tweaking thresh and enc distribution binning

anaUltraScurve.py

@@ -93,7 +93,7 @@ def plotAllSCurvesOnCanvas(vfatHistos, vfatHistosPanPin2=None, obsName="canvScur
                 g_scurve.Draw("AP")
             else:
                 g_scurve.Draw("sameP")
-        canv_dict[vfat].Update()
+        #canv_dict[vfat].Update()
     if vfatHistosPanPin2 is not None:
         for vfat,histo in vfatHistosPanPin2.iteritems():
             canv_dict[vfat].cd()
@@ -104,7 +104,7 @@ def plotAllSCurvesOnCanvas(vfatHistos, vfatHistosPanPin2=None, obsName="canvScur
 
                 g_scurve = r.TGraph(h_scurve)
                 g_scurve.Draw("sameP")
-            canv_dict[vfat].Update()
+            #canv_dict[vfat].Update()
 
     return canv_dict
 
@@ -145,8 +145,10 @@ def plotAllSCurvesOnCanvas(vfatHistos, vfatHistosPanPin2=None, obsName="canvScur
     outfilename = options.outfilename
     GEBtype = options.GEBtype
 
-    # Create the output File
+    # Create the output File and TTree
     outF = r.TFile(filename+'/'+outfilename, 'recreate')
+    if options.performFit:
+        myT = r.TTree('scurveFitTree','Tree Holding FitData')
 
     # Set the CAL DAC to fC conversion
     calDAC2Q_Intercept = np.zeros(24)
@@ -330,7 +332,7 @@ def plotAllSCurvesOnCanvas(vfatHistos, vfatHistosPanPin2=None, obsName="canvScur
                 effectivePedestals[vfat][chan] = fitter.scanFuncs[vfat][chan].Eval(0.0)
 
                 # Compute the value to apply MAD on for each channel
-                trimValue[chan] = scanFitResults[0][vfat][chan] - ztrim[0] * scanFitResults[1][vfat][chan]
+                trimValue[chan] = scanFitResults[0][vfat][chan] - options.ztrim * scanFitResults[1][vfat][chan]
                 pass
             fitFailed = np.logical_not(fitter.fitValid[vfat])
 
@@ -368,9 +370,10 @@ def plotAllSCurvesOnCanvas(vfatHistos, vfatHistosPanPin2=None, obsName="canvScur
                 calDAC2Q_m=calDAC2Q_Slope, 
                 calDAC2Q_b=calDAC2Q_Intercept)
 
-    # Create the output TTree and store the results
+    # Set the branches of the TTree and store the results
     if options.performFit:
-        myT = r.TTree('scurveFitTree','Tree Holding FitData')
+        # Due to weird ROOT black magic this cannot be done here
+        #myT = r.TTree('scurveFitTree','Tree Holding FitData')
 
         chi2 = array( 'f', [ 0 ] )
         myT.Branch( 'chi2', chi2, 'chi2/F')
@@ -419,13 +422,13 @@ def plotAllSCurvesOnCanvas(vfatHistos, vfatHistosPanPin2=None, obsName="canvScur
         threshSummaryPlots = {}
         encSummaryPlots = {}
         for vfat in range (0,24):
-            fitThr = np.zeroes(128)        
-            fitENC = np.zeroes(128)
-            stripPinOrChanArray = np.zeroes(128)
+            fitThr = np.zeros(128)        
+            fitENC = np.zeros(128)
+            stripPinOrChanArray = np.zeros(128)
             for chan in range (0, 128):
                 # Store stripChanOrPinType to use as x-axis of fit summary plots
-                stripPinOrChan = float( dict_vfatChanLUT[vfat][stripChanOrPinType][chan] )
-
+                stripPinOrChan = dict_vfatChanLUT[vfat][stripChanOrPinType][chan]
+                 
                 # Store Values for making fit summary plots
                 fitThr[stripPinOrChan] = scanFitResults[0][vfat][chan]
                 fitENC[stripPinOrChan] = scanFitResults[1][vfat][chan]
@@ -474,7 +477,7 @@ def plotAllSCurvesOnCanvas(vfatHistos, vfatHistosPanPin2=None, obsName="canvScur
                     len(fitThr),
                     stripPinOrChanArray,
                     fitThr,
-                    len(fitThr),
+                    np.zeros(len(fitThr)),
                     fitENC
                     )
             fitSummaryPlots[vfat].SetTitle("VFAT %i Fit Summary;Channel;Threshold [fC]"%vfat)
@@ -491,23 +494,25 @@ def plotAllSCurvesOnCanvas(vfatHistos, vfatHistosPanPin2=None, obsName="canvScur
 
