def prepareDYScaleFactors(filename, plotfile, inputdir, options,
dytag="DY+Jets",
unmergedfile="MC8TeV_TTJets_MSDecays_172v5.root"):
"""
Prepares a dictionary of (key, process title) -> scale factor,
to be used in the runPlotter script to scale DY by a data-derived
scale factor.
A list of keys to be scaled is generated from a plotfile (for merged
histograms), and from one file of an input directory (for unmerged ones).
"""
## First calculate the actual scalefactors
DYSFs = extractFactors(filename, options=options)
## Get a list of all keys that are to be scaled
allkeys = []
## Need the list of processes
tagsToFilter = ['_ee', '_mm', '_mumu']
## From merged plots:
allkeys += getAllPlotsFrom(openTFile(plotfile),
tagsToFilter=tagsToFilter,
filterByProcsFromJSON=options.json)
## From unmerged plots:
allkeys += getAllPlotsFrom(openTFile(os.path.join(inputdir,
unmergedfile)),
tagsToFilter=tagsToFilter)
## Prepare dictionary of (key,tag) -> scale factor
scaleFactors = {}
if options.verbose > 0:
print 'Will scale the following histograms for %s:' % dytag
for key in allkeys:
if '_ee' in key:
scaleFactors[(key, dytag)] = DYSFs['ee']
if options.verbose > 0:
print ' %-25s: %5.3f' % (key, DYSFs['ee'])
else:
scaleFactors[(key, dytag)] = DYSFs['mm']
if options.verbose > 0:
print ' %-25s: %5.3f' % (key, DYSFs['mm'])
return scaleFactors
def main(args, opt):
try:
cachefile = open(".eventyields.pck", 'r')
yields = pickle.load(cachefile)
cachefile.close()
return printYields(yields)
except EOFError:
print "Error loading pickle file, please delete it and try again"
return -1
except IOError:
pass
try:
tfile = openTFile(args[0])
except ReferenceError:
print "Please provide a valid input file"
return -1
ROOT.gStyle.SetOptTitle(0)
ROOT.gStyle.SetOptStat(0)
ROOT.gROOT.SetBatch(1)
allkeys = getAllPlotsFrom(tfile, tagsToFilter=[opt.dist])
channels = ['e', 'ee', 'em', 'm', 'mm']
print 30 * '-'
yields = {}
for chan in channels:
for key in allkeys:
histname = "%s_%s" % (opt.dist, chan)
if not key.endswith(histname): continue
hist = tfile.Get(key)
err = ROOT.Double(0.0)
firstBin = opt.firstBin
lastBin = hist.GetXaxis().GetNbins(
) if opt.lastBin < opt.firstBin else opt.lastBin
integral = hist.IntegralAndError(firstBin, lastBin, err)
err = float(err)
proc = key.rsplit('/', 1)[1]
proc = proc.replace('_%s' % histname, '')
yields[(chan, proc)] = (integral, err)
cachefile = open(".eventyields.pck", 'w')
pickle.dump(yields, cachefile, pickle.HIGHEST_PROTOCOL)
cachefile.close()
return printYields(yields)
def extractFactors(inputFile, options):
try:
cachefile = open('.svldyscalefactors.pck','r')
scaleFactors = pickle.load(cachefile)
cachefile.close()
print ('>>> Read Drell-Yan scale factors from cache '
'(.svldyscalefactors.pck)')
return scaleFactors
except IOError:
pass
try:
tfile = openTFile(inputFile)
except ReferenceError:
print "Please provide a valid input file"
allkeys = getAllPlotsFrom(tfile, tagsToFilter=HISTSTOPROCESS)
scaleFactors = {}
for histname in HISTSTOPROCESS:
tot_data = 0
tot_background = 0
tot_signal = 0
for key in allkeys:
