def main():
# User controlled arguments
parser = argparse.ArgumentParser()
parser.add_argument("--inFileName",
type=str,
default="",
help="The path to the input file from runBumpHunter")
parser.add_argument(
"--outPath",
type=str,
default="",
help="The path prefix (directory) where you want the output plots")
parser.add_argument("--lumi", type=float, default=1, help="Luminosity")
parser.add_argument("--overlaidSignal",
action='store_true',
help="Overlaid Signal on Background")
parser.add_argument("--signalFileName",
type=str,
default="",
help="Signal histogram overlaid on background")
parser.add_argument("--drawMCComparison",
action='store_true',
help="Draw the comparison between data and MC")
parser.add_argument("--mcFileName",
type=str,
default="",
help="MC File Name")
args = parser.parse_args()
inFileName = args.inFileName
outPath = args.outPath
luminosity = args.lumi * 1000
overlaidSignal = args.overlaidSignal
signalFileName = args.signalFileName
drawMCComparison = args.drawMCComparison
mcFileName = args.mcFileName
print "==================================="
print "inputFile : ", inFileName
print "outPath : ", outPath
print "Luminosity : ", args.lumi
print "OverlaidSignal : ", overlaidSignal
print "Signal File : ", signalFileName
print "drawMCComparison: ", drawMCComparison
print "mcFileName: ", mcFileName
print "==================================="
# Get input root file
inFile = ROOT.TFile.Open(inFileName, "READ")
if not inFile:
print inFileName, " doesn't exist."
return
# make plots folder i.e. make folder extension
if not os.path.exists(outPath):
os.makedirs(outPath)
# Define necessary quantities.
Ecm = 13
# Initialize painter
myPainter = Morisot()
myPainter.setColourPalette("Teals")
#myPainter.setEPS(True)
myPainter.setLabelType(
1) # Sets label type i.e. Internal, Work in progress etc.
# 0 Just ATLAS
# 1 "Preliminary"
# 2 "Internal"
# 3 "Simulation Preliminary"
# 4 "Simulation Internal"
# 5 "Simulation"
# 6 "Work in Progress"
# Retrieve search phase inputs
basicData = inFile.Get("basicData")
basicBkg = inFile.Get("basicBkg")
residualHist = inFile.Get("residualHist")
logLikelihoodPseudoStatHist = inFile.Get("logLikelihoodStatHistNullCase")
chi2PseudoStatHist = inFile.Get("chi2StatHistNullCase")
bumpHunterStatHist = inFile.Get("bumpHunterStatHistNullCase")
bumpHunterTomographyPlot = inFile.Get(
'bumpHunterTomographyFromPseudoexperiments')
bumpHunterStatOfFitToData = inFile.Get('bumpHunterStatOfFitToData')
logLOfFitToDataVec = inFile.Get('logLOfFitToData')
chi2OfFitToDataVec = inFile.Get('chi2OfFitToData')
statOfFitToData = inFile.Get('bumpHunterPLowHigh')
logLOfFitToData = logLOfFitToDataVec[0]
logLPVal = logLOfFitToDataVec[1]
chi2OfFitToData = chi2OfFitToDataVec[0]
chi2PVal = chi2OfFitToDataVec[1]
bumpHunterStatFitToData = statOfFitToData[0]
bumpHunterPVal = bumpHunterStatOfFitToData[1]
bumpLowEdge = statOfFitToData[1]
bumpHighEdge = statOfFitToData[2]
print "logL of fit to data is", logLOfFitToData
print "logL pvalue is", logLPVal
print "chi2 of fit to data is", chi2OfFitToData
print "chi2 pvalue is", chi2PVal
print "bump hunter stat of fit to data is", bumpHunterStatFitToData
print "bumpLowEdge, bumpHighEdge are", bumpLowEdge, bumpHighEdge
print "BumpHunter pvalue is", bumpHunterPVal
print "which is Z value of", GetZVal(bumpHunterPVal, True)
# Find range
# Calculate from fit range
fitRange = inFile.Get("FitRange")
firstBin = basicData.FindBin(fitRange[0]) - 1
lastBin = basicData.FindBin(fitRange[1])
print "firstbin, lastbin: ", firstBin, lastBin
print "First bin = ", firstBin, ": lower edge at", basicData.GetBinLowEdge(
firstBin)
print "Last bin = ", lastBin, ": higher edge at", basicData.GetBinLowEdge(
lastBin) + basicData.GetBinWidth(lastBin)
# Convert plots into desired final form
standardbins = basicData.GetXaxis().GetXbins()
newbins = [] #ROOT.TArrayD(standardbins.GetSize())
for np in range(standardbins.GetSize()):
newbins.append(standardbins[np] / 1000)
# Make never versions of old plots
newbasicdata = ROOT.TH1D("basicData_TeV", "basicData_TeV",
len(newbins) - 1, array('d', newbins))
newbasicBkg = ROOT.TH1D("basicBkg_TeV", "basicBkg_TeV",
len(newbins) - 1, array('d', newbins))
newresidualHist = ROOT.TH1D("residualHist_TeV", "residualHist_TeV",
len(newbins) - 1, array('d', newbins))
for histnew, histold in [[newbasicdata, basicData],
[newbasicBkg, basicBkg],
[newresidualHist, residualHist]]:
for bin in range(histnew.GetNbinsX() + 2):
histnew.SetBinContent(bin, histold.GetBinContent(bin))
histnew.SetBinError(bin, histold.GetBinError(bin))
# Significances for Todd
ToddSignificancesHist = ROOT.TH1D("ToddSignificancesHist",
"ToddSignificancesHist", 100, -5, 5)
for bin in range(0, newresidualHist.GetNbinsX() + 1):
if bin < firstBin: continue
if bin > lastBin: continue
residualValue = newresidualHist.GetBinContent(bin)
#print residualValue
ToddSignificancesHist.Fill(residualValue)
myPainter.drawBasicHistogram(ToddSignificancesHist, -1, -1, "Residuals",
"Entries",
"{0}/ToddSignificancesHist".format(outPath))
# Search phase plots
myPainter.drawDataAndFitOverSignificanceHist(newbasicdata,newbasicBkg,newresidualHist,\
'm_{jj} [TeV]','Events','Significance','{0}/figure1'.format(outPath),\
luminosity,13,fitRange[0],fitRange[1],firstBin,lastBin,True,\
bumpLowEdge/1000.0,bumpHighEdge/1000.0,[],True,False,[],True,bumpHunterPVal)
myPainter.drawDataAndFitOverSignificanceHist(newbasicdata,newbasicBkg,newresidualHist,\
'm_{jj} [TeV]','Events','Significance','{0}/figure1_nobump'.format(outPath),\
luminosity,13,fitRange[0],fitRange[1],firstBin,lastBin,False,\
bumpLowEdge,bumpHighEdge,[],True,False,[],True,bumpHunterPVal)
myPainter.drawPseudoExperimentsWithObservedStat(logLikelihoodPseudoStatHist,\
float(logLOfFitToData),logLPVal,0,luminosity,13,\
'logL statistic','Pseudo-exeperiments',"{0}/logLStatPlot".format(outPath))
myPainter.drawPseudoExperimentsWithObservedStat(chi2PseudoStatHist,
float(chi2OfFitToData),chi2PVal,0,luminosity,13,\
"#chi^{2}",'Pseudo-exeperiments',"{0}/chi2StatPlot".format(outPath))
myPainter.drawPseudoExperimentsWithObservedStat(bumpHunterStatHist,
float(bumpHunterStatFitToData),bumpHunterPVal,0,luminosity,13,\
'BumpHunter','Pseudo-exeperiments',"{0}/bumpHunterStatPlot".format(outPath))
myPainter.drawBumpHunterTomographyPlot(
bumpHunterTomographyPlot,
"{0}/bumpHunterTomographyPlot".format(outPath))
####### Draw Signal Template overlaid on Background###########
if overlaidSignal:
signalFile = ROOT.TFile.Open(signalFileName, "read")
if not signalFile:
print signalFileName, " doesn't exist!!!!"
return
# setup signal information
signalTitles = {"QStar": "#it{q}*"}
signalTypes = ["QStar"]
signalsMasses = {"QStar": [4000, 5000]}
signalScalingFactors = {"QStar": 0.1}
signalAxes = {
"QStar": {
"X": "M_{#it{q}*} [GeV]",
"Y": "#sigma #times #it{A} #times BR [pb]"
}
}
for signalType in signalTypes:
print "in signal", signalType
signalMasses = signalsMasses[signalType]
signalMassesTeV = signalsMasses[signalType][:]
for index in range(len(signalMasses)):
signalMassesTeV[index] = signalMasses[index] / 1000.0
print signalMassesTeV
signalPlotsTeV = []
legendlistTeV = []
for mass in signalMasses:
#sigplot = signalFile.Get("mjj_Scaled_"+signalType+"{0}_10fb".format(mass))
sigplot = signalFile.Get("h_mjj_{0}".format(mass))
#sigplot.Scale(luminosity/1000.)
