def makeControlPlot(systhistos, syst, tag, seltag, opt):
hists = tuple([systhistos[(tag, syst, c)] for c in COMBINATIONS.keys()])
h_tot, h_cor, h_wro, h_unm = hists
h_tot.Scale(1./h_tot.Integral())
ctrlplot = RatioPlot('ctrlplot_%s'%tag)
ctrlplot.add(h_cor, 'Correct')
ctrlplot.add(h_wro, 'Wrong')
ctrlplot.add(h_unm, 'Unmatched')
ctrlplot.reference = h_tot
ctrlplot.tag = ''
ctrlplot.subtag = seltag
ctrlplot.ratiotitle = 'Ratio wrt Total'
ctrlplot.ratiorange = (0., 3.0)
ctrlplot.colors = [ROOT.kBlue-3, ROOT.kRed-4, ROOT.kOrange-3]
ctrlplot.show("control_%s_%s"%(syst,tag), os.path.join(opt.outDir, 'syst_plots'))
ctrlplot.reset()
def main(args, opt):
os.system('mkdir -p %s'%opt.outDir)
systfiles = {} # procname -> filename
try:
for fname in os.listdir(os.path.join(args[0],'syst')):
if not os.path.splitext(fname)[1] == '.root': continue
for syst,_,systfile,_ in SYSTSFROMFILES:
if fname in systfile:
systfiles[syst] = [os.path.join(args[0], 'syst', fname)]
# Get the split nominal files
systfiles['nominal'] = []
for fname in os.listdir(os.path.join(args[0],'Chunks')):
if not os.path.splitext(fname)[1] == '.root': continue
isdata,procname,splitno = resolveFilename(fname)
if not procname == 'TTJets_MSDecays_172v5': continue
if not splitno: continue # file is split
systfiles['nominal'].append(os.path.join(args[0],'Chunks',fname))
if len(systfiles['nominal']) < 20:
print "ERROR >>> Missing files for split nominal sample?"
return -1
except IndexError:
print "Please provide a valid input directory"
exit(-1)
hname_to_keys = {} # hname -> (tag, syst, comb)
tasklist = {} # treefile -> tasklist
for fsyst in systfiles.keys():
if not fsyst in tasklist: tasklist[fsyst] = []
for tag,sel,_ in SELECTIONS:
if fsyst == 'nominal':
for syst,_,weight,combs in SYSTSFROMWEIGHTS:
tasks = makeSystTask(tag, sel, syst,
hname_to_keys, weight=weight,
combs=combs)
tasklist[fsyst] += tasks
tasks = []
for comb,combsel in COMBINATIONS.iteritems():
for var,nbins,xmin,xmax,titlex in CONTROLVARS:
hname = "%s_%s_%s" % (var, comb, tag)
finalsel = "%s*(%s&&%s)"%(COMMONWEIGHT, sel, combsel)
tasks.append((hname, var, finalsel,
nbins, xmin, xmax, titlex))
hname_to_keys[hname] = (tag, var, comb)
tasklist[fsyst] += tasks
tasks = []
for name, nus in [('nu', 1), ('nonu', 0), ('nuunm', -1)]:
hname = "SVLMass_%s_%s_%s" % ('tot', tag, name)
finalsel = "%s*(%s&&BHadNeutrino==%d)"%(
COMMONWEIGHT, sel, nus)
tasks.append((hname, 'SVLMass', finalsel,
NBINS, XMIN, XMAX, MASSXAXISTITLE))
hname_to_keys[hname] = (tag, name, 'tot')
tasklist[fsyst] += tasks
else:
tasks = makeSystTask(tag, sel, fsyst, hname_to_keys)
tasklist[fsyst] += tasks
if not opt.cache:
