def beginLoop(self, setup):
super(VHbbAnalyzer, self).beginLoop(setup)
if "outputfile" in setup.services:
setup.services["outputfile"].file.cd()
self.inputCounter = ROOT.TH1F("Count", "Count", 1, 0, 2)
self.inputCounterFullWeighted = ROOT.TH1F(
"CountFullWeighted", "Count with gen weight and pu weight", 1,
0, 2)
self.inputCounterWeighted = ROOT.TH1F(
"CountWeighted", "Count with sign(gen weight) and pu weight",
1, 0, 2)
self.inputCounterPosWeight = ROOT.TH1F("CountPosWeight",
"Count genWeight>0", 1, 0,
2)
self.inputCounterNegWeight = ROOT.TH1F("CountNegWeight",
"Count genWeight<0", 1, 0,
2)
#for LHE_scale in range(6):
setattr(
self, "inputCounterWeightedLHEWeightScale",
ROOT.TH1F(
"CountWeightedLHEWeightScale",
"Count with gen weight x LHE_weights_scale and pu weight",
6, -0.5, 5.5))
setattr(
self, "inputCounterWeightedLHEWeightPdf",
ROOT.TH1F(
"CountWeightedLHEWeightPdf",
"Count with gen weight x LHE_weights_pdf and pu weight",
103, -0.5, 102.5))
#for LHE_pdf in range(2):
# setattr(self, "inputCounterWeightedLHEWeightPdf_"+str(LHE_pdf), ROOT.TH1F("CountWeightedLHEWeightPdf_"+str(LHE_pdf),"Count with gen weight x LHE_weights_pdf["+str(LHE_pdf)+"] and pu weight",1,0,2))
self.regressions = {}
self.regressionVBF = {}
for re in self.cfg_ana.regressionVBF:
print "Initialize regression ", re
regression_VBF = JetRegression(re["weight"], re["name"])
for i in re["vtypes"]:
self.regressionVBF[i] = regression_VBF
for re in self.cfg_ana.regressions:
print "Initialize regression ", re
regression = JetRegression(re["weight"], re["name"])
for i in re["vtypes"]:
self.regressions[i] = regression
blike = self.cfg_ana.VBFblikelihood
print "Initialize VBF blikelihood ", blike
self.blikelihood = VBFblikelihood(blike["weight"], blike["name"])
def beginLoop(self,setup):
super(VHbbAnalyzer,self).beginLoop(setup)
if "outputfile" in setup.services :
setup.services["outputfile"].file.cd()
self.inputCounter = ROOT.TH1F("Count","Count",1,0,2)
self.inputCounterFullWeighted = ROOT.TH1F("CountFullWeighted","Count with gen weight and pu weight",1,0,2)
self.inputCounterWeighted = ROOT.TH1F("CountWeighted","Count with sign(gen weight) and pu weight",1,0,2)
self.inputCounterPosWeight = ROOT.TH1F("CountPosWeight","Count genWeight>0",1,0,2)
self.inputCounterNegWeight = ROOT.TH1F("CountNegWeight","Count genWeight<0",1,0,2)
#for LHE_scale in range(6):
setattr(self, "inputCounterWeightedLHEWeightScale", ROOT.TH1F("CountWeightedLHEWeightScale","Count with gen weight x LHE_weights_scale and pu weight",6,-0.5,5.5))
setattr(self, "inputCounterWeightedLHEWeightPdf", ROOT.TH1F("CountWeightedLHEWeightPdf","Count with gen weight x LHE_weights_pdf and pu weight",103,-0.5,102.5))
#for LHE_pdf in range(2):
# setattr(self, "inputCounterWeightedLHEWeightPdf_"+str(LHE_pdf), ROOT.TH1F("CountWeightedLHEWeightPdf_"+str(LHE_pdf),"Count with gen weight x LHE_weights_pdf["+str(LHE_pdf)+"] and pu weight",1,0,2))
self.regressions={}
self.regressionVBF={}
for re in self.cfg_ana.regressionVBF :
print "Initialize regression ",re
regression_VBF = JetRegression(re["weight"],re["name"])
for i in re["vtypes"] :
self.regressionVBF[i] = regression_VBF
for re in self.cfg_ana.regressions :
print "Initialize regression ",re
regression = JetRegression(re["weight"],re["name"])
for i in re["vtypes"] :
self.regressions[i] = regression
blike=self.cfg_ana.VBFblikelihood
print "Initialize VBF blikelihood ", blike
self.blikelihood = VBFblikelihood(blike["weight"],blike["name"])
def beginLoop(self,setup):
super(VHbbAnalyzer,self).beginLoop(setup)
if "outputfile" in setup.services :
setup.services["outputfile"].file.cd()
self.inputCounter = ROOT.TH1F("Count","Count",1,0,2)
self.inputCounterWeighted = ROOT.TH1F("CountWeighted","Count with gen weight and pu weight",1,0,2)
self.inputCounterPosWeight = ROOT.TH1F("CountPosWeight","Count genWeight>0",1,0,2)
self.inputCounterNegWeight = ROOT.TH1F("CountNegWeight","Count genWeight<0",1,0,2)
self.regressions={}
self.regressionVBF={}
for re in self.cfg_ana.regressionVBF :
print "Initialize regression ",re
regression_VBF = JetRegression(re["weight"],re["name"])
for i in re["vtypes"] :
self.regressionVBF[i] = regression_VBF
for re in self.cfg_ana.regressions :
print "Initialize regression ",re
regression = JetRegression(re["weight"],re["name"])
for i in re["vtypes"] :
self.regressions[i] = regression
blike=self.cfg_ana.VBFblikelihood
print "Initialize VBF blikelihood ", blike
self.blikelihood = VBFblikelihood(blike["weight"],blike["name"])
class VHbbAnalyzer( Analyzer ):
'''Analyze VH events
'''
def declareHandles(self):
super(VHbbAnalyzer, self).declareHandles()
self.handles['rhoN'] = AutoHandle( 'fixedGridRhoFastjetCentralNeutral','double' )
self.handles['rhoCHPU'] = AutoHandle( 'fixedGridRhoFastjetCentralChargedPileUp','double' )
self.handles['rhoCentral'] = AutoHandle( 'fixedGridRhoFastjetCentral','double' )
if getattr(self.cfg_ana,"doSoftActivityVH", False) or getattr(self.cfg_ana,"doVBF", True):
self.handles['pfCands'] = AutoHandle( 'packedPFCandidates', 'std::vector<pat::PackedCandidate>' )
if self.cfg_comp.isMC:
self.handles['GenInfo'] = AutoHandle( ('generator','',''), 'GenEventInfoProduct' )
def addNewBTag(self,event):
newtags = self.handles['btag'].product()
for i in xrange(0,len(newtags)) :
for j in event.cleanJets :
if j.physObj == newtags.key(i).get() :
j.btagnew=newtags.value(i)
newtags = self.handles['btagcsv'].product()
for i in xrange(0,len(newtags)) :
for j in event.cleanJets :
if j.physObj == newtags.key(i).get() :
j.btagcsv=newtags.value(i)
def beginLoop(self,setup):
super(VHbbAnalyzer,self).beginLoop(setup)
if "outputfile" in setup.services :
setup.services["outputfile"].file.cd()
self.inputCounter = ROOT.TH1F("Count","Count",1,0,2)
self.inputCounterFullWeighted = ROOT.TH1F("CountFullWeighted","Count with gen weight and pu weight",1,0,2)
self.inputCounterWeighted = ROOT.TH1F("CountWeighted","Count with sign(gen weight) and pu weight",1,0,2)
self.inputCounterPosWeight = ROOT.TH1F("CountPosWeight","Count genWeight>0",1,0,2)
self.inputCounterNegWeight = ROOT.TH1F("CountNegWeight","Count genWeight<0",1,0,2)
#for LHE_scale in range(6):
setattr(self, "inputCounterWeightedLHEWeightScale", ROOT.TH1F("CountWeightedLHEWeightScale","Count with gen weight x LHE_weights_scale and pu weight",6,-0.5,5.5))
setattr(self, "inputCounterWeightedLHEWeightPdf", ROOT.TH1F("CountWeightedLHEWeightPdf","Count with gen weight x LHE_weights_pdf and pu weight",103,-0.5,102.5))
#for LHE_pdf in range(2):
# setattr(self, "inputCounterWeightedLHEWeightPdf_"+str(LHE_pdf), ROOT.TH1F("CountWeightedLHEWeightPdf_"+str(LHE_pdf),"Count with gen weight x LHE_weights_pdf["+str(LHE_pdf)+"] and pu weight",1,0,2))
self.regressions={}