#sigplot.Scale(luminosity/1000./10*139)
sigplot.SetDirectory(0)
sigplottev = newbasicdata.Clone()
sigplottev.SetName("sigplot_{0}_{1}_TeV".format(
signalType, mass))
for bins in range(sigplot.GetNbinsX() + 2):
for bin in range(sigplottev.GetNbinsX() + 2):
if sigplot.GetBinLowEdge(
bins) / 1000. == sigplottev.GetBinLowEdge(bin):
sigplottev.SetBinContent(
bin, sigplot.GetBinContent(bins))
sigplottev.SetBinError(bin,
sigplot.GetBinError(bins))
sigplotforfitplusbkg = sigplottev.Clone()
sigplotforfitplusbkg.SetDirectory(0)
sigplotforfitplusbkg.SetName(sigplottev.GetName() +
"_forfitplusbkg_TeV")
sigplotforfitplusbkg.Scale(signalScalingFactors[signalType])
signalPlotsTeV.append(sigplotforfitplusbkg)
thistitle = signalTitles[signalType] + ", {0}= {1} TeV".format(
signalAxes[signalType]["X"].split("[GeV]")[0].replace(
"M", "m"), mass / 1000.0)
legendlistTeV.append(thistitle)
extLastBin = lastBin
for bin in range(sigplotforfitplusbkg.GetNbinsX()):
if bin > extLastBin and sigplotforfitplusbkg.GetBinContent(
bin) > 0.01:
extLastBin = bin
if sigplotforfitplusbkg.GetBinLowEdge(
bin) > 1.3 * mass / 1000.0:
continue
if extLastBin < lastBin:
extLastBin = lastBin
UserScaleText = signalTitles[signalType]
if signalScalingFactors[signalType] == 1:
UserScaleText = signalTitles[signalType]
else:
UserScaleText = UserScaleText + ", #sigma #times " + str(
signalScalingFactors[signalType])
outputName = outPath + "FancyFigure1_" + signalType
myPainter.drawDataAndFitWithSignalsOverSignificances(newbasicdata,newbasicBkg, None,\
newresidualHist,signalPlotsTeV, None, signalMassesTeV,legendlistTeV,\
"m_{jj} [TeV]","Events","","Significance ", outputName,luminosity,\
Ecm, firstBin,extLastBin,\
True, bumpLowEdge/1000,bumpHighEdge/1000,\
True,False,True,UserScaleText,True,bumpHunterPVal, True, \
fitRange[0], fitRange[1])
outputName = outPath + "FancyFigure1_" + signalType + "_nologx"
myPainter.drawDataAndFitWithSignalsOverSignificances(newbasicdata,newbasicBkg, None,\
newresidualHist,signalPlotsTeV, None, signalMassesTeV,legendlistTeV,\
"m_{jj} [TeV]","Events","","Significance ", outputName,luminosity,\
Ecm, firstBin,extLastBin,\
True, bumpLowEdge/1000,bumpHighEdge/1000,\
False,False,True,UserScaleText,True,bumpHunterPVal, True, \
fitRange[0], fitRange[1])
outputName = outPath + "FancyFigure1_" + signalType + "_noBump"
myPainter.drawDataAndFitWithSignalsOverSignificances(newbasicdata,newbasicBkg, None,\
newresidualHist,signalPlotsTeV, None, signalMassesTeV,legendlistTeV,\
"m_{jj} [TeV]","Events","","Significance ", outputName,luminosity,\
Ecm, firstBin,extLastBin,\
False, bumpLowEdge/1000,bumpHighEdge/1000,\
True,False,True,UserScaleText,True,bumpHunterPVal, True, \
fitRange[0], fitRange[1])
outputName = outPath + "FancyFigure1_" + signalType + "_noBump_nologx"
myPainter.drawDataAndFitWithSignalsOverSignificances(newbasicdata,newbasicBkg, None,\
newresidualHist,signalPlotsTeV, None, signalMassesTeV,legendlistTeV,\
"m_{jj} [TeV]","Events","","Significance ", outputName,luminosity,\
Ecm, firstBin,extLastBin,\
False, bumpLowEdge/1000,bumpHighEdge/1000,\
False,False,True,UserScaleText,True,bumpHunterPVal, True, \
fitRange[0], fitRange[1])
###############################
# Draw the comparison between data and MC in the bottom panel
if drawMCComparison:
mcFile = ROOT.TFile(mcFileName, "read")
if not mcFile:
print "Can not open: ", mcFileName
return
mchist_nominal = mcFile.Get("djet_mjj_nominal")
mchist_jesup = mcFile.Get("djet_mjj_JES_up")
mchist_jesdown = mcFile.Get("djet_mjj_JES_down")
newmchist_nominal = ROOT.TH1D("djet_mjj_nominal_TeV",
"djet_mjj_nominal_TeV",
len(newbins) - 1,
array('d', newbins))
newmchist_jesup = ROOT.TH1D("djet_mjj_jesup_TeV",
"djet_mjj_jesup_TeV",
len(newbins) - 1,
array('d', newbins))
newmchist_jesdown = ROOT.TH1D("djet_mjj_jesdown_TeV",
"djet_mjj_jesdown_TeV",
len(newbins) - 1,
array('d', newbins))
for iBin1 in range(1, newmchist_nominal.GetNbinsX() + 1):
for iBin2 in range(1, mchist_nominal.GetNbinsX() + 1):
if newmchist_nominal.GetBinLowEdge(
iBin1) * 1000 == mchist_nominal.GetBinLowEdge(
iBin2):
newmchist_nominal.SetBinContent(
iBin1, mchist_nominal.GetBinContent(iBin2))
newmchist_nominal.SetBinError(
iBin1, mchist_nominal.GetBinError(iBin2))
continue
for iBin1 in range(1, newmchist_jesup.GetNbinsX() + 1):
for iBin2 in range(1, mchist_jesup.GetNbinsX() + 1):
if newmchist_jesup.GetBinLowEdge(
iBin1) * 1000 == mchist_jesup.GetBinLowEdge(
iBin2):
newmchist_jesup.SetBinContent(
iBin1, mchist_jesup.GetBinContent(iBin2))
newmchist_jesup.SetBinError(
iBin1, mchist_jesup.GetBinError(iBin2))
continue
for iBin1 in range(1, newmchist_jesdown.GetNbinsX() + 1):
for iBin2 in range(1, mchist_jesdown.GetNbinsX() + 1):
if newmchist_jesdown.GetBinLowEdge(
iBin1) * 1000 == mchist_jesdown.GetBinLowEdge(
iBin2):
newmchist_jesdown.SetBinContent(
iBin1, mchist_jesdown.GetBinContent(iBin2))
newmchist_jesdown.SetBinError(
iBin1, mchist_jesdown.GetBinError(iBin2))
continue
tmpRatioHist = newbasicdata.Clone()
tmpRatioHist.SetMarkerColor(ROOT.kBlack)
tmpRatioHist.Add(newmchist_nominal, -1)
tmpRatioHist.Divide(newmchist_nominal)
## If data is 0 then there should be no ratio drawn
for iBin in range(1, tmpRatioHist.GetNbinsX() + 1):
if newbasicdata.GetBinContent(iBin) == 0:
tmpRatioHist.SetBinContent(iBin, 0)
tmpRatioHist.SetBinError(iBin, 0)
UpDownRatioHists = []
if mchist_jesup.GetEntries() >= 0:
tmpJESRatioHist = newmchist_jesup
tmpJESRatioHist.Add(newmchist_nominal, -1.)
tmpJESRatioHist.Divide(newmchist_nominal)
tmpJESRatioHist.SetMarkerColorAlpha(ROOT.kBlue, 0.15)
tmpJESRatioHist.SetLineColorAlpha(ROOT.kBlue, 0.15)
tmpJESRatioHist.SetFillColorAlpha(ROOT.kBlue, 0.15)
tmpJESRatioHist.SetFillStyle(1001)
UpDownRatioHists.append(tmpJESRatioHist)
if mchist_jesdown.GetEntries() >= 0:
tmpJESRatioHist = newmchist_jesdown
tmpJESRatioHist.Add(newmchist_nominal, -1.)
tmpJESRatioHist.Divide(newmchist_nominal)
tmpJESRatioHist.SetMarkerColorAlpha(ROOT.kBlue, 0.15)
tmpJESRatioHist.SetLineColorAlpha(ROOT.kBlue, 0.15)
tmpJESRatioHist.SetFillColorAlpha(ROOT.kBlue, 0.15)
tmpJESRatioHist.SetFillStyle(1001)
UpDownRatioHists.append(tmpJESRatioHist)
outputName = outPath + "FancyFigure1_" + signalType + "_WithMCRatio"
myPainter.drawDataAndFitWithSignalsOverSignificancesWithMCRatio(newbasicdata,newbasicBkg,None,\
newresidualHist, signalPlotsTeV, [], signalMassesTeV,legendlistTeV,\
"m_{jj} [TeV]","Events","#frac{Data-MC}{MC}","Significance",\
outputName,luminosity,Ecm,firstBin,lastBin,True,bumpLowEdge/1000.0,bumpHighEdge/1000.0,\
True,False,False, UserScaleText,True,bumpHunterPVal,True,fitRange[0],fitRange[1],\
newmchist_nominal,tmpRatioHist,UpDownRatioHists[0],UpDownRatioHists[1])
outputName = outPath + "FancyFigure1_" + signalType + "_WithMCRatio_nologx"
myPainter.drawDataAndFitWithSignalsOverSignificancesWithMCRatio(newbasicdata,newbasicBkg,None, \
newresidualHist, signalPlotsTeV, [],signalMassesTeV,legendlistTeV, \
"m_{jj} [TeV]","Events","#frac{Data-MC}{MC}","Significance",\
outputName,luminosity,Ecm,firstBin,lastBin,True,bumpLowEdge/1000.0,bumpHighEdge/1000.0,\
False,False,False, UserScaleText,True,bumpHunterPVal,True,fitRange[0],fitRange[1],\
newmchist_nominal,tmpRatioHist,UpDownRatioHists[0],UpDownRatioHists[1])
inFile.Close()
print "Done."
