def main(args, opt):
trees = {}
for filename in os.listdir(args[0]):
if not os.path.splitext(filename)[1] == '.root': continue
isdata, pname, splitno = resolveFilename(filename)
treefile = os.path.join(args[0], filename)
tfile = ROOT.TFile.Open(treefile, 'READ')
trees[pname] = tfile.Get(TREENAME)
try:
cachefile = open('genmlbhists.pck', 'r')
hists = pickle.load(cachefile)
print '>>> Read syst histos from cache (.svlsysthistos.pck)'
cachefile.close()
except IOError:
hists = {}
for tag, sel in MLBSELECTIONS:
for pname in trees.keys():
print '... processing', pname, tag
hists[(pname, tag)] = getHistoFromTree(trees[pname],
sel=sel,
var='GenMlb',
hname="GenMlb_%s_%s" %
(pname, tag),
nbins=100,
xmin=0,
xmax=200,
titlex=MLBAXISTITLE)
cachefile = open('genmlbhists.pck', 'w')
pickle.dump(hists, cachefile, pickle.HIGHEST_PROTOCOL)
print '>>> Dumped histos to cachefile (genmlbhists.pck)'
cachefile.close()
ROOT.gStyle.SetOptTitle(0)
ROOT.gStyle.SetOptStat(0)
ROOT.gROOT.SetBatch(1)
for tag, _ in MLBSELECTIONS:
plot = RatioPlot('genmlb')
plot.normalized = False
plot.add(hists[('TTJets_MSDecays_172v5', tag)], 'Nominal')
plot.add(hists[('TTJets_MSDecays_scaleup', tag)], 'Q^{2} scale up')
plot.add(hists[('TTJets_MSDecays_scaledown', tag)], 'Q^{2} scale down')
plot.tag = "Generator level m_{lb} shape"
if tag == 'cor':
plot.subtag = "Correct combinations"
else:
plot.subtag = "Wrong combinations"
plot.ratiotitle = 'Ratio wrt nominal'
plot.titlex = MLBAXISTITLE
plot.tagpos = (0.22, 0.85)
plot.subtagpos = (0.22, 0.78)
plot.legpos = (0.20, 0.55)
plot.ratiorange = (0.85, 1.15)
plot.colors = [ROOT.kBlue - 3, ROOT.kRed - 4, ROOT.kOrange - 3]
plot.show("genmlb_scale_%s" % tag, opt.outDir)
return 0
def main(args, opt):
trees = {}
for filename in os.listdir(args[0]):
if not os.path.splitext(filename)[1] == '.root': continue
isdata, pname, splitno = resolveFilename(filename)
treefile = os.path.join(args[0], filename)
tfile = ROOT.TFile.Open(treefile,'READ')
trees[pname] = tfile.Get(TREENAME)
try:
cachefile = open('genmlbhists.pck', 'r')
hists = pickle.load(cachefile)
print '>>> Read syst histos from cache (.svlsysthistos.pck)'
cachefile.close()
except IOError:
hists = {}
for tag,sel in MLBSELECTIONS:
for pname in trees.keys():
print '... processing', pname, tag
hists[(pname,tag)] = getHistoFromTree(trees[pname],
sel=sel,
var='GenMlb',
hname="GenMlb_%s_%s"%(pname,tag),
nbins=100,xmin=0,xmax=200,
titlex=MLBAXISTITLE)
cachefile = open('genmlbhists.pck', 'w')
pickle.dump(hists, cachefile, pickle.HIGHEST_PROTOCOL)
print '>>> Dumped histos to cachefile (genmlbhists.pck)'
cachefile.close()
ROOT.gStyle.SetOptTitle(0)
ROOT.gStyle.SetOptStat(0)
ROOT.gROOT.SetBatch(1)
for tag,_ in MLBSELECTIONS:
plot = RatioPlot('genmlb')
plot.normalized = False
