TreeMaker):

python lumiSummary.py
python calcLumi.py

The script lumiSummary.py loops over a list of data samples (by default, a list of Run2015C and Run2015D samples) and creates a JSON file for each sample consisting of the lumisections which were actually processed. Run python lumiSummary.py --help to see the available options. (This script is based on the CRAB3 client job report scripts.)

The resulting JSON file can be run through brilcalc using calcLumi.py to determine the integrated luminosity for the dataset. Run python calcLumi.py --help to see the available options. (NB: this only works on lxplus with brilcalc installed.)

Calculate Pileup Corrections

A script is available to calculate the pileup corrections for MC:

python pileupCorr.py

A ROOT file containing the data, MC, and data/MC histograms (with uncertainty variations) is produced. Run python pileupCorr.py --help to see the available options.

Info for New Samples

The script get_mcm.py can search the McM database for given samples (with wildcard support) to discern the status of the sample (whether it has finished running), the generator cross section, and the full dataset path for the sample ("/X/Y/Z" format). Command line options exist to specify campaign names and other information, which can be viewed with the --help option. An example dictionary of samples and extensions to check can be found at dict_mcm.py. This script requires cern-get-sso-cookie to access the McM database, which is installed on lxplus and cmslpc. Kerberos-based access is used by default, but certificate-based access is also available. Kerberos access requires a ticket for CERN.CH, and currently only works on lxplus. To configure your certificate for access, see the "User certificates" section of the cern-get-sso-cookie documentation. The same arguments --cert and --key can be used with the get_mcm.py script.

The script get_py.py will automatically create the "_cff.py" python file containing the list of ROOT files for samples specified in a Python ordered dictionary, e.g. dict.py (enabled with -p). For MC samples, it can also automatically generate the appropriate configuration line to add the sample to getWeightProducer_cff.py. This is enabled with the -w option and is used for MC samples only. The script can also check to see which sites (if any) have 100% dataset presence for the sample (enabled with -s). (You may also need export SSL_CERT_DIR='/etc/pki/tls/certs:/etc/grid-security/certificates' (bash) or setenv SSL_CERT_DIR '/etc/pki/tls/certs:/etc/grid-security/certificates' (tcsh) to avoid the error SSL: CERTIFICATE_VERIFY_FAILED from urllib2.)

Before running the script for the first time, some environment settings are necessary:

source /cvmfs/cms.cern.ch/crab3/crab.csh

To run the script:

python get_py.py -d dict.py [options]

To check for new samples, use the above script get_mcm.py or query DAS, e.g.:

dasgoclient -query="dataset=/*/RunIISpring16MiniAOD*/MINIAODSIM"

Samples with Negative Weight Events

Samples produced at NLO by amcatnlo have events with negative weights, which must be handled correctly. To get the effective number of events used to weight the sample, there is a multi-step process. This process also produces a histogram of TrueNumInteractions for pileup reweighting.

Step 1: Get the "_cff.py" files, without generating WeightProducer lines (assuming samples are listed in dictNLO.py).

python get_py.py dict=dictNLO.py wp=False

Step 2: Run NeffFinder, a simple analyzer which calculates the effective number of events for a sample. The analyzer should be submitted as a Condor batch job for each sample (assuming samples are listed in dict_neff.py), because the xrootd I/O bottleneck is prohibitive when running interactively. Be sure to sanity-check the results, as xrootd failures can cause jobs to terminate early.