residualValue = newresidualHist.GetBinContent(bin)
#print residualValue
ToddSignificancesHist.Fill(residualValue)
myPainter.drawBasicHistogram(ToddSignificancesHist,-1,-1,"Residuals","Entries","{0}/ToddSignificancesHist".format(folderextension))
print("fitRange0: ", fitRange[0])
print("fitRange1: ", fitRange[1])
# Search phase plots
myPainter.drawDataAndFitOverSignificanceHist(newbasicdata,newbasicBkgFrom4ParamFit,newresidualHist,\
'm_{jj} [TeV]','Events','Significance','{0}/figure1'.format(folderextension),\
luminosity,13,fitRange[0],fitRange[1],firstBin,lastBin-1,True,bumpLowEdge/1000.0,bumpHighEdge/1000.0,[],True,False,[],True,bumpHunterPVal,1,2E12)
myPainter.drawDataAndFitOverSignificanceHist(newbasicdata,newbasicBkgFrom4ParamFit,newresidualHist,\
'm_{jj} [TeV]','Prescale-weighted events','Significance','{0}/figure1_nobump'.format(folderextension),\
luminosity,13,fitRange[0],fitRange[1],firstBin,lastBin-1,False,bumpLowEdge,bumpHighEdge,[],True,False,[],True,bumpHunterPVal,5E3,5E6)
myPainter.drawPseudoExperimentsWithObservedStat(logLikelihoodPseudoStatHist,float(logLOfFitToData),logLPVal,0,luminosity,13,\
'logL statistic','Pseudo-exeperiments',"{0}/logLStatPlot".format(folderextension))
myPainter.drawPseudoExperimentsWithObservedStat(chi2PseudoStatHist,float(chi2OfFitToData),chi2PVal,0,luminosity,13,\
"#chi^{2}",'Pseudo-exeperiments',"{0}/chi2StatPlot".format(folderextension))
#myPainter.drawPseudoExperimentsWithObservedStat(bumpHunterStatHist,float(bumpHunterStatFitToData),bumpHunterPVal,0,luminosity,13,\
# 'BumpHunter','Pseudo-exeperiments',"{0}/bumpHunterStatPlot".format(folderextension))
myPainter.drawPseudoExperimentsWithObservedStat(bumpHunterStatHist,float(bumpHunterStatFitToData),bumpHunterPVal,0,luminosity,13,\
'BumpHunter test statistic','Pseudo-experiments',"{0}/bumpHunterStatPlot".format(folderextension))
myPainter.drawBumpHunterTomographyPlot(bumpHunterTomographyPlot,"{0}/bumpHunterTomographyPlot".format(folderextension),luminosity,13,303,611)
print "For Note: Bin range, N(Obs),N(exp)_Nom,N(exp_Alt)"
for bin in range(newbasicdata.GetNbinsX()+2) :
sys.stdout.write('{0}-{1}'.format(int(newbasicdata.GetBinLowEdge(bin)*1000),int(newbasicdata.GetBinLowEdge(bin+1)*1000)))
sys.stdout.write(" & ")
sys.stdout.write('{0}'.format(int(newbasicdata.GetBinContent(bin))))
sys.stdout.write(" & ")
def main():
# User controlled arguments
parser = argparse.ArgumentParser()
parser.add_argument("--inFileName", type=str, default="", help="The path to the input file from SearchPhase")
parser.add_argument("--outPath", type=str, default="./plotting/SearchPhase/plots/", help="The path prefix (directory) where you want the output plots")
parser.add_argument("--lumi", type=float, default=1, help="Luminosity in fb-1")
parser.add_argument("--doAlternate", action='store_true', help="Compare Nominal and Alternate Fit")
parser.add_argument("--overlaidSignal", action='store_true', help="Overlaid Signal on Background")
parser.add_argument("--signalFileName", type=str, default="", help="Signal histogram overlaid on background")
parser.add_argument("--drawMCComparison", action='store_true', help="Draw the comparison between data and MC")
parser.add_argument("--mcFileName", type=str, default="", help="MC File Name")
args = parser.parse_args()
inFileName = args.inFileName
outPath = args.outPath
luminosity = 1000*args.lumi
doAlternate = args.doAlternate
overlaidSignal = args.overlaidSignal
signalFileName = args.signalFileName
drawMCComparison = args.drawMCComparison
mcFileName = args.mcFileName
print "==================================="
print "Executing Run_SearchPhase.py with :"
print "inFileName : ",inFileName
print "outPath : ",outPath
print "Lumi : ", args.lumi
print "doAlternate: ", doAlternate
print "overlaidSignal: ", overlaidSignal
if overlaidSignal:
print " signalFileName: ", signalFileName
print "drawMCComparison: ", drawMCComparison
print "mcFileName: ", mcFileName
print "==================================="
# Get input (the rootfile name should match that specified in the SearchPhase.config file)
searchInputFile = ROOT.TFile(inFileName, "READ")
# make plots folder i.e. make folder extension
if not os.path.exists(outPath):
os.makedirs(outPath)
# Define necessary quantities.
Ecm = 13
# Get input
doStatSearch = False
#doAlternate = True # You can use this if you included the alternate fit function in the search phase
# Initialize painter
myPainter = Morisot()
myPainter.setColourPalette("Teals")
#myPainter.setEPS(True)
myPainter.setLabelType(2) # Sets label type i.e. Internal, Work in progress etc.
# See below for label explanation
# 0 Just ATLAS
# 1 "Preliminary"
# 2 "Internal"
# 3 "Simulation Preliminary"
# 4 "Simulation Internal"
# 5 "Simulation"
# 6 "Work in Progress"
# Retrieve search phase inputs
basicData = searchInputFile.Get("basicData")
basicBkg = searchInputFile.Get("basicBkgFrom4ParamFit")
#normalizedData = searchInputFile.Get("normalizedData")
#normalizedBkgFrom4ParamFit = searchInputFile.Get("normalizedBkgFrom4ParamFit")
residualHist = searchInputFile.Get("residualHist")
relativeDiffHist = searchInputFile.Get("relativeDiffHist")
sigOfDiffHist = searchInputFile.Get("sigOfDiffHist")
logLikelihoodPseudoStatHist = searchInputFile.Get("logLikelihoodStatHistNullCase")
chi2PseudoStatHist = searchInputFile.Get("chi2StatHistNullCase")
bumpHunterStatHist = searchInputFile.Get("bumpHunterStatHistNullCase")
#theFitFunction = searchInputFile.Get('theFitFunction')
bumpHunterTomographyPlot = searchInputFile.Get('bumpHunterTomographyFromPseudoexperiments')
bumpHunterStatOfFitToData = searchInputFile.Get('bumpHunterStatOfFitToData')
# nominal background +- statistical uncertainty(uncertainty from fitting parameters)
nomPlus1 = searchInputFile.Get("nominalBkgFromFit_plus1Sigma")
nomMinus1 = searchInputFile.Get("nominalBkgFromFit_minus1Sigma")
#nominal background +- uncertainties from fitting function choice
if doAlternate :
alternateBkg = searchInputFile.Get("alternateFitOnRealData")
nomWithNewFuncErrSymm = searchInputFile.Get("nomOnDataWithSymmetricRMSScaleFuncChoiceErr")
valueNewFuncErrDirected = searchInputFile.Get("nomOnDataWithDirectedRMSScaleFuncChoiceErr")
logLOfFitToDataVec = searchInputFile.Get('logLOfFitToData')
chi2OfFitToDataVec = searchInputFile.Get('chi2OfFitToData')
statOfFitToData = searchInputFile.Get('bumpHunterPLowHigh')
logLOfFitToData = logLOfFitToDataVec[0]
logLPVal = logLOfFitToDataVec[1]
chi2OfFitToData = chi2OfFitToDataVec[0]
chi2PVal = chi2OfFitToDataVec[1]
bumpHunterStatFitToData = statOfFitToData[0]
bumpHunterPVal = bumpHunterStatOfFitToData[1]
bumpLowEdge = statOfFitToData[1]
bumpHighEdge = statOfFitToData[2]
#NDF = searchInputFile.Get('NDF')[0]
#fitparams = searchInputFile.Get('fittedParameters')
print "logL of fit to data is",logLOfFitToData
print "logL pvalue is",logLPVal
print "chi2 of fit to data is",chi2OfFitToData
#print "NDF is",NDF
#print "chi2/NDF is",chi2OfFitToData/NDF
print "chi2 pvalue is",chi2PVal
print "bump hunter stat of fit to data is",bumpHunterStatFitToData
print "bumpLowEdge, bumpHighEdge are",bumpLowEdge,bumpHighEdge
print "BumpHunter pvalue is",bumpHunterPVal
print "which is Z value of",GetZVal(bumpHunterPVal,True)
#print "Fitted parameters were:",fitparams
# Find range
firstBin = 1000
lastBin = basicData.GetNbinsX()
print lastBin
while (basicData.GetBinContent(lastBin)==0 and lastBin > 0) :
lastBin = lastBin - 1
# Calculate from fit range
fitRange = searchInputFile.Get("FitRange")
firstBin = basicData.FindBin(fitRange[0])
