JESUncertainty_GraphUtils Namespace Reference

Functions
def	flagLowStatPoints (graph, debug)
	Avoid low statistics points. More...
def	flagLowStatPoints_noLargeErrorCheck (graph, debug)
	Avoid low statistics points, but do not check on upper bound of error. More...
def	eliminateLowStatPoints (graph, listOfBadPoints)
def	rescaleToGeV (graph)
def	printGraphContent (graph)
def	removeXErrors (graph)
def	removeYErrors (graph)
def	scaleGraph (graphToScale, scalingInputGraph, oneMinus=False)
def	extendLastGoodBin_2011 (graph, warn)
def	TGraphToTH1 (graph)
def	TGraphToTH1_withError (graph)

Function Documentation

eliminateLowStatPoints()

def JESUncertainty_GraphUtils.eliminateLowStatPoints	(		graph,
			listOfBadPoints
	)

Definition at line 91 of file JESUncertainty_GraphUtils.py.

def eliminateLowStatPoints(graph, listOfBadPoints) :
  
  #loop on all data points of the graph
  nPoints = graph.GetN()
 
  for iPoint in xrange(0,nPoints) :
    dataPointX = Double(0)
    dataPointY = Double(0)
    error = Double(0)
    errorX = graph.GetErrorX(iPoint)
    errorY = graph.GetErrorY(iPoint)
      
    graph.GetPoint(iPoint,dataPointX,dataPointY)
    
    if iPoint in listOfBadPoints :
      
       graph.SetPoint(iPoint, dataPointX, -999)
       graph.SetPointError(iPoint, errorX, 0)
 

extendLastGoodBin_2011()

def JESUncertainty_GraphUtils.extendLastGoodBin_2011	(		graph,
			warn
	)

Definition at line 214 of file JESUncertainty_GraphUtils.py.

def extendLastGoodBin_2011(graph, warn) :
  
  #loop on all data points of the graph
  nPoints = graph.GetN()
  
  #it should not happen that the first bin is bad, but just in case we keep it obvious
  lastGoodBin = -999
  lastGoodError = -999
  
  for iPoint in xrange(0,nPoints) :
    dataPointX = Double(0)
    dataPointY = Double(0)
    error = Double(0)
    errorX = graph.GetErrorX(iPoint)
    errorY = graph.GetErrorY(iPoint)
      
    graph.GetPoint(iPoint,dataPointX,dataPointY)
    
    #HACK: this is rather arbitrary!
    if dataPointY > 2. or errorY > 0.1:
      if warn : 
        print "WARNING in JESUncertainty_RatioUtils::extendLastGoodBin: "
        print "found bad point for graph:", graph.GetName(), "at point: ", iPoint, ", X:", dataPointX, "Y:", dataPointY
        print "extending previous point:", lastGoodBin
      newDataPointY=lastGoodBin
      newErrorY = lastGoodError
      graph.SetPoint(iPoint,dataPointX,newDataPointY)
      graph.SetPointError(iPoint, errorX, newErrorY)
    else : 
      lastGoodBin = dataPointY
      lastGoodError = errorY
 
 
#this fcn works with TGraphs with symmetric x errors

flagLowStatPoints()

def JESUncertainty_GraphUtils.flagLowStatPoints	(		graph,
			debug
	)

Avoid low statistics points.

Definition at line 14 of file JESUncertainty_GraphUtils.py.

def flagLowStatPoints (graph, debug) : 
 
  #stick all the indices of the bad points in this vector
  points = []
 
  #loop on all data points of the graph
  nPoints = graph.GetN()
 
  for iPoint in xrange(0,nPoints) :
    dataPointX = Double(0)
    dataPointY = Double(0)
    error = Double(0)
    errorX = graph.GetErrorX(iPoint)
    errorY = graph.GetErrorY(iPoint)
      
    graph.GetPoint(iPoint,dataPointX,dataPointY)
      
    
    if isnan(dataPointY) : 
      dataPointY = 0.000001
    #if dataPointY < 0.000001 : 
      #dataPointY = 0.000001
      points.append(iPoint)
      
    
    if errorY < 0.00000001 :
      points.append(iPoint)
 
    
    if errorY > 0.01 :
      points.append(iPoint)
      
  if debug == True:
    print "DEBUG: List of bad points for plot", graph.GetName(), ":"
    print points
  
  return points
 

flagLowStatPoints_noLargeErrorCheck()

def JESUncertainty_GraphUtils.flagLowStatPoints_noLargeErrorCheck	(		graph,
			debug
	)

Avoid low statistics points, but do not check on upper bound of error.