./lnbatch.sh myNeff
cd myNeff
python submitJobsNeff.py -p -d neff -N 50 -o root://cmseos.fnal.gov//store/user/YOURUSERNAME/myNeff
(after jobs are finished)
python getResults.py
./haddEOS.sh -d /store/user/YOURUSERNAME/myNeff -g _part -r

Step 3: Update dictNLO.py with the newly-obtained Neff values and generate WeightProducer lines. Combine all pileup distributions into a single file.

cd $CMSSW_BASE/src/TreeMaker/Production/test/myNeff
./haddWrongPU.sh -L /store/user/YOURUSERNAME/myNeff
cd $CMSSW_BASE/src/TreeMaker/Production/python
python get_py.py dict=dictNLO.py py=False

Options

Brief explanation of the options in makeTree.py

• scenario: the scenario name, in case of special requirements (default="")
• inputFilesConfig: name of the python file with a list of ROOT files for a sample, used for Condor production (automatically appended with "_cff") (default="")
• nstart: first file to use in above file list (default=0)
• nfiles: number of files to use in above file list, -1 includes all files (default=-1)
• dataset: direct list of input files (alternative to above 3 parameters) (default=[])
• numevents: number of input events to process, -1 processes all events (default=-1)
• outfile: name of the ROOT output file that will be created by the TFileService (appended with outfilesuff below) (default="test_run")
• outfilesuff: suffix to append to outfile above (default="_RA2AnalysisTree")
• treename: name of output ROOT TTree (default="PreSelection")
• lostlepton: switch to enable the lost lepton background estimation processes (default=True)
• hadtau: switch to enable the hadronic tau background estimation processes (storing soft jets and reclustering) (default=False)
• hadtaurecluster: switch to enable the hadronic tau reclustering to include jets with pT < 10 GeV, options: 0 = never, 1 = only TTJets/WJets MC, 2 = all MC, 3 = always (default=0)
• doZinv: switch to enable the Z->invisible background estimation processes (default=True)
• systematics: switch to enable JEC- and JER-related systematics (default=True)
• semivisible: switch to enable variables for semi-visible jets (default=True)
• deepAK8: switch to enable variables from the DeepAK8 tagger (default=True)
• deepDoubleB: switch to enable variables from the DeepDoubleB tagger (default=True)
• doPDFs: switch to enable the storage of PDF weights and scale variation weights from LHEEventInfo (default=True)
  The scale variations stored are: [mur=1, muf=1], [mur=1, muf=2], [mur=1, muf=0.5], [mur=2, muf=1], [mur=2, muf=2], [mur=2, muf=0.5], [mur=0.5, muf=1], [mur=0.5, muf=2], [mur=0.5, muf=0.5]
• debugtracks: store information for all PF candidates in every event (default=False) (use with caution, increases run time and output size by ~10x)
• applybaseline: switch to apply the baseline HT selection (default=False) The following parameters take their default values from the specified scenario:
• globaltag: global tag for CMSSW database conditions (ref. FrontierConditions)
• tagname: tag name for collections that can have different tags for data or MC
• geninfo: switch to enable use of generator information, should only be used for MC
• fastsim: switch to enable special settings for SUSY signal scans produced with FastSim
• pmssm: switch to enable special settings for pMSSM signal scans
• signal: switch to enable assessment of signal systematics (currently unused)
• jsonfile: name of JSON file to apply to data
• jecfile: name of a database file from which to get JECs
• jerfile: name of a database file from which to get JERs
• residual: switch to enable residual JECs for data
• pufile: name of a ROOT file from which to get pileup weights
• wrongpufile: name of a ROOT file from which to get per-sample pileup distributions (for Fall17 samples produced w/ wrong PU)
• era: CMS detector era for the dataset
• redir: xrootd redirector, storage element address, or site name (default="root://cmsxrootd.fnal.gov/") (fastsim default="root://cmseos.fnal.gov/")
• verbose: print messages from modules in the TreeMaker category (and from JetToolbox) (default=True)

Extra options in runMakeTreeFromMiniAOD_cfg.py:

• reportfreq: frequency of CMSSW log output (default=1000)
• dump: equivalent to edmConfigDump, but accounts for all command-line settings; exits without running (default=False)
• mp: enable igprof hooks for memory profiling (default=False)
• threads: run in multithreaded mode w/ specified number of threads (default=1)
• streams: run w/ specified number of streams (default=0 -> streams=threads)
• tmi: enable TimeMemoryInfo for simple profiling (default=False)
• trace: enable the tracer for debugging (default=False)