| 1 |
from __future__ import division
|
| 2 |
|
| 3 |
from HistGetter import *
|
| 4 |
from tdrStyle import *
|
| 5 |
from ROOT import *
|
| 6 |
from math import pow, exp, sqrt
|
| 7 |
from copy import deepcopy
|
| 8 |
from array import array
|
| 9 |
|
| 10 |
class QCDEstimator:
|
| 11 |
luminosity = 36.145#pb-1
|
| 12 |
mc_luminosity = 36.145#pb-1
|
| 13 |
luminosity_unit = 'pb-1'
|
| 14 |
scale = luminosity / mc_luminosity
|
| 15 |
jetBins = ['0jet', 'allJets', '1jet', '1orMoreJets', '2jets', '2orMoreJets', '3jets', '3orMoreJets', '4orMoreJets']
|
| 16 |
jetBinsLatex = {'0jet':'0 jet', 'allJets':'#geq 0 jets', '1jet':'1 jet', '1orMoreJets':'#geq 1 jet',
|
| 17 |
'2jets':'2 jets', '2orMoreJets':'#geq 2 jets', '3jets':'3 jets', '3orMoreJets':'#geq 3 jets',
|
| 18 |
'4orMoreJets':'#geq 4 jets'}
|
| 19 |
binWidth = 0.01
|
| 20 |
rebin = 10
|
| 21 |
fitRangesClosureTest = [ ( 0.1, 0.9 ), ( 0.1, 1.0 ), ( 0.1, 1.1 ),
|
| 22 |
( 0.2, 0.9 ), ( 0.2, 1.0 ), ( 0.2, 1.1 ),
|
| 23 |
( 0.3, 0.9 ), ( 0.3, 1.0 ), ( 0.3, 1.1 )]
|
| 24 |
|
| 25 |
fitRangesEstimation = [ ( 0.1, 1.1 ), ( 0.2, 1.1 )]#, ( 0.3, 1.1 )]
|
| 26 |
signalRegion = ( 0, 0.1 )
|
| 27 |
maxValue = 1.6
|
| 28 |
pfIsoHistogramPrefix = 'QCDest_PFIsolation_WithMETCutAndAsymJetCuts_'
|
| 29 |
pfIsoControlRegionHistogramPrefix = 'QCDest_PFIsolation_controlRegion2_WithMETCutAndAsymJetCuts_'
|
| 30 |
relIsoHistogramPrefix = 'QCDest_CombRelIso_'
|
| 31 |
pfIsoResults = {}
|
| 32 |
relIsoResults = {}
|
| 33 |
allPfIsoResults = {}
|
| 34 |
allRelIsoResults = {}
|
| 35 |
|
| 36 |
useEntryAsData = 'data'
|
| 37 |
constrainFit = False
|
| 38 |
numberOfFreeParameters = 0
|
| 39 |
|
| 40 |
currentFitRange = ( 0.1, 1.6 )
|
| 41 |
currentFitFuntion = 'gaus'
|
| 42 |
currentJetBin = jetBins[-1]
|
| 43 |
|
| 44 |
outputFormat = 'pdf'
|
| 45 |
outputFolder = ''
|
| 46 |
|
| 47 |
def __init__( self, files ):
|
| 48 |
self.files = files
|
| 49 |
HistGetter.samplefiles = files
|
| 50 |
self.histograms = {}
|
| 51 |
self.histGetter = HistGetter()
|
| 52 |
self.histGetter.setStyle()
|
| 53 |
self.getHistograms()
|
| 54 |
self.applyStyleAndCreateStack()
|
| 55 |
|
| 56 |
for bin in self.jetBins:
|
| 57 |
self.pfIsoResults[bin] = {'actualNumberOfQCDEvents': 0, 'estimatedNumberOfQCDEvents':0,
|
| 58 |
'fitFunction': None, 'fitParameters': {}, 'numberOfAllDataEvents':0,
|
| 59 |
'numberOfAllMCEvents':0}
|
| 60 |
self.relIsoResults[bin] = {'actualNumberOfQCDEvents': 0, 'estimatedNumberOfQCDEvents':0,
|
| 61 |
'fitFunction': None, 'fitParameters': {}, 'numberOfAllDataEvents':0,
|
| 62 |
'numberOfAllMCEvents':0}
|
| 63 |
|
| 64 |
def getHistograms( self ):
|
| 65 |
relIsoHists = [self.relIsoHistogramPrefix + jetbin for jetbin in self.jetBins]
|
| 66 |
pfIsoHists = [self.pfIsoHistogramPrefix + jetbin for jetbin in self.jetBins]
|
| 67 |
pfIsoControlHists = [self.pfIsoControlRegionHistogramPrefix + jetbin for jetbin in self.jetBins]
|
| 68 |
allHists = relIsoHists