lastBin = basicData.FindBin(fitRange[1])
print "New firstbin, lastbin",firstBin,lastBin
# Convert plots into desired final form
standardbins = basicData.GetXaxis().GetXbins()
newbins = []#ROOT.TArrayD(standardbins.GetSize())
for np in range(standardbins.GetSize()) :
newbins.append(standardbins[np]/1000)
# Make new versions of old plots
newbasicdata = ROOT.TH1D("basicData_TeV","basicData_TeV",len(newbins)-1,array('d',newbins))
newbasicBkg = ROOT.TH1D("basicBkg_TeV","basicBkg_TeV",len(newbins)-1,array('d',newbins))
newresidualHist = ROOT.TH1D("residualHist_TeV","residualHist_TeV",len(newbins)-1,array('d',newbins))
newrelativeDiffHist = ROOT.TH1D("relativeDiffHist_TeV","relativeDiffHist_TeV",len(newbins)-1,array('d',newbins))
newsigOfDiffHist = ROOT.TH1D("sigOfDiffHist_TeV","sigOfDiffHist_TeV",len(newbins)-1,array('d',newbins))
newNomPlus1= ROOT.TH1D("nomPlus1_TeV","nomPlus1_TeV",len(newbins)-1,array('d',newbins))
newNomMinus1= ROOT.TH1D("nomMinus1_TeV","nomMinus1_TeV",len(newbins)-1,array('d',newbins))
newAlternateBkg = ROOT.TH1D("alternateBkg_TeV","alternateBkg_TeV",len(newbins)-1,array('d',newbins))
newnomWithNewFuncErrSymm = ROOT.TH1D("nomWithNewFuncErrSymm_TeV","nomWithNewFuncErrSymm_TeV",len(newbins)-1,array('d',newbins))
newValueNewFuncErrDirected= ROOT.TH1D("nomWithNewFuncErrDirected_TeV","nomWithNewFuncErrDirected_TeV",len(newbins)-1,array('d',newbins))
for histnew,histold in [[newbasicdata,basicData],[newbasicBkg,basicBkg],\
[newNomPlus1,nomPlus1],[newNomMinus1,nomMinus1], \
[newresidualHist,residualHist],[newrelativeDiffHist,relativeDiffHist],\
[newsigOfDiffHist,sigOfDiffHist]] :
for bin in range(histnew.GetNbinsX()+2) :
histnew.SetBinContent(bin,histold.GetBinContent(bin))
histnew.SetBinError(bin,histold.GetBinError(bin))
if doAlternate:
icount=0
for histnew,histold in [[newAlternateBkg,alternateBkg], [newnomWithNewFuncErrSymm,nomWithNewFuncErrSymm],\
[newValueNewFuncErrDirected,valueNewFuncErrDirected]] :
print "Count : ",icount
print histnew.GetName()
print histold.GetName()
icount+=1
for bin in range(histnew.GetNbinsX()+2) :
histnew.SetBinContent(bin,histold.GetBinContent(bin))
histnew.SetBinError(bin,histold.GetBinError(bin))
# Significances for Todd
ToddSignificancesHist = ROOT.TH1D("ToddSignificancesHist","ToddSignificancesHist",100,-5,5)
for bin in range(0,newresidualHist.GetNbinsX()+1):
if bin < firstBin: continue
if bin > lastBin: continue
residualValue = newresidualHist.GetBinContent(bin)
ToddSignificancesHist.Fill(residualValue)
# Search phase plots
myPainter.drawDataAndFitOverSignificanceHist(newbasicdata,newbasicBkg,newresidualHist, 'm_{jj} [TeV]','Events','Significance','{0}/figure1'.format(outPath), luminosity,13,fitRange[0],fitRange[1], firstBin,lastBin,True,bumpLowEdge/1000.0,bumpHighEdge/1000.0,[],True,False,[],True,bumpHunterPVal)
myPainter.drawDataAndFitOverSignificanceHist(newbasicdata,newbasicBkg,newresidualHist, 'm_{jj} [TeV]','Events','Significance','{0}/figure1_nologx'.format(outPath), luminosity,13,fitRange[0],fitRange[1], firstBin,lastBin,True,bumpLowEdge/1000.0,bumpHighEdge/1000.0,[],False,False,[],True,bumpHunterPVal)
myPainter.drawDataAndFitOverSignificanceHist(newbasicdata,newbasicBkg,newresidualHist, 'm_{jj} [TeV]','Events','Significance','{0}/figure1_nobump'.format(outPath), luminosity,13,fitRange[0],fitRange[1],firstBin,lastBin+2,False,bumpLowEdge,bumpHighEdge,[],True,False,[],True,bumpHunterPVal)
myPainter.drawDataAndFitOverSignificanceHist(newbasicdata,newbasicBkg,newresidualHist, 'm_{jj} [TeV]','Events','Significance','{0}/figure1_nobump_nologx'.format(outPath), luminosity,13,fitRange[0],fitRange[1],firstBin,lastBin,False,bumpLowEdge,bumpHighEdge,[],False,False,[],True,bumpHunterPVal)
myPainter.drawPseudoExperimentsWithObservedStat(logLikelihoodPseudoStatHist,float(logLOfFitToData),logLPVal,0,luminosity,13, 'logL statistic','Pseudo-exeperiments',"{0}/logLStatPlot".format(outPath))
myPainter.drawPseudoExperimentsWithObservedStat(chi2PseudoStatHist,float(chi2OfFitToData),chi2PVal,0,luminosity,13, "#chi^{2}",'Pseudo-exeperiments',"{0}/chi2StatPlot".format(outPath))
myPainter.drawPseudoExperimentsWithObservedStat(bumpHunterStatHist,float(bumpHunterStatFitToData),bumpHunterPVal,0,luminosity,13, 'BumpHunter','Pseudo-exeperiments',"{0}/bumpHunterStatPlot".format(outPath))
myPainter.drawBumpHunterTomographyPlot(bumpHunterTomographyPlot,"{0}/bumpHunterTomographyPlot".format(outPath))
# Various significance plots
myPainter.drawBasicHistogram(ToddSignificancesHist,-1,-1,"Residuals","Entries","{0}/ToddSignificancesHist".format(outPath))
myPainter.drawSignificanceHistAlone(newrelativeDiffHist,"m_{jj} [TeV]","(D - B)/B","{0}/significanceonlyplot".format(outPath))
myPainter.drawSignificanceHistAlone(newsigOfDiffHist,"m_{jj} [TeV]","(D - B)/#sqrt{Derr^{2}+Berr^{2}}","{0}/sigofdiffonlyplot".format(outPath))
# Now to make the one comparing uncertainties
placeHolderNom = newbasicBkg.Clone()
placeHolderNom.SetName("placeHolderNom")
nomPlusSymmFuncErr = newbasicBkg.Clone()
nomPlusSymmFuncErr.SetName("nomPlusNewFuncErr")
nomMinusSymmFuncErr = newbasicBkg.Clone()
nomMinusSymmFuncErr.SetName("nomMinusNewFuncErr")
for bin in range(nomPlusSymmFuncErr.GetNbinsX()+2) :
nomPlusSymmFuncErr.SetBinContent(bin,newnomWithNewFuncErrSymm.GetBinContent(bin) + newnomWithNewFuncErrSymm.GetBinError(bin))
nomMinusSymmFuncErr.SetBinContent(bin,newnomWithNewFuncErrSymm.GetBinContent(bin) - newnomWithNewFuncErrSymm.GetBinError(bin))
if doAlternate:
myPainter.drawDataWithFitAsHistogram(newbasicdata,newbasicBkg,luminosity,13,"m_{jj} [TeV]","Events",["Data","Fit","Fit uncertainty","Function choice"],"{0}/compareFitQualityAndAlternateFit".format(outPath),True,[[newNomPlus1,newNomMinus1],[placeHolderNom,newAlternateBkg]],firstBin,lastBin,True,True,True,False,False)
myPainter.drawDataWithFitAsHistogram(newbasicdata,newbasicBkg,luminosity,13,"m_{jj} [TeV]","Events",["Data","Fit","Function choice","Alternate function"],"{0}/compareFitChoiceAndAlternateFit".format(outPath),True,[[nomPlusSymmFuncErr,nomMinusSymmFuncErr],[placeHolderNom,newAlternateBkg]],firstBin,lastBin,True,True,True,False,False)
myPainter.drawDataWithFitAsHistogram(newbasicdata,newbasicBkg,luminosity,13,"m_{jj} [TeV]","Events",["Data","Fit","Fit uncertainty","Function choice"],"{0}/compareFitQualtiyAndFitChoice".format(outPath),True,[[newNomPlus1,newNomMinus1],[nomPlusSymmFuncErr,nomMinusSymmFuncErr]],firstBin,lastBin,True,True,True,False,False)
myPainter.drawDataWithFitAsHistogram(newbasicdata,newbasicBkg,luminosity,13,"m_{jj} [TeV]","Events",["Data","Fit","Statistical fit uncertainty","Function choice"],"{0}/compareFitQualityAndFitChoice_Asymm".format(outPath),True,[[newNomPlus1,newNomMinus1],[placeHolderNom,newValueNewFuncErrDirected]],firstBin,lastBin,True,True,True,False,False)
# Overlay nominal and alternate fit functions
myPainter.drawDataWithFitAsHistogram(newbasicdata,newbasicBkg,luminosity,13,"m_{jj} [TeV]","Events",["Data","Fit","Alternate function"],"{0}/compareFitChoices".format(outPath),True,[[placeHolderNom,newAlternateBkg]],firstBin,lastBin,True,True,True,False,False)
# Make a ratio histogram for bottom plot.
altFitRatio = ROOT.TH1D("altFitRatio","altFitRatio",len(newbins)-1,array('d',newbins))
for bin in range(0,altFitRatio.GetNbinsX()+1) :
if newbasicBkg.GetBinContent(bin) == 0 :
altFitRatio.SetBinContent(bin,0)
else :
altFitRatio.SetBinContent(bin,(valueNewFuncErrDirected.GetBinContent(bin)-newbasicBkg.GetBinContent(bin))/newbasicBkg.GetBinContent(bin))
