| 1 |
nmohr |
1.1 |
#!/usr/bin/env python
|
| 2 |
|
|
from samplesclass import sample
|
| 3 |
|
|
from printcolor import printc
|
| 4 |
|
|
import pickle
|
| 5 |
|
|
import sys
|
| 6 |
|
|
import os
|
| 7 |
|
|
import ROOT
|
| 8 |
|
|
import math
|
| 9 |
|
|
import shutil
|
| 10 |
|
|
from ROOT import TFile
|
| 11 |
|
|
import ROOT
|
| 12 |
|
|
from array import array
|
| 13 |
|
|
import warnings
|
| 14 |
nmohr |
1.12 |
from optparse import OptionParser
|
| 15 |
|
|
from BetterConfigParser import BetterConfigParser
|
| 16 |
nmohr |
1.1 |
warnings.filterwarnings( action='ignore', category=RuntimeWarning, message='creating converter.*' )
|
| 17 |
|
|
|
| 18 |
|
|
|
| 19 |
|
|
#usage: ./write_regression_systematic.py path
|
| 20 |
|
|
|
| 21 |
|
|
path=sys.argv[1]
|
| 22 |
nmohr |
1.10 |
namelistIN=sys.argv[2]
|
| 23 |
|
|
namelist=namelistIN.split(',')
|
| 24 |
nmohr |
1.1 |
|
| 25 |
|
|
#load info
|
| 26 |
|
|
infofile = open(path+'/samples.info','r')
|
| 27 |
|
|
info = pickle.load(infofile)
|
| 28 |
|
|
infofile.close()
|
| 29 |
|
|
#os.mkdir(path+'/sys')
|
| 30 |
nmohr |
1.12 |
argv = sys.argv[3:]
|
| 31 |
|
|
parser = OptionParser()
|
| 32 |
|
|
parser.add_option("-C", "--config", dest="config", default=[], action="append",
|
| 33 |
|
|
help="configuration defining the plots to make")
|
| 34 |
|
|
(opts, args) = parser.parse_args(argv)
|
| 35 |
|
|
if opts.config ==[]:
|
| 36 |
|
|
opts.config = "config"
|
| 37 |
|
|
print opts.config
|
| 38 |
|
|
config = BetterConfigParser()
|
| 39 |
|
|
config.read(opts.config)
|
| 40 |
nmohr |
1.1 |
|
| 41 |
|
|
def deltaPhi(phi1, phi2):
|
| 42 |
|
|
result = phi1 - phi2
|
| 43 |
|
|
while (result > math.pi): result -= 2*math.pi
|
| 44 |
|
|
while (result <= -math.pi): result += 2*math.pi
|
| 45 |
|
|
return result
|
| 46 |
|
|
|
| 47 |
nmohr |
1.10 |
def resolutionBias(eta):
|
| 48 |
|
|
if(eta< 1.1): return 0.05
|
| 49 |
|
|
if(eta< 2.5): return 0.10
|
| 50 |
|
|
if(eta< 5): return 0.30
|
| 51 |
|
|
return 0
|
| 52 |
|
|
|
| 53 |
|
|
def corrPt(pt,eta,mcPt):
|
| 54 |
|
|
return (pt+resolutionBias(math.fabs(eta))*(pt-mcPt))/pt
|
| 55 |
|
|
|
| 56 |
nmohr |
1.1 |
def corrCSV(btag, csv, flav):
|
| 57 |
|
|
if(csv < 0.): return csv
|
| 58 |
|
|
if(csv > 1.): return csv;
|
| 59 |
|
|
if(flav == 0): return csv;
|
| 60 |
|
|
if(math.fabs(flav) == 5): return btag.ib.Eval(csv)
|
| 61 |
|
|
if(math.fabs(flav) == 4): return btag.ic.Eval(csv)
|
| 62 |
|
|
if(math.fabs(flav) != 4 and math.fabs(flav) != 5): return btag.il.Eval(csv)
|
| 63 |
|
|
return -10000
|
| 64 |
|
|
|
| 65 |
|
|
|
| 66 |
nmohr |
1.10 |
def csvReshape(sh, pt, eta, csv, flav):
|
| 67 |
|
|
return sh.reshape(float(eta), float(pt), float(csv), int(flav))
|
| 68 |
|
|
|
| 69 |
|
|
|
| 70 |
nmohr |
1.1 |
for job in info:
|
| 71 |
nmohr |
1.10 |
if not job.name in namelist: continue
|
| 72 |
nmohr |
1.1 |
#print job.name
|
| 73 |
nmohr |
1.10 |
#if job.name != 'ZH120': continue
|
| 74 |
nmohr |
1.1 |
ROOT.gROOT.ProcessLine(
|
| 75 |
|
|
"struct H {\
|
| 76 |
|
|
int HiggsFlag;\
|
| 77 |
|
|
float mass;\
|
| 78 |
|
|
float pt;\
|
| 79 |
|
|
float eta;\
|
| 80 |
|
|
float phi;\
|
| 81 |
|
|
float dR;\
|
| 82 |
|
|
float dPhi;\
|
| 83 |
|
|
float dEta;\
|
| 84 |
|
|
} ;"
|
| 85 |
|
|
)
|
| 86 |
nmohr |
1.10 |
#ROOT.gROOT.LoadMacro('../interface/btagshape.h+')
|
| 87 |
|
|
#from ROOT import BTagShape
|
| 88 |
|
|
#btagNom = BTagShape("../data/csvdiscr.root")
|
| 89 |
|
|
#btagNom.computeFunctions()
|
| 90 |
|
|
#btagUp = BTagShape("../data/csvdiscr.root")
|
| 91 |
|
|
#btagUp.computeFunctions(+1.,0.)