plot.add(hists[('TTJets_MSDecays_172v5', tag)], 'Nominal')
plot.add(hists[('TTJets_MSDecays_scaleup', tag)], 'Q^{2} scale up')
plot.add(hists[('TTJets_MSDecays_scaledown', tag)], 'Q^{2} scale down')
plot.tag = "Generator level m_{lb} shape"
if tag == 'cor':
plot.subtag = "Correct combinations"
else:
plot.subtag = "Wrong combinations"
plot.ratiotitle = 'Ratio wrt nominal'
plot.titlex = MLBAXISTITLE
plot.tagpos = (0.22, 0.85)
plot.subtagpos = (0.22, 0.78)
plot.legpos = (0.20, 0.55)
plot.ratiorange = (0.85, 1.15)
plot.colors = [ROOT.kBlue-3, ROOT.kRed-4, ROOT.kOrange-3]
plot.show("genmlb_scale_%s"%tag, opt.outDir)
return 0
def makeHistos((url, name)):
print "... processing", url
tree = ROOT.TFile.Open(url,'READ').Get(TREENAME)
histos = {}
# selection = "SVMassWeight*(SVLCombRank>0 && %s)"%INCSEL
selection = "(SVLCombRank>0 && %s)"%INCSEL
# selection = "(SVLCombRank>0 && %s)"%EMSEL
for varname, nbins, xmin, xmax, titlex in VARS:
for ntk1,ntk2 in NTRKBINS:
tksel = "(SVNtrk>=%d && SVNtrk<%d)"%(ntk1,ntk2)
finalsel = "%s*%s" %(selection,tksel)
histos[(varname, ntk1)] = getHistoFromTree(
tree, sel=finalsel, var=varname,
hname='%s_%s_%d'%(name, varname, ntk1),
nbins=nbins, xmin=xmin, xmax=xmax,
titlex=titlex)
return name,histos
def makeHistos((url, name)):
print "... processing", url
tree = ROOT.TFile.Open(url, 'READ').Get(TREENAME)
histos = {}
# selection = "SVMassWeight*(SVLCombRank>0 && %s)"%INCSEL
selection = "(SVLCombRank>0 && %s)" % INCSEL
# selection = "(SVLCombRank>0 && %s)"%EMSEL
for varname, nbins, xmin, xmax, titlex in VARS:
for ntk1, ntk2 in NTRKBINS:
tksel = "(SVNtrk>=%d && SVNtrk<%d)" % (ntk1, ntk2)
finalsel = "%s*%s" % (selection, tksel)
histos[(varname, ntk1)] = getHistoFromTree(tree,
sel=finalsel,
var=varname,
hname='%s_%s_%d' %
(name, varname, ntk1),
nbins=nbins,
xmin=xmin,
xmax=xmax,
titlex=titlex)
return name, histos
def main(args, options):
os.system('mkdir -p %s'%options.outDir)
try:
treefiles = {} # procname -> filename
for filename in os.listdir(args[0]):
if not os.path.splitext(filename)[1] == '.root': continue
procname = filename.split('_',1)[1][:-5]
treefiles[procname] = os.path.join(args[0],filename)
except OSError:
print "Not a valid input directory: %s" % args[0]
return -1
except IndexError:
print "Need to provide an input directory"
return -1
## Collect all the trees
svltrees = {} # proc -> tree
for proc in treefiles.keys():
tfile = ROOT.TFile.Open(treefiles[proc], 'READ')
if not filterUseless(treefiles[proc]): continue
svltrees[proc] = tfile.Get(TREENAME)
## Produce all the relevant histograms
if not options.cached:
masshistos = {} # (selection tag, process) -> histo
methistos = {} # (selection tag, process) -> histo
fittertkhistos = {} # (selection tag, process) -> [h_ntk1, h_ntk2, ...]
