forked from cp3-llbb/HHbbWWAnalysis
-
Notifications
You must be signed in to change notification settings - Fork 0
/
mvaEvaluatorDL_nonres.py
191 lines (174 loc) · 19.7 KB
/
mvaEvaluatorDL_nonres.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
from bamboo import treefunctions as op
def returnLeptonsMVAInputs(self,l1,l2,channel):
if channel == "ElEl":
cone_l1 = self.getElectronConeP4(l1)
cone_l2 = self.getElectronConeP4(l2)
elif channel == "MuMu":
cone_l1 = self.getMuonConeP4(l1)
cone_l2 = self.getMuonConeP4(l2)
elif channel == "ElMu":
cone_l1 = self.getElectronConeP4(l1)
cone_l2 = self.getMuonConeP4(l2)
else:
raise RuntimeError("Wrong channel")
return {('l1_E', 'Lead lepton E [GeV]', (50,0.,500.)) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , cone_l1.E(), cone_l2.E()),
('l1_Px', 'Lead lepton P_x [GeV]', (40,-200.,200.)) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , cone_l1.Px(), cone_l2.Px()),
('l1_Py', 'Lead lepton P_y [GeV]', (40,-200.,200.)) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , cone_l1.Py(), cone_l2.Py()),
('l1_Pz', 'Lead lepton P_z [GeV]', (40,-200.,200.)) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , cone_l1.Pz(), cone_l2.Pz()),
('l1_pdgId', 'Lead lepton pdg ID', (45,-22.,22.) ) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , l1.pdgId, l2.pdgId),
('l1_charge', 'Lead lepton charge', (2,0.,2.) ) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , l1.charge, l2.charge),
('l2_E', 'Sublead lepton E [GeV]', (50,0.,500.)) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , cone_l2.E(), cone_l1.E()),
('l2_Px', 'Sublead lepton P_x [GeV]', (40,-200.,200.)) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , cone_l2.Px(), cone_l1.Px()),
('l2_Py', 'Sublead lepton P_y [GeV]', (40,-200.,200.)) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , cone_l2.Py(), cone_l1.Py()),
('l2_Pz', 'Sublead lepton P_z [GeV]', (40,-200.,200.)) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , cone_l2.Pz(), cone_l1.Pz()),
('l2_pdgId', 'Sublead lepton pdg ID', (45,-22.,22.) ) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , l2.pdgId, l1.pdgId),
('l2_charge', 'Sublead lepton charge', (2,0.,2.) ) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , l2.charge, l1.charge)}
def returnJetsMVAInputs(self,jets):
return {('j1_E', 'Lead jet E [GeV]', (50,0.,500.)) : op.switch(op.rng_len(jets)>0,jets[0].p4.E(),op.c_float(0.)),
('j1_Px', 'Lead jet P_x [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(jets)>0,jets[0].p4.Px(),op.c_float(0.)),
('j1_Py', 'Lead jet P_y [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(jets)>0,jets[0].p4.Py(),op.c_float(0.)),
('j1_Pz', 'Lead jet P_z [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(jets)>0,jets[0].p4.Pz(),op.c_float(0.)),
('j1_btag', 'Lead jet btag score', (50,0.,1.) ) : op.switch(op.rng_len(jets)>0,jets[0].btagDeepFlavB,op.c_float(0.)),
('j2_E', 'Sublead jet E [GeV]', (50,0.,500.)) : op.switch(op.rng_len(jets)>1,jets[1].p4.E(),op.c_float(0.)),
('j2_Px', 'Sublead jet P_x [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(jets)>1,jets[1].p4.Px(),op.c_float(0.)),