# Make a ratio histogram for paper plot.
PlusNomRatio = ROOT.TH1D("PlusNomRatio","PlusNomRatio",len(newbins)-1,array('d',newbins))
for bin in range(0,PlusNomRatio.GetNbinsX()+1) :
if newbasicBkg.GetBinContent(bin) == 0 :
PlusNomRatio.SetBinContent(bin,0)
else :
PlusNomRatio.SetBinContent(bin,(newNomPlus1.GetBinContent(bin)-newbasicBkg.GetBinContent(bin))/newbasicBkg.GetBinContent(bin))
MinusNomRatio = ROOT.TH1D("MinusNomRatio","MinusNomRatio",len(newbins)-1,array('d',newbins))
for bin in range(0,MinusNomRatio.GetNbinsX()+1) :
if newbasicBkg.GetBinContent(bin) == 0 :
MinusNomRatio.SetBinContent(bin,0)
else :
MinusNomRatio.SetBinContent(bin,(newNomMinus1.GetBinContent(bin)-newbasicBkg.GetBinContent(bin))/newbasicBkg.GetBinContent(bin))
myPainter.drawDataWithFitAsHistogramAndResidual(newbasicdata,newbasicBkg,luminosity,13,"m_{jj} [TeV]","Events",["Data","Fit","Statistical uncertainty on fit","Function choice"],"{0}/compareFitQualityAndFitChoice_Asymm_WithRatio".format(outPath),True,[[newNomPlus1,newNomMinus1],[placeHolderNom,newValueNewFuncErrDirected]],[altFitRatio,MinusNomRatio,PlusNomRatio],firstBin,lastBin,True,True,True,False,False,True,False,True,bumpHunterPVal,True,fitRange[0],fitRange[1])
myPainter.drawDataWithFitAsHistogramAndResidualPaper(newbasicdata,newbasicBkg,luminosity,13,"m_{jj} [TeV]","Events",["Data","Fit","Statistical uncertainty on fit","Function choice"],"{0}/compareFitQualityAndFitChoice_Asymm_WithRatioPaper".format(outPath),True,[[newNomPlus1,newNomMinus1],[placeHolderNom,newValueNewFuncErrDirected]],[altFitRatio,MinusNomRatio,PlusNomRatio],firstBin,lastBin,True,True,True,False,False)
myPainter.drawMultipleFitsAndResiduals(newbasicdata,[newbasicBkg,newValueNewFuncErrDirected],[altFitRatio],["Nominal fit","Function choice"],"m_{jj} [TeV]","Events",["(alt-nom)/nom"],"{0}/directedFuncChoiceVersusNominal_withRatio".format(outPath),luminosity,13,firstBin,lastBin)
####### Draw Signal Template overlaid on Background###########
if overlaidSignal:
signalFile = ROOT.TFile.Open(signalFileName, "read")
if not signalFile:
print signalFileName, " doesn't exist!!!!"
return
# setup signal information
#signalTitles = {"QStar": "#it{q}*"}
#signalTypes = ["QStar"]
#signalsMasses = {"QStar":[4000, 5000]}
#signalScalingFactors = {"QStar": 0.1}
#signalAxes = {"QStar": {"X" : "M_{#it{q}*} [GeV]", "Y": "#sigma #times #it{A} #times BR [pb]"} }
signalTitles = {"DMZprime": "DM #it{Z}'"}
signalTypes = ["DMZprime"]
signalsMasses = {"DMZprime":[4000, 5000]}
signalScalingFactors = {"DMZprime": 1000}
signalAxes = {"DMZprime": {"X" : "M_{#it{q}*} [GeV]", "Y": "#sigma #times #it{A} #times BR [pb]"} }
for signalType in signalTypes:
print "in signal",signalType
signalMasses = signalsMasses[signalType]
signalMassesTeV = signalsMasses[signalType][:]
for index in range(len(signalMasses)) :
signalMassesTeV[index] = signalMasses[index]/1000.0
print signalMassesTeV
signalPlotsTeV = []
legendlistTeV = []
for mass in signalMasses :
#sigplot = signalFile.Get("h_mjj_{0}".format(mass))
sigplot = signalFile.Get("{0}_{1}".format(signalType, mass))
sigplot.SetDirectory(0)
sigplottev = newbasicdata.Clone()
sigplottev.SetName("sigplot_{0}_{1}_TeV".format(signalType,mass))
for bins in range(sigplot.GetNbinsX()+2) :
for bin in range(sigplottev.GetNbinsX()+2) :
if sigplot.GetBinLowEdge(bins)/1000.==sigplottev.GetBinLowEdge(bin) :
sigplottev.SetBinContent(bin,sigplot.GetBinContent(bins))
sigplottev.SetBinError(bin,sigplot.GetBinError(bins))
sigplotforfitplusbkg = sigplottev.Clone()
sigplotforfitplusbkg.SetDirectory(0)
sigplotforfitplusbkg.SetName(sigplottev.GetName()+"_forfitplusbkg_TeV")
sigplotforfitplusbkg.Scale(signalScalingFactors[signalType])
signalPlotsTeV.append(sigplotforfitplusbkg)
thistitle = signalTitles[signalType] + ", {0}= {1} TeV".format(signalAxes[signalType]["X"].split("[GeV]")[0].replace("M","m"),mass/1000.0)
legendlistTeV.append(thistitle)
extLastBin = lastBin
for bin in range(sigplotforfitplusbkg.GetNbinsX()) :
if bin > extLastBin and sigplotforfitplusbkg.GetBinContent(bin) > 0.01 :
extLastBin = bin
if sigplotforfitplusbkg.GetBinLowEdge(bin) > 1.3*mass/1000.0 :
continue
if extLastBin < lastBin :
extLastBin = lastBin
UserScaleText = signalTitles[signalType]
if signalScalingFactors[signalType] == 1 :
UserScaleText = signalTitles[signalType]
else :
UserScaleText = UserScaleText+", #sigma #times "+str(signalScalingFactors[signalType])
outputName = outPath+"FancyFigure1_"+signalType
myPainter.drawDataAndFitWithSignalsOverSignificances(newbasicdata,newbasicBkg, None,\
newresidualHist,signalPlotsTeV, None, signalMassesTeV,legendlistTeV,\
"m_{jj} [TeV]","Events","","Significance ", outputName,luminosity,\
Ecm, firstBin,extLastBin,\
True, bumpLowEdge/1000,bumpHighEdge/1000,\
True,False,True,UserScaleText,True,bumpHunterPVal, True, \
fitRange[0], fitRange[1])
outputName = outPath+"FancyFigure1_"+signalType+"_nologx"
myPainter.drawDataAndFitWithSignalsOverSignificances(newbasicdata,newbasicBkg, None,\
newresidualHist,signalPlotsTeV, None, signalMassesTeV,legendlistTeV,\
"m_{jj} [TeV]","Events","","Significance ", outputName,luminosity,\
Ecm, firstBin,extLastBin,\
True, bumpLowEdge/1000,bumpHighEdge/1000,\
False,False,True,UserScaleText,True,bumpHunterPVal, True, \
fitRange[0], fitRange[1])
outputName = outPath+"FancyFigure1_"+signalType+"_noBump"
myPainter.drawDataAndFitWithSignalsOverSignificances(newbasicdata,newbasicBkg, None,\
newresidualHist,signalPlotsTeV, None, signalMassesTeV,legendlistTeV,\
"m_{jj} [TeV]","Events","","Significance ", outputName,luminosity,\
Ecm, firstBin,extLastBin,\
False, bumpLowEdge/1000,bumpHighEdge/1000,\
True,False,True,UserScaleText,True,bumpHunterPVal, True, \
fitRange[0], fitRange[1])
outputName = outPath+"FancyFigure1_"+signalType+"_noBump_nologx"
myPainter.drawDataAndFitWithSignalsOverSignificances(newbasicdata,newbasicBkg, None,\
newresidualHist,signalPlotsTeV, None, signalMassesTeV,legendlistTeV,\
"m_{jj} [TeV]","Events","","Significance ", outputName,luminosity,\
Ecm, firstBin,extLastBin,\
False, bumpLowEdge/1000,bumpHighEdge/1000,\
False,False,True,UserScaleText,True,bumpHunterPVal, True, \
fitRange[0], fitRange[1])
###############################
# Draw the comparison between data and MC in the bottom panel
if drawMCComparison:
mcFile = ROOT.TFile(mcFileName, "read") ;
if not mcFile:
print "Can not open: ", mcFileName
return
mchist_nominal = mcFile.Get("djet_mjj_nominal")
mchist_jesup = mcFile.Get("djet_mjj_JES_up")
mchist_jesdown = mcFile.Get("djet_mjj_JES_down")
newmchist_nominal=ROOT.TH1D("djet_mjj_nominal_TeV","djet_mjj_nominal_TeV",len(newbins)-1,array('d',newbins))