             # Make thresh summary plot - bin size is variable
             histThresh = r.TH1F("scurveMean_vfat%i"%vfat,"VFAT %i;S-Curve Mean #left(fC#right);N"%vfat,
-                                100, np.mean(fitThr) - 5. * np.std(fitThr), np.mean(fitThr) + 5. * np.std(fitThr) )
+                                40, np.mean(fitThr) - 5. * np.std(fitThr), np.mean(fitThr) + 5. * np.std(fitThr) )
             histThresh.Sumw2()
-            for thresh in fitThr:
-                histThresh.Fill(thresh)
+            if np.std(fitThr) != 0: # Don't fill if we are still at initial values
+                for thresh in fitThr:
+                    histThresh.Fill(thresh)
             gThresh = r.TGraphErrors(histThresh)
             gThresh.SetName("gScurveMeanDist_vfat%i"%vfat)
             threshSummaryPlots[vfat] = gThresh
 
             # Make enc summary plot - bin size is variable
             histENC = r.TH1F("scurveSigma_vfat%i"%vfat,"VFAT %i;S-Curve Sigma #left(fC#right);N"%vfat,
-                                100, np.mean(fitENC) - 5. * np.std(fitENC), np.mean(fitENC) + 5. * np.std(fitENC) )
+                                40, np.mean(fitENC) - 5. * np.std(fitENC), np.mean(fitENC) + 5. * np.std(fitENC) )
             histENC.Sumw2()
-            for enc in fitENC:
-                histENC.Fill(enc)
+            if np.std(fitENC) != 0: # Don't fill if we are still at initial values
+                for enc in fitENC:
+                    histENC.Fill(enc)
             gENC = r.TGraphErrors(histENC)
             gENC.SetName("gScurveSigmaDist_vfat%i"%vfat)
-            encSummaryPlots[vfat] = gThresh
+            encSummaryPlots[vfat] = gENC
             pass
         pass
 
@@ -538,9 +543,9 @@ def plotAllSCurvesOnCanvas(vfatHistos, vfatHistosPanPin2=None, obsName="canvScur
             saveSummary(vSummaryPlotsNoHotChan, vSummaryPlotsNoHotChanPanPin2, '%s/PrunedSummary.png'%filename, trimVcal)
         else:
             saveSummary(vSummaryPlotsNoHotChan, None, '%s/PrunedSummary.png'%filename, trimVcal)
-        saveSummary(fitSummaryPlots, None, '%s/fitSummary.png'%filename, None)
-        saveSummary(threshSummaryPlots, None, '%s/ScurveMeanSummary.png'%filename, None)
-        saveSummary(encSummaryPlots, None, '%s/ScurveWidthSummary.png'%filename, None)
+        saveSummary(fitSummaryPlots, None, '%s/fitSummary.png'%filename, None, drawOpt="APE1")
+        saveSummary(threshSummaryPlots, None, '%s/ScurveMeanSummary.png'%filename, None, drawOpt="AP")
+        saveSummary(encSummaryPlots, None, '%s/ScurveWidthSummary.png'%filename, None, drawOpt="AP")
 
         confF = open(filename+'/chConfig.txt','w')
         confF.write('vfatN/I:vfatID/I:vfatCH/I:trimDAC/I:mask/I\n')
@@ -576,9 +581,9 @@ def plotAllSCurvesOnCanvas(vfatHistos, vfatHistosPanPin2=None, obsName="canvScur
             canvOfScurveFits[vfat].cd()
             for chan in range (0,128):
                 if chan == 0:
-                    fitter.scanFuncs[vfat][ch].Draw()
+                    fitter.scanFuncs[vfat][chan].Draw()
                 else:
-                    fitter.scanFuncs[vfat][ch].Draw("same")
+                    fitter.scanFuncs[vfat][chan].Draw("same")
             canvOfScurveFits[vfat].Update()
 