for tag,sel,_ in SELECTIONS:
for proc, tree in svltrees.iteritems():
if not filterUseless(treefiles[proc], sel): continue
htag = ("%s_%s"%(tag, proc)).replace('.','')
print ' ... processing %-30s %s htag=%s' % (proc, sel, htag)
masshistos[(tag, proc)] = getHistoFromTree(tree, sel=sel,
var='SVLMass',
hname="SVLMass_%s"%(htag),
titlex=MASSXAXISTITLE)
methistos[(tag, proc)] = getHistoFromTree(tree, sel=sel,
var='MET',
hname="MET_%s"%(htag),
xmin=0,xmax=200,
titlex="Missing E_{T} [GeV]")
fittertkhistos[(tag,proc)] = getNTrkHistos(tree, sel=sel,
tag=htag,
var='SVLMass',
titlex=MASSXAXISTITLE)
cachefile = open(".svlqcdmasshistos.pck", 'w')
pickle.dump(masshistos, cachefile, pickle.HIGHEST_PROTOCOL)
pickle.dump(methistos, cachefile, pickle.HIGHEST_PROTOCOL)
pickle.dump(fittertkhistos, cachefile, pickle.HIGHEST_PROTOCOL)
cachefile.close()
else:
cachefile = open(".svlqcdmasshistos.pck", 'r')
masshistos = pickle.load(cachefile)
methistos = pickle.load(cachefile)
fittertkhistos = pickle.load(cachefile)
cachefile.close()
#########################################################
## Write out the histograms, make data/mc plots
outputFileName = os.path.join(options.outDir,'qcd_DataMCHists.root')
ofi = ROOT.TFile(outputFileName, 'recreate')
ofi.cd()
for hist in [h for h in masshistos.values() + methistos.values()]:
hist.Write(hist.GetName())
for hist in [h for hists in fittertkhistos.values() for h in hists]:
hist.Write(hist.GetName())
ofi.Write()
ofi.Close()
## Run the plotter to get scaled MET plots
## Can then use those to subtract non-QCD backgrounds from data template
## Overwrite some of the options
options.filter = 'SVLMass,MET' ## only run SVLMass and MET plots
# options.excludeProcesses = 'QCD'
options.outFile = 'scaled_met_inputs.root'
options.cutUnderOverFlow = True
os.system('rm %s' % os.path.join(options.outDir, options.outFile))
runPlotter(outputFileName, options)
#########################################################
## Build the actual templates from the single histograms
templates = {} # selection tag -> template histo
inputfile = openTFile(os.path.join(options.outDir, options.outFile))
for tag,sel,_ in SELECTIONS:
categories = ['MET', 'SVLMass']
for x,_ in NTRKBINS:
categories.append('SVLMass_tot_%d'%int(x))
for category in categories:
plotdirname = '%s_%s'%(category,tag)
plotdir = inputfile.Get(plotdirname)
h_bg = None
h_data = None
for tkey in plotdir.GetListOfKeys():
key = tkey.GetName()
if key.startswith('Graph_from'): continue
if key.startswith('MC8TeV_QCD'): continue
hist = inputfile.Get('%s/%s'%(plotdirname,key))
if key.startswith('MC8TeV'):
if not h_bg:
h_bg = hist.Clone("%s_BGs" % tag)
else:
h_bg.Add(hist)
if key.startswith('Data8TeV'):
h_data = hist.Clone("%s_Data" % tag)
## Determine a suitable output name
histname = '%s_template'%tag
if category == 'MET': histname = "%s_%s" % ('met',histname)
if 'tot' in category:
tkbin = int(category.split('_')[2])
histname = "%s_%d" % (histname, tkbin)
h_data_subtr = h_data.Clone(histname)
## Now subtract the combined MC from the data
h_data_subtr.Add(h_bg, -1)
dicttag = tag
if category == 'MET':
dicttag = 'met_%s'%tag
if '_tot_' in category:
dicttag = '%s_%d' % (tag, tkbin)
templates[dicttag] = h_data_subtr
ofi = ROOT.TFile(os.path.join(options.outDir,'qcd_templates.root'),