|
| 92 |
|
|
#btagDown = BTagShape("../data/csvdiscr.root")
|
| 93 |
|
|
#btagDown.computeFunctions(-1.,0.)
|
| 94 |
|
|
#btagFUp = BTagShape("../data/csvdiscr.root")
|
| 95 |
|
|
#btagFUp.computeFunctions(0.,+1.)
|
| 96 |
|
|
#btagFDown = BTagShape("../data/csvdiscr.root")
|
| 97 |
|
|
#btagFDown.computeFunctions(0.,-1.)
|
| 98 |
nmohr |
1.11 |
ROOT.gSystem.Load('/shome/nmohr/CMSSW_5_2_4_patch4/src/UserCode/VHbb/interface/BTagReshaping_h.so')
|
| 99 |
nmohr |
1.10 |
from ROOT import BTagShapeInterface
|
| 100 |
|
|
btagNom = BTagShapeInterface("../data/csvdiscr.root",0,0)
|
| 101 |
|
|
btagUp = BTagShapeInterface("../data/csvdiscr.root",+1,0)
|
| 102 |
|
|
btagDown = BTagShapeInterface("../data/csvdiscr.root",-1,0)
|
| 103 |
|
|
btagFUp = BTagShapeInterface("../data/csvdiscr.root",0,+1.)
|
| 104 |
|
|
btagFDown = BTagShapeInterface("../data/csvdiscr.root",0,-1.)
|
| 105 |
nmohr |
1.1 |
|
| 106 |
|
|
print '\t - %s' %(job.name)
|
| 107 |
|
|
input = TFile.Open(job.getpath(),'read')
|
| 108 |
|
|
output = TFile.Open(job.path+'/sys/'+job.prefix+job.identifier+'.root','recreate')
|
| 109 |
|
|
|
| 110 |
|
|
input.cd()
|
| 111 |
|
|
obj = ROOT.TObject
|
| 112 |
|
|
for key in ROOT.gDirectory.GetListOfKeys():
|
| 113 |
|
|
input.cd()
|
| 114 |
|
|
obj = key.ReadObj()
|
| 115 |
|
|
#print obj.GetName()
|
| 116 |
|
|
if obj.GetName() == job.tree:
|
| 117 |
|
|
continue
|
| 118 |
|
|
output.cd()
|
| 119 |
|
|
#print key.GetName()
|
| 120 |
|
|
obj.Write(key.GetName())
|
| 121 |
|
|
|
| 122 |
nmohr |
1.3 |
input.cd()
|
| 123 |
nmohr |
1.1 |
tree = input.Get(job.tree)
|
| 124 |
|
|
nEntries = tree.GetEntries()
|
| 125 |
|
|
|
| 126 |
|
|
job.addpath('/sys')
|
| 127 |
|
|
if job.type != 'DATA':
|
| 128 |
|
|
job.SYS = ['Nominal','JER_up','JER_down','JES_up','JES_down','beff_up','beff_down','bmis_up','bmis_down']
|
| 129 |
|
|
|
| 130 |
|
|
H = ROOT.H()
|
| 131 |
|
|
HNoReg = ROOT.H()
|
| 132 |
|
|
tree.SetBranchStatus('H',0)
|
| 133 |
|
|
output.cd()
|
| 134 |
|
|
newtree = tree.CloneTree(0)
|
| 135 |
|
|
|
| 136 |
|
|
hJ0 = ROOT.TLorentzVector()
|
| 137 |
|
|
hJ1 = ROOT.TLorentzVector()
|
| 138 |
|
|
|
| 139 |
nmohr |
1.12 |
regWeight = config.get("Regression","regWeight")
|
| 140 |
|
|
regDict = eval(config.get("Regression","regDict"))
|
| 141 |
|
|
regVars = eval(config.get("Regression","regVars"))
|
| 142 |
|
|
useMET = eval(config.get("Regression","useMET"))
|