self.regressionVBF={}
for re in self.cfg_ana.regressionVBF :
print "Initialize regression ",re
regression_VBF = JetRegression(re["weight"],re["name"])
for i in re["vtypes"] :
self.regressionVBF[i] = regression_VBF
for re in self.cfg_ana.regressions :
print "Initialize regression ",re
regression = JetRegression(re["weight"],re["name"])
for i in re["vtypes"] :
self.regressions[i] = regression
blike=self.cfg_ana.VBFblikelihood
print "Initialize VBF blikelihood ", blike
self.blikelihood = VBFblikelihood(blike["weight"],blike["name"])
def doVBF(self,event) :
event.jetsForVBF = [x for x in event.cleanJetsAll if self.cfg_ana.higgsJetsPreSelectionVBF(x) ]
if event.Vtype in self.regressionVBF :
for analysis in ["","corrJECUp", "corrJECDown", "corrJERUp", "corrJERDown"]:
self.regressionVBF[event.Vtype].evaluateRegression(event,"pt_regVBF", analysis)
#compute only for events passing VBF selection
if len(event.jetsForVBF) < 4 or event.jetsForVBF[0] < 70 or event.jetsForVBF[1] < 55 or event.jetsForVBF[2] < 35 or event.jetsForVBF[3] < 20 :
return
event.jetsForVBF.sort(key=lambda x:x.pt(),reverse=True)
map(lambda x :x.qgl(),event.jetsForVBF[:6])
event.jetsForVBF=event.jetsForVBF[:4]
#compute QGL here for VBF jets if passing VBF pre-selection
event.bJetsForVBF=sorted(event.jetsForVBF,key = lambda jet : jet.btag(getattr(self.cfg_ana,"btagDiscriminator",'pfCombinedInclusiveSecondaryVertexV2BJetTags')), reverse=True)[:2]
j1=event.bJetsForVBF[0]
j2=event.bJetsForVBF[1]
#print "VBF"
#print event.selectedElectrons,event.selectedMuons,
event.softActivityJets=self.softActivity(event,j1,j2,event.jetsForVBF+event.selectedElectrons+event.selectedMuons)
def doSoftActivityVH(self,event) :
if not len(event.jetsForHiggs) >= 2 :
return
j1=event.hJetsCSV[0]
j2=event.hJetsCSV[1]
#print "VH"
excludedJets=event.hJetsCSV+event.selectedElectrons+event.selectedMuons
if event.isrJetVH >= 0 :
excludedJets+=[event.cleanJetsAll[event.isrJetVH]]
event.softActivityVHJets=[x for x in self.softActivity(event,j1,j2,excludedJets,-1000) if x.pt() > 2.0 ]
def addPullVector(self,event) :
""" Add color flow properties to the jetsForHiggs jet collections.
It calculates the pull angle for each jet of jetsForHiggs jet collection.
It stores the phi angle and the magnitude of the pull angle as property of a jet.
Contact: [email protected]
Args: event: the event.
Returns: nothing.
"""
# the pull vector (t_vect) is a 2D vector living in a (Phi,Rapidity) plane.
t_vect=ROOT.TVector2()
# pt threashold for PFCandidates to be considered in the pull angle calculation
min_pfCands_pt = 0.3
pfCandsForColorFlow=ROOT.std.vector(ROOT.reco.Particle.LorentzVector)()
for jet in event.jetsForHiggs:
# The color flow is calculated using only charged PFCandidates over a certain threashold of pt
pfCands=[jet.sourceCandidatePtr(i) for i in range(0,jet.numberOfSourceCandidatePtrs())]
goodPFCands = [pfCand for pfCand in pfCands if pfCand.charge() != 0 and pfCand.pt() > min_pfCands_pt ]
# put goodPFCandidates in a vector of LorentzVector and pass it to the C++ wrapper (colorFlow). The C++ wrapper improves the speed of the algorithm by about a fator 4.
map(lambda x:pfCandsForColorFlow.push_back(x.p4()), goodPFCands)
colorFlow = ROOT.heppy.ColorFlow(pfCandsForColorFlow)
t_vect=colorFlow.get_pull_vector()
# add to the jets properties the pull vector phi angle and magnitude.
jet.pullVectorPhi=t_vect.Phi()
jet.pullVectorMag=t_vect.Mod() # if there are less than 2 charged of candidates then the magnitude will be 0.
def softActivity(self,event,j1,j2,excludedJets,dR0=0.4) :
if not hasattr(event,"pfCands") :
event.pfCands = list(self.handles['pfCands'].product())
inputs=ROOT.std.vector(ROOT.heppy.ReclusterJets.LorentzVector)()
used=[]
for j in excludedJets :
for i in xrange(0,j.numberOfSourceCandidatePtrs()) :
if j.sourceCandidatePtr(i).isAvailable() :
used.append(j.sourceCandidatePtr(i))
#get the pointed objects
used = [x.get() for x in used]
remainingPF = [x for x in event.pfCands if x.charge() != 0 and abs(x.eta()) < 2.5 and x.pt() > 0.3 and x.fromPV() >= 2 and abs(x.dz()) < 0.2 and x not in used]
dRbb = deltaR(j1.eta(),j1.phi(),j2.eta(),j2.phi())
map(lambda x:inputs.push_back(x.p4()), remainingPF)
softActivity=ROOT.heppy.FastSoftActivity(inputs,-1,0.4,j1.p4(),j2.p4(),dRbb+2*dR0)
jets=softActivity.getGrouping(1)
softActivityJets = [ ROOT.reco.Particle.LorentzVector(p4) for p4 in jets ]
softActivityJets.sort(key=lambda x:x.pt(), reverse=True)
return softActivityJets
def searchISRforVH(self,event):
p4VH=event.HCSV+event.V
if p4VH.pt() > 30 :
phi=pi+p4VH.phi()
matchedJets=[x for x in event.cleanJetsAll if abs(deltaPhi(phi,x.phi())) < 0.4 and x.puJetId() > 0 and x.jetID('POG_PFID_Loose') and x not in event.hJetsCSV and x not in event.hJetsaddJetsdR08]
if len(matchedJets) > 0 :
event.isrJetVH=event.cleanJetsAll.index(sorted(matchedJets, key=lambda x:x.pt(),reverse=True)[0])
def makeJets(self,event,b):
inputs=ROOT.std.vector(ROOT.heppy.ReclusterJets.LorentzVector)()
event.pfCands = list(self.handles['pfCands'].product())
# print "original jets pt,eta,phi \t\t",map(lambda x:"%s,%s,%s --"%(x.pt(),x.eta(),x.phi()),event.cleanJets)
# print "BquarksFromH pt,eta,phi \t\t",map(lambda x:"%s,%s,%s -- "%(x.pt(),x.eta(),x.phi()),event.genbquarksFromH)
# print "Inv mass",(event.genbquarksFromH[0].p4()+event.genbquarksFromH[1].p4()).M()
# print "pt from sum ",(event.genbquarksFromH[0].p4()+event.genbquarksFromH[1].p4()).Pt()
# print "pt from h",event.genHiggsBoson[0].pt()
copyhb=map(lambda x: deepcopy(x.p4()),event.genbquarksFromH)
map(lambda x: Boost(x,b),copyhb)
# print "BquarksFromH(boost) pt,eta,phi,p \t\t",map(lambda x:"%s,%s,%s,%s -- "%(x.pt(),x.eta(),x.phi(),x.P()),copyhb)
# print "Inv mass (boost)",(copyhb[0]+copyhb[1]).M()
# print "pt (boost)",(copyhb[0]+copyhb[1]).Pt()
# print "boost",b.X(),b.Y(),b.Z(),"phi",b.Phi()
for pf in event.pfCands :
if pf.fromPV() or pf.charge()==0 :
p4copy=ROOT.heppy.ReclusterJets.LorentzVector(pf.p4())
bst=Boost(p4copy,b)
# if bst.pt() > 20 :
# print " candidate orig,boost",pf.pt(),bst.pt(),pf.phi(),bst.phi()
inputs.push_back(bst)
clusterizer=ROOT.heppy.ReclusterJets(inputs,-1,0.4)
jets = clusterizer.getGrouping(10)
#event.jee = list(self.handles['jee'].product())
# print "Boosted jets:",
for j in list(jets):
oldpt=j.pt()
oldeta=j.eta()
oldphi=j.phi()
Boost(j,-b)
# if j.pt() > 20 :
# print oldpt,oldeta,oldphi," -> ",j.pt(),j.eta(),j.phi(),"|",
# print " "
def doFakeMET(self,event):
#fake MET from Zmumu
event.fakeMET = ROOT.reco.Particle.LorentzVector(0.,0.,0.,0.)
