def skimRecoFilesParallel(thisfile):
f = TFile(thisfile)
t = f.Get('tree')
newname = thisfile.rstrip('.root') + '_skim.root'
newFile = TFile(newname, 'recreate')
#newTree = t.CopyTree('weight!=0.')
#newTree = t.CopyTree('metfilters==1 && trigger==1 && onelepton==1 && btags==1 && ak4jetmult==1 && ak4jetcuts==1 && validminimization==1')
newTree = t.CopyTree(
'fullselection==1 || wjets_cr_selection==1 || qcd_A_SR_selection==1 || qcd_B_SR_selection==1 || qcd_C_SR_selection==1 || qcd_A_CR_selection==1 || qcd_B_CR_selection==1 || qcd_C_CR_selection==1'
)
#newTree = t.CopyTree('fullselection==1')
#newTree = t.CopyTree('eventTopology==1')
newTree.Write()
newFile.Purge()
newFile.Close()
def produce(self):
from ROOT import TFile, TChain, TH1
from os.path import exists as pathExists
from os.path import split as splitPath
outFilePath = "%s/%s.%s.%s.root" % (self.outPath, "".join(
self.flags), "processed", self.name)
#if pathExists(outFilePath):
# return
outFile = TFile(
"%s/%s.%s.%s.root" %
(self.outPath, "".join(self.flags), "processed", self.name),
"recreate")
for section in self.config.sections():
trees = None
if section.startswith("dileptonTree:"):
treeProducerName = self.config.get(section, "treeProducerName")
trees = self._getDileptonTrees(section)
treeName = "DileptonTree"
subDirName = "%s%s" % (section.split("dileptonTree:")[1],
treeProducerName)
if "PFHTHLT" in subDirName:
subDirName = "%sFinalTrees" % subDirName.split(
treeProducerName)[0]
if section.startswith("isoTree:"):
treeProducerName = self.config.get(section, "treeProducerName")
trees = self._getIsoTrees(section)
treeName = "Iso"
subDirName = "%s%s" % (section.split("isoTree:")[1],
treeProducerName)
if not trees == None:
outDir = None
srcTree = {}
for object in trees:
srcTree[object] = TChain("%s%s" % (object, treeName))
processors = self.config.get(section, "%sProcessors" %
object).split()
filter = " and ".join(processors)
if self.config.has_option(section, "%sFilter" % object):
filter = self.config.get(section, "%sFilter" % object)
for treePath in trees[object]:
#srcFile = TFile(filePath,"r")
#srcTree = srcFile.Get(treePath)
filePath = "%s.root" % treePath.split(".root")[0]
inFile = TFile(filePath, "READ")
makeCounterSum = eval(
self.config.get("general", "counterSum"))
print "Add counter sum: %s" % makeCounterSum
if not self.counterSum and makeCounterSum:
if not outFile.GetDirectory(
"%sCounters" %
section.split("dileptonTree:")[1]):
outFile.mkdir(
"%sCounters" %
section.split("dileptonTree:")[1])
outFile.cd("%sCounters" %
section.split("dileptonTree:")[1])
task = None
if "vtxWeighter" in processors:
t = self.config.get("general", "tasks")
if ".%s." % t in splitPath(filePath)[1]:
assert task == None, "unable to disambiguate tasks '%s' matches both '%s' and '%s'" (
filePath, task, t)
task = t
else:
task = t
else:
for t in self.tasks:
if ".%s." % t in splitPath(filePath)[1]:
assert task == None, "unable to disambiguate tasks '%s' matches both '%s' and '%s'" (
filePath, task, t)
task = t
self.counterSum = inFile.Get(
"%sCounters/analysis paths" % task).Clone()
## also add 3D weights
#outFile.mkdir("%sWeightSummer" % section.split("dileptonTree:")[1])
#outFile.cd("%sWeightSummer" % section.split("dileptonTree:")[1])
## task has been defined above
#self.weightSum = inFile.Get("%sWeightSummer/Weights"%task).Clone()
else:
pass
#need to cope with different lumis :(
#h = inFile.Get("%sCounters/analysis paths"%task)
#print inFile, "%sCounters/analysis paths"%task, h
#self.counterSum.Add( h,1. )
inFile.Close()
srcTree[object].Add(treePath)
print "adding", treePath
srcTree[object].SetBranchStatus("*", 1)
### signal code
if (self.treeProcessors[filter].__class__.__name__
== SimpleSelector.__name__
and self.config.has_option(section,
"%sFilter" % object)):
expression = self.treeProcessors[filter].getExpression(
object)
print "Cutting tree down to: '%s'" % (expression)
outFile.Write()
outFile.Close()
srcTree[object] = srcTree[object].CopyTree(expression)
outFile = TFile(
"%s/%s.%s.%s.root" % (self.outPath, "".join(
self.flags), "processed", self.name), "UPDATE")
#####
for processorName in processors:
if processorName == "vtxWeighter":
srcTree[object].SetBranchStatus("weight", 0)