| 143 |
|
|
usePtRaw = eval(config.get("Regression","usePtRaw"))
|
| 144 |
|
|
useRho25 = eval(config.get("Regression","useRho25"))
|
| 145 |
nmohr |
1.1 |
|
| 146 |
|
|
#Regression branches
|
| 147 |
|
|
applyRegression = True
|
| 148 |
|
|
hJet_pt = array('f',[0]*2)
|
| 149 |
|
|
hJet_e = array('f',[0]*2)
|
| 150 |
|
|
newtree.Branch( 'H', H , 'HiggsFlag/I:mass/F:pt/F:eta:phi/F:dR/F:dPhi/F:dEta/F' )
|
| 151 |
|
|
newtree.Branch( 'HNoReg', HNoReg , 'HiggsFlag/I:mass/F:pt/F:eta:phi/F:dR/F:dPhi/F:dEta/F' )
|
| 152 |
|
|
Event = array('f',[0])
|
| 153 |
|
|
METet = array('f',[0])
|
| 154 |
|
|
rho25 = array('f',[0])
|
| 155 |
|
|
METphi = array('f',[0])
|
| 156 |
|
|
fRho25 = ROOT.TTreeFormula("rho25",'rho25',tree)
|
| 157 |
|
|
fEvent = ROOT.TTreeFormula("Event",'EVENT.event',tree)
|
| 158 |
|
|
fMETet = ROOT.TTreeFormula("METet",'METnoPU.et',tree)
|
| 159 |
|
|
fMETphi = ROOT.TTreeFormula("METphi",'METnoPU.phi',tree)
|
| 160 |
|
|
hJet_MET_dPhi = array('f',[0]*2)
|
| 161 |
|
|
hJet_regWeight = array('f',[0]*2)
|
| 162 |
|
|
hJet_MET_dPhiArray = [array('f',[0]),array('f',[0])]
|
| 163 |
|
|
newtree.Branch('hJet_MET_dPhi',hJet_MET_dPhi,'hJet_MET_dPhi[2]/F')
|
| 164 |
|
|
newtree.Branch('hJet_regWeight',hJet_regWeight,'hJet_regWeight[2]/F')
|
| 165 |
|
|
readerJet0 = ROOT.TMVA.Reader("!Color:!Silent" )
|
| 166 |
|
|
readerJet1 = ROOT.TMVA.Reader("!Color:!Silent" )
|
| 167 |
|
|
|
| 168 |
|
|
theForms = {}
|
| 169 |
|
|
theVars0 = {}
|
| 170 |
|
|
for var in regVars:
|
| 171 |
|
|
theVars0[var] = array( 'f', [ 0 ] )
|
| 172 |
|
|
readerJet0.AddVariable(var,theVars0[var])
|
| 173 |
|
|
theForms['form_reg_%s_0'%(regDict[var])] = ROOT.TTreeFormula("form_reg_%s_0"%(regDict[var]),'%s[0]' %(regDict[var]),tree)
|
| 174 |
nmohr |
1.12 |
if useMET:
|
| 175 |
|
|
readerJet0.AddVariable( "Jet_MET_dPhi", hJet_MET_dPhiArray[0] )
|
| 176 |
|
|
readerJet0.AddVariable( "METet", METet )
|
| 177 |
|
|
if useRho25: readerJet0.AddVariable( "rho25", rho25 )
|
| 178 |
nmohr |
1.1 |
|
| 179 |
|
|
theVars1 = {}
|
| 180 |
|
|
for var in regVars:
|
| 181 |
|
|
theVars1[var] = array( 'f', [ 0 ] )
|
| 182 |
|
|
readerJet1.AddVariable(var,theVars1[var])
|
| 183 |
|
|
theForms['form_reg_%s_1'%(regDict[var])] = ROOT.TTreeFormula("form_reg_%s_1"%(regDict[var]),'%s[1]' %(regDict[var]),tree)
|
| 184 |
nmohr |
1.12 |
if useMET:
|
| 185 |
|
|
readerJet1.AddVariable( "Jet_MET_dPhi", hJet_MET_dPhiArray[1] )
