Classes
class	TauValidationInfo
Functions
def	PlotPerformanceCurves
Variables
	HideMuonAndElectronRej = True
string	KinematicVar = 'pt'
int	KinematicVarMax = 50
int	KinematicVarMin = 20
string	LatexVar = "True p_{T}"
string	VarUnit = "GeV/c"

Detailed Description

  PerformanceCurvePlotter

  Author: Evan K. Friis, UC Davis (friis@physics.ucdavis.edu)

  Plot the efficiency for a given fake rate 
  for a series of cuts in a set pt range.

  With the default cuts (20 < pt < 50), this is designed
  to reproduce (independently) the TancBenchmark plots produced in
  RecoTauTag/TauTagTools/test/MVABenchmarks.py


      Instructions:

      Add the releases to as shown.  You must specify and signal 
      and backround file for each release, and a descriptive label.
      To choose the chain of discriminators to plot, select the appropriate
      EDProducer defined in Validation.RecoTau.TauTagValidation_cfi

      PreFourTanc = TauValidationInfo("Signal.root", 
                                      "Background.root",
                                      "310pre5 TaNC",
                                      Validation.RecoTau.RecoTauValidation_cfi.RunTancValidation)

      PreFourIso = TauValidationInfo("Signal.root", 
                                      "Background.root",
                                      "310pre4 Iso",
                                      Validation.RecoTau.RecoTauValidation_cfi.PFTausHighEfficiencyLeadingPionBothProngs)


      ReleasesToBuildFrom = []
      ReleasesToBuildFrom.append(PreFourTanc)
      ReleasesToBuildFrom.append(PreFourIso)

      PlotPerformanceCurves(ReleasesToBuildFrom, "output.png")

Function Documentation

def PerformanceCurvePlotter::PlotPerformanceCurves	(	ReleasesToBuildFrom,
		OutputFile
	)

Definition at line 184 of file PerformanceCurvePlotter.py.

00185                                                           :
00186    # Setup
00187    myCanvas = TCanvas("Validation", "Validation", 800, 800)
00188    # Cut label on the plots
00189    CutLabel = TPaveText(0.6, 0.1, 0.9, 0.27, "brNDC")
00190    CutLabel.AddText("%0.1f%s < %s < %0.1f%s" % (KinematicVarMin, VarUnit, LatexVar, KinematicVarMax, VarUnit))
00191    CutLabel.SetFillStyle(0)
00192    CutLabel.SetBorderSize(0)
00193 
00194    # Build the TGraphs of sigEff versus bkgFakeRate for each release
00195    for aRelease in ReleasesToBuildFrom:
00196       aRelease.LoadHistograms()
00197       aRelease.ComputeEfficiencies()
00198       aRelease.BuildTGraphSummary()
00199 
00200    CurrentHistogram = 0
00201 
00202    for aRelease in ReleasesToBuildFrom:
00203       if CurrentHistogram == 0:
00204          aRelease.SummaryTGraph.Draw("ALP")
00205          CurrentHistogram = aRelease.SummaryTGraph.GetHistogram()
00206       else:
00207          aRelease.SummaryTGraph.Draw("LP")
00208 
00209 
00210    CurrentHistogram.SetAxisRange(0, 1)
00211    CurrentHistogram.GetYaxis().SetRangeUser(0.0001,1)
00212    CurrentHistogram.GetXaxis().SetTitle("Efficiency")
00213    CurrentHistogram.GetYaxis().SetTitle("Fake Rate")
00214    CurrentHistogram.SetTitle("Performance Points")
00215 
00216    gPad.SetLogy(True)
00217 
00218    for aRelease in ReleasesToBuildFrom:
00219       for aPoint in aRelease.DiscriminatorPoints:
00220          aPoint['PerformancePoint'].Draw("P")
00221 
00222    TauValidationInfo.SummaryLegend.Draw()
00223    TauValidationInfo.DiscriminatorLegend.Draw()
00224    CutLabel.Draw()
00225 
00226    myCanvas.SaveAs(OutputFile)