Gen_tf.py

Go to the documentation of this file.

#! /usr/bin/env python
 
#  Copyright (C) 2002-2025 CERN for the benefit of the ATLAS collaboration
#
"""
# Run event generation and produce an EVNT file.
"""
 
import os, sys, time, shutil
from PyJobTransforms.trfLogger import msg
from PyJobTransforms.transform import transform
from PyJobTransforms.trfExe import athenaExecutor
from PyJobTransforms.trfArgs import addAthenaArguments
from PyJobTransforms.trfDecorators import stdTrfExceptionHandler, sigUsrStackTrace
from EvgenJobTransforms.evgenTrfArgs import addStdEvgenArgs
 

ListOfDefaultPositionalKeys=['--AMIConfig', '--AMITag', '--argJSON', '--asetup', '--athena', 
'--athenaMPMergeTargetSize', '--athenaopts', '--attempt', '--checkEventCount', '--command', 
'--dumpJSON', '--dumpPickle', '--ecmEnergy', '--env', '--eventAcceptanceEfficiency', 
'--evgenJobOpts', '--execOnly', '--fileValidation', '--firstEvent', '--ignoreErrors', 
'--ignoreFiles', '--ignorePatterns', '--imf', '--inputEVNT_PreFile', '--inputFileValidation', 
'--inputGenConfFile', '--inputGeneratorFile', '--jobConfig', '--jobid', '--maxEvents', '--orphanKiller', '--outputEVNTFile', '--outputEVNT_PreFile', '--outputHEPMCFile', '--outputFileValidation', '--outputNTUP_TRUTHFile', '--outputTXTFile', '--parallelFileValidation', '--postExec', '--postInclude', '--preExec', '--preInclude', '--wprintEvts', '--randomSeed', '--reportName', '--reportType', '--rivetAnas', '--runNumber', '--showGraph', '--showPath', '--showSteps', '--skipEvents', '--skipFileValidation', '--skipInputFileValidation', '--skipOutputFileValidation', '--steering', '--taskid', '--tcmalloc', '--valgrind', '--valgrindbasicopts', '--valgrindextraopts', '--lheOnly', '--localPath', '--cleanOut', '--saveList', '--avoidExtracting']
 
class EvgenExecutor(athenaExecutor):
  "Specialised trf executor class for event generation jobs"
  def __init__(self, name="generate", skeleton=None, skeletonCA=None, substep=None, inData=set(), outData=set()):
        athenaExecutor.__init__(self, name=name, skeletonFile=skeleton, skeletonCA=skeletonCA,
                                substep=substep, inputEventTest=False, tryDropAndReload=False,
                                inData=inData, outData=outData)
 
  def preExecute(self, input=set(), output=set()):
        "Get input tarball, unpack and set up env if an evgenJobOpts arg was provided."
 
        
        super(EvgenExecutor, self).preExecute(input, output)
 
        def expand_if_archive(filename):
            "Function to expand a file if it is a zip archive or tarball"
            if ".tar" in filename or ".tgz" in filename:
                import tarfile
                with tarfile.open(filename) as tf:
                    tf.extractall()
            elif filename.endswith(".zip"):
                import zipfile
                zf = zipfile.ZipFile(filename)
                zf.extractall()
                zf.close()
            elif ".gz" in filename:
                shouldExtract = True
                if "avoidExtracting" in self._trf.argdict:
                    shouldExtract = not self._trf.argdict["avoidExtracting"].value
                if shouldExtract:
                    import gzip
                    with gzip.open(filename, 'rb') as f_in:
                       lfilename=os.path.basename(filename)
                       with open(lfilename[:-3], 'wb') as f_out:
                           shutil.copyfileobj(f_in, f_out)
                else:
                    msg.info("Avoid extracting gz input generator file.")
        def get_immediate_subdirectories(a_dir):
            return [name for name in os.listdir(a_dir)
                    if os.path.isdir(os.path.join(a_dir, name))]
 