'recreate')
ofi.cd()
for hist in templates.values(): hist.Write(hist.GetName())
ofi.Write()
ofi.Close()
for key,hist in sorted(templates.iteritems()):
print key
#########################################################
## Make a plot comparing the templates
for tag,_,seltag in SELECTIONS:
if not tag.endswith('_qcd'): continue
templateplot = RatioPlot('qcdtemplates_%s'%tag)
for key,hist in sorted(templates.iteritems()):
if not tag in key: continue
if key.startswith('met'): continue
if key == tag:
templateplot.add(hist, 'Inclusive')
templateplot.reference = hist
else:
ntrkbin = int(key.rsplit('_',1)[1])
templateplot.add(hist, 'N_{track} = %d'%ntrkbin)
templateplot.tag = 'QCD templates'
templateplot.subtag = seltag
templateplot.tagpos = (0.90, 0.85)
templateplot.subtagpos = (0.90, 0.78)
templateplot.legpos = (0.75, 0.25)
templateplot.ratiotitle = 'Ratio wrt Inclusive'
templateplot.extratext = '' #Work in Progress'
templateplot.ratiorange = (0.2, 2.2)
templateplot.colors = [ROOT.kBlack, ROOT.kBlue-8, ROOT.kAzure-2,
ROOT.kCyan-3]
templateplot.show("qcdtemplates_%s"%tag, options.outDir)
templateplot.reset()
return 0
def main(args, options):
os.system('mkdir -p %s' % options.outDir)
try:
treefiles = {} # procname -> filename
for filename in os.listdir(args[0]):
if not os.path.splitext(filename)[1] == '.root': continue
procname = filename.split('_', 1)[1][:-5]
treefiles[procname] = os.path.join(args[0], filename)
except OSError:
print "Not a valid input directory: %s" % args[0]
return -1
except IndexError:
print "Need to provide an input directory"
return -1
## Collect all the trees
svltrees = {} # proc -> tree
for proc in treefiles.keys():
tfile = ROOT.TFile.Open(treefiles[proc], 'READ')
if not filterUseless(treefiles[proc]): continue
svltrees[proc] = tfile.Get(TREENAME)
## Produce all the relevant histograms
if not options.cached:
masshistos = {} # (selection tag, process) -> histo
methistos = {} # (selection tag, process) -> histo
fittertkhistos = {
} # (selection tag, process) -> [h_ntk1, h_ntk2, ...]
for tag, sel, _ in SELECTIONS:
for proc, tree in svltrees.iteritems():
if not filterUseless(treefiles[proc], sel): continue
htag = ("%s_%s" % (tag, proc)).replace('.', '')
print ' ... processing %-30s %s htag=%s' % (proc, sel, htag)
masshistos[(tag, proc)] = getHistoFromTree(
tree,
sel=sel,
var='SVLMass',
hname="SVLMass_%s" % (htag),
titlex=MASSXAXISTITLE)
methistos[(tag, proc)] = getHistoFromTree(
tree,
sel=sel,
var='MET',
hname="MET_%s" % (htag),
xmin=0,
xmax=200,
titlex="Missing E_{T} [GeV]")
fittertkhistos[(tag,
proc)] = getNTrkHistos(tree,
sel=sel,
tag=htag,
var='SVLMass',
titlex=MASSXAXISTITLE)
cachefile = open(".svlqcdmasshistos.pck", 'w')
pickle.dump(masshistos, cachefile, pickle.HIGHEST_PROTOCOL)
pickle.dump(methistos, cachefile, pickle.HIGHEST_PROTOCOL)
pickle.dump(fittertkhistos, cachefile, pickle.HIGHEST_PROTOCOL)
cachefile.close()
else:
cachefile = open(".svlqcdmasshistos.pck", 'r')
masshistos = pickle.load(cachefile)
methistos = pickle.load(cachefile)
fittertkhistos = pickle.load(cachefile)
cachefile.close()
#########################################################
## Write out the histograms, make data/mc plots
outputFileName = os.path.join(options.outDir, 'qcd_DataMCHists.root')
