def btagEfficiencyTreeProducer(stageName="Z+jet", muChannel=True, path='../testfiles/'):
#search for category number
stage=-1
for cat in categoryNames :
stage+=1
if cat==stageName : break
# prepare output
path='/nfs/user/llbb/Pat_8TeV_532p4/DYjets_Summer12_V2/'
ROOT.gROOT.ProcessLine(
"struct MyStruct {\
Float_t pt;\
Float_t eta;\
Int_t flavor;\
Float_t ssvhe;\
Float_t ssvhp;\
Float_t csv;\
Float_t eventWeight;\
};" )
from ROOT import MyStruct
mystruct = MyStruct()
f = ROOT.TFile( 'mybtagEfftree.root', 'RECREATE' )
tree = ROOT.TTree( 'btagEff', 'btag efficiency' )
tree.Branch( 'data', mystruct, 'pt/F:eta/F:flavor/I:ssvhe/F:ssvhp/F:csv/F:eventWeight/F' )
# input
if os.path.isdir(path):
dirList=os.listdir(path)
files=[]
for fname in dirList:
files.append(path+fname)
elif os.path.isfile(path):
files=[path]
else:
files=[]
events = AnalysisEvent(files)
EventSelection.prepareAnalysisEvent(events)
# event loop
eventCnt = 0
print "starting loop on events"
for event in events:
categoryData = event.catMu if muChannel else event.catEle
goodJets = event.goodJets_mu if muChannel else event.goodJets_ele
if EventSelection.isInCategory(stage, categoryData):
eventCnt = eventCnt +1
if eventCnt%100==0 : print ".",
if eventCnt%1000==0 : print ""
# event weight
mystruct.eventWeight = event.weight(weightList=["PileUp"])
# that's where we access the jets
for index,jet in enumerate(event.jets):
if not goodJets[index]: continue
mystruct.pt = jet.pt()
mystruct.eta = jet.eta()
mystruct.flavor = jet.partonFlavour()
mystruct.ssvhe = jet.bDiscriminator("simpleSecondaryVertexHighEffBJetTags")
mystruct.ssvhp = jet.bDiscriminator("simpleSecondaryVertexHighPurBJetTags")
mystruct.csv = jet.bDiscriminator("combinedSecondaryVertexBJetTags")
tree.Fill()
f.Write()
f.Close()
print ""
def process(self, event):
"""EventSelectionControlPlots"""
result = { }
## event category
categoryData = event.category
result["category"] = [ ]
for category in range(self.eventCategories):
if EventSelection.isInCategory(category, categoryData):
result["category"].append(category)
result["event"] = event.event()
return result
def process(self, event):
"""EventSelectionControlPlots"""
result = {}
## event category
categoryData = event.category
result["category"] = []
for category in range(self.eventCategories):
if EventSelection.isInCategory(category, categoryData):
result["category"].append(category)
result["event"] = event.event()
return result
def DumpEventList(category, path="./", output="eventlist.txt"):
"""Dump a list of events in a given category"""
# input
if os.path.isdir(path):
dirList = os.listdir(path)
files = []
for fname in dirList:
files.append(path + fname)
elif os.path.isfile(path):
files = [path]
else:
files = []
# events
events = AnalysisEvent(files)
# output
event_list = open(output, "w")
# collections and producers used in the analysis
EventSelection.prepareAnalysisEvent(events)
for event in events:
# check category
if EventSelection.isInCategory(category, event.category):
# print
print >> event_list, "Event", event.event()
def DumpEventList(category, path="./", output="eventlist.txt"):
"""Dump a list of events in a given category"""
# input
if os.path.isdir(path):
dirList=os.listdir(path)
files=[]
for fname in dirList:
files.append(path+fname)
elif os.path.isfile(path):
files=[path]
else:
files=[]
# events
events = AnalysisEvent(files)
# output
event_list = open(output,"w")
# collections and producers used in the analysis
EventSelection.prepareAnalysisEvent(events)
for event in events:
# check category
if EventSelection.isInCategory(category, event.category):
# print
print >> event_list , "Event", event.event()
def runAnalysis(path,
levels,
outputname="controlPlots.root",
Njobs=1,
jobNumber=1):
"""produce all the plots in one go"""
# inputs
if os.path.isdir(path):
dirList = list(
itertools.islice(os.listdir(path), jobNumber, None, Njobs))
files = []
for fname in dirList:
files.append(path + "/" + fname)
elif os.path.isfile(path):
files = [path]
else:
files = []
# output
output = ROOT.TFile(outputname, "RECREATE")
if configuration.runningMode == "dataset":
ROOT.RooAbsData.setDefaultStorageType(ROOT.RooAbsData.Tree)
rds = ROOT.RooDataSet(configuration.RDSname, configuration.RDSname,
ROOT.RooArgSet())
# events iterator, plus configuration of standard collections and producers
events = AnalysisEvent(files)
EventSelection.prepareAnalysisEvent(events)
# prepare the plots
controlPlots = []
if configuration.runningMode == "plots":
leafList = [None] * EventSelection.eventCategories()
createDirectory(EventSelection.categoriesHierarchy(), output, leafList)
for levelDir in leafList:
levelPlots = []
for cp in configuration.controlPlots:
levelPlots.append(
getattr(import_module(cp.module),
cp.classname)(dir=levelDir.mkdir(cp.label),
mode="plots"))
controlPlots.append(levelPlots)
else:
for cp in configuration.controlPlots:
controlPlots.append(
getattr(import_module(cp.module), cp.classname)(dir=None,
mode="dataset",
dataset=rds))
# book histograms (separate iteration for clarity)
if configuration.runningMode == "plots":
for level in levels:
for conf, cp in zip(configuration.controlPlots,
controlPlots[level]):
cp.beginJob(**conf.kwargs)
else:
for conf, cp in zip(configuration.controlPlots, controlPlots):
cp.beginJob(**conf.kwargs)
for cp in controlPlots[:1]:
cp.defineCategories(EventSelection.categoryNames)
# process events
i = 0
t0 = time.time()
for event in events:
# printout
if i % 100 == 0:
print "Processing... event %d. Last batch in %f s." % (i, (
time.time() - t0))
t0 = time.time()
if configuration.runningMode == "plots":
# loop on channels
plots = filter(
lambda x: EventSelection.isInCategory(x, event.category),
levels)
# process the event once (for the first level)
selectionPlotsData = []
for level in plots[:1]:
for cp in controlPlots[level]:
selectionPlotsData.append(cp.process(event))
# fill the histograms
for level in plots:
for cp, data in zip(controlPlots[level], selectionPlotsData):
cp.fill(data, event.weight(category=level))
else:
for cp in controlPlots[:1]:
# set categories (first CP only)
cp.setCategories(
map(
lambda c: EventSelection.isInCategory(
c, event.category),
range(EventSelection.eventCategories())))
for cp in controlPlots:
# process event (all CP)
cp.processEvent(event)
# add to the dataset
rds.add(getArgSet(controlPlots))
i += 1
# save all
if configuration.runningMode == "plots":
for level in levels:
for cp in controlPlots[level]:
cp.endJob()
else:
for cp in controlPlots:
cp.endJob()
# for dataset, write the merged RDS to file
if configuration.runningMode == "dataset":
output.cd()
ws_ras = ROOT.RooWorkspace(configuration.WSname, configuration.WSname)
getattr(ws_ras, 'import')(rds.get())
output.Add(ws_ras)
ws_ras.Write()
rds.tree().Write()
# close the file
output.Close()
def runAnalysis(path, levels, outputname="controlPlots.root", Njobs=1, jobNumber=1):
"""produce all the plots in one go"""
# inputs
if os.path.isdir(path):
dirList=list(itertools.islice(os.listdir(path), jobNumber, None, Njobs))
files=[]