|
| 186 |
|
|
readerJet1.AddVariable( "METet", METet )
|
| 187 |
|
|
if useRho25: readerJet1.AddVariable( "rho25", rho25 )
|
| 188 |
nmohr |
1.1 |
readerJet0.BookMVA( "jet0Regression", regWeight );
|
| 189 |
|
|
readerJet1.BookMVA( "jet1Regression", regWeight );
|
| 190 |
|
|
|
| 191 |
bortigno |
1.4 |
#Add training Flag
|
| 192 |
|
|
EventForTraining = array('f',[0])
|
| 193 |
|
|
newtree.Branch('EventForTraining',EventForTraining,'EventForTraining/F')
|
| 194 |
|
|
EventForTraining[0]=0
|
| 195 |
bortigno |
1.6 |
|
| 196 |
bortigno |
1.5 |
TFlag=ROOT.TTreeFormula("EventForTraining","EVENT.event%2",tree)
|
| 197 |
nmohr |
1.1 |
|
| 198 |
|
|
if job.type != 'DATA':
|
| 199 |
|
|
#CSV branches
|
| 200 |
|
|
hJet_flavour = array('f',[0]*2)
|
| 201 |
|
|
hJet_csv = array('f',[0]*2)
|
| 202 |
|
|
hJet_csvOld = array('f',[0]*2)
|
| 203 |
|
|
hJet_csvUp = array('f',[0]*2)
|
| 204 |
|
|
hJet_csvDown = array('f',[0]*2)
|
| 205 |
|
|
hJet_csvFUp = array('f',[0]*2)
|
| 206 |
|
|
hJet_csvFDown = array('f',[0]*2)
|
| 207 |
|
|
newtree.Branch('hJet_csvOld',hJet_csvOld,'hJet_csvOld[2]/F')
|
| 208 |
|
|
newtree.Branch('hJet_csvUp',hJet_csvUp,'hJet_csvUp[2]/F')
|
| 209 |
|
|
newtree.Branch('hJet_csvDown',hJet_csvDown,'hJet_csvDown[2]/F')
|
| 210 |
|
|
newtree.Branch('hJet_csvFUp',hJet_csvFUp,'hJet_csvFUp[2]/F')
|
| 211 |
|
|
newtree.Branch('hJet_csvFDown',hJet_csvFDown,'hJet_csvFDown[2]/F')
|
| 212 |
|
|
|
| 213 |
|
|
#JER branches
|
| 214 |
|
|
hJet_pt_JER_up = array('f',[0]*2)
|
| 215 |
|
|
newtree.Branch('hJet_pt_JER_up',hJet_pt_JER_up,'hJet_pt_JER_up[2]/F')
|
| 216 |
|
|
hJet_pt_JER_down = array('f',[0]*2)
|
| 217 |
|
|
newtree.Branch('hJet_pt_JER_down',hJet_pt_JER_down,'hJet_pt_JER_down[2]/F')
|
| 218 |
|
|
hJet_e_JER_up = array('f',[0]*2)
|
| 219 |
|
|
newtree.Branch('hJet_e_JER_up',hJet_e_JER_up,'hJet_e_JER_up[2]/F')
|
| 220 |
|
|
hJet_e_JER_down = array('f',[0]*2)
|
| 221 |
|
|
newtree.Branch('hJet_e_JER_down',hJet_e_JER_down,'hJet_e_JER_down[2]/F')
|
| 222 |
|
|
H_JER = array('f',[0]*4)
|
| 223 |
|
|
newtree.Branch('H_JER',H_JER,'mass_up:mass_down:pt_up:pt_down/F')
|
| 224 |
|
|
|
| 225 |
|
|
#JES branches
|
| 226 |
|
|
hJet_pt_JES_up = array('f',[0]*2)
|
| 227 |
|
|
newtree.Branch('hJet_pt_JES_up',hJet_pt_JES_up,'hJet_pt_JES_up[2]/F')
|
| 228 |
|
|
hJet_pt_JES_down = array('f',[0]*2)
|
| 229 |
|
|
newtree.Branch('hJet_pt_JES_down',hJet_pt_JES_down,'hJet_pt_JES_down[2]/F')