event.fakeMET.sumet = 0
if event.Vtype == 0 :
event.fakeMET=event.met.p4() + event.V
event.fakeMET.sumet = event.met.sumEt() - event.V.pt()
def doHtMhtJets30(self,event):
## with Central Jets
objects30 = [ j for j in event.cleanJets if j.pt() > 30 ] + event.selectedLeptons
event.htJet30 = sum([x.pt() for x in objects30])
event.mhtJet30vec = ROOT.reco.Particle.LorentzVector(-1.*(sum([x.px() for x in objects30])) , -1.*(sum([x.py() for x in objects30])), 0, 0 )
event.mhtJet30 = event.mhtJet30vec.pt()
event.mhtPhiJet30 = event.mhtJet30vec.phi()
def doHiggsHighCSV(self,event) :
#leading csv interpretation
if ( len(event.jetsForHiggs) >= 2):
event.hJetsCSV=sorted(event.jetsForHiggs,key = lambda jet : jet.btag(getattr(self.cfg_ana,"btagDiscriminator",'pfCombinedInclusiveSecondaryVertexV2BJetTags')), reverse=True)[0:2]
event.aJetsCSV = [x for x in event.cleanJets if x not in event.hJetsCSV]
event.hjidxCSV=[event.cleanJetsAll.index(x) for x in event.hJetsCSV ]
event.ajidxCSV=[event.cleanJetsAll.index(x) for x in event.aJetsCSV ]
event.aJetsCSV+=event.cleanJetsFwd
event.HCSV = event.hJetsCSV[0].p4()+event.hJetsCSV[1].p4()
def doHiggsHighPt(self,event) :
#highest pair interpretations
if ( len(event.jetsForHiggs) >= 2):
event.hJets=list(max(itertools.combinations(event.jetsForHiggs,2), key = lambda x : (x[0].p4()+x[1].p4()).pt() ))
event.aJets = [x for x in event.cleanJets if x not in event.hJets]
event.hjidx=[event.cleanJetsAll.index(x) for x in event.hJets ]
event.ajidx=[event.cleanJetsAll.index(x) for x in event.aJets ]
event.aJets+=event.cleanJetsFwd
hJetsByCSV = sorted(event.hJets , key = lambda jet : jet.btag(getattr(self.cfg_ana,"btagDiscriminator",'pfCombinedInclusiveSecondaryVertexV2BJetTags')), reverse=True)
event.hjidxDiJetPtByCSV = [event.cleanJetsAll.index(x) for x in hJetsByCSV]
event.H = event.hJets[0].p4()+event.hJets[1].p4()
def doHiggsAddJetsdR08(self,event) :
if ( len(event.jetsForHiggs) >= 2):
event.hJetsaddJetsdR08 = [x for x in event.hJetsCSV]
event.dRaddJetsdR08 = []
event.aJetsaddJetsdR08 = [x for x in event.aJetsCSV]
event.hjidxaddJetsdR08 = [x for x in event.hjidxCSV]
event.ajidxaddJetsdR08 = [x for x in event.ajidxCSV]
#multiple jets interpretations, for central jets closest to dR<0.8 from higgs jets
jetsForHiggsAddJetsdR08 = [x for x in event.cleanJetsAll if (x.pt()>15 and abs(x.eta())<3.0 and x.puJetId() > 0 and x.jetID('POG_PFID_Loose') ) ]
if (len(jetsForHiggsAddJetsdR08) > 2):
addJetsForHiggs = [x for x in jetsForHiggsAddJetsdR08 if ( x not in event.hJetsCSV and min(deltaR( x.eta(), x.phi(), event.hJetsCSV[0].eta(), event.hJetsCSV[0].phi()),deltaR( x.eta(), x.phi(), event.hJetsCSV[1].eta(), event.hJetsCSV[1].phi()))<0.8 ) ]
for x in addJetsForHiggs:
event.hJetsaddJetsdR08.append(x)
event.dRaddJetsdR08.append( min(deltaR( x.eta(), x.phi(), event.hJetsCSV[0].eta(), event.hJetsCSV[0].phi()),deltaR( x.eta(), x.phi(), event.hJetsCSV[1].eta(), event.hJetsCSV[1].phi() )) )
event.hjidxaddJetsdR08=[event.cleanJetsAll.index(x) for x in event.hJetsaddJetsdR08 ]
event.aJetsaddJetsdR08 = [x for x in event.cleanJets if x not in event.hJetsaddJetsdR08]
event.aJetsaddJetsdR08+=event.cleanJetsFwd
event.ajidxaddJetsdR08=[event.cleanJetsAll.index(x) for x in event.aJetsaddJetsdR08 ]
event.HaddJetsdR08 = sum(map(lambda x:x.p4(), event.hJetsaddJetsdR08), ROOT.reco.Particle.LorentzVector(0.,0.,0.,0.))
def doVHRegression(self, event):
if (len(event.jetsForHiggs) >=2 ):
# ""=nominal, the other correspond to jet energies up/down for JEC and JER
for analysis in ["","corrJECUp", "corrJECDown", "corrJERUp", "corrJERDown"]:
self.regressions[event.Vtype].evaluateRegression(event, "pt_reg", analysis)
hJetCSV_reg0 =ROOT.reco.Particle.LorentzVector( event.hJetsCSV[0].p4())
hJetCSV_reg1 =ROOT.reco.Particle.LorentzVector( event.hJetsCSV[1].p4())
hJetCSV_reg0*=getattr(event.hJetsCSV[0],"pt_reg"+analysis)/event.hJetsCSV[0].pt()
hJetCSV_reg1*=getattr(event.hJetsCSV[1],"pt_reg"+analysis)/event.hJetsCSV[1].pt()
setattr(event,"HCSV_reg"+("_"+analysis if analysis!="" else ""), hJetCSV_reg0+hJetCSV_reg1)
hJet_reg0=ROOT.reco.Particle.LorentzVector(event.hJets[0].p4())
hJet_reg1=ROOT.reco.Particle.LorentzVector(event.hJets[1].p4())
hJet_reg0*=getattr(event.hJets[0],"pt_reg"+analysis)/event.hJets[0].pt()
hJet_reg1*=getattr(event.hJets[1],"pt_reg"+analysis)/event.hJets[1].pt()
setattr(event,"H_reg"+("_"+analysis if analysis!="" else ""), hJet_reg0+hJet_reg1)
#print "H_reg"+("_"+analysis if analysis!="" else ""), getattr(event, "H_reg"+("_"+analysis if analysis!="" else "")).Pt()
else:
for analysis in ["","corrJECUp", "corrJECDown", "corrJERUp", "corrJERDown"]:
setattr(event,"HCSV_reg"+("_"+analysis if analysis!="" else ""), ROOT.reco.Particle.LorentzVector() )
setattr(event,"H_reg"+("_"+analysis if analysis!="" else ""), ROOT.reco.Particle.LorentzVector() )
def doVBFblikelihood(self, event):
self.blikelihood.evaluateBlikelihood(event)
def classifyMCEvent(self,event):
if self.cfg_comp.isMC:
event.VtypeSim = -1
if len(event.genvbosons) == 1:
# ZtoLL events, same flavour leptons
if event.genvbosons[0].pdgId()==23 and event.genvbosons[0].numberOfDaughters()>1 and abs(event.genvbosons[0].daughter(0).pdgId()) == abs(event.genvbosons[0].daughter(1).pdgId()):
if abs(event.genvbosons[0].daughter(0).pdgId()) == 11:
#Ztoee
event.VtypeSim = 1
if abs(event.genvbosons[0].daughter(0).pdgId()) == 13:
#ZtoMuMu
event.VtypeSim = 0
if abs(event.genvbosons[0].daughter(0).pdgId()) == 15:
#ZtoTauTau
event.VtypeSim = 5
if abs(event.genvbosons[0].daughter(0).pdgId()) in [12,14,16]:
event.VtypeSim = 4
#WtoLNu events
if abs(event.genvbosons[0].pdgId())==24 and event.genvbosons[0].numberOfDaughters()==2:
if abs(event.genvbosons[0].daughter(0).pdgId()) == 11 and abs(event.genvbosons[0].daughter(1).pdgId()) == 12:
#WtoEleNu_e
event.VtypeSim = 3
if abs(event.genvbosons[0].daughter(0).pdgId()) == 13 and abs(event.genvbosons[0].daughter(1).pdgId()) == 14:
#WtoMuNu_mu
event.VtypeSim = 2
#to be added: WtoTauNu
if len(event.genvbosons)>1:
#print 'more than one W/Zbosons?'