if not key.endswith(histname): continue
integral = tfile.Get(key).Integral()
if 'Data8TeV' in key:
tot_data += integral
elif SIGNAL in key:
tot_signal += integral
else:
tot_background += integral
channame = histname.rsplit('_',1)[1]
if options.verbose>0:
print 30*'-'
print histname
print ("Total data: %6.1f, total bg: %6.1f, total signal: %6.1f "
% (tot_data, tot_background, tot_signal))
# tot_background + SF * tot_signal = tot_data
SF = (tot_data-tot_background)/tot_signal
if options.verbose>0:
print " -> scale factor: %5.3f" % SF
scaleFactors[channame] = SF
if options.verbose>0: print 30*'-'
cachefile = open(".svldyscalefactors.pck", 'w')
pickle.dump(scaleFactors, cachefile, pickle.HIGHEST_PROTOCOL)
cachefile.close()
return scaleFactors
def main(args, opt):
try:
cachefile = open(".eventyields.pck", 'r')
yields = pickle.load(cachefile)
cachefile.close()
return printYields(yields)
except EOFError:
print "Error loading pickle file, please delete it and try again"
return -1
except IOError:
pass
try:
tfile = openTFile(args[0])
except ReferenceError:
print "Please provide a valid input file"
return -1
ROOT.gStyle.SetOptTitle(0)
ROOT.gStyle.SetOptStat(0)
ROOT.gROOT.SetBatch(1)
allkeys = getAllPlotsFrom(tfile, tagsToFilter=[opt.dist])
channels = ['e', 'ee', 'em', 'm', 'mm']
print 30*'-'
yields = {}
for chan in channels:
for key in allkeys:
histname = "%s_%s"%(opt.dist,chan)
if not key.endswith(histname): continue
hist = tfile.Get(key)
err = ROOT.Double(0.0)
firstBin=opt.firstBin
lastBin=hist.GetXaxis().GetNbins() if opt.lastBin<opt.firstBin else opt.lastBin
integral = hist.IntegralAndError(firstBin,lastBin, err)
err = float(err)
proc = key.rsplit('/',1)[1]
proc = proc.replace('_%s'%histname, '')
yields[(chan,proc)] = (integral, err)
cachefile = open(".eventyields.pck", 'w')
pickle.dump(yields, cachefile, pickle.HIGHEST_PROTOCOL)
cachefile.close()
return printYields(yields)
def extractLJNTrkWeights(inputFile=None, verbose=0):
try:
tfile = openTFile(inputFile)
except ReferenceError:
print "[extractLJNTrkWeights] Please provide a valid input file"
allkeys = getAllPlotsFrom(tfile, tagsToFilter=HISTSTOPROCESS)
ofile = ROOT.TFile.Open('ljntkweights.root','recreate')
ofile.cd()
for histname in HISTSTOPROCESS:
tot_data, tot_mc = None, None
for key in allkeys:
if not key.endswith(histname): continue
hist = tfile.Get(key)
integral = hist.Integral()
if 'Data8TeV' in key:
if not tot_data:
tot_data = hist.Clone("SVNtrk_data")
tot_data.SetDirectory(0)
else:
print "WARNING: Found two data histograms"
tot_data = hist.Clone("SVNtrk_data")
tot_data.SetDirectory(0)
if 'MC8TeV_TTJets_MSDecays_172v5' in key:
if not tot_mc:
tot_mc = hist.Clone("SVNtrk_mc")
tot_mc.SetDirectory(0)
else:
tot_mc.Add(hist)
## Normalize them
tot_data.Scale(1.0/tot_data.Integral())
tot_mc.Scale(1.0/tot_mc.Integral())
ratio = tot_data.Clone("%s_weights"%histname)
ratio.Divide(tot_mc)
ofile.cd()
ratio.Write(ratio.GetName())
ofile.Write()
ofile.Close()
print ">>> Wrote light jet ntrk weights to ljntkweights.root"
def extractLJNTrkWeights(inputFile=None, verbose=0):