#### default code
#~ if (self.treeProcessors[processorName].__class__.__name__ == SimpleSelector.__name__ and not self.config.has_option(section,"%sFilter"%object)):
#~ print "Requirements met, applying simple selection boosting =)"
#~ expression = self.treeProcessors[processorName].getExpression(object)
#~ print "Cutting tree down to: '%s'" % (expression)
#~ srcTree[object] = srcTree[object].CopyTree(expression)
######
self.treeProcessors[processorName].prepareSrc(
srcTree[object], object, self.treeProcessors)
for object in trees:
processors = self.config.get(section, "%sProcessors" %
object).split()
filter = " and ".join(processors)
if self.config.has_option(section, "%sFilter" % object):
filter = self.config.get(section, "%sFilter" % object)
if not outDir:
outDir = outFile.mkdir(subDirName)
outFile.cd(subDirName)
destTree = srcTree[object].CloneTree(0)
destTree.SetAutoSave(5000000000)
#print processors
for processorName in processors:
self.treeProcessors[processorName].prepareDest(
destTree, object)
#~ print "%s: %d" % (str(processorName), self.treeProcessors[processorName].nEntries)
endOfLine = 1000
for i in srcTree[object]:
if endOfLine < 1:
pass
#continue
endOfLine -= 1
processingResults = {}
### default code
#~ for processorName in processors:
#~ processingResults[processorName] = self.treeProcessors[processorName].processEvent(srcTree[object], object)
#~ if filter == "" or eval(filter, processingResults):
#~ destTree.Fill()
####
### signal code
keep = False
for processorName in processors:
keep = self.treeProcessors[
processorName].processEvent(
srcTree[object], object)
if keep:
destTree.Fill()
####
#srcFile.Close()
outFile.Write()
#from pprint import pprint
#pprint( trees)
outFile.Purge()
outFile.Close()
def produce(self):
from ROOT import TFile, TChain, TH1
from os.path import exists as pathExists
from os.path import split as splitPath
outFilePath = "%s/%s.%s.%s.root" % (self.outPath, "".join(
self.flags), "processed", self.name)
#if pathExists(outFilePath):
# return
outFile = TFile(
"%s/%s.%s.%s.root" %
(self.outPath, "".join(self.flags), "processed", self.name),
"recreate")
for section in self.config.sections():
trees = None
if section.startswith("dileptonTree:"):
treeProducerName = self.config.get(section, "treeProducerName")
trees = self._getDileptonTrees(section)
treeName = "DileptonTree"
subDirName = "%s%s" % (section.split("dileptonTree:")[1],
treeProducerName)
if not trees == None:
outDir = None
srcTree = {}
for object in trees:
srcTree[object] = TChain("%s%s" % (object, treeName))
processors = self.config.get(section, "%sProcessors" %
object).split()
filter = " and ".join(processors)
if self.config.has_option(section, "%sFilter" % object):
filter = self.config.get(section, "%sFilter" % object)
### Quick and dirty workaround to prefer certain primary datasets (double muon > double electron > emu)
### To be in synch with ETH group
### Remove duplication of loops and conditions on path if this is to be removed again
for treePath in trees[object]:
### Use events from dimuon primary dataset first
if not "DoubleMuon" in treePath:
continue
#srcFile = TFile(filePath,"r")
#srcTree = srcFile.Get(treePath)
filePath = "%s.root" % treePath.split(".root")[0]
inFile = TFile(filePath, "READ")
makeCounterSum = eval(
self.config.get("general", "counterSum"))
print "Add counter sum: %s" % makeCounterSum
if not self.counterSum and makeCounterSum:
if not outFile.GetDirectory(
"%sCounters" %
section.split("dileptonTree:")[1]):
outFile.mkdir(
"%sCounters" %
section.split("dileptonTree:")[1])
outFile.cd("%sCounters" %
section.split("dileptonTree:")[1])
task = None
if "vtxWeighter" in processors:
t = self.config.get("general", "tasks")
if ".%s." % t in splitPath(filePath)[1]:
assert task == None, "unable to disambiguate tasks '%s' matches both '%s' and '%s'" (
filePath, task, t)
task = t
else:
task = t
else:
for t in self.tasks:
if ".%s." % t in splitPath(filePath)[1]:
assert task == None, "unable to disambiguate tasks '%s' matches both '%s' and '%s'" (
filePath, task, t)
task = t
self.counterSum = inFile.Get(
"%sCounters/analysis paths" % task).Clone()
else:
pass
inFile.Close()
srcTree[object].Add(treePath)