event.VtypeSim = -2
# if event.VtypeSim == -1:
# print '===================================='
# print ' --------- Debug VtypeSim -1 --------'
# print '# genVbosons: ',len(event.genvbosons), '| #daughters ', event.genvbosons[0].numberOfDaughters()
# for i in xrange (0, event.genvbosons[0].numberOfDaughters() ) :
# print 'daughter ',i ,'| pdgId', event.genvbosons[0].daughter(i).pdgId()
def classifyEvent(self,event):
#assign events to analysis (Vtype)
#enum CandidateType{Zmumu, Zee, Wmun, Wen, Znn, Zemu, Ztaumu, Ztaue, Wtaun, Ztautau, Zbb, UNKNOWN};
event.Vtype=-1
nLep=len(event.selectedLeptons)
event.vLeptons=[]
#WH requires exactly one selected lepton
wElectrons=[x for x in event.selectedElectrons if self.cfg_ana.wEleSelection(x) ]
wMuons=[x for x in event.selectedMuons if self.cfg_ana.wMuSelection(x) ]
zElectrons=[x for x in event.selectedElectrons if self.cfg_ana.zEleSelection(x) ]
zMuons=[x for x in event.selectedMuons if self.cfg_ana.zMuSelection(x) ]
zMuons.sort(key=lambda x:x.pt(),reverse=True)
zElectrons.sort(key=lambda x:x.pt(),reverse=True)
if len(zMuons) >= 2 :
# print zMuons[0].pt()
if zMuons[0].pt() > self.cfg_ana.zLeadingMuPt :
for i in xrange(1,len(zMuons)):
if zMuons[0].charge()*zMuons[i].charge()<0 :
event.Vtype = 0
event.vLeptons =[zMuons[0],zMuons[i]]
break
elif len(zElectrons) >= 2 :
# for i in zElectrons[0].electronIDs() :
# print i.first,i.second
if zElectrons[0].pt() > self.cfg_ana.zLeadingElePt :
for i in xrange(1,len(zElectrons)):
if zElectrons[0].charge()*zElectrons[i].charge()<0 :
event.Vtype = 1
event.vLeptons =[zElectrons[0],zElectrons[i]]
break
elif len(wElectrons) + len(wMuons) == 1:
if abs(event.selectedLeptons[0].pdgId())==13 :
event.Vtype = 2
event.vLeptons =wMuons
if abs(event.selectedLeptons[0].pdgId())==11 :
event.Vtype = 3
event.vLeptons =wElectrons
elif len(zElectrons) + len(zMuons) > 0 :
event.Vtype = 5 #there are some loose (Z selection) leptons but not matching the W/Z above requirements
else :
event.Vtype = 4 #no leptons at all, apply MET cut
#apply MET cut
if event.met.pt() < 80 :
event.Vtype = -1
event.V=sum(map(lambda x:x.p4(), event.vLeptons),ROOT.reco.Particle.LorentzVector(0.,0.,0.,0.))
if event.Vtype > 1 :
event.V+=ROOT.reco.Particle.LorentzVector(event.met.p4().x(),event.met.p4().y(),0,event.met.p4().pt())
if event.V.Et() > event.V.pt() :
event.V.goodMt = sqrt(event.V.Et()**2-event.V.pt()**2)
else :
event.V.goodMt = -sqrt(-event.V.Et()**2+event.V.pt()**2)
event.aLeptons = [x for x in event.inclusiveLeptons if x not in event.vLeptons]
return True
def fillTauIndices(self,event) :
for j in event.cleanJetsAll :
j.tauIdxs = [event.inclusiveTaus.index(x) for x in j.taus if j.taus in event.inclusiveTaus]
for t in event.inclusiveTaus :
dRmin = 1.e+3
t.jetIdx = -1
for jIdx, j in enumerate(event.cleanJetsAll) :
dR = None
if t.isPFTau():
dR = deltaR(t.p4Jet().eta(),t.p4Jet().phi(),j.eta(),j.phi())
else:
dR = deltaR(t.pfEssential().p4CorrJet_.eta(),t.pfEssential().p4CorrJet_.phi(),j.eta(),j.phi())
if dR < 0.3 and dR < dRmin :
t.jetIdx = jIdx
dRmin = dR
def additionalRhos(self,event) :
event.rhoN= self.handles['rhoN'].product()
event.rhoCHPU= self.handles['rhoCHPU'].product()
event.rhoCentral= self.handles['rhoCentral'].product()
def initOutputs (self,event) :
event.hJets = []
event.aJets = []
event.hjidx = []
event.ajidx = []
event.hJetsCSV = []
event.aJetsCSV = []
event.hjidxCSV = []
event.ajidxCSV = []
event.hJetsaddJetsdR08 = []
event.dRaddJetsdR08 = []
event.aJetsaddJetsdR08 = []
event.hjidxaddJetsdR08 = []
event.ajidxaddJetsdR08 = []
event.aLeptons = []
event.vLeptons = []
event.isrJetVH=-1
event.H = ROOT.reco.Particle.LorentzVector(0.,0.,0.,0.)
event.HCSV = ROOT.reco.Particle.LorentzVector(0.,0.,0.,0.)
event.HaddJetsdR08 = ROOT.reco.Particle.LorentzVector(0.,0.,0.,0.)
event.H_reg = ROOT.reco.Particle.LorentzVector(0.,0.,0.,0.)
event.HCSV_reg = ROOT.reco.Particle.LorentzVector(0.,0.,0.,0.)
for analysis in ["","corrJECUp", "corrJECDown", "corrJERUp", "corrJERDown"]:
setattr(event,"HCSV_reg"+("_"+analysis if analysis!="" else ""), ROOT.reco.Particle.LorentzVector() )
setattr(event,"H_reg"+("_"+analysis if analysis!="" else ""), ROOT.reco.Particle.LorentzVector() )
event.V = ROOT.reco.Particle.LorentzVector(0.,0.,0.,0.)
event.minDr3=-1
event.V.goodMt=0
event.hjidxDiJetPtByCSV = []
event.softActivityJets=[]
event.softActivityVHJets=[]
event.rhoN= -1
event.rhoCHPU= -1
event.rhoCentral= -1
def process(self, event):
#print "Event number",event.iEv
self.readCollections( event.input )
self.inputCounter.Fill(1)
if self.cfg_comp.isMC:
genWeight = self.handles['GenInfo'].product().weight()
self.inputCounterWeighted.Fill(1,copysign(1.0,genWeight)*event.puWeight)
self.inputCounterFullWeighted.Fill(1,genWeight*event.puWeight)
for LHE_scale in range(len(event.LHE_weights_scale)):
getattr(self, "inputCounterWeightedLHEWeightScale").Fill(LHE_scale,copysign(1.0, genWeight)*event.puWeight*(event.LHE_weights_scale[LHE_scale]).wgt)
for LHE_pdf in range(len(event.LHE_weights_pdf)):
getattr(self, "inputCounterWeightedLHEWeightPdf").Fill(LHE_pdf,copysign(1.0, genWeight)*event.puWeight*(event.LHE_weights_pdf[LHE_pdf]).wgt)
if genWeight > 0:
self.inputCounterPosWeight.Fill(1)
elif genWeight < 0:
self.inputCounterNegWeight.Fill(1)
self.initOutputs(event)
# event.pfCands = self.handles['pfCands'].product()
# met = event.met
#self.addNewBTag(event)
self.classifyMCEvent(event)
self.classifyEvent(event)
self.doFakeMET(event)
self.doHtMhtJets30(event)
self.fillTauIndices(event)
#Add CSV ranking
csvSortedJets=sorted(event.cleanJetsAll, key = lambda jet : jet.btag(getattr(self.cfg_ana,"btagDiscriminator",'pfCombinedInclusiveSecondaryVertexV2BJetTags')),reverse=True)
for j in event.cleanJetsAll:
j.btagIdx=csvSortedJets.index(j)
for j in event.discardedJets:
j.btagIdx=-1
#substructure threshold, make configurable
ssThreshold = 200.