        
        os.environ['LOCAL_INSTALL_DIR'] = (os.environ['JOBOPTSEARCHPATH']).split(":")[0]
        if os.path.exists('/cvmfs/atlas.cern.ch/repo/sw/Generators/MCJobOptions/common'):
           CommonCvmfsDir = '/cvmfs/atlas.cern.ch/repo/sw/Generators/MCJobOptions/common'
           os.environ["JOBOPTSEARCHPATH"] = CommonCvmfsDir+":"+os.environ["JOBOPTSEARCHPATH"]
        if os.path.exists('/cvmfs/atlas.cern.ch/repo/sw/Generators/MCJobOptions/common/MadGraphControl/dat/'):
            datCvmfsDir = '/cvmfs/atlas.cern.ch/repo/sw/Generators/MCJobOptions/common/MadGraphControl/dat/'
            os.environ["DATAPATH"] = datCvmfsDir+":"+os.environ["DATAPATH"]
#        dsidparam = (self._trf.argdict["jobConfig"].value).values()[0][0]
        dsidpar = (self._trf.argdict["jobConfig"].value).values()
        dsidparam = list(dsidpar)[0][0]
        # Adding cvmfs path to JOBOPTSEARCHPATH
        if os.path.exists('/cvmfs/atlas.cern.ch/repo/sw/Generators/MCJobOptions/'):
           BaseCvmfsPath = "/cvmfs/atlas.cern.ch/repo/sw/Generators/MCJobOptions/"
 
        if dsidparam.isdigit() and (len(dsidparam)==6 or len(dsidparam)==7):
            #only dsid is provided, add cvmfs folder like 123xxx to JOBOPTSEARCHPATH
            Jodir = dsidparam[:3]+'xxx'
            if len(dsidparam)==7:
               Jodir = dsidparam[0] + '/' + dsidparam[:4]+'xxx'
            cwdir = os.getcwd()
            cwd_ful = os.path.join(cwdir, dsidparam)
            if (os.path.isdir(cwd_ful)):
               os.environ["JOBOPTSEARCHPATH"] = cwd_ful+":"+os.environ["JOBOPTSEARCHPATH"]
               os.environ["DATAPATH"] = cwd_ful+":"+os.environ["DATAPATH"]
            else:
               cwd_Jodir = os.path.join(cwdir,Jodir)
               cwd_Jodir_ful = os.path.join(cwd_Jodir,dsidparam)
               if (os.path.isdir(cwd_Jodir_ful)):
                  os.environ["JOBOPTSEARCHPATH"] = cwd_Jodir_ful+":"+os.environ["JOBOPTSEARCHPATH"]
                  os.environ["DATAPATH"] = cwd_Jodir_ful+":"+os.environ["DATAPATH"]
               else:
                  if (os.path.isdir(BaseCvmfsPath)):
                     JoCvmfsPath = os.path.join(BaseCvmfsPath, Jodir)
                     JoCvmfsPath_ful = os.path.join(JoCvmfsPath, dsidparam)
                     os.environ["JOBOPTSEARCHPATH"] = JoCvmfsPath_ful+":"+os.environ["JOBOPTSEARCHPATH"]
                     os.environ["DATAPATH"] = JoCvmfsPath_ful+":"+os.environ["DATAPATH"]
                  else:
                     msg.error("No access to JOs cvmfs location. JOs should be placed in working directory or full path should be given.")
                     raise RuntimeError("JOs not found - no cvmfs access")            
        else:  #Suppose full path of dsid folder is provided(/afs/.../123xxx/123456), add cvmfs floder and local path(/afs/.../123xxx) to JOBOPTSEARCHPATH
            if dsidparam.startswith("Test"):  # for testing
                for token in (os.environ['JOBOPTSEARCHPATH']).split(":"):
                    if "jobOptions" in token:
                        search_token = token
                        break
                search_token += "/EvgenJobTransforms/EvgenTest/" + dsidparam.split("Test")[-1]
                os.environ["JOBOPTSEARCHPATH"] = search_token+":"+os.environ["JOBOPTSEARCHPATH"]
                os.environ["DATAPATH"] = search_token+":"+os.environ["DATAPATH"]
            elif (os.path.isdir(dsidparam)):
                os.environ["JOBOPTSEARCHPATH"] = dsidparam+":"+os.environ["JOBOPTSEARCHPATH"]
                os.environ["DATAPATH"] = dsidparam+":"+os.environ["DATAPATH"]
            else:
                msg.error("JOs not found, please check = '%s'", dsidparam)
                raise RuntimeError("JOs not found")
 
        msg.info("Using JOBOPTSEARCHPATH = '%s'", os.environ["JOBOPTSEARCHPATH"])
        msg.info("Using DATAPATH = '%s'", os.environ["DATAPATH"])
 
        if "evgenJobOpts" in self._trf.argdict: 
            tarball = self._trf.argdict["evgenJobOpts"].value
            
            if tarball.startswith("http"):
                url = tarball
                tarball = os.basename(tarball)
            else:
                url = "http://cern.ch/atlas-computing/links/kitsDirectory/EvgenJobOpts/" + tarball
            
            if not os.path.exists(tarball):
                from EvgenJobTransforms.download import downloadUsingProxy
                status, output = downloadUsingProxy(url)
                if status != 0:
                    raise EnvironmentError('Error downloading tarball %s. Downloader reports: %s' % (tarball, output))
                msg.info('Evgen tarball download success: %s', output)
            
            expand_if_archive(tarball)
            