newmchist_jesup=ROOT.TH1D("djet_mjj_jesup_TeV","djet_mjj_jesup_TeV",len(newbins)-1,array('d',newbins))
newmchist_jesdown=ROOT.TH1D("djet_mjj_jesdown_TeV","djet_mjj_jesdown_TeV",len(newbins)-1,array('d',newbins))
for iBin1 in range(1, newmchist_nominal.GetNbinsX()+1):
for iBin2 in range(1, mchist_nominal.GetNbinsX()+1):
if newmchist_nominal.GetBinLowEdge(iBin1)*1000==mchist_nominal.GetBinLowEdge(iBin2):
newmchist_nominal.SetBinContent(iBin1, mchist_nominal.GetBinContent(iBin2))
newmchist_nominal.SetBinError(iBin1, mchist_nominal.GetBinError(iBin2))
continue
for iBin1 in range(1, newmchist_jesup.GetNbinsX()+1):
for iBin2 in range(1, mchist_jesup.GetNbinsX()+1):
if newmchist_jesup.GetBinLowEdge(iBin1)*1000==mchist_jesup.GetBinLowEdge(iBin2):
newmchist_jesup.SetBinContent(iBin1, mchist_jesup.GetBinContent(iBin2))
newmchist_jesup.SetBinError(iBin1, mchist_jesup.GetBinError(iBin2))
continue
for iBin1 in range(1, newmchist_jesdown.GetNbinsX()+1):
for iBin2 in range(1, mchist_jesdown.GetNbinsX()+1):
if newmchist_jesdown.GetBinLowEdge(iBin1)*1000==mchist_jesdown.GetBinLowEdge(iBin2):
newmchist_jesdown.SetBinContent(iBin1, mchist_jesdown.GetBinContent(iBin2))
newmchist_jesdown.SetBinError(iBin1, mchist_jesdown.GetBinError(iBin2))
continue
tmpRatioHist = newbasicdata.Clone()
tmpRatioHist.SetMarkerColor(ROOT.kBlack)
tmpRatioHist.Add(newmchist_nominal,-1)
tmpRatioHist.Divide(newmchist_nominal)
## If data is 0 then there should be no ratio drawn
for iBin in range(1, tmpRatioHist.GetNbinsX()+1):
if newbasicdata.GetBinContent(iBin) == 0:
tmpRatioHist.SetBinContent(iBin, 0)
tmpRatioHist.SetBinError(iBin, 0)
UpDownRatioHists = []
if mchist_jesup.GetEntries() >= 0:
tmpJESRatioHist = newmchist_jesup
tmpJESRatioHist.Add( newmchist_nominal, -1. )
tmpJESRatioHist.Divide( newmchist_nominal )
tmpJESRatioHist.SetMarkerColorAlpha( ROOT.kBlue,0.15)
tmpJESRatioHist.SetLineColorAlpha( ROOT.kBlue,0.15)
tmpJESRatioHist.SetFillColorAlpha( ROOT.kBlue, 0.15)
tmpJESRatioHist.SetFillStyle(1001)
UpDownRatioHists.append(tmpJESRatioHist)
if mchist_jesdown.GetEntries() >= 0:
tmpJESRatioHist = newmchist_jesdown
tmpJESRatioHist.Add( newmchist_nominal, -1. )
tmpJESRatioHist.Divide( newmchist_nominal )
tmpJESRatioHist.SetMarkerColorAlpha( ROOT.kBlue,0.15)
tmpJESRatioHist.SetLineColorAlpha( ROOT.kBlue,0.15)
tmpJESRatioHist.SetFillColorAlpha( ROOT.kBlue, 0.15)
tmpJESRatioHist.SetFillStyle(1001)
UpDownRatioHists.append(tmpJESRatioHist)
outputName = outPath+"FancyFigure1_"+signalType+"_WithMCRatio"
myPainter.drawDataAndFitWithSignalsOverSignificancesWithMCRatio(newbasicdata,newbasicBkg,None,\
newresidualHist, signalPlotsTeV, [], signalMassesTeV,legendlistTeV,\
"m_{jj} [TeV]","Events","#frac{Data-MC}{MC}","Significance",\
outputName,luminosity,Ecm,firstBin,lastBin,True,bumpLowEdge/1000.0,bumpHighEdge/1000.0,\
True,False,False, UserScaleText,True,bumpHunterPVal,True,fitRange[0],fitRange[1],\
newmchist_nominal,tmpRatioHist,UpDownRatioHists[0],UpDownRatioHists[1])
outputName = outPath+"FancyFigure1_"+signalType+"_WithMCRatio_nologx"
myPainter.drawDataAndFitWithSignalsOverSignificancesWithMCRatio(newbasicdata,newbasicBkg,None, \
newresidualHist, signalPlotsTeV, [],signalMassesTeV,legendlistTeV, \
"m_{jj} [TeV]","Events","#frac{Data-MC}{MC}","Significance",\
outputName,luminosity,Ecm,firstBin,lastBin,True,bumpLowEdge/1000.0,bumpHighEdge/1000.0,\
False,False,False, UserScaleText,True,bumpHunterPVal,True,fitRange[0],fitRange[1],\
newmchist_nominal,tmpRatioHist,UpDownRatioHists[0],UpDownRatioHists[1])
searchInputFile.Close()
del searchInputFile
print "Done."
logLOfFitToData = searchInputFile.Get('logLOfFitToData')[0]
chi2OfFitToData = searchInputFile.Get('chi2OfFitToData')[0]
statOfFitToData = searchInputFile.Get('bumpHunterPLowHigh')
bumpHunterPvalFitToData = statOfFitToData[0]
bumpLowEdge = statOfFitToData[1]
bumpHighEdge = statOfFitToData[2]
fitparams = searchInputFile.Get('fittedParameters')
print "logL of fit to data is", logLOfFitToData
print "chi2 of fit to data is", chi2OfFitToData
print "bump hunter stat of fit to data is", bumpHunterPvalFitToData
print "bumpLowEdge, bumpHighEdge are", bumpLowEdge, bumpHighEdge
# Search phase plots
myPainter.drawDataAndFitOverSignificanceHist(basicData,basicBkgFrom4ParamFit,residualHist,\
'Mass[GeV]','Events','Significance',folderextension+'figure1',luminosity,8,250,6000,True,bumpLowEdge,bumpHighEdge)
myPainter.drawDataAndFitOverSignificanceHist(basicData,basicBkgFrom4ParamFit,residualHist,\
'Mass[GeV]','Events','Significance',folderextension+'figure1_nobump',luminosity,8,200,5000,False,bumpLowEdge,bumpHighEdge)
myPainter.drawPseudoExperimentsWithObservedStat(logLikelihoodPseudoStatHist,float(logLOfFitToData),luminosity,8,\
'logL statistic','Pseudo-exeperiments',folderextension+"logLStatPlot")
myPainter.drawPseudoExperimentsWithObservedStat(chi2PseudoStatHist,float(chi2OfFitToData),luminosity,8,\
"#chi^{2}",'Pseudo-exeperiments',folderextension+"chi2StatPlot")
myPainter.drawPseudoExperimentsWithObservedStat(bumpHunterStatHist,float(bumpHunterPvalFitToData),luminosity,8,\
'BumpHunter','Pseudo-exeperiments',folderextension+"bumpHunterStatPlot")
myPainter.drawBumpHunterTomographyPlot(
bumpHunterTomographyPlot, folderextension + "bumpHunterTomographyPlot")
def main():
# User controlled arguments
parser = argparse.ArgumentParser()
parser.add_argument("--inFileName",
type=str,
default="",
help="The path to the input file from SearchPhase")
parser.add_argument(
"--outPath",
type=str,
default="./plotting/SearchPhase/plots/",
help="The path prefix (directory) where you want the output plots")
parser.add_argument("--lumi", type=float, default=1, help="Luminosity")
args = parser.parse_args()
inFileName = args.inFileName
outPath = args.outPath
print "==================================="
print "Executing Run_SearchPhase.py with :"
print "inFileName : ", inFileName
print "outPath : ", outPath
print "Lumi : ", args.lumi
print "==================================="
# Get input (the rootfile name should match that specified in the SearchPhase.config file)
searchInputFile = ROOT.TFile(inFileName, "READ")
# make plots folder i.e. make folder extension
if not os.path.exists(outPath):
os.makedirs(outPath)
# Define necessary quantities.
luminosity = 1000 * args.lumi
Ecm = 13
# Get input
doStatSearch = False
doAlternate = True # You can use this if you included the alternate fit function in the search phase
# Initialize painter
myPainter = Morisot()
myPainter.setColourPalette("Teals")
#myPainter.setEPS(True)
myPainter.setLabelType(
2) # Sets label type i.e. Internal, Work in progress etc.