# filter events with less than 2 jets with pt 20
event.jetsForHiggs = [x for x in event.cleanJets if self.cfg_ana.higgsJetsPreSelection(x) ]
if not ( len(event.jetsForHiggs) >= 2 or (len(event.cleanJets) == 1 and event.cleanJets[0].pt() > ssThreshold ) ) :
return self.cfg_ana.passall
if event.Vtype < 0 and not ( sum(x.pt() > 30 for x in event.jetsForHiggs) >= 4 or sum(x.pt() for x in event.jetsForHiggs[:4]) > 160 ):
return self.cfg_ana.passall
map(lambda x :x.qgl(),event.jetsForHiggs[:6])
map(lambda x :x.qgl(),(x for x in event.jetsForHiggs if x.pt() > 30) )
self.doHiggsHighCSV(event)
self.doHiggsHighPt(event)
self.doHiggsAddJetsdR08(event)
self.searchISRforVH(event)
self.doVHRegression(event)
self.doVBFblikelihood(event)
self.fillTauIndices(event)
self.addPullVector(event)
if getattr(self.cfg_ana,"doVBF", True) :
self.doVBF(event)
if getattr(self.cfg_ana,"doSoftActivityVH", False) :
self.doSoftActivityVH(event)
# event.jee = list(self.handles['jee'].product())
#for j in list(jets)[0:3]:
# print j.pt(),
#print " "
#to implement
#if some threshold:
# self.computeSubStructuresStuff()
#self.doIVFHiggs()
#self.computePullAngle()
#
#perhaps in different producers:
# LHE weights
# Trigger weights
# gen level VH specific info
# add soft jet info to jets
# PU weights
# SIM B hadrons information
# MET corrections (in MET analyzer)
# trigger flags
# Hbalance = ROOT.TLorentzVector()
# Hbalance.SetPtEtaPhiM(event.V.pt(),0,event.V.phi(),125.)
# hh=event.genHiggsBoson[0]
# Hbalance2 = ROOT.TLorentzVector()
# Hbalance2.SetPtEtaPhiM(hh.pt(),hh.eta(),hh.phi(),125.)
# print "Hbalance pt,e,phi",Hbalance.Pt(),Hbalance.E(),Hbalance.Phi()
# print "gen H pt,e,phi",hh.pt(),hh.p4().E(),hh.phi()
# print "Hbalance2 pt,e,phi",Hbalance2.Pt(),Hbalance2.E(),Hbalance2.Phi()
# print "boostVector:",Hbalance.BoostVector().X(),Hbalance.BoostVector().Y(),Hbalance.BoostVector().Z()
# print "boostVector2:",Hbalance2.BoostVector().X(),Hbalance2.BoostVector().Y(),Hbalance2.BoostVector().Z()
#print "Unboosted Hbalance",Boost(Hbalance,-Hbalance.BoostVector()).Pt()
# print "hbalance boost vector",Hbalance.BoostVector()
# hhh=deepcopy(Hbalance2)
# Boost(hhh,-Hbalance2.BoostVector())
# print "unboosted pt",hhh.Pt()
# self.makeJets(event,-Hbalance2.BoostVector())
return True
class VHbbAnalyzer(Analyzer):
'''Analyze VH events
'''
def declareHandles(self):
super(VHbbAnalyzer, self).declareHandles()
self.handles['rhoN'] = AutoHandle('fixedGridRhoFastjetCentralNeutral',
'double')
self.handles['rhoCHPU'] = AutoHandle(
'fixedGridRhoFastjetCentralChargedPileUp', 'double')
self.handles['rhoCentral'] = AutoHandle('fixedGridRhoFastjetCentral',
'double')
if getattr(self.cfg_ana, "doSoftActivityVH", False) or getattr(
self.cfg_ana, "doVBF", True):
self.handles['pfCands'] = AutoHandle(
'packedPFCandidates', 'std::vector<pat::PackedCandidate>')
if self.cfg_comp.isMC:
self.handles['GenInfo'] = AutoHandle(('generator', '', ''),
'GenEventInfoProduct')
def addNewBTag(self, event):
newtags = self.handles['btag'].product()
for i in xrange(0, len(newtags)):
for j in event.cleanJets:
if j.physObj == newtags.key(i).get():
j.btagnew = newtags.value(i)
newtags = self.handles['btagcsv'].product()
for i in xrange(0, len(newtags)):
for j in event.cleanJets:
if j.physObj == newtags.key(i).get():
j.btagcsv = newtags.value(i)
def beginLoop(self, setup):
super(VHbbAnalyzer, self).beginLoop(setup)
if "outputfile" in setup.services:
setup.services["outputfile"].file.cd()
self.inputCounter = ROOT.TH1F("Count", "Count", 1, 0, 2)
self.inputCounterFullWeighted = ROOT.TH1F(
"CountFullWeighted", "Count with gen weight and pu weight", 1,
0, 2)
self.inputCounterWeighted = ROOT.TH1F(
"CountWeighted", "Count with sign(gen weight) and pu weight",
1, 0, 2)
self.inputCounterPosWeight = ROOT.TH1F("CountPosWeight",
"Count genWeight>0", 1, 0,
2)
self.inputCounterNegWeight = ROOT.TH1F("CountNegWeight",
"Count genWeight<0", 1, 0,
2)
#for LHE_scale in range(6):
setattr(
self, "inputCounterWeightedLHEWeightScale",
ROOT.TH1F(
"CountWeightedLHEWeightScale",
"Count with gen weight x LHE_weights_scale and pu weight",
6, -0.5, 5.5))
setattr(
self, "inputCounterWeightedLHEWeightPdf",
ROOT.TH1F(
"CountWeightedLHEWeightPdf",
"Count with gen weight x LHE_weights_pdf and pu weight",
103, -0.5, 102.5))
#for LHE_pdf in range(2):
# setattr(self, "inputCounterWeightedLHEWeightPdf_"+str(LHE_pdf), ROOT.TH1F("CountWeightedLHEWeightPdf_"+str(LHE_pdf),"Count with gen weight x LHE_weights_pdf["+str(LHE_pdf)+"] and pu weight",1,0,2))
self.regressions = {}
self.regressionVBF = {}
for re in self.cfg_ana.regressionVBF:
print "Initialize regression ", re
regression_VBF = JetRegression(re["weight"], re["name"])
for i in re["vtypes"]:
self.regressionVBF[i] = regression_VBF
for re in self.cfg_ana.regressions:
print "Initialize regression ", re
regression = JetRegression(re["weight"], re["name"])
for i in re["vtypes"]:
self.regressions[i] = regression
blike = self.cfg_ana.VBFblikelihood
print "Initialize VBF blikelihood ", blike
self.blikelihood = VBFblikelihood(blike["weight"], blike["name"])
def doVBF(self, event):
event.jetsForVBF = [
x for x in event.cleanJetsAll
if self.cfg_ana.higgsJetsPreSelectionVBF(x)
]
if event.Vtype in self.regressionVBF:
for analysis in [
"", "corrJECUp", "corrJECDown", "corrJERUp", "corrJERDown"
]:
self.regressionVBF[event.Vtype].evaluateRegression(
event, "pt_regVBF", analysis)
#compute only for events passing VBF selection
if len(event.jetsForVBF) < 4 or event.jetsForVBF[
0] < 70 or event.jetsForVBF[1] < 55 or event.jetsForVBF[
2] < 35 or event.jetsForVBF[3] < 20:
return
event.jetsForVBF.sort(key=lambda x: x.pt(), reverse=True)
map(lambda x: x.qgl(), event.jetsForVBF[:6])
event.jetsForVBF = event.jetsForVBF[:4]
#compute QGL here for VBF jets if passing VBF pre-selection
event.bJetsForVBF = sorted(
event.jetsForVBF,
key=lambda jet: jet.btag(
getattr(self.cfg_ana, "btagDiscriminator",
'pfCombinedInclusiveSecondaryVertexV2BJetTags')),
reverse=True)[:2]
j1 = event.bJetsForVBF[0]
j2 = event.bJetsForVBF[1]
#print "VBF"
#print event.selectedElectrons,event.selectedMuons,
event.softActivityJets = self.softActivity(
event, j1, j2,
event.jetsForVBF + event.selectedElectrons + event.selectedMuons)
def doSoftActivityVH(self, event):
if not len(event.jetsForHiggs) >= 2:
return
j1 = event.hJetsCSV[0]
j2 = event.hJetsCSV[1]
#print "VH"
excludedJets = event.hJetsCSV + event.selectedElectrons + event.selectedMuons
if event.isrJetVH >= 0:
excludedJets += [event.cleanJetsAll[event.isrJetVH]]
event.softActivityVHJets = [
x for x in self.softActivity(event, j1, j2, excludedJets, -1000)
if x.pt() > 2.0
]
def addPullVector(self, event):
""" Add color flow properties to the jetsForHiggs jet collections.