# print ' Will process the following tasks:'
# for filename,tasks in sorted(tasklist.iteritems()):
# print filename
# for task in tasks:
# print task
# raw_input("Press any key to continue...")
runTasks(systfiles, tasklist, opt, 'syst_histos')
systhistos = {} # (tag, syst, comb) -> histo
systhistos = gatherHistosFromFiles(tasklist, systfiles,
os.path.join(opt.outDir, 'syst_histos'),
hname_to_keys)
cachefile = open(".svlsysthistos.pck", 'w')
pickle.dump(systhistos, cachefile, pickle.HIGHEST_PROTOCOL)
cachefile.close()
# print "Wrote syst histos to cache file"
# raw_input("press key")
cachefile = open(".svlsysthistos.pck", 'r')
systhistos = pickle.load(cachefile)
print '>>> Read syst histos from cache (.svlsysthistos.pck)'
cachefile.close()
ROOT.gStyle.SetOptTitle(0)
ROOT.gStyle.SetOptStat(0)
ROOT.gROOT.SetBatch(1)
for var,_,_,_,_ in CONTROLVARS:
for sel,tag in [
#('inclusive', 'Fully Inclusive'),
#('inclusive_mrank1', 'Mass ranked, leading p_{T}'),
#('inclusive_mrank1dr', 'Mass ranked, #DeltaR<2, leading p_{T}'),
#('inclusive_drrank1dr', '#DeltaR ranked, #DeltaR<2, leading p_{T}'),
('inclusive_optmrank', 'Optimized mass rank')]:
try: makeControlPlot(systhistos, var, sel, tag, opt)
except KeyError:
print 'control plots for %s selection not found' % sel
for tag,_,_ in SELECTIONS:
if not 'inclusive' in tag : continue
print "... processing %s"%tag
# Make plot of mass with and without neutrino:
# for comb in COMBINATIONS.keys():
# plot = RatioPlot('neutrino_%s'%tag)
# plot.rebin = 2
# plot.add(systhistos[(tag,'nonu', 'tot')], 'Without neutrino')
# plot.add(systhistos[(tag,'nu', 'tot')], 'With neutrino')
# plot.add(systhistos[(tag,'nuunm','tot')], 'Unmatched')
# plot.reference = systhistos[(tag,'nominal','tot')]
# plot.tag = "Mass shape with and without neutrinos"
# plot.subtag = SELNAMES[tag] + COMBNAMES['tot']
# plot.ratiotitle = 'Ratio wrt Total'
# plot.ratiorange = (0.7, 1.3)
# plot.colors = [ROOT.kBlue-3, ROOT.kRed-4, ROOT.kOrange-3]
# plot.show("neutrino_%s_%s"%(tag,'tot'),
# os.path.join(opt.outDir, 'syst_plots'))
# plot.reset()
for name, title, systs, colors, comb in SYSTPLOTS:
print name, title, systs, colors, comb
plot = RatioPlot('%s_%s'%(name,comb))
plot.rebin = 2
for syst in systs:
try:
plot.add(systhistos[(tag,syst,comb)], SYSTNAMES[syst])
except:
print 'failed to add',(tag,syst,comb),syst
plot.tag = title
subtag = SELNAMES[tag] + COMBNAMES[comb]
plot.subtag = subtag
plot.ratiotitle = 'Ratio wrt %s' % SYSTNAMES[systs[0]]
plot.ratiorange = (0.85, 1.15)
plot.colors = colors
filename = "%s_%s"%(name,tag)
if comb != 'tot': filename += '_%s'%comb
plot.show(filename, os.path.join(opt.outDir,'syst_plots'))
plot.reset()
# Make top pt plot with both correct and wrong
plot = RatioPlot('toppt_paper_cor_wro')
plot.canvassize = (600,600)
plot.tag = 'Top p_{T} mis-modeling'
plot.rebin = 2
plot.subtag = 'Inclusive channels'
plot.ratiotitle = '1 / Nominal'
plot.ratiorange = (0.85, 1.15)
plot.legpos = (0.55, 0.30)