Definition at line 56 of file JESUncertainty_GraphUtils.py.

def flagLowStatPoints_noLargeErrorCheck (graph, debug) : 
 
  #stick all the indices of the bad points in this vector
  points = []
 
  #loop on all data points of the graph
  nPoints = graph.GetN()
 
  for iPoint in xrange(0,nPoints) :
    dataPointX = Double(0)
    dataPointY = Double(0)
    error = Double(0)
    errorX = graph.GetErrorX(iPoint)
    errorY = graph.GetErrorY(iPoint)
      
    graph.GetPoint(iPoint,dataPointX,dataPointY)
      
    
    if isnan(dataPointY) : 
      dataPointY = 0.000001
    #if dataPointY < 0.000001 : 
      #dataPointY = 0.000001
      points.append(iPoint)
      
    
    if errorY < 0.00000001 :
      points.append(iPoint)
 
      
  if debug == True:
    print "DEBUG: List of bad points for plot", graph.GetName(), ":"
    print points
  
  return points
 

printGraphContent()

def JESUncertainty_GraphUtils.printGraphContent

(

graph

)

Definition at line 127 of file JESUncertainty_GraphUtils.py.

def printGraphContent(graph) :
  
  #loop on all data points of the graph
  nPoints = graph.GetN()
  
  print graph.GetName()
  
  for iPoint in xrange(0,nPoints) :
    dataPointX = Double(0)
    dataPointY = Double(0)
    error = Double(0)
    errorX = graph.GetErrorX(iPoint)
    errorY = graph.GetErrorY(iPoint)
      
    graph.GetPoint(iPoint,dataPointX,dataPointY)
    
    print dataPointX, ":", dataPointY
    

removeXErrors()

def JESUncertainty_GraphUtils.removeXErrors

(

graph

)

Definition at line 145 of file JESUncertainty_GraphUtils.py.

def removeXErrors(graph) :
  
  #loop on all data points of the graph
  nPoints = graph.GetN()
 
  for iPoint in xrange(0,nPoints) :
    errorY = graph.GetErrorY(iPoint)      
    graph.SetPointError(iPoint, 0, errorY)
 

removeYErrors()

def JESUncertainty_GraphUtils.removeYErrors

(

graph

)

Definition at line 154 of file JESUncertainty_GraphUtils.py.

def removeYErrors(graph) :
  
  #loop on all data points of the graph
  nPoints = graph.GetN()
 
  for iPoint in xrange(0,nPoints) :
    errorX = graph.GetErrorX(iPoint)      
    graph.SetPointError(iPoint, errorX, 0.00000001)
 
 

rescaleToGeV()

def JESUncertainty_GraphUtils.rescaleToGeV

(

graph

)

Definition at line 110 of file JESUncertainty_GraphUtils.py.

def rescaleToGeV(graph) :
  
  #loop on all data points of the graph
  nPoints = graph.GetN()
 
  for iPoint in xrange(0,nPoints) :
    dataPointX = Double(0)
    dataPointY = Double(0)
    error = Double(0)
    errorX = graph.GetErrorX(iPoint)
    errorY = graph.GetErrorY(iPoint)
      
    graph.GetPoint(iPoint,dataPointX,dataPointY)
    
    graph.SetPoint(iPoint, dataPointX/1000, dataPointY)
    graph.SetPointError(iPoint, errorX/1000, errorY)
 

scaleGraph()

def JESUncertainty_GraphUtils.scaleGraph	(		graphToScale,
			scalingInputGraph,
			oneMinus = False
	)

Definition at line 164 of file JESUncertainty_GraphUtils.py.

def scaleGraph(graphToScale, scalingInputGraph, oneMinus = False) :
  
  #get number of points for the two graphs
  nPointsGraphToScale= graphToScale.GetN()
  nPointsScalingInputGraph = scalingInputGraph.GetN()
 
  if nPointsGraphToScale!=nPointsScalingInputGraph :
    print "ERROR in JESUncertainty_RatioUtils::scaleGraph: graphs don't have the same number of points"
    return None
   
  #prepare the final graph
  scaledGraph = graphToScale.Clone()
  scaledGraph.SetName(graphToScale.GetName()+"_Scaled")
  
  #loop on points of the graphs
  for iPoint in xrange(0,nPointsGraphToScale) :
    
    #make the new point
    maxPoint = Double(0)
    maxPointError = Double(0)
    
    #retrieve the point
    dataPointXgraphToScale = Double(0)
    dataPointYgraphToScale = Double(0)
    errorXgraphToScale= graphToScale.GetErrorX(iPoint)
    errorYgraphToScale = graphToScale.GetErrorY(iPoint)
    graphToScale.GetPoint(iPoint,dataPointXgraphToScale,dataPointYgraphToScale)
    
    dataPointXscalingInputGraph = Double(0)
    dataPointYscalingInputGraph = Double(0)
    errorXscalingInputGraph = scalingInputGraph.GetErrorX(iPoint)
    errorYscalingInputGraph = scalingInputGraph.GetErrorY(iPoint)
    scalingInputGraph.GetPoint(iPoint,dataPointXscalingInputGraph,dataPointYscalingInputGraph)
 
    if abs(dataPointXgraphToScale-dataPointXscalingInputGraph) > 0.0001:
      print "ERROR in JESUncertainty_RatioUtils::scaleGraph: x coordinates of data points do not match"
      return None
 