|
| 69 |
allHists.extend( pfIsoHists )
|
| 70 |
allHists.extend( pfIsoControlHists )
|
| 71 |
HistGetter.hists = allHists
|
| 72 |
|
| 73 |
self.histograms = self.histGetter.getHistsFromFiles()
|
| 74 |
self.histograms = self.histGetter.addSampleSum( self.histograms )
|
| 75 |
|
| 76 |
def applyStyleAndCreateStack( self ):
|
| 77 |
samplesOfInterest = ['data', 'qcd', 'zjets', 'wjets', 'singleTop', 'ttbar', 'allMC']
|
| 78 |
colors = {'ttbar' : kRed + 1,
|
| 79 |
'wjets' : kGreen - 3,
|
| 80 |
'zjets' : kAzure - 2,
|
| 81 |
'qcd' : kYellow,
|
| 82 |
'singleTop' : kMagenta}
|
| 83 |
|
| 84 |
mcStack = {}
|
| 85 |
|
| 86 |
for sample in samplesOfInterest:#sample
|
| 87 |
for histname in self.histograms[sample].keys():
|
| 88 |
self.histograms[sample][histname].Rebin( self.rebin )
|
| 89 |
if not sample in ['data', 'allMC']:
|
| 90 |
self.histograms[sample][histname].Scale( self.scale )
|
| 91 |
self.histograms[sample][histname].SetFillStyle( 1001 )
|
| 92 |
self.histograms[sample][histname].SetFillColor( colors[sample] )
|
| 93 |
if not mcStack.has_key( histname ):
|
| 94 |
mcStack[histname] = THStack( "MC_" + histname, "MC_" + histname );
|
| 95 |
mcStack[histname].Add( self.histograms[sample][histname] )
|
| 96 |
else:
|
| 97 |
self.histograms[sample][histname].SetMarkerStyle( 8 );
|
| 98 |
|
| 99 |
|
| 100 |
self.histograms['MCStack'] = mcStack
|
| 101 |
self.setStyle()
|
| 102 |
print "=" * 40
|
| 103 |
print "data integrated luminosity:", self.luminosity, self.luminosity_unit
|
| 104 |
print "MC integrated luminosity:", self.mc_luminosity, self.luminosity_unit
|
| 105 |
print "MC scale factor: ", self.scale
|
| 106 |
print '=' * 40
|
| 107 |
|
| 108 |
|
| 109 |
def setStyle( self ):
|
| 110 |
tdrstyle = setTDRStyle();
|
| 111 |
tdrstyle.SetPadRightMargin( 0.05 )#originally was 0.02, too narrow!
|
| 112 |
tdrStyle.SetStatH( 0.2 );
|
| 113 |
tdrStyle.SetOptStat( 0 );#off title
|
| 114 |
tdrStyle.SetOptFit( 0 );#off title
|
| 115 |
tdrStyle.cd();
|
| 116 |
gROOT.ForceStyle();
|
| 117 |
gStyle.SetTitleH( 0.1 );
|
| 118 |
gStyle.SetStatH( 0.22 ); #0.24);
|
| 119 |
gStyle.SetStatW( 0.22 ); #0.26);
|
| 120 |
gStyle.SetOptStat( 1 ); #on stat
|
| 121 |
gStyle.SetLineScalePS( 2 ); #D=3
|
| 122 |
gStyle.SetOptFit( 112 );
|
| 123 |
|
| 124 |
|
| 125 |
def doClosureTests( self, function = 'gaus' ):
|
| 126 |
self.useEntryAsData = 'allMC'
|
| 127 |
self.currentFitFuntion = function
|
| 128 |
|
| 129 |
for fitRange in self.fitRangesClosureTest:
|
| 130 |
self.currentFitRange = fitRange
|
| 131 |
for bin in self.jetBins:
|
| 132 |
self.currentJetBin = bin
|
| 133 |
self.EstimateJetBin( bin )
|
| 134 |
self.plot( self.pfIsoHistogramPrefix + bin, self.pfIsoResults[bin] )
|
| 135 |
self.plot( self.relIsoHistogramPrefix + bin, self.relIsoResults[bin] )
|
| 136 |
self.allPfIsoResults['%1.1f-%1.1f' % fitRange] = deepcopy( self.pfIsoResults )
|
| 137 |