plot.colors = [ROOT.kGreen+2, ROOT.kGreen-6, ROOT.kRed+2, ROOT.kRed-6]
plot.add(systhistos[(tag,'nominal','cor')], 'Nominal (correct)', includeInRatio=False)
plot.add(systhistos[(tag,'toppt','cor')], 'p_{T} weighted (correct)', includeInRatio=True)
plot.add(systhistos[(tag,'nominal','wro')], 'Nominal (wrong)', includeInRatio=False)
plot.add(systhistos[(tag,'toppt','wro')], 'p_{T} weighted (wrong)', includeInRatio=True)
plot.reference = [systhistos[(tag,'nominal','cor')], systhistos[(tag,'nominal','wro')]]
plot.show('toppt_cor_wro_forpaper_%s'%tag, os.path.join(opt.outDir,'syst_plots'))
plot.reset()
# Make b fragmentation plot for paper
plot = RatioPlot('bfrag_paper')
plot.canvassize = (600,600)
plot.tag = 'b fragmentation'
plot.rebin = 2
plot.subtag = 'Inclusive channels'
plot.ratiotitle = '1 / Z2* rb LEP'
plot.ratiorange = (0.85, 1.15)
plot.legpos = (0.65, 0.15)
plot.colors = [ROOT.kMagenta, ROOT.kMagenta+2, ROOT.kMagenta-9, ROOT.kAzure+7]
plot.add(systhistos[(tag,'nominal', 'tot')], 'Z2* rb LEP', includeInRatio=False)
plot.add(systhistos[(tag,'bfragdn', 'tot')], 'Z2* rb LEP soft')
plot.add(systhistos[(tag,'bfragup', 'tot')], 'Z2* rb LEP hard')
plot.add(systhistos[(tag,'bfragz2s','tot')], 'Z2* nominal')
plot.reference = [systhistos[(tag,'nominal','tot')]]
plot.show('bfrag_paper_%s'%tag, os.path.join(opt.outDir,'syst_plots'))
plot.reset()
# for tag,sel,seltag in SELECTIONS:
# print 70*'-'
# print '%-10s: %s' % (tag, sel)
# fcor, fwro, funm = {}, {}, {}
# for mass in sorted(massfiles.keys()):
# # mass = 172.5
# hists = masshistos[(tag, mass)]
# n_tot, n_cor, n_wro, n_unm = (x.GetEntries() for x in hists)
# fcor[mass] = 100.*(n_cor/float(n_tot))
# fwro[mass] = 100.*(n_wro/float(n_tot))
# funm[mass] = 100.*(n_unm/float(n_tot))
# print (' %5.1f GeV: %7d entries \t'
# '(%4.1f%% corr, %4.1f%% wrong, %4.1f%% unmatched)' %
# (mass, n_tot, fcor[mass], fwro[mass], funm[mass]))
# oname = os.path.join(opt.outDir, 'fracvsmt_%s'%tag)
# plotFracVsTopMass(fcor, fwro, funm, tag, seltag, oname)
# print 112*'-'
# print 'Estimated systematics (from a crude chi2 fit)'
# print '%20s | %-15s | %-15s | %-15s | %-15s | %-15s' % (
# 'selection', 'bfrag', 'scale',
# 'toppt', 'matching', 'uecr')
# for tag,_,_ in SELECTIONS:
# sys.stdout.write("%20s | " % tag)
# for syst in ['bfrag', 'scale', 'toppt', 'matching', 'uecr']:
# err, chi2 = systematics[(tag,syst)]
# sys.stdout.write('%4.1f (%4.1f GeV)' % (chi2*1e5, err))
# # sys.stdout.write('%4.1f (%4.1f GeV)' % (chi2, err))
# sys.stdout.write(' | ')
# sys.stdout.write('\n')
# print 112*'-'
return 0
def makeControlPlot(systhistos, syst, tag, seltag, opt):
hists = tuple([systhistos[(tag, syst, c)] for c in COMBINATIONS.keys()])
h_tot, h_cor, h_wro, h_unm = hists
h_tot.Scale(1. / h_tot.Integral())
ctrlplot = RatioPlot('ctrlplot_%s' % tag)
ctrlplot.add(h_cor, 'Correct')
ctrlplot.add(h_wro, 'Wrong')
ctrlplot.add(h_unm, 'Unmatched')
ctrlplot.reference = h_tot
ctrlplot.tag = ''
ctrlplot.subtag = seltag
ctrlplot.ratiotitle = 'Ratio wrt Total'