print "adding", treePath
for treePath in trees[object]:
### Then take ee primary dataset
if not "DoubleElectron" in treePath:
continue
#srcFile = TFile(filePath,"r")
#srcTree = srcFile.Get(treePath)
filePath = "%s.root" % treePath.split(".root")[0]
inFile = TFile(filePath, "READ")
makeCounterSum = eval(
self.config.get("general", "counterSum"))
print "Add counter sum: %s" % makeCounterSum
if not self.counterSum and makeCounterSum:
if not outFile.GetDirectory(
"%sCounters" %
section.split("dileptonTree:")[1]):
outFile.mkdir(
"%sCounters" %
section.split("dileptonTree:")[1])
outFile.cd("%sCounters" %
section.split("dileptonTree:")[1])
task = None
if "vtxWeighter" in processors:
t = self.config.get("general", "tasks")
if ".%s." % t in splitPath(filePath)[1]:
assert task == None, "unable to disambiguate tasks '%s' matches both '%s' and '%s'" (
filePath, task, t)
task = t
else:
task = t
else:
for t in self.tasks:
if ".%s." % t in splitPath(filePath)[1]:
assert task == None, "unable to disambiguate tasks '%s' matches both '%s' and '%s'" (
filePath, task, t)
task = t
self.counterSum = inFile.Get(
"%sCounters/analysis paths" % task).Clone()
else:
pass
inFile.Close()
srcTree[object].Add(treePath)
print "adding", treePath
for treePath in trees[object]:
### Now the rest, but avoid taking a tree twice
if "DoubleElectron" in treePath or "DoubleMuon" in treePath:
continue
#srcFile = TFile(filePath,"r")
#srcTree = srcFile.Get(treePath)
filePath = "%s.root" % treePath.split(".root")[0]
inFile = TFile(filePath, "READ")
makeCounterSum = eval(
self.config.get("general", "counterSum"))
print "Add counter sum: %s" % makeCounterSum
if not self.counterSum and makeCounterSum:
if not outFile.GetDirectory(
"%sCounters" %
section.split("dileptonTree:")[1]):
outFile.mkdir(
"%sCounters" %
section.split("dileptonTree:")[1])
outFile.cd("%sCounters" %
section.split("dileptonTree:")[1])
task = None
if "vtxWeighter" in processors:
t = self.config.get("general", "tasks")
if ".%s." % t in splitPath(filePath)[1]:
assert task == None, "unable to disambiguate tasks '%s' matches both '%s' and '%s'" (
filePath, task, t)
task = t
else:
task = t
else:
for t in self.tasks:
if ".%s." % t in splitPath(filePath)[1]:
assert task == None, "unable to disambiguate tasks '%s' matches both '%s' and '%s'" (
filePath, task, t)
task = t
self.counterSum = inFile.Get(
"%sCounters/analysis paths" % task).Clone()
else:
pass
inFile.Close()
srcTree[object].Add(treePath)
print "adding", treePath
srcTree[object].SetBranchStatus("*", 1)
for processorName in processors:
if processorName == "vtxWeighter":
srcTree[object].SetBranchStatus("weight", 0)
#### old code
if (self.treeProcessors[processorName].__class__.
__name__ == SimpleSelector.__name__
and not self.config.has_option(
section, "%sFilter" % object)):
print "Requirements met, applying simple selection boosting =)"
expression = self.treeProcessors[
processorName].getExpression(object)
print "Cutting tree down to: '%s'" % (expression)
srcTree[object] = srcTree[object].CopyTree(
expression)
######
self.treeProcessors[processorName].prepareSrc(
srcTree[object], object, self.treeProcessors)
for object in trees:
processors = self.config.get(section, "%sProcessors" %
object).split()
filter = " and ".join(processors)
if self.config.has_option(section, "%sFilter" % object):
filter = self.config.get(section, "%sFilter" % object)
if not outDir:
outDir = outFile.mkdir(subDirName)
outFile.cd(subDirName)
destTree = srcTree[object].CloneTree(0)
destTree.SetAutoSave(5000000000)
#print processors
for processorName in processors:
self.treeProcessors[processorName].prepareDest(
destTree, object)
#~ print "%s: %d" % (str(processorName), self.treeProcessors[processorName].nEntries)
endOfLine = 1000
for i in srcTree[object]:
if endOfLine < 1:
pass
#continue
endOfLine -= 1
processingResults = {}
### old code
for processorName in processors:
processingResults[
processorName] = self.treeProcessors[
processorName].processEvent(
srcTree[object], object)
if filter == "" or eval(filter, processingResults):
destTree.Fill()
####
#srcFile.Close()
outFile.Write()
#from pprint import pprint
#pprint( trees)
outFile.Purge()
outFile.Close()