It calculates the pull angle for each jet of jetsForHiggs jet collection.
It stores the phi angle and the magnitude of the pull angle as property of a jet.
Contact: [email protected]
Args: event: the event.
Returns: nothing.
"""
# the pull vector (t_vect) is a 2D vector living in a (Phi,Rapidity) plane.
t_vect = ROOT.TVector2()
# pt threashold for PFCandidates to be considered in the pull angle calculation
min_pfCands_pt = 0.3
pfCandsForColorFlow = ROOT.std.vector(
ROOT.reco.Particle.LorentzVector)()
for jet in event.jetsForHiggs:
# The color flow is calculated using only charged PFCandidates over a certain threashold of pt
pfCands = [
jet.sourceCandidatePtr(i)
for i in range(0, jet.numberOfSourceCandidatePtrs())
]
goodPFCands = [
pfCand for pfCand in pfCands
if pfCand.charge() != 0 and pfCand.pt() > min_pfCands_pt
]
# put goodPFCandidates in a vector of LorentzVector and pass it to the C++ wrapper (colorFlow). The C++ wrapper improves the speed of the algorithm by about a fator 4.
map(lambda x: pfCandsForColorFlow.push_back(x.p4()), goodPFCands)
colorFlow = ROOT.heppy.ColorFlow(pfCandsForColorFlow)
t_vect = colorFlow.get_pull_vector()
# add to the jets properties the pull vector phi angle and magnitude.
jet.pullVectorPhi = t_vect.Phi()
jet.pullVectorMag = t_vect.Mod(
) # if there are less than 2 charged of candidates then the magnitude will be 0.
def softActivity(self, event, j1, j2, excludedJets, dR0=0.4):
if not hasattr(event, "pfCands"):
event.pfCands = list(self.handles['pfCands'].product())
inputs = ROOT.std.vector(ROOT.heppy.ReclusterJets.LorentzVector)()
used = []
for j in excludedJets:
for i in xrange(0, j.numberOfSourceCandidatePtrs()):
if j.sourceCandidatePtr(i).isAvailable():
used.append(j.sourceCandidatePtr(i))
#get the pointed objects
used = [x.get() for x in used]
remainingPF = [
x for x in event.pfCands
if x.charge() != 0 and abs(x.eta()) < 2.5 and x.pt() > 0.3 and
(x.fromPV() >= 2 or x.pvAssociationQuality() >= x.CompatibilityDz)
and x not in used
]
dRbb = deltaR(j1.eta(), j1.phi(), j2.eta(), j2.phi())
map(lambda x: inputs.push_back(x.p4()), remainingPF)
softActivity = ROOT.heppy.FastSoftActivity(inputs, -1, 0.4, j1.p4(),
j2.p4(), dRbb + 2 * dR0)
jets = softActivity.getGrouping(1)
softActivityJets = [
ROOT.reco.Particle.LorentzVector(p4) for p4 in jets
]
softActivityJets.sort(key=lambda x: x.pt(), reverse=True)
return softActivityJets
def searchISRforVH(self, event):
p4VH = event.HCSV + event.V
if p4VH.pt() > 30:
phi = pi + p4VH.phi()
matchedJets = [
x for x in event.cleanJetsAll
if abs(deltaPhi(phi, x.phi())) < 0.4 and x.puJetId() > 0
and x.jetID('POG_PFID_Loose') and x not in event.hJetsCSV
and x not in event.hJetsaddJetsdR08
]
if len(matchedJets) > 0:
event.isrJetVH = event.cleanJetsAll.index(
sorted(matchedJets, key=lambda x: x.pt(), reverse=True)[0])
def makeJets(self, event, b):
inputs = ROOT.std.vector(ROOT.heppy.ReclusterJets.LorentzVector)()
event.pfCands = list(self.handles['pfCands'].product())
# print "original jets pt,eta,phi \t\t",map(lambda x:"%s,%s,%s --"%(x.pt(),x.eta(),x.phi()),event.cleanJets)
# print "BquarksFromH pt,eta,phi \t\t",map(lambda x:"%s,%s,%s -- "%(x.pt(),x.eta(),x.phi()),event.genbquarksFromH)
# print "Inv mass",(event.genbquarksFromH[0].p4()+event.genbquarksFromH[1].p4()).M()
# print "pt from sum ",(event.genbquarksFromH[0].p4()+event.genbquarksFromH[1].p4()).Pt()
# print "pt from h",event.genHiggsBoson[0].pt()
copyhb = map(lambda x: deepcopy(x.p4()), event.genbquarksFromH)
map(lambda x: Boost(x, b), copyhb)
# print "BquarksFromH(boost) pt,eta,phi,p \t\t",map(lambda x:"%s,%s,%s,%s -- "%(x.pt(),x.eta(),x.phi(),x.P()),copyhb)
# print "Inv mass (boost)",(copyhb[0]+copyhb[1]).M()
# print "pt (boost)",(copyhb[0]+copyhb[1]).Pt()
# print "boost",b.X(),b.Y(),b.Z(),"phi",b.Phi()
for pf in event.pfCands:
if pf.fromPV() or pf.charge() == 0:
p4copy = ROOT.heppy.ReclusterJets.LorentzVector(pf.p4())
bst = Boost(p4copy, b)
# if bst.pt() > 20 :
# print " candidate orig,boost",pf.pt(),bst.pt(),pf.phi(),bst.phi()
inputs.push_back(bst)
clusterizer = ROOT.heppy.ReclusterJets(inputs, -1, 0.4)
jets = clusterizer.getGrouping(10)
#event.jee = list(self.handles['jee'].product())
# print "Boosted jets:",
for j in list(jets):
oldpt = j.pt()
oldeta = j.eta()
oldphi = j.phi()
Boost(j, -b)
# if j.pt() > 20 :
# print oldpt,oldeta,oldphi," -> ",j.pt(),j.eta(),j.phi(),"|",
# print " "
def doFakeMET(self, event):
#fake MET from Zmumu
event.fakeMET = ROOT.reco.Particle.LorentzVector(0., 0., 0., 0.)
event.fakeMET.sumet = 0
if event.Vtype == 0:
event.fakeMET = event.met.p4() + event.V
event.fakeMET.sumet = event.met.sumEt() - event.V.pt()
def doHtMhtJets30(self, event):
## with Central Jets
objects30 = [j for j in event.cleanJets if j.pt() > 30
] + event.selectedLeptons
event.htJet30 = sum([x.pt() for x in objects30])
event.mhtJet30vec = ROOT.reco.Particle.LorentzVector(
-1. * (sum([x.px() for x in objects30])),
-1. * (sum([x.py() for x in objects30])), 0, 0)
event.mhtJet30 = event.mhtJet30vec.pt()
event.mhtPhiJet30 = event.mhtJet30vec.phi()
def doHiggsHighCSV(self, event):
#leading csv interpretation
if (len(event.jetsForHiggs) >= 2):
event.hJetsCSV = sorted(
event.jetsForHiggs,
key=lambda jet: jet.btag(
getattr(self.cfg_ana, "btagDiscriminator",
'pfCombinedInclusiveSecondaryVertexV2BJetTags')),
reverse=True)[0:2]
event.aJetsCSV = [
x for x in event.cleanJets if x not in event.hJetsCSV
]
event.hjidxCSV = [
event.cleanJetsAll.index(x) for x in event.hJetsCSV
]
event.ajidxCSV = [
event.cleanJetsAll.index(x) for x in event.aJetsCSV
]
event.aJetsCSV += event.cleanJetsFwd
event.HCSV = event.hJetsCSV[0].p4() + event.hJetsCSV[1].p4()
def doHiggsHighPt(self, event):
#highest pair interpretations
if (len(event.jetsForHiggs) >= 2):
event.hJets = list(
max(itertools.combinations(event.jetsForHiggs, 2),
key=lambda x: (x[0].p4() + x[1].p4()).pt()))
event.aJets = [x for x in event.cleanJets if x not in event.hJets]
event.hjidx = [event.cleanJetsAll.index(x) for x in event.hJets]
event.ajidx = [event.cleanJetsAll.index(x) for x in event.aJets]
event.aJets += event.cleanJetsFwd
hJetsByCSV = sorted(
event.hJets,
key=lambda jet: jet.btag(
getattr(self.cfg_ana, "btagDiscriminator",
'pfCombinedInclusiveSecondaryVertexV2BJetTags')),
reverse=True)
event.hjidxDiJetPtByCSV = [
event.cleanJetsAll.index(x) for x in hJetsByCSV
]
event.H = event.hJets[0].p4() + event.hJets[1].p4()
def doHiggsAddJetsdR08(self, event):
if (len(event.jetsForHiggs) >= 2):
event.hJetsaddJetsdR08 = [x for x in event.hJetsCSV]
event.dRaddJetsdR08 = []
event.aJetsaddJetsdR08 = [x for x in event.aJetsCSV]
event.hjidxaddJetsdR08 = [x for x in event.hjidxCSV]
event.ajidxaddJetsdR08 = [x for x in event.ajidxCSV]
#multiple jets interpretations, for central jets closest to dR<0.8 from higgs jets
jetsForHiggsAddJetsdR08 = [
x for x in event.cleanJetsAll
if (x.pt() > 15 and abs(x.eta()) < 3.0 and x.puJetId() > 0
and x.jetID('POG_PFID_Loose'))
]
if (len(jetsForHiggsAddJetsdR08) > 2):
addJetsForHiggs = [
x for x in jetsForHiggsAddJetsdR08
if (x not in event.hJetsCSV and min(
deltaR(x.eta(), x.phi(), event.hJetsCSV[0].eta(),
event.hJetsCSV[0].phi()),
deltaR(x.eta(), x.phi(), event.hJetsCSV[1].eta(),
event.hJetsCSV[1].phi())) < 0.8)
]
for x in addJetsForHiggs:
event.hJetsaddJetsdR08.append(x)
event.dRaddJetsdR08.append(
min(
deltaR(x.eta(), x.phi(), event.hJetsCSV[0].eta(),
event.hJetsCSV[0].phi()),
deltaR(x.eta(), x.phi(), event.hJetsCSV[1].eta(),
event.hJetsCSV[1].phi())))
event.hjidxaddJetsdR08 = [
event.cleanJetsAll.index(x) for x in event.hJetsaddJetsdR08
]
event.aJetsaddJetsdR08 = [
x for x in event.cleanJets
if x not in event.hJetsaddJetsdR08
]
event.aJetsaddJetsdR08 += event.cleanJetsFwd
event.ajidxaddJetsdR08 = [
event.cleanJetsAll.index(x) for x in event.aJetsaddJetsdR08
]
event.HaddJetsdR08 = sum(
map(lambda x: x.p4(), event.hJetsaddJetsdR08),
ROOT.reco.Particle.LorentzVector(0., 0., 0., 0.))