ctrlplot.ratiorange = (0., 3.0)
ctrlplot.colors = [ROOT.kBlue - 3, ROOT.kRed - 4, ROOT.kOrange - 3]
ctrlplot.show("control_%s_%s" % (syst, tag), opt.outDir)
ctrlplot.reset()
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):
os.system('mkdir -p %s' % opt.outDir)
masstrees, massfiles = getMassTrees(args[0], verbose=True)
masspoints = sorted(list(set([mass for mass, _ in masstrees.keys()])))
hname_to_keys = {} # hname -> (tag, chan, mass, comb, [ntk1])
tasklist = {} ## (mass,chan) -> tasks
for (mass, chan) in masstrees.keys():
tasks = []
for tag, sel, _ in SELECTIONS:
#htag = ("%s_%5.1f_tt"%(tag,mass)).replace('.','')
#if not chan == 'tt':
htag = ("%s_%5.1f_%s" % (tag, mass, chan)).replace('.', '')
for comb, combsel in COMBINATIONS.iteritems():
hname = "SVLMass_%s_%s" % (comb, htag)
finalsel = "%s*(%s&&%s)" % (COMMONWEIGHT, sel, combsel)
tasks.append((hname, 'SVLMass', finalsel, NBINS, XMIN, XMAX,
MASSXAXISTITLE))
hname_to_keys[hname] = (tag, chan, mass, comb)
for ntk1, ntk2 in NTRKBINS:
tksel = "(SVNtrk>=%d && SVNtrk<%d)" % (ntk1, ntk2)
finalsel = "%s*(%s&&%s&&%s)" % (COMMONWEIGHT, sel, combsel,
tksel)
hname = "SVLMass_%s_%s_%d" % (comb, htag, ntk1)
tasks.append((hname, 'SVLMass', finalsel, NBINS, XMIN,
XMAX, MASSXAXISTITLE))
hname_to_keys[hname] = (tag, chan, mass, comb, ntk1)
tasklist[(mass, chan)] = tasks
if not opt.cache:
runTasks(massfiles, tasklist, opt, 'mass_histos')
## Retrieve the histograms from the individual files
# (tag, chan, mass, comb) -> histo
# (tag, chan, mass, comb, ntk) -> histo
masshistos = gatherHistosFromFiles(
tasklist, massfiles, os.path.join(opt.outDir, 'mass_histos'),
hname_to_keys)
cachefile = open(".svlmasshistos.pck", 'w')
pickle.dump(masshistos, cachefile, pickle.HIGHEST_PROTOCOL)
cachefile.close()
print 'Wrote .svlmasshistos.pck with all the mass histos'
else:
## Read mass scan histos:
cachefile = open(".svlmasshistos.pck", 'r')
masshistos = pickle.load(cachefile)
print '>>> Read mass scan histograms from cache (.svlmasshistos.pck)'
cachefile.close()
# ofi = ROOT.TFile(os.path.join(opt.outDir,'masshistos.root'),
# 'recreate')
# ofi.cd()
# for key in masshistos.keys():
# tag, chan, mass = key[0],key[1],key[2]
# if not ofi.cd(tag):
# outDir = ofi.mkdir(tag)
# outDir.cd()
# for comb in COMBINATIONS.keys():
# masshistos[(tag,chan,mass,comb)].Write()
# for ntk,_ in NTRKBINS:
# masshistos[(tag,chan,mass,comb,ntk)].Write()
# ofi.Write()
# ofi.Close()
ROOT.gStyle.SetOptTitle(0)
ROOT.gStyle.SetOptStat(0)
ROOT.gROOT.SetBatch(1)
cachefile = open(".xsecweights.pck", 'r')
xsecweights = pickle.load(cachefile)
cachefile.close()
print '>>> Read xsec weights from cache (.xsecweights.pck)'
## Scale all the histograms for the plotting:
for key, hist in masshistos.iteritems():
hist.Scale(LUMI * xsecweights[CHANMASSTOPROCNAME[(key[1], key[2])]])
errorGetters = {} # tag -> function(chi2) -> error
systematics = {} # (seltag, systname) -> error
mass_scan_dir = os.path.join(opt.outDir, 'mass_scans')