|
| 230 |
|
|
hJet_e_JES_up = array('f',[0]*2)
|
| 231 |
|
|
newtree.Branch('hJet_e_JES_up',hJet_e_JES_up,'hJet_e_JES_up[2]/F')
|
| 232 |
|
|
hJet_e_JES_down = array('f',[0]*2)
|
| 233 |
|
|
newtree.Branch('hJet_e_JES_down',hJet_e_JES_down,'hJet_e_JES_down[2]/F')
|
| 234 |
|
|
H_JES = array('f',[0]*4)
|
| 235 |
|
|
newtree.Branch('H_JES',H_JES,'mass_up:mass_down:pt_up:pt_down/F')
|
| 236 |
|
|
|
| 237 |
|
|
|
| 238 |
|
|
#iter=0
|
| 239 |
|
|
|
| 240 |
|
|
|
| 241 |
|
|
for entry in range(0,nEntries):
|
| 242 |
|
|
tree.GetEntry(entry)
|
| 243 |
|
|
|
| 244 |
|
|
#fill training flag
|
| 245 |
|
|
#iter+=1
|
| 246 |
|
|
#if (iter%2==0):
|
| 247 |
|
|
# EventForTraining[0]=1
|
| 248 |
|
|
#else:
|
| 249 |
|
|
# EventForTraining[0]=0
|
| 250 |
|
|
#iter+=1
|
| 251 |
|
|
|
| 252 |
bortigno |
1.4 |
# if job.type != 'DATA':
|
| 253 |
|
|
# EventForTraining=int(not TFlag.EvalInstance())
|
| 254 |
nmohr |
1.10 |
#EventForTraining[0]=int(not TFlag.EvalInstance())
|
| 255 |
bortigno |
1.9 |
|
| 256 |
nmohr |
1.1 |
|
| 257 |
|
|
#get
|
| 258 |
|
|
hJet_pt = tree.hJet_pt
|
| 259 |
|
|
hJet_e = tree.hJet_e
|
| 260 |
|
|
hJet_pt0 = tree.hJet_pt[0]
|
| 261 |
|
|
hJet_pt1 = tree.hJet_pt[1]
|
| 262 |
|
|
hJet_eta0 = tree.hJet_eta[0]
|
| 263 |
|
|
hJet_eta1 = tree.hJet_eta[1]
|
| 264 |
|
|
hJet_genPt0 = tree.hJet_genPt[0]
|
| 265 |
|
|
hJet_genPt1 = tree.hJet_genPt[1]
|
| 266 |
|
|
hJet_e0 = tree.hJet_e[0]
|
| 267 |
|
|
hJet_e1 = tree.hJet_e[1]
|
| 268 |
|
|
hJet_phi0 = tree.hJet_phi[0]
|
| 269 |
|
|
hJet_phi1 = tree.hJet_phi[1]
|
| 270 |
|
|
hJet_JECUnc0 = tree.hJet_JECUnc[0]
|
| 271 |
|
|
hJet_JECUnc1 = tree.hJet_JECUnc[1]
|
| 272 |
|
|
|
| 273 |
|
|
Event[0]=fEvent.EvalInstance()
|
| 274 |
|
|
METet[0]=fMETet.EvalInstance()
|
| 275 |
|
|
rho25[0]=fRho25.EvalInstance()
|
| 276 |
|
|
METphi[0]=fMETphi.EvalInstance()
|
| 277 |
|
|
for key, value in regDict.items():
|
| 278 |
|
|
theVars0[key][0] = theForms["form_reg_%s_0" %(value)].EvalInstance()
|
| 279 |
|
|
theVars1[key][0] = theForms["form_reg_%s_1" %(value)].EvalInstance()
|
| 280 |
|
|
for i in range(2):
|
| 281 |
|
|
hJet_MET_dPhi[i] = deltaPhi(METphi[0],tree.hJet_phi[i])
|
| 282 |
|
|
hJet_MET_dPhiArray[i][0] = deltaPhi(METphi[0],tree.hJet_phi[i])
|
| 283 |
|
|
|
| 284 |
nmohr |
1.12 |
if not job.type == 'DATA' and usePtRaw:
|
| 285 |
nmohr |
1.10 |