 
        # copy config files to cwd
        FIRST_DIR = (os.environ['JOBOPTSEARCHPATH']).split(":")[0]
        configFiles = [f for f in os.listdir(FIRST_DIR) if ( "GRID" in f)]
        confFile=None
        if len(configFiles) == 1:
            msg.info("gridpack for only one energy available ")
        elif len(configFiles) >1:
            msg.info("more then one gridpack ! ")
        if len(configFiles) >=1:
          if "--ecmEnergy" in str(sys.argv[1:]):
             split_args=str(sys.argv[1:]).split("ecmEnergy",1)[1]
             split_args=split_args.lstrip("\',=")
             ener_GeV=split_args.split(",")[0].strip(" ,\']")
             energy=str(float(ener_GeV)/1000.0).replace('.','p').strip(r"=0\p']")
             msg.info("Should be used gridpack for energy "+energy)
          else:
             msg.info("no ecm energy given, assuming 13.6 TeV ")
             energy="13p6"
          for x in configFiles:
              gridS="mc_"+energy+"TeV"
              msg.info("Gridpack should start from "+gridS)
              if x.startswith(gridS):
                 confFile = os.path.join(FIRST_DIR, x)
                 msg.info("using gridpack = "+confFile)
          if confFile is None:
             msg.error("No *GRID* config files, for requested energy = '%s'  please check = '%s'", energy, dsidparam)
             sys.exit(1)
 
        if confFile is not None:
           expand_if_archive(confFile)
#       os.system("cp %s ." % confFile)
           msg.info("Configuration input gridpack found " + confFile)
 
        #Expand if a tarball is found in local directory
        loc_files = os.listdir(os.getcwd())
        for loc_file in loc_files:
            if "GRID" not in loc_file:
               expand_if_archive(loc_file)
 
        
        if "inputGeneratorFile" in self._trf.argdict:
#           expand_if_archive(self._trf.argdict["inputGeneratorFile"].value)
            myinputfiles = self._trf.argdict["inputGeneratorFile"].value
            genInputFiles = myinputfiles.split(',')
            for file in genInputFiles:
               expand_if_archive(file)
        if "inputGenConfFile" in self._trf.argdict:
            expand_if_archive(self._trf.argdict["inputGenConfFile"].value)
 
def move_files(main_dir,tmp_dir,allowedlist):
    files = os.listdir(tmp_dir)
    files.sort()
    for f in files:
       for i in allowedlist:
            if i in f:
                src = tmp_dir+"/"+f
                dest = main_dir+"/"+f
                os.rename(src,dest)
                break
 
 
def getTransform():
    exeSet = set()
    msg.info("Transform arguments %s", sys.argv[1:])
    if "--outputEVNTFile" in str(sys.argv[1:]):
       exeSet.add(EvgenExecutor(name="generate", skeleton="EvgenJobTransforms/skel.GENtoEVGEN.py", skeletonCA="EvgenJobTransforms.GEN_Skeleton", inData=["inNULL"], outData=["YODA", "EVNT", "EVNT_Pre", "TXT"]))
       msg.info("Output EVNT file")
    elif "--outputYODAFile" in str(sys.argv[1:]):
       exeSet.add(EvgenExecutor(name="generate", skeleton="EvgenJobTransforms/skel.GENtoEVGEN.py", skeletonCA="EvgenJobTransforms.GEN_Skeleton", inData=["inNULL"], outData=["YODA", "TXT"]))
       msg.info("Output EVNT file")
    elif "--outputTXTFile" in str(sys.argv[1:]):
       exeSet.add(EvgenExecutor(name="generate", skeleton="EvgenJobTransforms/skel.GENtoTXT.py", skeletonCA="EvgenJobTransforms.GEN_Skeleton", inData=["inNULL"], outData=["TXT"]))
       msg.info("Output TXT file")
    elif "--outputHEPMCFile" not in str(sys.argv[1:]):
       msg.error("Output cannot be recognised")
 