for tag, sel, seltag in SELECTIONS:
print "... processing %s" % tag
pairing = 'inclusive'
try:
pairing = tag.split('_', 1)[1]
except IndexError:
pass
for chan in ['tt', 't', 'tbar', 'tW', 'tbarW']:
# print " %s channel" % chan
## Skip some useless combinations:
chanTitle = chan
if chanTitle == 'tt': chanTitle = 'ttbar'
chanTitle = chanTitle.replace('tbar', '#bar{t}')
if chan in ['t', 'tbar'] and ('ee' in tag or 'mm' in tag
or 'em' in tag):
continue
if chan in ['tW', 'tbarW'] and tag in [
'e', 'm', 'eplus', 'eminus', 'mplus', 'mminus',
'e_optmrank', 'm_optmrank', 'eplus_optmrank',
'eminus_optmrank', 'mplus_optmrank', 'mminus_optmrank'
]:
continue
ratplot = RatioPlot('ratioplot')
ratplot.normalized = False
ratplot.ratiotitle = "Ratio wrt 172.5 GeV"
ratplot.ratiorange = (0.5, 1.5)
ratplot.rebin = 2
reference = masshistos[(tag, chan, 172.5, 'tot')]
ratplot.reference = reference
for mass in masspoints:
legentry = 'm_{t} = %5.1f GeV' % mass
try:
histo = masshistos[(tag, chan, mass, 'tot')]
ratplot.add(histo, legentry)
except KeyError:
pass
# print "Can't find ", (tag,chan,mass,'tot')
ratplot.tag = 'All combinations'
ratplot.subtag = '%s %s' % (seltag, chanTitle)
ratplot.show("massscan_%s_%s_tot" % (tag, chan), mass_scan_dir)
# chi2s = ratplot.getChiSquares(rangex=FITRANGE)
# chi2stofit = []
# for legentry in sorted(chi2s.keys()):
# chi2stofit.append((float(legentry[8:-4]), chi2s[legentry]))
# errorGetters[tag] = fitChi2(chi2stofit,
# tag=seltag,
# oname=os.path.join(mass_scan_dir,
# "chi2_simple_fit_%s.pdf"%tag),
# drawfit=False)
ratplot.reset()
ratplot.reference = masshistos[(tag, chan, 172.5, 'cor')]
for mass in masspoints:
legentry = 'm_{t} = %5.1f GeV' % mass
try:
ratplot.add(masshistos[(tag, chan, mass, 'cor')], legentry)
except KeyError:
pass
ratplot.tag = 'Correct combinations'
ratplot.subtag = '%s %s' % (seltag, chanTitle)
ratplot.show("massscan_%s_%s_cor" % (tag, chan), mass_scan_dir)
ratplot.reset()
ratplot.reference = masshistos[(tag, chan, 172.5, 'wro')]
for mass in masspoints:
legentry = 'm_{t} = %5.1f GeV' % mass
try:
ratplot.add(masshistos[(tag, chan, mass, 'wro')], legentry)
except KeyError:
pass
ratplot.tag = 'Wrong combinations'
ratplot.subtag = '%s %s' % (seltag, chanTitle)
ratplot.show("massscan_%s_%s_wro" % (tag, chan), mass_scan_dir)
ratplot.reset()
ratplot.reference = masshistos[(tag, chan, 172.5, 'unm')]
for mass in masspoints:
legentry = 'm_{t} = %5.1f GeV' % mass
try:
ratplot.add(masshistos[(tag, chan, mass, 'unm')], legentry)
except KeyError:
pass
ratplot.tag = 'Unmatched combinations'
ratplot.subtag = '%s %s' % (seltag, chanTitle)
ratplot.show("massscan_%s_%s_unm" % (tag, chan), mass_scan_dir)
ratplot.reset()
## ntkscan plot
if 'inclusive' in tag:
ntkmassplot = RatioPlot('ntkmassplot_%s' % tag)
ntkmassplot.rebin = 2
ntkmassplot.reference = [masshistos[(tag, 'tt', 172.5, 'tot')]]
ntkmassplot.add(masshistos[(tag, 'tt', 172.5, 'tot')],
'Sum',
includeInRatio=False)
for ntk1, ntk2 in NTRKBINS:
title = "%d #leq N_{trk.} < %d" % (ntk1, ntk2)