def doVHRegression(self, event):
if (len(event.jetsForHiggs) >= 2):
# ""=nominal, the other correspond to jet energies up/down for JEC and JER
for analysis in [
"", "corrJECUp", "corrJECDown", "corrJERUp", "corrJERDown"
]:
self.regressions[event.Vtype].evaluateRegression(
event, "pt_reg", analysis)
hJetCSV_reg0 = ROOT.reco.Particle.LorentzVector(
event.hJetsCSV[0].p4())
hJetCSV_reg1 = ROOT.reco.Particle.LorentzVector(
event.hJetsCSV[1].p4())
hJetCSV_reg0 *= getattr(event.hJetsCSV[0], "pt_reg" +
analysis) / event.hJetsCSV[0].pt()
hJetCSV_reg1 *= getattr(event.hJetsCSV[1], "pt_reg" +
analysis) / event.hJetsCSV[1].pt()
setattr(
event,
"HCSV_reg" + ("_" + analysis if analysis != "" else ""),
hJetCSV_reg0 + hJetCSV_reg1)
hJet_reg0 = ROOT.reco.Particle.LorentzVector(
event.hJets[0].p4())
hJet_reg1 = ROOT.reco.Particle.LorentzVector(
event.hJets[1].p4())
hJet_reg0 *= getattr(
event.hJets[0], "pt_reg" + analysis) / event.hJets[0].pt()
hJet_reg1 *= getattr(
event.hJets[1], "pt_reg" + analysis) / event.hJets[1].pt()
setattr(event,
"H_reg" + ("_" + analysis if analysis != "" else ""),
hJet_reg0 + hJet_reg1)
#print "H_reg"+("_"+analysis if analysis!="" else ""), getattr(event, "H_reg"+("_"+analysis if analysis!="" else "")).Pt()
else:
for analysis in [
"", "corrJECUp", "corrJECDown", "corrJERUp", "corrJERDown"
]:
setattr(
event,
"HCSV_reg" + ("_" + analysis if analysis != "" else ""),
ROOT.reco.Particle.LorentzVector())
setattr(event,
"H_reg" + ("_" + analysis if analysis != "" else ""),
ROOT.reco.Particle.LorentzVector())
def doVBFblikelihood(self, event):
self.blikelihood.evaluateBlikelihood(event)
def classifyMCEvent(self, event):
if self.cfg_comp.isMC:
event.VtypeSim = -1
if len(event.genvbosons) == 1:
# ZtoLL events, same flavour leptons
if event.genvbosons[0].pdgId() == 23 and event.genvbosons[
0].numberOfDaughters() > 1 and abs(
event.genvbosons[0].daughter(0).pdgId()) == abs(
event.genvbosons[0].daughter(1).pdgId()):
if abs(event.genvbosons[0].daughter(0).pdgId()) == 11:
#Ztoee
event.VtypeSim = 1
if abs(event.genvbosons[0].daughter(0).pdgId()) == 13:
#ZtoMuMu
event.VtypeSim = 0
if abs(event.genvbosons[0].daughter(0).pdgId()) == 15:
#ZtoTauTau
event.VtypeSim = 5
if abs(event.genvbosons[0].daughter(0).pdgId()) in [
12, 14, 16
]:
event.VtypeSim = 4
#WtoLNu events
if abs(event.genvbosons[0].pdgId(
)) == 24 and event.genvbosons[0].numberOfDaughters() == 2:
if abs(event.genvbosons[0].daughter(
0).pdgId()) == 11 and abs(
event.genvbosons[0].daughter(1).pdgId()) == 12:
#WtoEleNu_e
event.VtypeSim = 3
if abs(event.genvbosons[0].daughter(0).pdgId()) == 13 and abs(
event.genvbosons[0].daughter(1).pdgId()) == 14:
#WtoMuNu_mu
event.VtypeSim = 2
#to be added: WtoTauNu
if len(event.genvbosons) > 1:
#print 'more than one W/Zbosons?'
event.VtypeSim = -2
# if event.VtypeSim == -1:
# print '===================================='
# print ' --------- Debug VtypeSim -1 --------'
# print '# genVbosons: ',len(event.genvbosons), '| #daughters ', event.genvbosons[0].numberOfDaughters()
# for i in xrange (0, event.genvbosons[0].numberOfDaughters() ) :
# print 'daughter ',i ,'| pdgId', event.genvbosons[0].daughter(i).pdgId()
def classifyEvent(self, event):
#assign events to analysis (Vtype)
#enum CandidateType{Zmumu, Zee, Wmun, Wen, Znn, Zemu, Ztaumu, Ztaue, Wtaun, Ztautau, Zbb, UNKNOWN};
event.Vtype = -1
nLep = len(event.selectedLeptons)
event.vLeptons = []
#WH requires exactly one selected lepton
wElectrons = [
x for x in event.selectedElectrons if self.cfg_ana.wEleSelection(x)
]
wMuons = [
x for x in event.selectedMuons if self.cfg_ana.wMuSelection(x)
]
zElectrons = [
x for x in event.selectedElectrons if self.cfg_ana.zEleSelection(x)
]
zMuons = [
x for x in event.selectedMuons if self.cfg_ana.zMuSelection(x)
]
zMuons.sort(key=lambda x: x.pt(), reverse=True)
zElectrons.sort(key=lambda x: x.pt(), reverse=True)
if len(zMuons) >= 2:
# print zMuons[0].pt()
if zMuons[0].pt() > self.cfg_ana.zLeadingMuPt:
for i in xrange(1, len(zMuons)):
if zMuons[0].charge() * zMuons[i].charge() < 0:
event.Vtype = 0
event.vLeptons = [zMuons[0], zMuons[i]]
break
elif len(zElectrons) >= 2:
# for i in zElectrons[0].electronIDs() :
# print i.first,i.second
if zElectrons[0].pt() > self.cfg_ana.zLeadingElePt:
for i in xrange(1, len(zElectrons)):
if zElectrons[0].charge() * zElectrons[i].charge() < 0:
event.Vtype = 1
event.vLeptons = [zElectrons[0], zElectrons[i]]
break
elif len(wElectrons) + len(wMuons) == 1:
if abs(event.selectedLeptons[0].pdgId()) == 13:
event.Vtype = 2
event.vLeptons = wMuons
if abs(event.selectedLeptons[0].pdgId()) == 11:
event.Vtype = 3
event.vLeptons = wElectrons
elif len(zElectrons) + len(zMuons) > 0:
event.Vtype = 5 #there are some loose (Z selection) leptons but not matching the W/Z above requirements
else:
event.Vtype = 4 #no leptons at all, apply MET cut
#apply MET cut
if event.met.pt() < 80:
event.Vtype = -1
event.V = sum(map(lambda x: x.p4(), event.vLeptons),
ROOT.reco.Particle.LorentzVector(0., 0., 0., 0.))