try:
tfile = openTFile(inputFile)
except ReferenceError:
print "[extractLJNTrkWeights] Please provide a valid input file"
allkeys = getAllPlotsFrom(tfile, tagsToFilter=HISTSTOPROCESS)
ofile = ROOT.TFile.Open('ljntkweights.root', 'recreate')
ofile.cd()
for histname in HISTSTOPROCESS:
tot_data, tot_mc = None, None
for key in allkeys:
if not key.endswith(histname): continue
hist = tfile.Get(key)
integral = hist.Integral()
if 'Data8TeV' in key:
if not tot_data:
tot_data = hist.Clone("SVNtrk_data")
tot_data.SetDirectory(0)
else:
print "WARNING: Found two data histograms"
tot_data = hist.Clone("SVNtrk_data")
tot_data.SetDirectory(0)
if 'MC8TeV_TTJets_MSDecays_172v5' in key:
if not tot_mc:
tot_mc = hist.Clone("SVNtrk_mc")
tot_mc.SetDirectory(0)
else:
tot_mc.Add(hist)
## Normalize them
tot_data.Scale(1.0 / tot_data.Integral())
tot_mc.Scale(1.0 / tot_mc.Integral())
ratio = tot_data.Clone("%s_weights" % histname)
ratio.Divide(tot_mc)
ofile.cd()
ratio.Write(ratio.GetName())
ofile.Write()
ofile.Close()
print ">>> Wrote light jet ntrk weights to ljntkweights.root"
def compareNtrk(inputFile=None, verbose=0):
xsecweights=readXSecWeights()
try:
tfile = openTFile(inputFile)
except ReferenceError:
print "Please provide a valid input file"
for reg in ['regA','regB','regC']:
for ch in ['e','m','mm','em','ee']:
histToProcess='SVNtrk_%s_optmrank' % ch
allkeys = getAllPlotsFrom(tfile, tagsToFilter=[histToProcess])
data=None
mc=None
for key in allkeys:
if not (reg in key) : continue
weight=1.0
for xsecKey in xsecweights:
procName=xsecKey.replace('MC8TeV','')+'_'
if procName in key : weight =xsecweights[xsecKey]*LUMI
if '2012' in key:
try:
data.Add( tfile.Get(key), weight )
except:
data=tfile.Get(key).Clone('data')
data.Scale(weight)
else:
try:
mc.Add( tfile.Get(key), weight )
except:
mc=tfile.Get(key).Clone('mc')
mc.Scale(weight)
varPlot = Plot('%s_%s_ntrks'%(ch,reg),False)
varPlot.add(mc, 'expected', ROOT.kGray, False)
varPlot.add(data, 'data', 1, True)
varPlot.show('~/www/tmp')
varPlot.reset()
def compareNtrk(inputFile=None, verbose=0):
xsecweights = readXSecWeights()
try:
tfile = openTFile(inputFile)
except ReferenceError:
print "Please provide a valid input file"
for reg in ['regA', 'regB', 'regC']:
for ch in ['e', 'm', 'mm', 'em', 'ee']:
histToProcess = 'SVNtrk_%s_optmrank' % ch
allkeys = getAllPlotsFrom(tfile, tagsToFilter=[histToProcess])
data = None
mc = None
for key in allkeys:
if not (reg in key): continue
weight = 1.0
for xsecKey in xsecweights:
procName = xsecKey.replace('MC8TeV', '') + '_'
if procName in key: weight = xsecweights[xsecKey] * LUMI
if '2012' in key:
try:
data.Add(tfile.Get(key), weight)
except:
data = tfile.Get(key).Clone('data')
data.Scale(weight)
else:
try:
mc.Add(tfile.Get(key), weight)
except:
mc = tfile.Get(key).Clone('mc')
mc.Scale(weight)
varPlot = Plot('%s_%s_ntrks' % (ch, reg), False)
varPlot.add(mc, 'expected', ROOT.kGray, False)
varPlot.add(data, 'data', 1, True)
varPlot.show('~/www/tmp')
varPlot.reset()
def extractNTrkWeights(inputFile=None, verbose=0):
try:
cachefile = open('.svntrkweights.pck', 'r')
ntkWeights = pickle.load(cachefile)
cachefile.close()
print('>>> Read SV ntrk weights from cache ' '(.svntrkweights.pck)')
return ntkWeights
except IOError:
pass
try:
tfile = openTFile(inputFile)
except ReferenceError:
print "[extractNTrkWeights] Please provide a valid input file"
allkeys = getAllPlotsFrom(tfile, tagsToFilter=HISTSTOPROCESS)
ntkWeights = {}
for histname in HISTSTOPROCESS:
tot_data, tot_mc = None, None
for key in allkeys:
if not key.endswith(histname): continue
hist = tfile.Get(key)
integral = hist.Integral()
if 'Data8TeV' in key:
if not tot_data:
tot_data = hist.Clone("SVNtrk_data")
else:
print "WARNING: Found two data histograms"
tot_data = hist.Clone("SVNtrk_data")
else:
if not tot_mc:
tot_mc = hist.Clone("SVNtrk_mc")
else:
tot_mc.Add(hist)
ratio = tot_data.Clone("ratio")
ratio.Divide(tot_mc)
key = 'inclusive'
if len(histname.split('_')) > 2:
key = histname.rsplit('_', 1)[1]
if not key in ntkWeights:
ntkWeights[key] = {}
if verbose > 0:
print '---------------------------'
print histname
for x in xrange(1, ratio.GetNbinsX() + 1):
if verbose > 0:
print("Ntrk=%1d: Ndata: %6d / Nmc: %8.1f , weight = %5.3f" %
(x + 1, tot_data.GetBinContent(x),
tot_mc.GetBinContent(x), ratio.GetBinContent(x)))
ntkWeights[key][x + 1] = ratio.GetBinContent(x)
cachefile = open(".svntrkweights.pck", 'w')
pickle.dump(ntkWeights, cachefile, pickle.HIGHEST_PROTOCOL)
cachefile.close()
print ">>> Dumped SV ntrk weights to .svntrkweights.pck"
return ntkWeights
def extractNTrkWeights(inputFile=None, verbose=0):
try:
cachefile = open('.svntrkweights.pck','r')
ntkWeights = pickle.load(cachefile)
cachefile.close()
print ('>>> Read SV ntrk weights from cache '
'(.svntrkweights.pck)')
return ntkWeights
except IOError:
pass
try:
tfile = openTFile(inputFile)
except ReferenceError:
print "[extractNTrkWeights] Please provide a valid input file"
allkeys = getAllPlotsFrom(tfile, tagsToFilter=HISTSTOPROCESS)
ntkWeights = {}
for histname in HISTSTOPROCESS:
tot_data, tot_mc = None, None
for key in allkeys:
if not key.endswith(histname): continue
hist = tfile.Get(key)
integral = hist.Integral()
if 'Data8TeV' in key:
if not tot_data:
tot_data = hist.Clone("SVNtrk_data")
else:
print "WARNING: Found two data histograms"
tot_data = hist.Clone("SVNtrk_data")
else:
if not tot_mc:
tot_mc = hist.Clone("SVNtrk_mc")
else:
tot_mc.Add(hist)
ratio = tot_data.Clone("ratio")
ratio.Divide(tot_mc)
key = 'inclusive'
if len(histname.split('_'))>2:
key = histname.rsplit('_',1)[1]
if not key in ntkWeights:
ntkWeights[key] = {}
if verbose>0:
print '---------------------------'
print histname
for x in xrange(1,ratio.GetNbinsX()+1):
if verbose>0:
print ("Ntrk=%1d: Ndata: %6d / Nmc: %8.1f , weight = %5.3f" %
(x+1,
tot_data.GetBinContent(x),
tot_mc.GetBinContent(x),
ratio.GetBinContent(x)))
ntkWeights[key][x+1] = ratio.GetBinContent(x)
cachefile = open(".svntrkweights.pck", 'w')
pickle.dump(ntkWeights, cachefile, pickle.HIGHEST_PROTOCOL)
cachefile.close()
print ">>> Dumped SV ntrk weights to .svntrkweights.pck"
return ntkWeights