    exeSet.add(EvgenExecutor(name="afterburn", skeleton="EvgenJobTransforms/skel.ABtoEVGEN.py", skeletonCA="EvgenJobTransforms.GEN_Skeleton", inData=["EVNT_Pre"], outData=["EVNT"]))
    exeSet.add(athenaExecutor(name = "AODtoDPD", skeletonFile = "PATJobTransforms/skeleton.AODtoDPD_tf.py",
                              substep = "a2d", inData = ["EVNT"], outData = ["NTUP_TRUTH"], perfMonFile = "ntuple_AODtoDPD.pmon.gz"))
    exeSet.add(athenaExecutor(name = 'EVNTtoHEPMC', skeletonCA = 'EvgenJobTransforms.POOLtoHEPMC_Skeleton',
                              substep = "a2d", perfMonFile = 'ntuple.pmon.gz', inData=['EVNT'], outData=['HEPMC']))
    trf = transform(executor=exeSet)
    addAthenaArguments(trf.parser, maxEventsDefaultSubstep='all')
    addStdEvgenArgs(trf.parser)
    return trf
 
 
@stdTrfExceptionHandler
@sigUsrStackTrace
def main():
    msg.info("This is %s", sys.argv[0])
 
    main_dir = os.getcwd()
    trf = getTransform()
    trf.parseCmdLineArgs(sys.argv[1:])
    if (("cleanOut" in trf.argdict) and (trf.argdict["cleanOut"].value != 0)):
      name_tmpdir = "tmprun"
      tmp_dir = os.path.join(main_dir, name_tmpdir)
      if os.path.isdir(tmp_dir):
         shutil.rmtree(tmp_dir, ignore_errors=True)
      os.mkdir("tmprun")
      os.chdir("tmprun")
      tmp_dir = os.getcwd()
      allowedlist_in = ['MC','group','TXT']
      move_files(tmp_dir,main_dir,allowedlist_in)
 
    trf.execute()
    trf.generateReport()
    msg.info("%s stopped at %s, trf exit code %d", sys.argv[0], time.asctime(), trf.exitCode)
 
 
# read files/dirs that should be saved and if present in cwd - remove
 
    if (("cleanOut" in trf.argdict) and (trf.argdict["cleanOut"].value!=0)):
       allowedlist_out = ['log.generate','.root']
       if "outputTXTFile" in trf.argdict:
         allowedlist_out.append('TXT')
       if "saveList" in trf.argdict:
         saveList_dic= trf.argdict["saveList"].value
         saveList_str= str(saveList_dic)
         saveList_str=saveList_str[10:-3]
         saveList= saveList_str.split(",")
         for item in saveList:
           test_ex = os.path.join(main_dir,str(item))
           if os.path.isdir(test_ex):
             shutil.rmtree(test_ex, ignore_errors=True)
           elif os.path.isfile(test_ex):
             os.remove(test_ex)
         if not saveList[0].isdigit():
             allowedlist_out=allowedlist_out+saveList
 
       move_files(main_dir,tmp_dir,allowedlist_out)
       os.chdir(main_dir)
       if "saveList" not in trf.argdict:
         shutil.rmtree(tmp_dir, ignore_errors=True)
       elif not saveList[0].isdigit():
         shutil.rmtree(tmp_dir, ignore_errors=True)
# if cleanOut is not defined and multipleinput preset, remove the merged file
    elif ("inputGeneratorFile" in trf.argdict):
        myinputfiles = trf.argdict["inputGeneratorFile"].value
        genInputFiles = myinputfiles.split(',')
        numberOfFiles = len(genInputFiles)
        merge_file = 'merged_lhef._0.events'
        if((numberOfFiles>1) and (os.path.exists(merge_file))):
           os.remove(merge_file)
#
    if (("lheOnly" in trf.argdict ) and (trf.argdict["lheOnly"].value == 1)):
          outputName = ''.join(trf.argdict["outputEVNTFile"].value)
          os.remove(outputName)
    sys.exit(trf.exitCode)
 
 
# TODO: Open resulting EVNT file to extract cross-section, generator names+versions, etc. from the HepMC::GenRun or whatever... in an executor postExecute?
 
 
if __name__ == "__main__":
    main()