if ntk2 > 100:
title = "%d #leq N_{trk.}" % (ntk1)
ntkmassplot.add(masshistos[(tag, 'tt', 172.5, 'tot', ntk1)],
title)
ntkmassplot.colors = [
ROOT.kOrange + 8, ROOT.kGreen + 3, ROOT.kGreen + 1,
ROOT.kGreen, ROOT.kGreen - 10, ROOT.kYellow - 3,
ROOT.kYellow - 5
]
ntkmassplot.ratiorange = (0, 3.0)
ntkmassplot.ratiotitle = "Ratio wrt Sum"
ntkmassplot.tag = 'm_{t} = 172.5 GeV'
ntkmassplot.subtag = '%s, %s' % (pairing, seltag)
ntkmassplot.show("ntkscan_%s" % tag, opt.outDir)
ntkmassplot.reset()
#fractions
for tag, sel, seltag in SELECTIONS:
print 70 * '-'
print '%-10s: %s %s' % (tag, sel, seltag)
if not ('inclusive' in tag): continue
fcor, fwro, funm = {}, {}, {}
for mass, proc in sorted(massfiles.keys()):
# mass = 172.5
ncount = {}
for comb in ['tot', 'cor', 'wro', 'unm']:
hist = masshistos[(tag, proc, mass, comb)]
ncount[comb] = hist.Integral()
fcor[mass] = 100. * (ncount['cor'] / float(ncount['tot']))
fwro[mass] = 100. * (ncount['wro'] / float(ncount['tot']))
funm[mass] = 100. * (ncount['unm'] / float(ncount['tot']))
print(
' %-6s %5.1f GeV: %7d entries \t'
'(%4.1f%% corr, %4.1f%% wrong, %4.1f%% unmatched)' %
(proc, mass, ncount['tot'], fcor[mass], fwro[mass],
funm[mass]))
oname = os.path.join(opt.outDir, 'fracvsmt_%s' % tag)
plotFracVsTopMass(fcor, fwro, funm, tag, seltag, oname)
return 0
def makeMassPlot(histos, outname, tag='', subtag=''):
ratplot = RatioPlot('ratioplot')
ratplot.normalized = False
ratplot.ratiotitle = "Ratio wrt 172.5 GeV"
ratplot.extratext = ''
ratplot.tag = tag
ratplot.subtag = subtag
ratplot.rebin = 1
ratplot.legpos = (0.65, 0.15)
ratplot.ratiorange = (0.7, 1.3)
ratplot.reference = [histos[172.5]]
ratplot.titley = "Combinations / %3.1f GeV" %histos['data'].GetXaxis().GetBinWidth(1)
####################################################
# Construct dummy data for now
# dummy = histos[172.5].Clone('dummy_data')
# dummy.Reset('ICE')
# dummy.FillRandom(histos[173.5], ROOT.gRandom.Poisson(histos[172.5].Integral()))
# ratplot.add(dummy,'Pseudo Data (@173.5 GeV)')
####################################################
####################################################
# REAL DATA:
ratplot.add(histos['data'], 'Data')
####################################################
# histos[MASSESTOPLOT[0]].SetLineStyle(2)
# histos[MASSESTOPLOT[-1]].SetLineStyle(3)
for mass in MASSESTOPLOT:
legentry = 'm_{top} = %5.1f GeV' % mass
try:
histo = histos[mass]
ratplot.add(histo, legentry, includeInRatio=(mass != 172.5))
except KeyError: pass
# ratplot.colors = [ROOT.kBlack, ROOT.kAzure-4, ROOT.kAzure+1, ROOT.kAzure-6]
# ratplot.colors = [ROOT.kBlack, ROOT.kAzure-2, ROOT.kGray, ROOT.kOrange+9]
ratplot.colors = [ROOT.kBlack, ROOT.kSpring-8, ROOT.kGray, ROOT.kOrange+9]
# ratplot.colors = [ROOT.kBlack, ROOT.kOrange-4, ROOT.kOrange-3, ROOT.kOrange+9]
ratplot.drawoptions = ['PE', 'hist', 'hist', 'hist']
ratplot.markerstyles = [20,1,1,1]
ratplot.markersizes = [1.5,1,1,1]
ratplot.show("massscan_paper_%s"%outname, opt.outDir)