if event.Vtype > 1:
event.V += ROOT.reco.Particle.LorentzVector(
event.met.p4().x(),
event.met.p4().y(), 0,
event.met.p4().pt())
if event.V.Et() > event.V.pt():
event.V.goodMt = sqrt(event.V.Et()**2 - event.V.pt()**2)
else:
event.V.goodMt = -sqrt(-event.V.Et()**2 + event.V.pt()**2)
event.aLeptons = [
x for x in event.inclusiveLeptons if x not in event.vLeptons
]
return True
def fillTauIndices(self, event):
for j in event.cleanJetsAll:
j.tauIdxs = [
event.inclusiveTaus.index(x) for x in j.taus
if j.taus in event.inclusiveTaus
]
for t in event.inclusiveTaus:
dRmin = 1.e+3
t.jetIdx = -1
for jIdx, j in enumerate(event.cleanJetsAll):
dR = None
if t.isPFTau():
dR = deltaR(t.p4Jet().eta(),
t.p4Jet().phi(), j.eta(), j.phi())
else:
dR = deltaR(t.pfEssential().p4CorrJet_.eta(),
t.pfEssential().p4CorrJet_.phi(), j.eta(),
j.phi())
if dR < 0.3 and dR < dRmin:
t.jetIdx = jIdx
dRmin = dR
def additionalRhos(self, event):
event.rhoN = self.handles['rhoN'].product()
event.rhoCHPU = self.handles['rhoCHPU'].product()
event.rhoCentral = self.handles['rhoCentral'].product()
def initOutputs(self, event):
event.hJets = []
event.aJets = []
event.hjidx = []
event.ajidx = []
event.hJetsCSV = []
event.aJetsCSV = []
event.hjidxCSV = []
event.ajidxCSV = []
event.hJetsaddJetsdR08 = []
event.dRaddJetsdR08 = []
event.aJetsaddJetsdR08 = []
event.hjidxaddJetsdR08 = []
event.ajidxaddJetsdR08 = []
event.aLeptons = []
event.vLeptons = []
event.isrJetVH = -1
event.H = ROOT.reco.Particle.LorentzVector(0., 0., 0., 0.)
event.HCSV = ROOT.reco.Particle.LorentzVector(0., 0., 0., 0.)
event.HaddJetsdR08 = ROOT.reco.Particle.LorentzVector(0., 0., 0., 0.)
event.H_reg = ROOT.reco.Particle.LorentzVector(0., 0., 0., 0.)
event.HCSV_reg = ROOT.reco.Particle.LorentzVector(0., 0., 0., 0.)
event.V = ROOT.reco.Particle.LorentzVector(0., 0., 0., 0.)
event.minDr3 = -1
event.V.goodMt = 0
event.hjidxDiJetPtByCSV = []
event.softActivityJets = []
event.softActivityVHJets = []
event.rhoN = -1
event.rhoCHPU = -1
event.rhoCentral = -1
def process(self, event):
#print "Event number",event.iEv
self.readCollections(event.input)
self.inputCounter.Fill(1)
if self.cfg_comp.isMC:
genWeight = self.handles['GenInfo'].product().weight()
self.inputCounterWeighted.Fill(
1,
copysign(1.0, genWeight) * event.puWeight)
self.inputCounterFullWeighted.Fill(1, genWeight * event.puWeight)
for LHE_scale in range(len(event.LHE_weights_scale)):
getattr(self, "inputCounterWeightedLHEWeightScale").Fill(
LHE_scale,
copysign(1.0, genWeight) * event.puWeight *
(event.LHE_weights_scale[LHE_scale]).wgt)
for LHE_pdf in range(len(event.LHE_weights_pdf)):
getattr(self, "inputCounterWeightedLHEWeightPdf").Fill(
LHE_pdf,
copysign(1.0, genWeight) * event.puWeight *
(event.LHE_weights_pdf[LHE_pdf]).wgt)
if genWeight > 0:
self.inputCounterPosWeight.Fill(1)
elif genWeight < 0:
self.inputCounterNegWeight.Fill(1)
self.initOutputs(event)
# event.pfCands = self.handles['pfCands'].product()
# met = event.met
#self.addNewBTag(event)
self.classifyMCEvent(event)
self.classifyEvent(event)
self.doFakeMET(event)
self.doHtMhtJets30(event)
self.fillTauIndices(event)
#Add CSV ranking
csvSortedJets = sorted(
event.cleanJetsAll,
key=lambda jet: jet.btag(
getattr(self.cfg_ana, "btagDiscriminator",
'pfCombinedInclusiveSecondaryVertexV2BJetTags')),
reverse=True)
for j in event.cleanJetsAll:
j.btagIdx = csvSortedJets.index(j)
for j in event.discardedJets:
j.btagIdx = -1
#substructure threshold, make configurable
ssThreshold = 200.
# filter events with less than 2 jets with pt 20
event.jetsForHiggs = [
x for x in event.cleanJets if self.cfg_ana.higgsJetsPreSelection(x)
]
if not (len(event.jetsForHiggs) >= 2 or
(len(event.cleanJets) == 1
and event.cleanJets[0].pt() > ssThreshold)):
return self.cfg_ana.passall
if event.Vtype < 0 and not (
sum(x.pt() > 30 for x in event.jetsForHiggs) >= 4
or sum(x.pt() for x in event.jetsForHiggs[:4]) > 160):
return self.cfg_ana.passall
map(lambda x: x.qgl(), event.jetsForHiggs[:6])
map(lambda x: x.qgl(), (x for x in event.jetsForHiggs if x.pt() > 30))
self.doHiggsHighCSV(event)
self.doHiggsHighPt(event)
self.doHiggsAddJetsdR08(event)
self.searchISRforVH(event)
self.doVHRegression(event)
self.doVBFblikelihood(event)
self.fillTauIndices(event)
self.addPullVector(event)
if getattr(self.cfg_ana, "doVBF", True):
self.doVBF(event)
if getattr(self.cfg_ana, "doSoftActivityVH", False):
self.doSoftActivityVH(event)
# event.jee = list(self.handles['jee'].product())
#for j in list(jets)[0:3]:
# print j.pt(),
#print " "
#to implement
#if some threshold:
# self.computeSubStructuresStuff()
#self.doIVFHiggs()
#self.computePullAngle()
#
#perhaps in different producers:
# LHE weights
# Trigger weights
# gen level VH specific info
# add soft jet info to jets
# PU weights
# SIM B hadrons information
# MET corrections (in MET analyzer)
# trigger flags
# Hbalance = ROOT.TLorentzVector()
# Hbalance.SetPtEtaPhiM(event.V.pt(),0,event.V.phi(),125.)
# hh=event.genHiggsBoson[0]
# Hbalance2 = ROOT.TLorentzVector()
# Hbalance2.SetPtEtaPhiM(hh.pt(),hh.eta(),hh.phi(),125.)
# print "Hbalance pt,e,phi",Hbalance.Pt(),Hbalance.E(),Hbalance.Phi()
# print "gen H pt,e,phi",hh.pt(),hh.p4().E(),hh.phi()
# print "Hbalance2 pt,e,phi",Hbalance2.Pt(),Hbalance2.E(),Hbalance2.Phi()
# print "boostVector:",Hbalance.BoostVector().X(),Hbalance.BoostVector().Y(),Hbalance.BoostVector().Z()
# print "boostVector2:",Hbalance2.BoostVector().X(),Hbalance2.BoostVector().Y(),Hbalance2.BoostVector().Z()
#print "Unboosted Hbalance",Boost(Hbalance,-Hbalance.BoostVector()).Pt()
# print "hbalance boost vector",Hbalance.BoostVector()
# hhh=deepcopy(Hbalance2)
# Boost(hhh,-Hbalance2.BoostVector())
# print "unboosted pt",hhh.Pt()
# self.makeJets(event,-Hbalance2.BoostVector())
return True