for fname in dirList:
files.append(path+"/"+fname)
elif os.path.isfile(path):
files=[path]
else:
files=[]
# output
output = ROOT.TFile(outputname, "RECREATE")
if configuration.runningMode=="dataset":
ROOT.RooAbsData.setDefaultStorageType(ROOT.RooAbsData.Tree)
rds = ROOT.RooDataSet(configuration.RDSname, configuration.RDSname, ROOT.RooArgSet())
# events iterator, plus configuration of standard collections and producers
events = AnalysisEvent(files)
EventSelection.prepareAnalysisEvent(events)
# prepare the plots
controlPlots=[]
if configuration.runningMode=="plots":
leafList = [None]*EventSelection.eventCategories()
createDirectory(EventSelection.categoriesHierarchy(), output, leafList)
for levelDir in leafList:
levelPlots=[]
for cp in configuration.controlPlots:
levelPlots.append(getattr(import_module(cp.module),cp.classname)(dir=levelDir.mkdir(cp.label),mode="plots"))
controlPlots.append(levelPlots)
else:
for cp in configuration.controlPlots:
controlPlots.append(getattr(import_module(cp.module),cp.classname)(dir=None, mode="dataset", dataset=rds))
# book histograms (separate iteration for clarity)
if configuration.runningMode=="plots":
for level in levels:
for conf,cp in zip(configuration.controlPlots,controlPlots[level]):
cp.beginJob(**conf.kwargs)
else:
for conf,cp in zip(configuration.controlPlots,controlPlots):
cp.beginJob(**conf.kwargs)
for cp in controlPlots[:1]:
cp.defineCategories(EventSelection.categoryNames)
# process events
i = 0
t0 = time.time()
for event in events:
# printout
if i%100==0 :
print "Processing... event %d. Last batch in %f s." % (i,(time.time()-t0))
t0 = time.time()
if configuration.runningMode=="plots":
# loop on channels
plots = filter(lambda x: EventSelection.isInCategory(x,event.category) ,levels)
# process the event once (for the first level)
selectionPlotsData=[]
for level in plots[:1]:
for cp in controlPlots[level]:
selectionPlotsData.append(cp.process(event))
# fill the histograms
for level in plots:
for cp, data in zip(controlPlots[level],selectionPlotsData):
cp.fill(data, event.weight(category=level))
else:
for cp in controlPlots[:1]:
# set categories (first CP only)
cp.setCategories(map(lambda c:EventSelection.isInCategory(c, event.category),range(EventSelection.eventCategories())))
for cp in controlPlots:
# process event (all CP)
cp.processEvent(event)
# add to the dataset
rds.add(getArgSet(controlPlots))
i += 1
# save all
if configuration.runningMode=="plots":
for level in levels:
for cp in controlPlots[level]:
cp.endJob()
else:
for cp in controlPlots:
cp.endJob()
# for dataset, write the merged RDS to file
if configuration.runningMode=="dataset":
output.cd()
ws_ras = ROOT.RooWorkspace(configuration.WSname, configuration.WSname)
getattr(ws_ras,'import')(rds.get())
output.Add(ws_ras)
ws_ras.Write()
rds.tree().Write()
# close the file
output.Close()
def runAnalysis(path, txt, levels, outputname="controlPlots.root", Njobs=1, jobNumber=1):
"""produce all the plots in one go"""
# inputs
dcapDir = False
if "dcap://" in path:
split = path.split("/")
decapDir = os.path.isdir("/"+"/".join(split[3:]))
if os.path.isdir(path) or decapDir:
if txt: # IWN: use txt file with all names
f = open(txt)
dirList=list(itertools.islice([line[:-1] for line in f], jobNumber, None, Njobs))
f.close()
else:
dirList=list(itertools.islice(os.listdir(path), jobNumber, None, Njobs))
files=[]
for fname in dirList:
files.append(path+"/"+fname)
elif os.path.isfile(path):
files=[path]
else:
files=[path]
# output
output = ROOT.TFile(outputname, "RECREATE")