self.allRelIsoResults['%1.1f-%1.1f' % fitRange] = deepcopy( self.relIsoResults )
|
| 138 |
self.plotClosureTest( self.pfIsoHistogramPrefix, self.allPfIsoResults )
|
| 139 |
self.plotClosureTest( self.relIsoHistogramPrefix, self.allRelIsoResults )
|
| 140 |
|
| 141 |
|
| 142 |
def doEstimate( self, function = 'gaus' ):
|
| 143 |
self.useEntryAsData = 'data'
|
| 144 |
self.currentFitFuntion = function
|
| 145 |
|
| 146 |
for fitRange in self.fitRangesEstimation:
|
| 147 |
self.currentFitRange = fitRange
|
| 148 |
for bin in self.jetBins:
|
| 149 |
self.currentJetBin = bin
|
| 150 |
self.EstimateJetBin( bin )
|
| 151 |
self.plot( self.pfIsoHistogramPrefix + bin, self.pfIsoResults[bin] )
|
| 152 |
self.plot( self.relIsoHistogramPrefix + bin, self.relIsoResults[bin] )
|
| 153 |
self.allPfIsoResults['%1.1f-%1.1f' % fitRange] = deepcopy( self.pfIsoResults )
|
| 154 |
self.allRelIsoResults['%1.1f-%1.1f' % fitRange] = deepcopy( self.relIsoResults )
|
| 155 |
|
| 156 |
def EstimateJetBin( self, jetbin ):
|
| 157 |
# fitRange = self.currentFitRange
|
| 158 |
function = self.currentFitFuntion
|
| 159 |
|
| 160 |
QCD = self.histograms['qcd']
|
| 161 |
data = self.histograms[self.useEntryAsData]
|
| 162 |
allMC = self.histograms['allMC']
|
| 163 |
pfIsoHist = self.pfIsoHistogramPrefix + jetbin
|
| 164 |
relIsoHist = self.relIsoHistogramPrefix + jetbin
|
| 165 |
|
| 166 |
self.pfIsoResults[jetbin]['actualNumberOfQCDEvents'] = QCD[pfIsoHist].GetBinContent(1)
|
| 167 |
self.pfIsoResults[jetbin]['numberOfAllDataEvents'] = data[pfIsoHist].GetBinContent(1)
|
| 168 |
self.pfIsoResults[jetbin]['numberOfAllMCEvents'] = allMC[pfIsoHist].GetBinContent(1)
|
| 169 |
|
| 170 |
pfIsoFit = self.doFit( data[pfIsoHist] )
|
| 171 |
self.pfIsoResults[jetbin]['fitFunction'] = pfIsoFit
|
| 172 |
self.pfIsoResults[jetbin]['fitParameters'] = self.getFitParameters( pfIsoFit )
|
| 173 |
estimate = 0
|
| 174 |
if pfIsoFit:
|
| 175 |
estimate = pfIsoFit.Integral( self.signalRegion[0], self.signalRegion[1] ) / ( self.binWidth * self.rebin )
|
| 176 |
|
| 177 |
self.pfIsoResults[jetbin]['estimatedNumberOfQCDEvents'] = estimate
|
| 178 |
|
| 179 |
#----------------------------------------------------------------------------------------------------------------------
|
| 180 |
self.relIsoResults[jetbin]['actualNumberOfQCDEvents'] = QCD[relIsoHist].GetBinContent(1)
|
| 181 |
self.relIsoResults[jetbin]['numberOfAllDataEvents'] = data[relIsoHist].GetBinContent(1)
|
| 182 |
self.relIsoResults[jetbin]['numberOfAllMCEvents'] = allMC[relIsoHist].GetBinContent(1)
|
| 183 |
|
| 184 |
relIsoFit = self.doFit( data[relIsoHist] )
|
| 185 |
self.relIsoResults[jetbin]['fitFunction'] = relIsoFit
|
| 186 |
self.relIsoResults[jetbin]['fitParameters'] = self.getFitParameters( relIsoFit )
|
| 187 |
|
| 188 |
estimate = relIsoFit.Integral( self.signalRegion[0], self.signalRegion[1] ) / ( self.binWidth * self.rebin )
|
| 189 |
self.relIsoResults[jetbin]['estimatedNumberOfQCDEvents'] = estimate