('j2_Py', 'Sublead jet P_y [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(jets)>1,jets[1].p4.Py(),op.c_float(0.)),
('j2_Pz', 'Sublead jet P_z [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(jets)>1,jets[1].p4.Pz(),op.c_float(0.)),
('j2_btag', 'Sublead jet btag score', (50,0.,1.) ) : op.switch(op.rng_len(jets)>1,jets[1].btagDeepFlavB,op.c_float(0.)),
('j3_E', 'Subsublead jet E [GeV]', (50,0.,500.)) : op.switch(op.rng_len(jets)>2,jets[2].p4.E(),op.c_float(0.)),
('j3_Px', 'Subsublead jet P_x [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(jets)>2,jets[2].p4.Px(),op.c_float(0.)),
('j3_Py', 'Subsublead jet P_y [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(jets)>2,jets[2].p4.Py(),op.c_float(0.)),
('j3_Pz', 'Subsublead jet P_z [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(jets)>2,jets[2].p4.Pz(),op.c_float(0.)),
('j3_btag', 'Subsublead jet btag score', (50,0.,1.) ) : op.switch(op.rng_len(jets)>2,jets[2].btagDeepFlavB,op.c_float(0.)),
('j4_E', 'Subsubsublead jet E [GeV]', (50,0.,500.)) : op.switch(op.rng_len(jets)>3,jets[3].p4.E(),op.c_float(0.)),
('j4_Px', 'Subsubsublead jet P_x [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(jets)>3,jets[3].p4.Px(),op.c_float(0.)),
('j4_Py', 'Subsubsublead jet P_y [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(jets)>3,jets[3].p4.Py(),op.c_float(0.)),
('j4_Pz', 'Subsubsublead jet P_z [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(jets)>3,jets[3].p4.Pz(),op.c_float(0.)),
('j4_btag', 'Subsubsublead jet btag score', (50,0.,1.) ) : op.switch(op.rng_len(jets)>3,jets[3].btagDeepFlavB,op.c_float(0.))}
def returnFatjetMVAInputs(self,fatjets):
return {('fatjet_E', 'Fatjet E [GeV]', (50,0.,500.)) : op.switch(op.rng_len(fatjets)>0, fatjets[0].p4.E(), op.c_float(0.)),
('fatjet_Px', 'Fatjet P_x [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(fatjets)>0, fatjets[0].p4.Px(), op.c_float(0.)),
('fatjet_Py', 'Fatjet P_y [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(fatjets)>0, fatjets[0].p4.Py(), op.c_float(0.)),
('fatjet_Pz', 'Fatjet P_z [GeV]', (40,-200.,200.)) : op.switch(op.rng_len(fatjets)>0, fatjets[0].p4.Pz(), op.c_float(0.)),
('fatjet_tau1', 'Fatjet #tau_1', (50,0.,1.)) : op.switch(op.rng_len(fatjets)>0, fatjets[0].tau1, op.c_float(0.)),
('fatjet_tau2', 'Fatjet #tau_2', (50,0.,1.)) : op.switch(op.rng_len(fatjets)>0, fatjets[0].tau2, op.c_float(0.)),
('fatjet_tau3', 'Fatjet #tau_3', (50,0.,1.)) : op.switch(op.rng_len(fatjets)>0, fatjets[0].tau3, op.c_float(0.)),
('fatjet_tau4', 'Fatjet #tau_4', (50,0.,1.)) : op.switch(op.rng_len(fatjets)>0, fatjets[0].tau4, op.c_float(0.)),
('fatjet_softdrop', 'Fatjet softdrop mass [GeV]', (50,0.,1000.)) : op.switch(op.rng_len(fatjets)>0, fatjets[0].msoftdrop, op.c_float(0.))}
def returnMETMVAInputs(self,met):
return {('met_E', 'MET Energy', (50,0.,500.)) : met.p4.E(),
('met_Px', 'MET P_x', (40,-200.,200.)) : met.p4.Px(),
('met_Py', 'MET P_y', (40,-200.,200.)) : met.p4.Py(),
('met_Pz', 'MET P_z', (40,-200.,200.)) : met.p4.Pz()}
def returnHighLevelMVAInputs(self,l1,l2,met,jets,bjets,electrons,muons,channel):
if channel == "ElEl":
