POOLtoHEPMC_Skeleton.py

Go to the documentation of this file.

# Copyright (C) 2002-2025 CERN for the benefit of the ATLAS collaboration
# Job transform version of converting an EVNT file into a HEPMC file
 
# For the exit code at the end
import sys
import re
 
# For translating the run arguments into flags
from PyJobTransforms.CommonRunArgsToFlags import commonRunArgsToFlags
 
# For using pre include, pre-exec, etc; should only rarely be needed
from PyJobTransforms.TransformUtils import processPreExec, processPreInclude, processPostExec, processPostInclude
 
# Force no legacy job properties
from AthenaCommon import JobProperties
JobProperties.jobPropertiesDisallowed = True
 
def fromRunArgs(runArgs):
    # Start the logger and identify ourselves
    from AthenaCommon.Logging import logging
    log = logging.getLogger('POOLtoHEPMC')
    log.info('*** Starting POOLtoHEPMC translation ***')
 
    # Print some job information
    log.info('*** Transformation run arguments ***')
    log.info(str(runArgs))
 
    # Set up the flags we need
    log.info('*** Setting-up configuration flags ***')
    from AthenaConfiguration.AllConfigFlags import initConfigFlags
    flags = initConfigFlags()
    commonRunArgsToFlags(runArgs, flags)
 
    # Set ProductionStep
    from AthenaConfiguration.Enums import ProductionStep
    flags.Common.ProductionStep = ProductionStep.Derivation
 
    # Set the input file and configure the input collection key
    if hasattr(runArgs, 'inputEVNTFile'):
        flags.Input.Files = runArgs.inputEVNTFile
        McEventKey = 'GEN_EVENT'
    elif hasattr(runArgs, 'inputHITSFile'):
        flags.Input.Files = runArgs.inputHITSFile
        McEventKey = 'TruthEvent'
    elif hasattr(runArgs, 'inputRDOFile'):
        flags.Input.Files = runArgs.inputRDOFile
        McEventKey = 'TruthEvent'
    else:
        log.error('Input EVNT, HITS, or RDO file required for POOLtoHEPMC')
    # Set the output file
    if hasattr(runArgs, 'outputHEPMCFile'):
       if ('.tar' in runArgs.outputHEPMCFile):
          index = re.search(".tar",runArgs.outputHEPMCFile).span()[0]
          if hasattr(runArgs, 'extension') and ('events' in runArgs.extension):
               my_output_HepMCFile = runArgs.outputHEPMCFile[:index]+'.events'
          else:
               my_output_HepMCFile = runArgs.outputHEPMCFile[:index]+'.hepmc'
       else:
          log.error('Output should be a tar.gz file but it is '+runArgs.outputHEPMCFile)
    else:
        log.error('OutputHEPMCFile required for POOLtoHEPMC')
 
    hepMCFormat = 'hepmc2'
    if hasattr(runArgs, 'hepmcFormat'):
       hepMCFormat = runArgs.hepmcFormat
 
    # Setup perfmon flags from runargs
    from PerfMonComps.PerfMonConfigHelpers import setPerfmonFlagsFromRunArgs
    setPerfmonFlagsFromRunArgs(flags, runArgs)
 
    # Pre-include
    processPreInclude(runArgs, flags)
 
    # Pre-exec
    processPreExec(runArgs, flags)
 
    # To respect --athenaopts 
    flags.fillFromArgs()
 
    # Lock flags
    flags.lock()
 
    # Do the configuration of the main services
    from AthenaConfiguration.MainServicesConfig import MainServicesCfg
    cfg = MainServicesCfg(flags)
 
    # Set us up for reading a POOL file
    from AthenaPoolCnvSvc.PoolReadConfig import PoolReadCfg
    cfg.merge(PoolReadCfg(flags))
 
    # We need the component factory to build the job up
    from AthenaConfiguration.ComponentFactory import CompFactory
 
    # Add FixHepMC to remove loops here
    # This is a work-around for AGENE-2342, which needs a HepMC patch to fix
    cfg.addEventAlgo(CompFactory.FixHepMC("FixHepMC"))
 
    # Use the WriteHepMC AlgTool from TruthIO to do the conversion
    cfg.addEventAlgo( CompFactory.WriteHepMC( 'WriteHepMC',
                      OutputFile = my_output_HepMCFile,
                      Format = hepMCFormat,
                      McEventKey = McEventKey ) )
    # Here one should set the output format
    # Post-include
    processPostInclude(runArgs, flags, cfg)
 
    # Post-exec
    processPostExec(runArgs, flags, cfg)
 
    import time
    tic = time.time()
 
    # Run the final accumulator
    sc = cfg.run()
 
    # Compress the output file
    log.info('Compressing HEPMC output (may take a moment)')
    import tarfile
    with tarfile.open(runArgs.outputHEPMCFile,'w:gz') as out_tar:
        out_tar.add( my_output_HepMCFile )
 
    # And remove the uncompressed version
    log.debug('Deleting original (uncompressed) file')
    import os
    os.remove( my_output_HepMCFile )
 
    # All done, now just report back
    log.info("Ran POOLtoHEPMC in " + str(time.time()-tic) + " seconds")
 
    sys.exit(not sc.isSuccess())