|
| 190 |
|
| 191 |
|
| 192 |
|
| 193 |
def doFit( self, histogram ):
|
| 194 |
function = self.currentFitFuntion
|
| 195 |
fitRange = self.currentFitRange
|
| 196 |
|
| 197 |
if not self.constrainFit:
|
| 198 |
histogram.Fit( function, "Q0", "ah", fitRange[0], fitRange[1] )
|
| 199 |
|
| 200 |
else:
|
| 201 |
ff = TF1( function, function, 0, 1 );
|
| 202 |
self.numberOfFreeParameters = ff.GetNumberFreeParameters();
|
| 203 |
|
| 204 |
return histogram.GetFunction( function )
|
| 205 |
|
| 206 |
def plot( self, histname, results ):
|
| 207 |
data = self.histograms[self.useEntryAsData][histname]
|
| 208 |
mcStack = self.histograms['MCStack'][histname]
|
| 209 |
fitFunction = results['fitFunction']
|
| 210 |
if not fitFunction:
|
| 211 |
print 'no fitfunction found'
|
| 212 |
return;
|
| 213 |
|
| 214 |
data.GetXaxis().SetRangeUser( 0, self.maxValue - 0.01 );
|
| 215 |
fitFunction.SetLineColor( kRed );
|
| 216 |
fitFunction.SetLineWidth( 2 )
|
| 217 |
|
| 218 |
fitFunction2 = fitFunction.Clone()
|
| 219 |
fitFunction2.SetLineColor( kBlue );
|
| 220 |
fitFunction2.SetRange( self.signalRegion[0], self.signalRegion[1] );
|
| 221 |
|
| 222 |
fitFunction3 = fitFunction.Clone()
|
| 223 |
fitFunction3.SetLineColor( kBlue );
|
| 224 |
fitFunction3.SetLineStyle( kDashed );
|
| 225 |
fitFunction3.SetRange( self.signalRegion[1], self.currentFitRange[0] );
|
| 226 |
|
| 227 |
canvas = TCanvas( "c1", "Iso fit", 1920, 1080 )
|
| 228 |
data.Draw();
|
| 229 |
|
| 230 |
max = 0
|
| 231 |
if mcStack.GetMaximum() > data.GetBinContent( 1 ):
|
| 232 |
max = mcStack.GetMaximum()*1.1
|
| 233 |
else:
|
| 234 |
max = data.GetBinContent( 1 ) * 1.1
|
| 235 |
|
| 236 |
data.GetYaxis().SetRangeUser( 0, max );
|
| 237 |
data.SetXTitle( "Relative Isolation" );
|
| 238 |
data.SetYTitle( "Events/0.1" );
|
| 239 |
# draw mc
|
| 240 |
mcStack.Draw( "hist same" );
|
| 241 |
data.Draw( "ae same" );
|
| 242 |
|
| 243 |
fitFunction.Draw( "same" );
|
| 244 |
fitFunction2.Draw( "same" );
|
| 245 |
fitFunction3.Draw( "same" );
|
| 246 |
|
| 247 |
label = self.add_cms_label( self.currentJetBin )
|
| 248 |
label.Draw()
|
| 249 |
|
| 250 |
legend = self.add_legend( histname )
|
| 251 |
# legend.Draw()
|
| 252 |
|
| 253 |
if self.currentFitFuntion == "pol1":
|
| 254 |
out = "%s_fit_linear_from_0%.0f_%s" % ( histname, self.currentFitRange[0] * 10.0, self.useEntryAsData );
|
| 255 |
else:
|
| 256 |
out = "%s_fit_%s_from_%1.1f_to_%1.1f_%s" % ( histname, self.currentFitFuntion, self.currentFitRange[0],
|
| 257 |
self.currentFitRange[1], self.useEntryAsData );
|
| 258 |
# if self.outputFormat == 'pdf':
|
| 259 |
# canvas.SaveAs( '%s.eps' % out );
|
| 260 |
# gROOT.ProcessLine( ".!ps2pdf -dEPSCrop %s.eps" % out );
|
| 261 |
# gROOT.ProcessLine( ".!rm -f %s.eps" % out );
|
| 262 |
# else:
|
| 263 |
canvas.SaveAs( '%s/%s.%s' % ( self.outputFolder, out, self.outputFormat ) )
|
| 264 |
|
| 265 |
canvas.Close(); #crucial!