cone_l1 = self.getElectronConeP4(l1)
cone_l2 = self.getElectronConeP4(l2)
elif channel == "MuMu":
cone_l1 = self.getMuonConeP4(l1)
cone_l2 = self.getMuonConeP4(l2)
elif channel == "ElMu":
cone_l1 = self.getElectronConeP4(l1)
cone_l2 = self.getMuonConeP4(l2)
else:
raise RuntimeError("Wrong channel")
dijets = op.combine(jets, N=2)
import bamboo.treeoperations as _to
def rng_min(rng, fun=(lambda x : x), typeName="float"):
return op._to.Reduce.fromRngFun(rng, op.c_float(float("+inf"), typeName), ( lambda fn : (
lambda res, elm : op.extMethod("std::min", returnType="Float_t")(res, fn(elm))
) )(fun) )
if self.args.Boosted0Btag or self.args.Boosted1Btag:
VBFJetPairs = self.VBFJetPairsBoosted
elif self.args.Resolved0Btag or self.args.Resolved1Btag or self.args.Resolved2Btag:
VBFJetPairs = self.VBFJetPairsResolved
else:
raise RuntimeError("Wrong selection to be used by the DNN")
return {
('m_bb_bregcorr', 'Di-bjet invariant mass (regcorr) [GeV]', (100,0.,1000.)) : op.multiSwitch((op.rng_len(bjets) == 0, op.c_float(0.)),
(op.rng_len(bjets) == 1, self.HLL.getCorrBp4(bjets[0]).M()),
op.invariant_mass(self.HLL.getCorrBp4(bjets[0]),self.HLL.getCorrBp4(bjets[1]))),
('ht', 'HT(jets) [GeV]', (100,0.,1000.)) : op.rng_sum(jets, lambda j : j.pt),
('min_dr_jets_lep1', 'Min(#Delta R(lead lepton,jets))', (25,0.,5.)) : op.switch(op.rng_len(jets) > 0,
op.switch(cone_l1.Pt() >= cone_l2.Pt(),
self.HLL.MinDR_part1_partCont(cone_l1,jets),
self.HLL.MinDR_part1_partCont(cone_l2,jets)),
op.c_float(0.)),
('min_dr_jets_lep2', 'Min(#Delta R(sublead lepton,jets))', (25,0.,5.)) : op.switch(op.rng_len(jets) > 0,
op.switch(cone_l1.Pt() >= cone_l2.Pt(),
self.HLL.MinDR_part1_partCont(cone_l2,jets),
self.HLL.MinDR_part1_partCont(cone_l1,jets)),
op.c_float(0.)),
('m_ll', 'Dilepton invariant mass [GeV]', (100,0.,1000.)) : op.invariant_mass(cone_l1,cone_l2),
('dr_ll', 'Dilepton #Delta R', (25,0.,5.)) : op.deltaR(cone_l1,cone_l2),
('min_dr_jet', 'Min(#Delta R(jets))', (25,0.,5.)) : op.switch(op.rng_len(dijets) > 0,
op.rng_min(dijets,lambda dijet : op.deltaR(dijet[0].p4,dijet[1].p4)),
op.c_float(0.)),
('min_dhi_jet', 'Min(#Delta #Phi(jets))', (16,0.,3.2)) : op.switch(op.rng_len(dijets) > 0,
rng_min(dijets,lambda dijet : op.abs(op.deltaPhi(dijet[0].p4,dijet[1].p4)),typeName='double'),
op.c_float(0.)),
('m_hh_simplemet_bregcorr','M_{HH} (simple MET) (regcorr) [GeV]', (100,0.,1000.)) : op.invariant_mass(op.rng_sum(bjets,
lambda bjet : self.HLL.getCorrBp4(bjet),
start=self.HLL.empty_p4),
cone_l1,
cone_l2,
met.p4),
('met_ld', 'MET_{LD}', (100,0.,1000.)) : self.HLL.MET_LD_DL(met,jets,electrons,muons),
('dr_bb', 'Di-bjet #Delta R', (25,0.,5.)) : op.switch(op.rng_len(bjets)>=2,
op.deltaR(bjets[0].p4,bjets[1].p4),
op.c_float(0.)),
('min_dr_leps_b1', 'Min(#Delta R(lead bjet,dilepton))', (25,0.,5.)) : op.switch(op.rng_len(bjets)>=1,
self.HLL.MinDR_part1_dipart(bjets[0].p4,[cone_l1,cone_l2]),
op.c_float(0.)),
('min_dr_leps_b2', 'Min(#Delta R(sublead bjet,dilepton))', (25,0.,5.)) : op.switch(op.rng_len(bjets)>=2,
self.HLL.MinDR_part1_dipart(bjets[1].p4,[cone_l1,cone_l2]),
op.c_float(0.)),