# See below for label explanation
# 0 Just ATLAS
# 1 "Preliminary"
# 2 "Internal"
# 3 "Simulation Preliminary"
# 4 "Simulation Internal"
# 5 "Simulation"
# 6 "Work in Progress"
# Retrieve search phase inputs
basicData = searchInputFile.Get("basicData")
normalizedData = searchInputFile.Get("normalizedData")
basicBkgFrom4ParamFit = searchInputFile.Get("basicBkgFrom4ParamFit")
normalizedBkgFrom4ParamFit = searchInputFile.Get(
"normalizedBkgFrom4ParamFit")
residualHist = searchInputFile.Get("residualHist")
relativeDiffHist = searchInputFile.Get("relativeDiffHist")
sigOfDiffHist = searchInputFile.Get("sigOfDiffHist")
logLikelihoodPseudoStatHist = searchInputFile.Get(
"logLikelihoodStatHistNullCase")
chi2PseudoStatHist = searchInputFile.Get("chi2StatHistNullCase")
bumpHunterStatHist = searchInputFile.Get("bumpHunterStatHistNullCase")
theFitFunction = searchInputFile.Get('theFitFunction')
bumpHunterTomographyPlot = searchInputFile.Get(
'bumpHunterTomographyFromPseudoexperiments')
bumpHunterStatOfFitToData = searchInputFile.Get(
'bumpHunterStatOfFitToData')
if doAlternate:
searchInputFile.ls()
alternateBkg = searchInputFile.Get("alternateFitOnRealData")
nomPlus1 = searchInputFile.Get("nominalBkgFromFit_plus1Sigma")
nomMinus1 = searchInputFile.Get("nominalBkgFromFit_minus1Sigma")
nomWithNewFuncErrSymm = searchInputFile.Get(
"nomOnDataWithSymmetricRMSScaleFuncChoiceErr")
valueNewFuncErrDirected = searchInputFile.Get(
"nomOnDataWithDirectedRMSScaleFuncChoiceErr")
logLOfFitToDataVec = searchInputFile.Get('logLOfFitToData')
chi2OfFitToDataVec = searchInputFile.Get('chi2OfFitToData')
statOfFitToData = searchInputFile.Get('bumpHunterPLowHigh')
logLOfFitToData = logLOfFitToDataVec[0]
logLPVal = logLOfFitToDataVec[1]
chi2OfFitToData = chi2OfFitToDataVec[0]
chi2PVal = chi2OfFitToDataVec[1]
bumpHunterStatFitToData = statOfFitToData[0]
bumpHunterPVal = bumpHunterStatOfFitToData[1]
bumpLowEdge = statOfFitToData[1]
bumpHighEdge = statOfFitToData[2]
#NDF = searchInputFile.Get('NDF')[0]
fitparams = searchInputFile.Get('fittedParameters')
print "logL of fit to data is", logLOfFitToData
print "logL pvalue is", logLPVal
print "chi2 of fit to data is", chi2OfFitToData
#print "NDF is",NDF
#print "chi2/NDF is",chi2OfFitToData/NDF
print "chi2 pvalue is", chi2PVal
print "bump hunter stat of fit to data is", bumpHunterStatFitToData
print "bumpLowEdge, bumpHighEdge are", bumpLowEdge, bumpHighEdge
print "BumpHunter pvalue is", bumpHunterPVal
print "which is Z value of", GetZVal(bumpHunterPVal, True)
print "Fitted parameters were:", fitparams
# Find range
firstBin = 1000
lastBin = basicData.GetNbinsX()
while (basicData.GetBinContent(lastBin) == 0 and lastBin > 0):
lastBin -= 1
if (firstBin > lastBin):
firstBin = 1
lastBin = basicData.GetNbinsX()
print "First bin = ", firstBin, ": lower edge at", basicData.GetBinLowEdge(
firstBin)
# Calculate from fit range
fitRange = searchInputFile.Get("FitRange")
firstBin = basicData.FindBin(fitRange[0]) - 1
lastBin = basicData.FindBin(fitRange[1])
print "New firstbin, lastbin", firstBin, lastBin
#firstBin = basicData.FindBin(1100)
#print "and another firstbin is",firstBin
# Convert plots into desired final form
standardbins = basicData.GetXaxis().GetXbins()
newbins = [] #ROOT.TArrayD(standardbins.GetSize())
for np in range(standardbins.GetSize()):
newbins.append(standardbins[np] / 1000)
# Make never versions of old plots
newbasicdata = ROOT.TH1D("basicData_TeV", "basicData_TeV",
len(newbins) - 1, array('d', newbins))
newbasicBkgFrom4ParamFit = ROOT.TH1D("basicBkgFrom4ParamFit_TeV",
"basicBkgFrom4ParamFit_TeV",
len(newbins) - 1, array('d', newbins))
newresidualHist = ROOT.TH1D("residualHist_TeV", "residualHist_TeV",
len(newbins) - 1, array('d', newbins))
newrelativeDiffHist = ROOT.TH1D("relativeDiffHist_TeV",
"relativeDiffHist_TeV",
len(newbins) - 1, array('d', newbins))
newsigOfDiffHist = ROOT.TH1D("sigOfDiffHist_TeV", "sigOfDiffHist_TeV",
len(newbins) - 1, array('d', newbins))
newAlternateBkg = ROOT.TH1D("alternateBkg_TeV", "alternateBkg_TeV",
len(newbins) - 1, array('d', newbins))
newNomPlus1 = ROOT.TH1D("nomPlus1_TeV", "nomPlus1_TeV",
len(newbins) - 1, array('d', newbins))
newNomMinus1 = ROOT.TH1D("nomMinus1_TeV", "nomMinus1_TeV",
len(newbins) - 1, array('d', newbins))
newnomWithNewFuncErrSymm = ROOT.TH1D("nomWithNewFuncErrSymm_TeV",
"nomWithNewFuncErrSymm_TeV",
len(newbins) - 1, array('d', newbins))
newValueNewFuncErrDirected = ROOT.TH1D("nomWithNewFuncErrDirected_TeV",
"nomWithNewFuncErrDirected_TeV",
len(newbins) - 1,
array('d', newbins))
for histnew,histold in [[newbasicdata,basicData],[newbasicBkgFrom4ParamFit,basicBkgFrom4ParamFit],\
[newresidualHist,residualHist],[newrelativeDiffHist,relativeDiffHist],[newsigOfDiffHist,sigOfDiffHist]] :
for bin in range(histnew.GetNbinsX() + 2):
histnew.SetBinContent(bin, histold.GetBinContent(bin))
histnew.SetBinError(bin, histold.GetBinError(bin))
# Significances for Todd
ToddSignificancesHist = ROOT.TH1D("ToddSignificancesHist",
"ToddSignificancesHist", 100, -5, 5)
for bin in range(0, newresidualHist.GetNbinsX() + 1):
if bin < firstBin: continue
if bin > lastBin: continue
residualValue = newresidualHist.GetBinContent(bin)
#print residualValue
ToddSignificancesHist.Fill(residualValue)
myPainter.drawBasicHistogram(ToddSignificancesHist, -1, -1, "Residuals",
"Entries",
"{0}/ToddSignificancesHist".format(outPath))
# Search phase plots
myPainter.drawDataAndFitOverSignificanceHist(newbasicdata,newbasicBkgFrom4ParamFit,newresidualHist,\
'm_{jj} [TeV]','Events','Significance','{0}/figure1'.format(outPath),\
luminosity,13,fitRange[0],fitRange[1],firstBin,lastBin,True,bumpLowEdge/1000.0,bumpHighEdge/1000.0,[],True,False,[],True,bumpHunterPVal)
myPainter.drawDataAndFitOverSignificanceHist(newbasicdata,newbasicBkgFrom4ParamFit,newresidualHist,\
'm_{jj} [TeV]','Events','Significance','{0}/figure1_nologx'.format(outPath),\
luminosity,13,fitRange[0],fitRange[1],firstBin,lastBin,True,bumpLowEdge/1000.0,bumpHighEdge/1000.0,[],False,False,[],True,bumpHunterPVal)
myPainter.drawDataAndFitOverSignificanceHist(newbasicdata,newbasicBkgFrom4ParamFit,newresidualHist,\
'm_{jj} [TeV]','Prescale-weighted events','Significance','{0}/figure1_nobump'.format(outPath),\
luminosity,13,fitRange[0],fitRange[1],firstBin,lastBin,False,bumpLowEdge,bumpHighEdge,[],True,False,[],True,bumpHunterPVal)
myPainter.drawDataAndFitOverSignificanceHist(newbasicdata,newbasicBkgFrom4ParamFit,newresidualHist,\
'm_{jj} [TeV]','Prescale-weighted events','Significance','{0}/figure1_nobump_nologx'.format(outPath),\
luminosity,13,fitRange[0],fitRange[1],firstBin,lastBin,False,bumpLowEdge,bumpHighEdge,[],False,False,[],True,bumpHunterPVal)
myPainter.drawPseudoExperimentsWithObservedStat(logLikelihoodPseudoStatHist,float(logLOfFitToData),logLPVal,0,luminosity,13,\
'logL statistic','Pseudo-exeperiments',"{0}/logLStatPlot".format(outPath))
myPainter.drawPseudoExperimentsWithObservedStat(chi2PseudoStatHist,float(chi2OfFitToData),chi2PVal,0,luminosity,13,\
"#chi^{2}",'Pseudo-exeperiments',"{0}/chi2StatPlot".format(outPath))