     # Save TObjects
@@ -590,12 +595,12 @@ def plotAllSCurvesOnCanvas(vfatHistos, vfatHistosPanPin2=None, obsName="canvScur
         dirVFAT.cd()
         vSummaryPlots[vfat].Write()
         if options.PanPin:
-            vSummaryPlotsPanPin2.Write()
+            vSummaryPlotsPanPin2[vfat].Write()
         canvOfScurveHistos[vfat].Write()
         if options.performFit:
-            vSummaryPlotsNoHotChan.Write()
+            vSummaryPlotsNoHotChan[vfat].Write()
             if options.PanPin:
-                vSummaryPlotsNoHotChanPanPin2.Write()
+                vSummaryPlotsNoHotChanPanPin2[vfat].Write()
             fitSummaryPlots[vfat].Write()
             threshSummaryPlots[vfat].Write()
             encSummaryPlots[vfat].Write()

fitting/fitScanData.py

@@ -112,11 +112,11 @@ def fit(self):
 
         random = r.TRandom3()
         random.SetSeed(0)
-        fitTF1 = r.TF1('myERF','[3]*TMath::Erf((TMath::Max([2],x)-[0])/(TMath::Sqrt(2)*[1]))+[3]',
-                        self.calDAC2Q_m[vfat]*1+self.calDAC2Q_b[vfat],self.calDAC2Q_m[vfat]*253+self.calDAC2Q_b[vfat])
         for vfat in range(0,24):
-            print 'fitting vfat %i'%vfat
+            fitTF1 = r.TF1('myERF','[3]*TMath::Erf((TMath::Max([2],x)-[0])/(TMath::Sqrt(2)*[1]))+[3]',
+                        self.calDAC2Q_m[vfat]*1+self.calDAC2Q_b[vfat],self.calDAC2Q_m[vfat]*253+self.calDAC2Q_b[vfat])
             for ch in range(0,128):
+                print 'fitting vfat %i chan %i'%(vfat,ch)
                 if self.isDead[vfat][ch]:
                     continue # Don't try to fit dead channels
                 elif not (self.scanHistos[vfat][ch].Integral() > 0):
@@ -126,19 +126,24 @@ def fit(self):
                 stepN = 0
                 while(stepN < 15):
                     rand = random.Gaus(10, 5)
+
+                    # Make sure the input parameters are positive
                     if (rand < 0.0 or rand > 100): continue
+                    #if self.calDAC2Q_m[vfat]*(8+stepN*8)+self.calDAC2Q_b[vfat] < 0: continue
+                    #if self.calDAC2Q_m[vfat]*(rand)+self.calDAC2Q_b[vfat] < 0: continue
+
                     # Provide an initial guess
                     fitTF1.SetParameter(0, self.calDAC2Q_m[vfat]*(8+stepN*8)+self.calDAC2Q_b[vfat] )
                     fitTF1.SetParameter(1, self.calDAC2Q_m[vfat]*(rand)+self.calDAC2Q_b[vfat])
                     fitTF1.SetParameter(2, self.calDAC2Q_m[vfat]*(8+stepN*8)+self.calDAC2Q_b[vfat])
                     fitTF1.SetParameter(3, self.Nev[vfat][ch]/2.)
 
                     # Set Parameter Limits
-                    fitTF1.SetParLimits(0, 0.01, self.calDAC2Q_m[vfat]*(300.0)+self.calDAC2Q_b[vfat])
-                    fitTF1.SetParLimits(1, 0.0,  self.calDAC2Q_m[vfat]*(100.0)+self.calDAC2Q_b[vfat])
-                    fitTF1.SetParLimits(2, 0.0,  self.calDAC2Q_m[vfat]*(300.0)+self.calDAC2Q_b[vfat])
+                    fitTF1.SetParLimits(0, 0.01, self.calDAC2Q_m[vfat]*(256)+self.calDAC2Q_b[vfat])
+                    fitTF1.SetParLimits(1, 0.0,  self.calDAC2Q_m[vfat]*(128)+self.calDAC2Q_b[vfat])
+                    fitTF1.SetParLimits(2, 0.0,  self.calDAC2Q_m[vfat]*(256)+self.calDAC2Q_b[vfat])
                     fitTF1.SetParLimits(3, 0.0,  self.Nev[vfat][ch] * 2.)
-
+                    
                     # Fit
                     fitResult = self.scanHistos[vfat][ch].Fit('myERF','SQ')
                     fitEmpty = fitResult.IsEmpty()