theVars0['Jet_ptRaw'][0] = theForms["form_reg_hJet_ptRaw_0"].EvalInstance()*corrPt(tree.hJet_pt[0],tree.hJet_eta[0],tree.hJet_genPt[0])
|
| 286 |
|
|
theVars1['Jet_ptRaw'][0] = theForms["form_reg_hJet_ptRaw_1"].EvalInstance()*corrPt(tree.hJet_pt[1],tree.hJet_eta[1],tree.hJet_genPt[1])
|
| 287 |
|
|
|
| 288 |
nmohr |
1.1 |
if applyRegression:
|
| 289 |
|
|
hJ0.SetPtEtaPhiE(hJet_pt0,hJet_eta0,hJet_phi0,hJet_e0)
|
| 290 |
|
|
hJ1.SetPtEtaPhiE(hJet_pt1,hJet_eta1,hJet_phi1,hJet_e1)
|
| 291 |
|
|
HNoReg.HiggsFlag = 1
|
| 292 |
|
|
HNoReg.mass = (hJ0+hJ1).M()
|
| 293 |
|
|
HNoReg.pt = (hJ0+hJ1).Pt()
|
| 294 |
|
|
HNoReg.eta = (hJ0+hJ1).Eta()
|
| 295 |
|
|
HNoReg.phi = (hJ0+hJ1).Phi()
|
| 296 |
|
|
HNoReg.dR = hJ0.DeltaR(hJ1)
|
| 297 |
|
|
HNoReg.dPhi = hJ0.DeltaPhi(hJ1)
|
| 298 |
|
|
HNoReg.dEta = abs(hJ0.Eta()-hJ1.Eta())
|
| 299 |
nmohr |
1.10 |
rPt0 = max(0.0001,readerJet0.EvaluateRegression( "jet0Regression" )[0])
|
| 300 |
|
|
rPt1 = max(0.0001,readerJet1.EvaluateRegression( "jet1Regression" )[0])
|
| 301 |
nmohr |
1.1 |
hJet_regWeight[0] = rPt0/hJet_pt0
|
| 302 |
|
|
hJet_regWeight[1] = rPt1/hJet_pt1
|
| 303 |
|
|
rE0 = hJet_e0*hJet_regWeight[0]
|
| 304 |
|
|
rE1 = hJet_e1*hJet_regWeight[1]
|
| 305 |
|
|
hJ0.SetPtEtaPhiE(rPt0,hJet_eta0,hJet_phi0,rE0)
|
| 306 |
|
|
hJ1.SetPtEtaPhiE(rPt1,hJet_eta1,hJet_phi1,rE1)
|
| 307 |
|
|
tree.hJet_pt[0] = rPt0
|
| 308 |
|
|
tree.hJet_pt[1] = rPt1
|
| 309 |
|
|
tree.hJet_e[0] = rE0
|
| 310 |
|
|
tree.hJet_e[1] = rE1
|
| 311 |
|
|
H.HiggsFlag = 1
|
| 312 |
|
|
H.mass = (hJ0+hJ1).M()
|
| 313 |
|
|
H.pt = (hJ0+hJ1).Pt()
|
| 314 |
|
|
H.eta = (hJ0+hJ1).Eta()
|
| 315 |
|
|
H.phi = (hJ0+hJ1).Phi()
|
| 316 |
|
|
H.dR = hJ0.DeltaR(hJ1)
|
| 317 |
|
|
H.dPhi = hJ0.DeltaPhi(hJ1)
|
| 318 |
|
|
H.dEta = abs(hJ0.Eta()-hJ1.Eta())
|
| 319 |
|
|
if hJet_regWeight[0] > 5. or hJet_regWeight[1] > 5.:
|
| 320 |
|
|
print 'MET %.2f' %(METet[0])
|
| 321 |
|
|
print 'rho25 %.2f' %(rho25[0])
|
| 322 |
|
|
for key, value in regDict.items():
|
| 323 |
|
|
print '%s 0: %.2f'%(key, theVars0[key][0])
|
| 324 |
|
|
print '%s 0: %.2f'%(key, theVars1[key][0])
|
| 325 |
|
|
for i in range(2):
|
| 326 |
|
|
print 'dPhi %.0f %.2f' %(i,hJet_MET_dPhiArray[i][0])
|
| 327 |
|
|
print 'corr 0 %.2f' %(hJet_regWeight[0])
|
| 328 |
|
|
print 'corr 1 %.2f' %(hJet_regWeight[1])
|
| 329 |
|
|
print 'Event %.0f' %(Event[0])
|
| 330 |
|
|
print 'rPt0 %.2f' %(rPt0)
|
| 331 |
|
|
print 'rPt1 %.2f' %(rPt1)
|
| 332 |
|
|