def main(args, opt):
os.system('mkdir -p %s'%opt.outDir)
masstrees, massfiles = getMassTrees(args[0], verbose=True)
masspoints = sorted(list(set([mass for mass,_ in masstrees.keys()])))
hname_to_keys = {} # hname -> (tag, chan, mass, comb, [ntk1])
tasklist = {} ## (mass,chan) -> tasks
for (mass,chan) in masstrees.keys():
tasks = []
for tag,sel,_ in SELECTIONS:
#htag = ("%s_%5.1f_tt"%(tag,mass)).replace('.','')
#if not chan == 'tt':
htag = ("%s_%5.1f_%s"%(tag,mass,chan)).replace('.','')
for comb,combsel in COMBINATIONS.iteritems():
hname = "SVLMass_%s_%s" % (comb, htag)
finalsel = "%s*(%s&&%s)"%(COMMONWEIGHT,sel,combsel)
tasks.append((hname, 'SVLMass', finalsel,
NBINS, XMIN, XMAX, MASSXAXISTITLE))
hname_to_keys[hname] = (tag, chan, mass, comb)
for ntk1,ntk2 in NTRKBINS:
tksel = "(SVNtrk>=%d && SVNtrk<%d)"%(ntk1,ntk2)
finalsel = "%s*(%s&&%s&&%s)"%(COMMONWEIGHT, sel,
combsel,tksel)
hname = "SVLMass_%s_%s_%d" % (comb, htag, ntk1)
tasks.append((hname, 'SVLMass', finalsel,
NBINS, XMIN, XMAX, MASSXAXISTITLE))
hname_to_keys[hname] = (tag, chan, mass, comb, ntk1)
tasklist[(mass,chan)] = tasks
if not opt.cache:
runTasks(massfiles, tasklist, opt, 'mass_histos')
## Retrieve the histograms from the individual files
# (tag, chan, mass, comb) -> histo
# (tag, chan, mass, comb, ntk) -> histo
masshistos = gatherHistosFromFiles(tasklist, massfiles,
os.path.join(opt.outDir,
'mass_histos'),
hname_to_keys)
cachefile = open(".svlmasshistos.pck", 'w')
pickle.dump(masshistos, cachefile, pickle.HIGHEST_PROTOCOL)
cachefile.close()
print 'Wrote .svlmasshistos.pck with all the mass histos'
else:
## Read mass scan histos:
cachefile = open(".svlmasshistos.pck", 'r')
masshistos = pickle.load(cachefile)
print '>>> Read mass scan histograms from cache (.svlmasshistos.pck)'
cachefile.close()
# ofi = ROOT.TFile(os.path.join(opt.outDir,'masshistos.root'),
# 'recreate')
# ofi.cd()
# for key in masshistos.keys():
# tag, chan, mass = key[0],key[1],key[2]
# if not ofi.cd(tag):
# outDir = ofi.mkdir(tag)
# outDir.cd()
# for comb in COMBINATIONS.keys():
# masshistos[(tag,chan,mass,comb)].Write()
# for ntk,_ in NTRKBINS:
# masshistos[(tag,chan,mass,comb,ntk)].Write()
# ofi.Write()
# ofi.Close()
ROOT.gStyle.SetOptTitle(0)
ROOT.gStyle.SetOptStat(0)
ROOT.gROOT.SetBatch(1)
cachefile = open(".xsecweights.pck", 'r')
xsecweights = pickle.load(cachefile)
cachefile.close()
print '>>> Read xsec weights from cache (.xsecweights.pck)'
## Scale all the histograms for the plotting:
for key, hist in masshistos.iteritems():
hist.Scale(LUMI*xsecweights[CHANMASSTOPROCNAME[(key[1], key[2])]])
errorGetters = {} # tag -> function(chi2) -> error
systematics = {} # (seltag, systname) -> error
mass_scan_dir = os.path.join(opt.outDir, 'mass_scans')
for tag,sel,seltag in SELECTIONS:
print "... processing %s"%tag
pairing = 'inclusive'
try: pairing = tag.split('_',1)[1]
except IndexError: pass
for chan in ['tt', 't', 'tbar', 'tW', 'tbarW']:
# print " %s channel" % chan
## Skip some useless combinations:
chanTitle=chan