|
| 266 |
|
| 267 |
|
| 268 |
def plotClosureTest( self, histname, results ):
|
| 269 |
c2 = TCanvas( "c2", "QCD estimates", 1080, 1080 );
|
| 270 |
x = array( 'd', [2, 3, 4] )
|
| 271 |
jetBinsOfInterest = [#'1jet',
|
| 272 |
'2jets', '3jets', '4orMoreJets']
|
| 273 |
function = self.currentFitFuntion
|
| 274 |
|
| 275 |
gStyle.SetMarkerSize( 1.7 );
|
| 276 |
gStyle.SetMarkerStyle( 20 );
|
| 277 |
c2.SetTopMargin( 0.1 );
|
| 278 |
# c2.SetLeftMargin( 0.12 );
|
| 279 |
c2.SetRightMargin( 0.35 );
|
| 280 |
|
| 281 |
y = {}
|
| 282 |
for fitRange in self.fitRangesClosureTest:
|
| 283 |
range = '%1.1f-%1.1f' % fitRange
|
| 284 |
y[range] = []
|
| 285 |
for bin in jetBinsOfInterest:
|
| 286 |
|
| 287 |
est = results[range][bin]['estimatedNumberOfQCDEvents']
|
| 288 |
true = results[range][bin]['actualNumberOfQCDEvents']
|
| 289 |
variation = 0
|
| 290 |
if not true == 0:
|
| 291 |
variation = ( est - true ) / true
|
| 292 |
y[range].append( variation )
|
| 293 |
nbins = 3
|
| 294 |
gr1 = TGraph( nbins, x, array( 'd', y['%1.1f-%1.1f' % self.fitRangesClosureTest[0]] ) )
|
| 295 |
gr2 = TGraph( nbins, x, array( 'd', y['%1.1f-%1.1f' % self.fitRangesClosureTest[1]] ) )
|
| 296 |
gr3 = TGraph( nbins, x, array( 'd', y['%1.1f-%1.1f' % self.fitRangesClosureTest[2]] ) )
|
| 297 |
gr4 = TGraph( nbins, x, array( 'd', y['%1.1f-%1.1f' % self.fitRangesClosureTest[3]] ) )
|
| 298 |
gr5 = TGraph( nbins, x, array( 'd', y['%1.1f-%1.1f' % self.fitRangesClosureTest[4]] ) )
|
| 299 |
gr6 = TGraph( nbins, x, array( 'd', y['%1.1f-%1.1f' % self.fitRangesClosureTest[5]] ) )
|
| 300 |
gr7 = TGraph( nbins, x, array( 'd', y['%1.1f-%1.1f' % self.fitRangesClosureTest[6]] ) )
|
| 301 |
gr8 = TGraph( nbins, x, array( 'd', y['%1.1f-%1.1f' % self.fitRangesClosureTest[7]] ) )
|
| 302 |
gr9 = TGraph( nbins, x, array( 'd', y['%1.1f-%1.1f' % self.fitRangesClosureTest[8]] ) )
|
| 303 |
|
| 304 |
gr1.SetMarkerColor( kGreen + 1 );
|
| 305 |
gr2.SetMarkerColor( kGreen + 2 );
|
| 306 |
gr3.SetMarkerColor( kGreen + 3 );
|
| 307 |
gr4.SetMarkerColor( kAzure + 7 );
|
| 308 |
gr5.SetMarkerColor( kAzure - 3 );
|
| 309 |
gr6.SetMarkerColor( kBlue );
|
| 310 |
gr7.SetMarkerColor( kOrange );
|
| 311 |
gr8.SetMarkerColor( kOrange - 1 );
|
| 312 |
gr9.SetMarkerColor( kOrange - 6 );
|
| 313 |
|
| 314 |
gStyle.SetTitleW( 0.9 );
|
| 315 |
gStyle.SetTitleH( 0.05 )
|
| 316 |
|
| 317 |
h = None
|
| 318 |
if function == "gaus":
|
| 319 |
h = TH1D( "h", "Variation of QCD estimates with fit range (Gaussian)", 4, 0.5, 4.5 );
|
| 320 |
elif function == "pol3":
|
| 321 |
h = TH1D( "h", "Variation of QCD estimates with fit range (Pol3)", 4, 0.5, 4.5 );
|
| 322 |
elif function == "pol1":
|
| 323 |
h = TH1D( "h", "Variation of QCD estimates with fit range (Pol1)", 4, 0.5, 4.5 );
|
| 324 |
elif function == "landau":
|
| 325 |
h = TH1D( "h", "Variation of QCD estimates with fit range (Landau)", 4, 0.5, 4.5 );
|
| 326 |
|
| 327 |
h.SetStats( kFALSE ); # no statistics
|
| 328 |
h.Draw();
|
| 329 |
h.SetYTitle( "Deviation = (N_{QCD,est}-N_{QCD,true})/N_{QCD,true}" );
|
| 330 |
h.GetYaxis().SetRangeUser( -1, 1 );
|
| 331 |
h.GetXaxis().SetRangeUser( 1.5, 5.5 );
|
| 332 |
# h.GetXaxis().SetBinLabel( 1, "1j" );
|
| 333 |
h.GetXaxis().SetBinLabel( 2, "2j" );
|
| 334 |
h.GetXaxis().SetBinLabel( 3, "3j" );
|
| 335 |
h.GetXaxis().SetBinLabel( 4, "#geq4j" );
|
| 336 |
h.GetXaxis().SetLabelSize( 0.07 );
|
| 337 |
h.GetYaxis().SetTitleOffset( 1.3 );
|