print 'rE0 %.2f' %(rE0)
|
| 333 |
|
|
print 'rE1 %.2f' %(rE1)
|
| 334 |
|
|
print 'Mass %.2f' %(H.mass)
|
| 335 |
|
|
|
| 336 |
|
|
if job.type == 'DATA':
|
| 337 |
|
|
newtree.Fill()
|
| 338 |
|
|
continue
|
| 339 |
|
|
|
| 340 |
|
|
for i in range(2):
|
| 341 |
nmohr |
1.10 |
flavour = int(tree.hJet_flavour[i])
|
| 342 |
|
|
pt = float(tree.hJet_pt[i])
|
| 343 |
|
|
eta = float(tree.hJet_eta[i])
|
| 344 |
|
|
csv = float(tree.hJet_csv[i])
|
| 345 |
|
|
hJet_csvOld[i] = csv
|
| 346 |
|
|
tree.hJet_csv[i] = btagNom.reshape(eta,pt,csv,flavour)
|
| 347 |
|
|
hJet_csvDown[i] = btagDown.reshape(eta,pt,csv,flavour)
|
| 348 |
|
|
hJet_csvUp[i] = btagUp.reshape(eta,pt,csv,flavour)
|
| 349 |
|
|
hJet_csvFDown[i] = btagFDown.reshape(eta,pt,csv,flavour)
|
| 350 |
|
|
hJet_csvFUp[i] = btagFUp.reshape(eta,pt,csv,flavour)
|
| 351 |
|
|
#print pt, eta, flavour, csv
|
| 352 |
|
|
#tree.hJet_csv[i] = corrCSV(btagNom,csv,flavour)
|
| 353 |
|
|
#hJet_csvDown[i] = corrCSV(btagDown,csv,flavour)
|
| 354 |
|
|
#hJet_csvUp[i] = corrCSV(btagUp,csv,flavour)
|
| 355 |
|
|
#hJet_csvFDown[i] = corrCSV(btagFDown,csv,flavour)
|
| 356 |
|
|
#hJet_csvFUp[i] = corrCSV(btagFUp,csv,flavour)
|
| 357 |
nmohr |
1.1 |
|
| 358 |
|
|
for updown in ['up','down']:
|
| 359 |
|
|
#JER
|
| 360 |
|
|
if updown == 'up':
|
| 361 |
|
|
inner = 0.06
|
| 362 |
|
|
outer = 0.1
|
| 363 |
|
|
if updown == 'down':
|
| 364 |
|
|
inner = -0.06
|
| 365 |
|
|
outer = -0.1
|
| 366 |
|
|
#Calculate
|
| 367 |
|
|
if abs(hJet_eta0)<1.1: res0 = inner
|
| 368 |
|
|
else: res0 = outer
|
| 369 |
|
|
if abs(hJet_eta1)<1.1: res1 = inner
|
| 370 |
|
|
else: res1 = outer
|
| 371 |
|
|
rPt0 = hJet_pt0 + (hJet_pt0-hJet_genPt0)*res0
|
| 372 |
|
|
rPt1 = hJet_pt1 + (hJet_pt1-hJet_genPt1)*res1
|
| 373 |
|
|
rE0 = hJet_e0*rPt0/hJet_pt0
|
| 374 |
|
|
rE1 = hJet_e1*rPt1/hJet_pt1
|
| 375 |
|
|
if applyRegression:
|
| 376 |
|
|
theVars0['Jet_pt'][0] = rPt0
|
| 377 |
|
|
theVars1['Jet_pt'][0] = rPt1
|
| 378 |
|
|
theVars0['Jet_e'][0] = rE0
|
| 379 |
|
|
theVars1['Jet_e'][0] = rE1
|
| 380 |
nmohr |
1.10 |
rPt0 = max(0.0001,readerJet0.EvaluateRegression( "jet0Regression" )[0])
|
| 381 |
|
|
rPt1 = max(0.0001,readerJet1.EvaluateRegression( "jet1Regression" )[0])
|
| 382 |
nmohr |
1.1 |
rE0 = hJet_e0*rPt0/hJet_pt0
|
| 383 |
|
|
rE1 = hJet_e1*rPt1/hJet_pt1
|
| 384 |
|
|
hJ0.SetPtEtaPhiE(rPt0,hJet_eta0,hJet_phi0,rE0)
|
| 385 |
|
|
hJ1.SetPtEtaPhiE(rPt1,hJet_eta1,hJet_phi1,rE1)