myPainter.drawPseudoExperimentsWithObservedStat(bumpHunterStatHist,float(bumpHunterStatFitToData),bumpHunterPVal,0,luminosity,13,\
'BumpHunter','Pseudo-exeperiments',"{0}/bumpHunterStatPlot".format(outPath))
myPainter.drawBumpHunterTomographyPlot(
bumpHunterTomographyPlot,
"{0}/bumpHunterTomographyPlot".format(outPath))
# Various significance plots
myPainter.drawSignificanceHistAlone(
newrelativeDiffHist, "m_{jj} [TeV]", "(D - B)/B",
"{0}/significanceonlyplot".format(outPath))
myPainter.drawSignificanceHistAlone(
newsigOfDiffHist, "m_{jj} [TeV]", "(D - B)/#sqrt{Derr^{2}+Berr^{2}}",
"{0}/sigofdiffonlyplot".format(outPath))
# Now to make the one comparing uncertainties
placeHolderNom = newbasicBkgFrom4ParamFit.Clone()
placeHolderNom.SetName("placeHolderNom")
nomPlusSymmFuncErr = newbasicBkgFrom4ParamFit.Clone()
nomPlusSymmFuncErr.SetName("nomPlusNewFuncErr")
nomMinusSymmFuncErr = newbasicBkgFrom4ParamFit.Clone()
nomMinusSymmFuncErr.SetName("nomMinusNewFuncErr")
for bin in range(nomPlusSymmFuncErr.GetNbinsX() + 2):
nomPlusSymmFuncErr.SetBinContent(
bin,
newnomWithNewFuncErrSymm.GetBinContent(bin) +
newnomWithNewFuncErrSymm.GetBinError(bin))
nomMinusSymmFuncErr.SetBinContent(
bin,
newnomWithNewFuncErrSymm.GetBinContent(bin) -
newnomWithNewFuncErrSymm.GetBinError(bin))
myPainter.drawDataWithFitAsHistogram(
newbasicdata, newbasicBkgFrom4ParamFit, luminosity, 13, "m_{jj} [TeV]",
"Events", ["Data", "Fit", "Fit uncertainty", "Function choice"],
"{0}/compareFitQualityAndAlternateFit".format(outPath), True,
[[newNomPlus1, newNomMinus1], [placeHolderNom, newAlternateBkg]],
firstBin, lastBin + 2, True, True, True, False, False)
myPainter.drawDataWithFitAsHistogram(
newbasicdata, newbasicBkgFrom4ParamFit, luminosity, 13, "m_{jj} [TeV]",
"Events", ["Data", "Fit", "Function choice", "Alternate function"],
"{0}/compareFitChoiceAndAlternateFit".format(outPath), True,
[[nomPlusSymmFuncErr, nomMinusSymmFuncErr],
[placeHolderNom, newAlternateBkg]], firstBin, lastBin + 2, True, True,
True, False, False)
myPainter.drawDataWithFitAsHistogram(
newbasicdata, newbasicBkgFrom4ParamFit, luminosity, 13, "m_{jj} [TeV]",
"Events", ["Data", "Fit", "Fit uncertainty", "Function choice"],
"{0}/compareFitQualtiyAndFitChoice".format(outPath),
True, [[newNomPlus1, newNomMinus1],
[nomPlusSymmFuncErr, nomMinusSymmFuncErr]], firstBin,
lastBin + 2, True, True, True, False, False)
myPainter.drawDataWithFitAsHistogram(
newbasicdata, newbasicBkgFrom4ParamFit, luminosity, 13, "m_{jj} [TeV]",
"Events",
["Data", "Fit", "Statistical fit uncertainty", "Function choice"],
"{0}/compareFitQualityAndFitChoice_Asymm".format(outPath), True,
[[newNomPlus1, newNomMinus1],
[placeHolderNom, newValueNewFuncErrDirected]], firstBin, lastBin + 2,
True, True, True, False, False)
# Overlay 3 and 4 parameter fit functions
myPainter.drawDataWithFitAsHistogram(
newbasicdata, newbasicBkgFrom4ParamFit, luminosity, 13, "m_{jj} [TeV]",
"Events", ["Data", "Fit", "Alternate function"],
"{0}/compareFitChoices".format(outPath), True,
[[placeHolderNom, newAlternateBkg]], firstBin, lastBin + 2, True, True,
True, False, False)
#print "RATI"
#newfitfunctionratio = newAlternateBkg.Clone()
#newfitfunctionratio.Add(newAlternateBkg,-1)
#for bin in range(0,newAlternateBkg.GetNbinsX()+1) :
# if newAlternateBkg.GetBinContent(bin) == 0 :
# newfitfunctionratio.SetBinContent(bin,0)
# else :
# newfitfunctionratio.SetBinContent(bin,(newbasicBkgFrom4ParamFit.GetBinContent(bin)/newAlternateBkg.GetBinContent(bin)))
#myPainter.drawBasicHistogram(newfitfunctionratio,firstBin,lastBin-5,"m_{jj} [TeV]","3 par/4par","{0}/compareDiffFitChoices".format(outPath))
# BROKEN myPainter.drawManyOverlaidHistograms([newbasicBkgFrom4ParamFit],["3 par"],"m_{jj} [TeV]","Events","CompareFitFunctions",firstBin,lastBin+2,0,1E6)
# Make a ratio histogram for Sasha's plot.
# altFitRatio = ROOT.TH1D("altFitRatio","altFitRatio",len(newbins)-1,array('d',newbins))
# for bin in range(0,altFitRatio.GetNbinsX()+1) :
# if newbasicBkgFrom4ParamFit.GetBinContent(bin) == 0 :
# altFitRatio.SetBinContent(bin,0)
# else :
# altFitRatio.SetBinContent(bin,(valueNewFuncErrDirected.GetBinContent(bin)-newbasicBkgFrom4ParamFit.GetBinContent(bin))/newbasicBkgFrom4ParamFit.GetBinContent(bin))
# Make a ratio histogram for paper plot.
PlusNomRatio = ROOT.TH1D("PlusNomRatio", "PlusNomRatio",
len(newbins) - 1, array('d', newbins))
for bin in range(0, PlusNomRatio.GetNbinsX() + 1):
if newbasicBkgFrom4ParamFit.GetBinContent(bin) == 0:
PlusNomRatio.SetBinContent(bin, 0)
else:
PlusNomRatio.SetBinContent(
bin, (newNomPlus1.GetBinContent(bin) -
newbasicBkgFrom4ParamFit.GetBinContent(bin)) /
newbasicBkgFrom4ParamFit.GetBinContent(bin))
MinusNomRatio = ROOT.TH1D("MinusNomRatio", "MinusNomRatio",
len(newbins) - 1, array('d', newbins))
for bin in range(0, MinusNomRatio.GetNbinsX() + 1):
if newbasicBkgFrom4ParamFit.GetBinContent(bin) == 0:
MinusNomRatio.SetBinContent(bin, 0)
else:
MinusNomRatio.SetBinContent(
bin, (newNomMinus1.GetBinContent(bin) -
newbasicBkgFrom4ParamFit.GetBinContent(bin)) /
newbasicBkgFrom4ParamFit.GetBinContent(bin))
# myPainter.drawDataWithFitAsHistogramAndResidual(newbasicdata,newbasicBkgFrom4ParamFit,luminosity,13,"m_{jj} [TeV]","Events",["Data","Fit","Statistical uncertainty on fit","Function choice"],"{0}/compareFitQualityAndFitChoice_Asymm_WithRatio".format(outPath),True,[[newNomPlus1,newNomMinus1],[placeHolderNom,newValueNewFuncErrDirected]],[altFitRatio,MinusNomRatio,PlusNomRatio],firstBin,lastBin+2,True,True,True,False,False,True,False,True,bumpHunterPVal,True,fitRange[0],fitRange[1]) # changed from lastBin+2 to lastBin+18 to match FancyFigure
#
# myPainter.drawDataWithFitAsHistogramAndResidualPaper(newbasicdata,newbasicBkgFrom4ParamFit,luminosity,13,"m_{jj} [TeV]","Events",["Data","Fit","Statistical uncertainty on fit","Function choice"],"{0}/compareFitQualityAndFitChoice_Asymm_WithRatioPaper".format(outPath),True,[[newNomPlus1,newNomMinus1],[placeHolderNom,newValueNewFuncErrDirected]],[altFitRatio,MinusNomRatio,PlusNomRatio],firstBin,lastBin+2,True,True,True,False,False) # changed from lastBin+2 to lastBin+18 to match FancyFigure
#
# myPainter.drawMultipleFitsAndResiduals(newbasicdata,[newbasicBkgFrom4ParamFit,newValueNewFuncErrDirected],[altFitRatio],["Nominal fit","Func choice unc"],"m_{jj} [TeV]","Events",["(alt-nom)/nom"],"{0}/directedFuncChoiceVersusNominal_withRatio".format(outPath),luminosity,13,firstBin,lastBin+2)
if doStatSearch:
TomographyPlotWithStats = searchInputFile.Get(
"TomographyPlotWithStats")
bumpHunterStatsWSyst = searchInputFile.Get("bumpHunterStatsWSyst")
bumpPValStat = bumpHunterStatsWSyst[0]
bumpPValLow = bumpHunterStatsWSyst[1]
bumpPValHigh = bumpHunterStatsWSyst[2]
myPainter.drawPseudoExperimentsWithObservedStat(bumpHunterStatsWSyst,3.20359,0.7241,0,luminosity,13,\
'BumpHunter','Pseudo-experiments',"{0}/bumpHunterStatPlot_withUncertainties".format(outPath))
myPainter.drawBumpHunterTomographyPlot(
TomographyPlotWithStats,
"{0}/bumpHunterTomographyPlot_withStats".format(outPath))
print "with stats, bumpLowEdge, bumpHighEdge are", bumpPValLow, bumpPValHigh
print "BumpHunter pvalue is", bumpPValStat
searchInputFile.Close()
print "Done."