| 338 |
|
| 339 |
gr1.Draw( "P" );
|
| 340 |
gr2.Draw( "P" ); #to superimpose graphs, do not re-draw axis
|
| 341 |
gr3.Draw( "P" );
|
| 342 |
gr4.Draw( "P" );
|
| 343 |
gr5.Draw( "P" );
|
| 344 |
gr6.Draw( "P" );
|
| 345 |
gr7.Draw( "P" );
|
| 346 |
gr8.Draw( "P" );
|
| 347 |
gr9.Draw( "P" );
|
| 348 |
|
| 349 |
c2.SetGrid( 1, 1 );
|
| 350 |
|
| 351 |
leg = TLegend( 0.65, 0.1, 0.98, 0.9 );
|
| 352 |
leg.SetFillColor( 0 );
|
| 353 |
leg.AddEntry( gr1, "Free: 0.1-0.9", "p" );
|
| 354 |
leg.AddEntry( gr2, "Free: 0.1-1.0", "p" );
|
| 355 |
leg.AddEntry( gr3, "Free: 0.1-1.1", "p" );
|
| 356 |
leg.AddEntry( gr4, "Free: 0.2-0.9", "p" );
|
| 357 |
leg.AddEntry( gr5, "Free: 0.2-1.0", "p" );
|
| 358 |
leg.AddEntry( gr6, "Free: 0.2-1.1", "p" );
|
| 359 |
leg.AddEntry( gr7, "Free: 0.3-0.9", "p" );
|
| 360 |
leg.AddEntry( gr8, "Free: 0.3-1.0", "p" );
|
| 361 |
leg.AddEntry( gr9, "Free: 0.3-1.1", "p" );
|
| 362 |
|
| 363 |
leg.Draw();
|
| 364 |
|
| 365 |
c2.SaveAs( "%s/%s_qcd_estimate_%s.%s" % ( self.outputFolder, histname, function, self.outputFormat ) )
|
| 366 |
|
| 367 |
setRange = h.GetYaxis().SetRangeUser
|
| 368 |
saveAs = c2.SaveAs
|
| 369 |
for limit in [1, 2, 3, 6, 8]:
|
| 370 |
setRange( -limit, limit );
|
| 371 |
saveAs( "%s/%s_qcd_estimate_%s_zoom_%d.%s" % ( self.outputFolder, histname, function, limit, self.outputFormat ) );
|
| 372 |
|
| 373 |
def getFitParameters( self, fitFunction ):
|
| 374 |
fitParameters = {'chi2':-1, 'numberOfdegreesOfFreedom': 0, 'constrain1': 0, 'constrain2': 0,
|
| 375 |
'constrain3': 0, 'constrain4': 0}
|
| 376 |
if fitFunction:
|
| 377 |
fitParameters['chi2'] = fitFunction.GetChisquare()
|
| 378 |
fitParameters['numberOfdegreesOfFreedom'] = fitFunction.GetNDF()
|
| 379 |
return fitParameters
|
| 380 |
|
| 381 |
def doConstrainedFit( self, histogram, function = 'gaus', limits = ( 0.1, 1.6 ) ):
|
| 382 |
fitFunction = None
|
| 383 |
if function == 'gaus':
|
| 384 |
fitFunction = TF1( "gaus", "gaus", 0, 2 );
|
| 385 |
elif function == 'pol3':
|
| 386 |
fitFunction = TF1( "pol3", "[0] * ( 1 + [1]*x + [2]*x^2 + [3]*x^3 )", 0, 2 );
|
| 387 |
elif function == 'landau':
|
| 388 |
fitFunction = TF1( "landau", "landau", 0, 2 )
|
| 389 |
|
| 390 |
|
| 391 |
myFitResult = data.Fit( function, "Q0", "ah", limits[0], limits[1] );
|
| 392 |
|
| 393 |
def add_cms_label( self, njet = "" ):
|
| 394 |
|
| 395 |
mytext = TPaveText( 0.3, 0.8, 0.6, 0.93, "NDC" );
|
| 396 |
mytext.AddText( "CMS Preliminary" );
|
| 397 |
mytext.AddText( "%.1f pb^{-1} at #sqrt{s} = 7 TeV" % self.luminosity );
|
| 398 |
if njet != "":
|
| 399 |
mytext.AddText( "e+jets, %s" % self.jetBinsLatex[njet] )
|
| 400 |
mytext.SetFillStyle( 0 );
|
| 401 |
mytext.SetBorderSize( 0 );
|
| 402 |
mytext.SetTextFont( 42 );
|
| 403 |
mytext.SetTextAlign( 13 );
|
| 404 |
return mytext
|
| 405 |
|
| 406 |
def add_legend( self, histname ):
|
| 407 |
function = self.currentFitFuntion
|
| 408 |
|
| 409 |
tt = self.histograms['ttbar'][histname]
|
| 410 |
wj = self.histograms['wjets'][histname]
|
| 411 |
zj = self.histograms['zjets'][histname]
|
| 412 |
data = self.histograms['data'][histname]
|
| 413 |
QCD = self.histograms['qcd'][histname]
|
| 414 |
stop = self.histograms['singleTop'][histname]
|
| 415 |
|
| 416 |
leg = TLegend( 0.64, 0.4, 0.9, 0.9 );
|
| 417 |
leg.SetFillStyle( 0 );
|
| 418 |
leg.SetBorderSize( 0 );
|
| 419 |
leg.SetTextFont( 42 );
|
| 420 |
|
| 421 |
# Here I define coloured lines for use in the legend
|