|
| 386 |
|
|
#Set
|
| 387 |
|
|
if updown == 'up':
|
| 388 |
|
|
hJet_pt_JER_up[0]=rPt0
|
| 389 |
|
|
hJet_pt_JER_up[1]=rPt1
|
| 390 |
|
|
hJet_e_JER_up[0]=rE0
|
| 391 |
|
|
hJet_e_JER_up[1]=rE1
|
| 392 |
|
|
H_JER[0]=(hJ0+hJ1).M()
|
| 393 |
|
|
H_JER[2]=(hJ0+hJ1).Pt()
|
| 394 |
|
|
if updown == 'down':
|
| 395 |
|
|
hJet_pt_JER_down[0]=rPt0
|
| 396 |
|
|
hJet_pt_JER_down[1]=rPt1
|
| 397 |
|
|
hJet_e_JER_down[0]=rE0
|
| 398 |
|
|
hJet_e_JER_down[1]=rE1
|
| 399 |
|
|
H_JER[1]=(hJ0+hJ1).M()
|
| 400 |
|
|
H_JER[3]=(hJ0+hJ1).Pt()
|
| 401 |
|
|
|
| 402 |
|
|
#JES
|
| 403 |
|
|
if updown == 'up':
|
| 404 |
|
|
variation=1
|
| 405 |
|
|
if updown == 'down':
|
| 406 |
|
|
variation=-1
|
| 407 |
|
|
#calculate
|
| 408 |
|
|
rPt0 = hJet_pt0*(1+variation*hJet_JECUnc0)
|
| 409 |
|
|
rPt1 = hJet_pt1*(1+variation*hJet_JECUnc1)
|
| 410 |
|
|
rE0 = hJet_e0*(1+variation*hJet_JECUnc0)
|
| 411 |
|
|
rE1 = hJet_e1*(1+variation*hJet_JECUnc1)
|
| 412 |
|
|
if applyRegression:
|
| 413 |
|
|
theVars0['Jet_pt'][0] = rPt0
|
| 414 |
|
|
theVars1['Jet_pt'][0] = rPt1
|
| 415 |
|
|
theVars0['Jet_e'][0] = rE0
|
| 416 |
|
|
theVars1['Jet_e'][0] = rE1
|
| 417 |
nmohr |
1.10 |
rPt0 = max(0.0001,readerJet0.EvaluateRegression( "jet0Regression" )[0])
|
| 418 |
|
|
rPt1 = max(0.0001,readerJet1.EvaluateRegression( "jet1Regression" )[0])
|
| 419 |
nmohr |
1.1 |
rE0 = hJet_e0*rPt0/hJet_pt0
|
| 420 |
|
|
rE1 = hJet_e1*rPt1/hJet_pt1
|
| 421 |
|
|
hJ0.SetPtEtaPhiE(rPt0,hJet_eta0,hJet_phi0,rE0)
|
| 422 |
|
|
hJ1.SetPtEtaPhiE(rPt1,hJet_eta1,hJet_phi1,rE1)
|
| 423 |
|
|
#Fill
|
| 424 |
|
|
if updown == 'up':
|
| 425 |
|
|
hJet_pt_JES_up[0]=rPt0
|
| 426 |
|
|
hJet_pt_JES_up[1]=rPt1
|
| 427 |
|
|
hJet_e_JES_up[0]=rE0
|
| 428 |
|
|
hJet_e_JES_up[1]=rE1
|
| 429 |
|
|
H_JES[0]=(hJ0+hJ1).M()
|
| 430 |
|
|
H_JES[2]=(hJ0+hJ1).Pt()
|
| 431 |
|
|
if updown == 'down':
|
| 432 |
|
|
hJet_pt_JES_down[0]=rPt0
|
| 433 |
|
|
hJet_pt_JES_down[1]=rPt1
|
| 434 |
|
|
hJet_e_JES_down[0]=rE0
|
| 435 |
|
|
hJet_e_JES_down[1]=rE1
|
| 436 |
|
|
H_JES[1]=(hJ0+hJ1).M()
|
| 437 |
|
|
H_JES[3]=(hJ0+hJ1).Pt()
|
| 438 |
|
|
|
| 439 |
|
|
newtree.Fill()
|
| 440 |
|
|
|
| 441 |
|
|
newtree.AutoSave()
|
| 442 |
|
|
output.Close()
|
| 443 |
|
|
|
| 444 |
|
|
#dump info
|
| 445 |
nmohr |
1.11 |
#infofile = open(path+'/sys'+'/samples.info','w')
|
| 446 |
|
|
#pickle.dump(info,infofile)
|
| 447 |
|
|
#infofile.close()
|