('lep1_conept', 'Lead lepton cone-P_T [GeV]', (40,0.,200.)) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , cone_l1.Pt() , cone_l2.Pt()),
('lep2_conept', 'Sublead lepton cone-P_T [GeV]', (40,0.,200.)) : op.switch(cone_l1.Pt() >= cone_l2.Pt() , cone_l2.Pt() , cone_l1.Pt()),
('mww_simplemet', 'M_{WW} (simple MET) [GeV]', (100,0.,1000.)) : op.invariant_mass(cone_l1,cone_l2,met.p4),
('vbf_tag', 'VBF tag', (2,0.,2.)) : op.c_int(op.rng_len(VBFJetPairs)>0),
('boosted_tag', 'Boosted tag', (2,0.,2.)) : op.c_int(op.OR(op.rng_len(self.ak8BJets) > 0, # Boosted 1B
op.AND(op.rng_len(self.ak8BJets) == 0, # Boosted 0B
op.rng_len(self.ak8Jets) > 0,
op.rng_len(self.ak4BJets) == 0))),
('dphi_met_dilep', 'Dilepton-MET #Delta #Phi', (32,-3.2,3.2)) : op.abs(op.deltaPhi(met.p4,(cone_l1+cone_l2))),
('dphi_met_dibjet', 'Dibjet-MET #Delta #Phi', (32,-3.2,3.2)) : op.multiSwitch((op.rng_len(bjets) == 0, op.c_float(0.)),
(op.rng_len(bjets) == 1, op.abs(op.deltaPhi(met.p4,bjets[0].p4))),
op.abs(op.deltaPhi(met.p4,(bjets[0].p4+bjets[1].p4)))),
('dr_dilep_dijet', 'Dilepton-dijet #Delta R', (25,0.,5.)) : op.multiSwitch((op.rng_len(jets) == 0, op.c_float(0.)),
(op.rng_len(jets) == 1, op.deltaR((cone_l1+cone_l2),jets[0].p4)),
op.deltaR((cone_l1+cone_l2),(jets[0].p4+jets[1].p4))),
('dr_dilep_dibjet', 'Dilepton-dibjet #Delta R', (25,0.,5.)) : op.multiSwitch((op.rng_len(bjets) == 0, op.c_float(0.)),
(op.rng_len(bjets) == 1, op.deltaR((cone_l1+cone_l2),bjets[0].p4)),
op.deltaR((cone_l1+cone_l2),(bjets[0].p4+bjets[1].p4))),
('vbf_pair_mass', 'VBF pair M_{jj}', (100,0.,1000.)) : op.switch(op.rng_len(VBFJetPairs)>0,
op.invariant_mass(VBFJetPairs[0][0].p4,VBFJetPairs[0][1].p4),
op.c_float(0.)),
('vbf_pairs_absdeltaeta', 'VBF pair #Delta#eta', (25,0.,5.)) : op.switch(op.rng_len(VBFJetPairs)>0,
op.abs(VBFJetPairs[0][0].eta-VBFJetPairs[0][1].eta),
op.c_float(0.)),
('sphericity', 'None', (1,0.,1.)) : op.c_float(0.),
('sphericity_T', 'None', (1,0.,1.)) : op.c_float(0.),
('aplanarity', 'None', (1,0.,1.)) : op.c_float(0.),
('eventshape_C', 'None', (1,0.,1.)) : op.c_float(0.),
('eventshape_D', 'None', (1,0.,1.)) : op.c_float(0.),
('eventshape_Y', 'None', (1,0.,1.)) : op.c_float(0.),
('foxwolfram1', 'None', (1,0.,1.)) : op.c_float(0.),
('foxwolfram2', 'None', (1,0.,1.)) : op.c_float(0.),
('foxwolfram3', 'None', (1,0.,1.)) : op.c_float(0.),
('foxwolfram4', 'None', (1,0.,1.)) : op.c_float(0.),
('foxwolfram5', 'None', (1,0.,1.)) : op.c_float(0.),
('centrality', 'None', (1,0.,1.)) : op.c_float(0.),
('centrality_jets', 'None', (1,0.,1.)) : op.c_float(0.),
('eigenvalue1', 'None', (1,0.,1.)) : op.c_float(0.),
('eigenvalue2', 'None', (1,0.,1.)) : op.c_float(0.),
('eigenvalue3', 'None', (1,0.,1.)) : op.c_float(0.),
}
def returnParamMVAInputs(self):
if self.args.era is None:
return {('year', 'Year', (3,2016.,2019.)) : op.c_int(int(self.era))}
else:
return {('year', 'Year', (3,2016.,2019.)) : op.c_int(int(self.args.era))}
def returnEventNrMVAInputs(self,t):
return {('eventnr', 'Event number', (int(1e6+1),0.,1e6)) : t.event}
def inputStaticCast(inputDict,cast='float'):
return [op.static_cast(cast,v) for v in inputDict.values()]