| 422 |
blue = TF1( "blue", "pol0", 0, 1 );
|
| 423 |
red = TF1( "red", "pol0", 0, 1 );
|
| 424 |
|
| 425 |
blue.SetLineColor( kBlue );
|
| 426 |
red.SetLineColor( kRed );
|
| 427 |
|
| 428 |
red.SetLineWidth( 2 );
|
| 429 |
blue.SetLineWidth( 2 );
|
| 430 |
|
| 431 |
blue.SetLineStyle( kDashed );
|
| 432 |
|
| 433 |
# Add entry to legend
|
| 434 |
if self.useEntryAsData == 'data':
|
| 435 |
leg.AddEntry( data, "Data", "LP" );
|
| 436 |
else:
|
| 437 |
leg.AddEntry( data, "All MC events", "LP" );
|
| 438 |
if function == "pol1":
|
| 439 |
leg.AddEntry( red, "Linear Fit", "l" );
|
| 440 |
elif function == "expo":
|
| 441 |
leg.AddEntry( red, "Exponenetial Fit", "l" );
|
| 442 |
elif function == "gaus":
|
| 443 |
leg.AddEntry( red, "Gaussian Fit", "l" );
|
| 444 |
|
| 445 |
leg.AddEntry( blue, "Extrapolation", "l" );
|
| 446 |
leg.AddEntry( tt, "t#bar{t}", "F" );
|
| 447 |
leg.AddEntry( stop, "Single-Top", "F" );
|
| 448 |
leg.AddEntry( wj, "W#rightarrowl#nu", "F" );
|
| 449 |
leg.AddEntry( zj, "Z/#gamma*#rightarrowl^{+}l^{-}", "F" );
|
| 450 |
leg.AddEntry( QCD, "QCD & #gamma+jets", "F" );
|
| 451 |
leg.Draw()
|
| 452 |
return ( leg, red, blue )
|
| 453 |
|
| 454 |
def printResults( self, results ):
|
| 455 |
self.printJetBinResults( results, '3jets' )
|
| 456 |
print '=' * 60
|
| 457 |
self.printJetBinResults( results, '3orMoreJets' )
|
| 458 |
print '=' * 60
|
| 459 |
self.printJetBinResults( results, '4orMoreJets' )
|
| 460 |
print '=' * 60
|
| 461 |
|
| 462 |
|
| 463 |
def printJetBinResults( self, results, jetBin ):
|
| 464 |
estimate = 0
|
| 465 |
estimate2 = 0
|
| 466 |
predicted = results[results.keys()[0]][jetBin]['actualNumberOfQCDEvents']
|
| 467 |
allData = results[results.keys()[0]][jetBin]['numberOfAllDataEvents']
|
| 468 |
allMC = results[results.keys()[0]][jetBin]['numberOfAllMCEvents']
|
| 469 |
|
| 470 |
if jetBin == '4orMoreJets':
|
| 471 |
print 'Estimation for >= 4 jets'
|
| 472 |
elif jetBin == '3orMoreJets':
|
| 473 |
print 'Estimation for >= 3 jets'
|
| 474 |
elif jetBin == '3jets':
|
| 475 |
print 'Estimation for == 3 jets'
|
| 476 |
print 'predicted number of QCD events', predicted
|
| 477 |
for fitRange in self.fitRangesEstimation:
|
| 478 |
range = '%1.1f-%1.1f' % fitRange
|
| 479 |
est = results[range][jetBin]['estimatedNumberOfQCDEvents']
|
| 480 |
true = results[range][jetBin]['actualNumberOfQCDEvents']
|
| 481 |
variation = ( est - true ) / true
|
| 482 |
estimate += est
|
| 483 |
estimate2 += est * est
|
| 484 |
|
| 485 |
|
| 486 |
print
|
| 487 |
print 'estimated number of QCD events'
|
| 488 |
print 'for range', range, ': ',
|
| 489 |
print est
|
| 490 |
print
|
| 491 |
mean = estimate / len( self.fitRangesEstimation )
|
| 492 |
mean2 = estimate2 / len( self.fitRangesEstimation )
|
| 493 |
error = sqrt( ( mean2 - mean * mean ) / len( self.fitRangesEstimation ) )
|
| 494 |
print 'average estimate', mean, '+-', error
|
| 495 |
weight = estimate / len( self.fitRangesEstimation ) / predicted
|
| 496 |
print 'average weight factor', weight
|
| 497 |
print 'Total number of data in signal bin(<0.1)', allData
|
| 498 |
print 'Total number of MC in signal bin(<0.1) before reweighting QCD', allMC
|
| 499 |
print 'Total number of MC in signal bin(<0.1) after reweighting QCD', ( allMC - predicted ) + predicted * weight
|
| 500 |
|
| 501 |
def plotControlRegionComparison( self ):
|
| 502 |
for bin in self.jetBins:
|
| 503 |