if configuration.runningMode=="dataset":
ROOT.RooAbsData.setDefaultStorageType(ROOT.RooAbsData.Tree)
rds = ROOT.RooDataSet(configuration.RDSname, configuration.RDSname, ROOT.RooArgSet())
# events iterator, plus configuration of standard collections and producers
events = AnalysisEvent(files)
EventSelection.prepareAnalysisEvent(events)
# prepare the plots
controlPlots=[]
if configuration.runningMode=="plots":
leafList = [None]*EventSelection.eventCategories()
createDirectory(EventSelection.categoriesHierarchy(), output, leafList)
for levelDir in leafList:
levelPlots=[]
for cp in configuration.controlPlots:
levelPlots.append(getattr(import_module(cp.module),cp.classname)(dir=levelDir.mkdir(cp.label),mode="plots"))
controlPlots.append(levelPlots)
else:
for cp in configuration.controlPlots:
controlPlots.append(getattr(import_module(cp.module),cp.classname)(dir=None, mode="dataset", dataset=rds))
# book histograms (separate iteration for clarity)
if configuration.runningMode=="plots":
for level in levels:
for conf,cp in zip(configuration.controlPlots,controlPlots[level]):
cp.beginJob(**conf.kwargs)
else:
for conf,cp in zip(configuration.controlPlots,controlPlots):
cp.beginJob(**conf.kwargs)
for cp in controlPlots[:1]:
cp.defineCategories(EventSelection.categoryNames)
# process events
i = 0
DeltaTb = 0
tb = time.time()
n = events.GetEntries()
ETA = " "
print "\n\n\t__%s_events_to_process.__" %n
for event in events:
# printout
if i%1000==0 :
if i%10000==0 :
if DeltaTb>0:
# ETA = " ETA: %s" % time.strftime("%H h. %M min. %S s.", time.gmtime( DeltaTb/i*(n-i) ))
# runtime = "\n Running for %s" % time.strftime("%H h. %M min. %S s.", time.gmtime(time.time()-t0))
# print runtime.replace(" 00 h.","").replace("for 0","for ") + ETA.replace(" 00 h.","").replace(": 0",": ")
(m,s) = divmod(time.time()-t0,60)
(h,m) = divmod(m,60)
runtime = "\n Running for %i h. %i min. %i s." % (h,m,s)
(m,s) = divmod(DeltaTb/i*(n-i),60)
(h,m) = divmod(m,60)
ETA = " ETA: %i h. %i min. %i s." % (h,m,s)
print runtime.replace(" 0 h.","") + ETA.replace(" 0 h.","")
DeltaTb += time.time()-tb
print "%d%%: Processing... event %d. Last batch in %f s." % (i*100/n,i,(time.time()-tb))
tb = time.time()
i += 1
if configuration.runningMode=="plots":
# loop on channels
plots = filter(lambda x: EventSelection.isInCategory(x,event.category) ,levels)
# process the event once (for the first level)
selectionPlotsData=[]
for level in plots[:1]:
for cp in controlPlots[level]:
selectionPlotsData.append(cp.process(event))
# fill the histograms
for level in plots:
for cp, data in zip(controlPlots[level],selectionPlotsData):
cp.fill(data, event.weight(category=level))
else:
for cp in controlPlots[:1]:
# set categories (first CP only)
cp.setCategories(map(lambda c: EventSelection.isInCategory(c, event.category),range(EventSelection.eventCategories())))
for cp in controlPlots:
# process event (all CP)
cp.processEvent(event)
# add to the dataset
rds.add(getArgSet(controlPlots))
# save all
if configuration.runningMode=="plots":
for level in levels:
for cp in controlPlots[level]:
cp.endJob(level)
else:
for cp in controlPlots:
cp.endJob()
# for dataset, write the merged RDS to file
if configuration.runningMode=="dataset":
output.cd()
ws_ras = ROOT.RooWorkspace(configuration.WSname, configuration.WSname)
getattr(ws_ras,'import')(rds.get())
output.Add(ws_ras)
ws_ras.Write()
rds.tree().Write()
# close the file
output.Close()
print "\nDone. Only took %s!\n" % time.strftime("%H:%M:%S", time.gmtime(time.time()-t0))