    #set the point/error of the output graph to the scaled point/error of the input one (protect against negative scaling)
    if abs(dataPointYscalingInputGraph) < 0:
      print "WARNING in JESUncertainty_RatioUtils::scaleGraph: negative scale factor for point: ", iPoint, ", X:", dataPointXgraphToScale
    scaledGraph.SetPoint(iPoint, dataPointXgraphToScale, dataPointYscalingInputGraph*dataPointYgraphToScale)
    scaledGraph.SetPointError(iPoint, errorXgraphToScale, abs(dataPointYscalingInputGraph)*dataPointYgraphToScale)
    
    if oneMinus : 
      scaledGraph.SetPoint(iPoint, dataPointXgraphToScale, (1-dataPointYscalingInputGraph)*dataPointYgraphToScale)
      scaledGraph.SetPointError(iPoint, errorXgraphToScale, abs(dataPointYscalingInputGraph)*dataPointYgraphToScale)
    
  return scaledGraph
  

TGraphToTH1()

def JESUncertainty_GraphUtils.TGraphToTH1

(

graph

)

Definition at line 248 of file JESUncertainty_GraphUtils.py.

def TGraphToTH1 (graph) :
  
    #loop on all data points
    nPoints = graph.GetN()
    #print "nPoints", nPoints
    
    x=array('f')
    y=array('f')
    ey=array('f')
 
    dataPointX = None
    dataErrorX = None
    
    for iPoint in xrange(0,nPoints) :
        #get info on data point
        dataPointX = Double(0)
        dataPointY = Double(0)
        graph.GetPoint(iPoint,dataPointX,dataPointY)
        dataErrorX = graph.GetErrorX(iPoint)
        dataErrorY = graph.GetErrorY(iPoint)
 
        x.append(dataPointX-dataErrorX)
        y.append(dataPointY)
        ey.append(dataErrorY)
        
    #this needs to be done at the end of the loop, for the upper limit
    x.append(dataPointX+dataErrorX)
    
    #safety checks
    for iPoint in xrange(0,nPoints) :
        if isnan(y[iPoint]) : 
          y[iPoint]=0
        if isnan(x[iPoint]) : 
          x[iPoint]=0
        if y[iPoint]<0.00000001 :
          y[iPoint]=0
          ey[iPoint]=0
        
    myHist=TH1F("Hist"+graph.GetName(),graph.GetTitle(),nPoints,x)
 
    for iPoint in xrange(0,nPoints) :
 
        myHist.SetBinContent(iPoint+1,y[iPoint])#the +1 is needed because TH1 have overflows
        #myHist.SetBinError(iPoint+1,ey[iPoint])#the +1 is needed because TH1 have overflows
        myHist.SetBinError(iPoint+1,0)#avoid poissonian errors
 
    return myHist
 
#this fcn works with TGraphs with symmetric x errors

TGraphToTH1_withError()

def JESUncertainty_GraphUtils.TGraphToTH1_withError

(

graph

)

Definition at line 297 of file JESUncertainty_GraphUtils.py.

def TGraphToTH1_withError (graph) :
  
    #loop on all data points
    nPoints = graph.GetN()
    #print "nPoints", nPoints
    
    x=array('f')
    y=array('f')
    ey=array('f')
 
    dataPointX = None
    dataErrorX = None
    
    for iPoint in xrange(0,nPoints) :
        #get info on data point
        dataPointX = Double(0)
        dataPointY = Double(0)
        graph.GetPoint(iPoint,dataPointX,dataPointY)
        dataErrorX = graph.GetErrorX(iPoint)
        dataErrorY = graph.GetErrorY(iPoint)
 
        x.append(dataPointX-dataErrorX)
        y.append(dataPointY)
        ey.append(dataErrorY)
        
    #this needs to be done at the end of the loop, for the upper limit
    x.append(dataPointX+dataErrorX)
    
    #safety checks
    for iPoint in xrange(0,nPoints) :
        if isnan(y[iPoint]) : 
          y[iPoint]=0
        if isnan(x[iPoint]) : 
          x[iPoint]=0
        if y[iPoint]<0.00000001 :
          y[iPoint]=0
          ey[iPoint]=0
        
    myHist=TH1F("Hist"+graph.GetName(),graph.GetTitle(),nPoints,x)
 
    for iPoint in xrange(0,nPoints) :
 
        myHist.SetBinContent(iPoint+1,y[iPoint])#the +1 is needed because TH1 have overflows
        myHist.SetBinError(iPoint+1,ey[iPoint])#the +1 is needed because TH1 have overflows
        #myHist.SetBinError(iPoint+1,0)#avoid poissonian errors
 
    return myHist