ofi = ROOT.TFile(outputFileName, 'recreate')
ofi.cd()
for hist in [h for h in masshistos.values() + methistos.values()]:
hist.Write(hist.GetName())
for hist in [h for hists in fittertkhistos.values() for h in hists]:
hist.Write(hist.GetName())
ofi.Write()
ofi.Close()
## Run the plotter to get scaled MET plots
## Can then use those to subtract non-QCD backgrounds from data template
## Overwrite some of the options
options.filter = 'SVLMass,MET' ## only run SVLMass and MET plots
# options.excludeProcesses = 'QCD'
options.outFile = 'scaled_met_inputs.root'
options.cutUnderOverFlow = True
os.system('rm %s' % os.path.join(options.outDir, options.outFile))
runPlotter(outputFileName, options)
#########################################################
## Build the actual templates from the single histograms
templates = {} # selection tag -> template histo
inputfile = openTFile(os.path.join(options.outDir, options.outFile))
for tag, sel, _ in SELECTIONS:
categories = ['MET', 'SVLMass']
for x, _ in NTRKBINS:
categories.append('SVLMass_tot_%d' % int(x))
for category in categories:
plotdirname = '%s_%s' % (category, tag)
plotdir = inputfile.Get(plotdirname)
h_bg = None
h_data = None
for tkey in plotdir.GetListOfKeys():
key = tkey.GetName()
if key.startswith('Graph_from'): continue
if key.startswith('MC8TeV_QCD'): continue
hist = inputfile.Get('%s/%s' % (plotdirname, key))
if key.startswith('MC8TeV'):
if not h_bg:
h_bg = hist.Clone("%s_BGs" % tag)
else:
h_bg.Add(hist)
if key.startswith('Data8TeV'):
h_data = hist.Clone("%s_Data" % tag)
## Determine a suitable output name
histname = '%s_template' % tag
if category == 'MET': histname = "%s_%s" % ('met', histname)
if 'tot' in category:
tkbin = int(category.split('_')[2])
histname = "%s_%d" % (histname, tkbin)
h_data_subtr = h_data.Clone(histname)
## Now subtract the combined MC from the data
h_data_subtr.Add(h_bg, -1)
dicttag = tag
if category == 'MET':
dicttag = 'met_%s' % tag
if '_tot_' in category:
dicttag = '%s_%d' % (tag, tkbin)
templates[dicttag] = h_data_subtr
ofi = ROOT.TFile(os.path.join(options.outDir, 'qcd_templates.root'),
'recreate')
ofi.cd()
for hist in templates.values():
hist.Write(hist.GetName())
ofi.Write()
ofi.Close()
for key, hist in sorted(templates.iteritems()):
print key
#########################################################
## Make a plot comparing the templates
for tag, _, seltag in SELECTIONS:
if not tag.endswith('_qcd'): continue
templateplot = RatioPlot('qcdtemplates_%s' % tag)
for key, hist in sorted(templates.iteritems()):
if not tag in key: continue
if key.startswith('met'): continue
if key == tag:
templateplot.add(hist, 'Inclusive')
templateplot.reference = hist
else:
ntrkbin = int(key.rsplit('_', 1)[1])
templateplot.add(hist, 'N_{track} = %d' % ntrkbin)
templateplot.tag = 'QCD templates'
templateplot.subtag = seltag
templateplot.tagpos = (0.90, 0.85)
templateplot.subtagpos = (0.90, 0.78)
templateplot.legpos = (0.75, 0.25)
templateplot.ratiotitle = 'Ratio wrt Inclusive'
templateplot.extratext = '' #Work in Progress'
templateplot.ratiorange = (0.2, 2.2)
templateplot.colors = [
ROOT.kBlack, ROOT.kBlue - 8, ROOT.kAzure - 2, ROOT.kCyan - 3
]
templateplot.show("qcdtemplates_%s" % tag, options.outDir)
templateplot.reset()
return 0