hist = self.pfIsoHistogramPrefix + bin
|
| 504 |
histControl = self.pfIsoControlRegionHistogramPrefix + bin
|
| 505 |
QCD_control = self.histograms['qcd'][histControl]
|
| 506 |
QCD = self.histograms['qcd'][hist]
|
| 507 |
nQCD_Control = QCD_control.Integral()
|
| 508 |
nQCD = QCD.Integral()
|
| 509 |
if nQCD_Control > 0:
|
| 510 |
QCD_control.Scale( 1 / nQCD_Control )
|
| 511 |
if nQCD > 0:
|
| 512 |
QCD.Scale( 1 / nQCD )
|
| 513 |
QCD_control.SetFillStyle( 3004 )
|
| 514 |
QCD.GetXaxis().SetRangeUser( 0., self.maxValue - 0.01 )
|
| 515 |
|
| 516 |
max = 0
|
| 517 |
if QCD.GetMaximum() > QCD_control.GetMaximum():
|
| 518 |
max = QCD.GetMaximum()*1.1
|
| 519 |
else:
|
| 520 |
max = QCD_control.GetMaximum()*1.1
|
| 521 |
|
| 522 |
QCD.GetYaxis().SetRangeUser( 0., max )
|
| 523 |
|
| 524 |
canvas = TCanvas( "c1", "Shape comparision", 1920, 1080 )
|
| 525 |
QCD.Draw()
|
| 526 |
QCD_control.Draw( 'same' )
|
| 527 |
label = self.add_cms_label( bin )
|
| 528 |
label.Draw()
|
| 529 |
|
| 530 |
leg = TLegend( 0.64, 0.4, 0.9, 0.9 );
|
| 531 |
leg.SetFillStyle( 0 );
|
| 532 |
leg.SetBorderSize( 0 );
|
| 533 |
leg.SetTextFont( 42 );
|
| 534 |
leg.AddEntry( QCD_control, 'QCD control region' )
|
| 535 |
leg.AddEntry( QCD, 'QCD standard selection' )
|
| 536 |
leg.Draw()
|
| 537 |
|
| 538 |
canvas.SaveAs( '%s/%s_shapeComparison.%s' % ( self.outputFolder, hist, self.outputFormat ) )
|
| 539 |
|
| 540 |
if __name__ == '__main__':
|
| 541 |
gROOT.SetBatch( True )
|
| 542 |
gROOT.ProcessLine( 'gErrorIgnoreLevel = 1001;' )
|
| 543 |
|
| 544 |
path = '/storage/results/outputfiles/Preapproval2/'
|
| 545 |
files = {'data': path + "data_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 546 |
'ttbar' : path + "ttjet_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 547 |
'wjets' : path + "wj_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 548 |
'zjets' : path + "zj_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 549 |
'bce1' : path + "bce1_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 550 |
'bce2' : path + "bce2_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 551 |
'bce3' : path + "bce3_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 552 |
'enri1' : path + "enri1_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 553 |
'enri2' : path + "enri2_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 554 |
'enri3' : path + "enri3_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 555 |
'pj1' : path + "pj1_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 556 |
'pj2' : path + "pj2_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 557 |
'pj3' : path + "pj3_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 558 |
'tW' : path + "tW_36.145pb_PFElectron_PF2PATJets_PFMET.root",
|
| 559 |
'tchan' : path + "tchan_36.145pb_PFElectron_PF2PATJets_PFMET.root"}
|
| 560 |
q = QCDEstimator( files )
|
| 561 |
QCDEstimator.outputFolder = '/storage/results/QCD_Estimate/Preapproval2'
|
| 562 |
QCDEstimator.outputFormat = 'pdf'
|
| 563 |
|
| 564 |
q.doEstimate( 'gaus' )
|
| 565 |
print '=' * 60
|
| 566 |
print 'ParticleFlowIsolation results'
|
| 567 |
q.printResults( q.allPfIsoResults )
|
| 568 |
# print 'Relative isolation results'
|
| 569 |
# q.printResults( q.allRelIsoResults )
|
| 570 |
# print '=' * 60
|
| 571 |
|
| 572 |
|
| 573 |
print 'Starting closure tests'
|
| 574 |
q.doClosureTests( 'gaus' )
|
| 575 |
q.plotControlRegionComparison()
|
| 576 |
|