if chanTitle=='tt' : chanTitle='ttbar'
chanTitle=chanTitle.replace('tbar','#bar{t}')
if chan in ['t','tbar'] and ('ee' in tag or 'mm' in tag or 'em' in tag):
continue
if chan in ['tW','tbarW'] and tag in ['e', 'm',
'eplus','eminus', 'mplus','mminus',
'e_optmrank', 'm_optmrank',
'eplus_optmrank','eminus_optmrank', 'mplus_optmrank','mminus_optmrank']:
continue
ratplot = RatioPlot('ratioplot')
ratplot.normalized = False
ratplot.ratiotitle = "Ratio wrt 172.5 GeV"
ratplot.ratiorange = (0.5, 1.5)
ratplot.rebin = 2
reference = masshistos[(tag,chan,172.5,'tot')]
ratplot.reference = reference
for mass in masspoints:
legentry = 'm_{t} = %5.1f GeV' % mass
try:
histo = masshistos[(tag,chan,mass,'tot')]
ratplot.add(histo, legentry)
except KeyError: pass
# print "Can't find ", (tag,chan,mass,'tot')
ratplot.tag = 'All combinations'
ratplot.subtag = '%s %s' % (seltag, chanTitle)
ratplot.show("massscan_%s_%s_tot"%(tag,chan), mass_scan_dir)
# chi2s = ratplot.getChiSquares(rangex=FITRANGE)
# chi2stofit = []
# for legentry in sorted(chi2s.keys()):
# chi2stofit.append((float(legentry[8:-4]), chi2s[legentry]))
# errorGetters[tag] = fitChi2(chi2stofit,
# tag=seltag,
# oname=os.path.join(mass_scan_dir,
# "chi2_simple_fit_%s.pdf"%tag),
# drawfit=False)
ratplot.reset()
ratplot.reference = masshistos[(tag,chan,172.5,'cor')]
for mass in masspoints:
legentry = 'm_{t} = %5.1f GeV' % mass
try: ratplot.add(masshistos[(tag,chan,mass,'cor')], legentry)
except KeyError: pass
ratplot.tag = 'Correct combinations'
ratplot.subtag = '%s %s' % (seltag, chanTitle)
ratplot.show("massscan_%s_%s_cor"%(tag,chan), mass_scan_dir)
ratplot.reset()
ratplot.reference = masshistos[(tag,chan,172.5,'wro')]
for mass in masspoints:
legentry = 'm_{t} = %5.1f GeV' % mass
try: ratplot.add(masshistos[(tag,chan,mass,'wro')], legentry)
except KeyError: pass
ratplot.tag = 'Wrong combinations'
ratplot.subtag = '%s %s' % (seltag, chanTitle)
ratplot.show("massscan_%s_%s_wro"%(tag,chan), mass_scan_dir)
ratplot.reset()
ratplot.reference = masshistos[(tag,chan,172.5,'unm')]
for mass in masspoints:
legentry = 'm_{t} = %5.1f GeV' % mass
try: ratplot.add(masshistos[(tag,chan,mass,'unm')], legentry)
except KeyError: pass
ratplot.tag = 'Unmatched combinations'
ratplot.subtag = '%s %s' % (seltag, chanTitle)
ratplot.show("massscan_%s_%s_unm"%(tag,chan), mass_scan_dir)
ratplot.reset()
## ntkscan plot
if 'inclusive' in tag:
ntkmassplot = RatioPlot('ntkmassplot_%s'%tag)
ntkmassplot.rebin = 2
ntkmassplot.reference = [masshistos[(tag, 'tt', 172.5, 'tot')]]
ntkmassplot.add(masshistos[(tag, 'tt', 172.5, 'tot')], 'Sum', includeInRatio=False)
for ntk1,ntk2 in NTRKBINS:
title = "%d #leq N_{trk.} < %d" %(ntk1, ntk2)
if ntk2 > 100:
title = "%d #leq N_{trk.}" %(ntk1)
ntkmassplot.add(masshistos[(tag, 'tt', 172.5, 'tot', ntk1)],
title)
ntkmassplot.colors = [ROOT.kOrange+8,
ROOT.kGreen+3, ROOT.kGreen+1,
ROOT.kGreen, ROOT.kGreen-10,
ROOT.kYellow-3, ROOT.kYellow-5]
ntkmassplot.ratiorange = (0,3.0)
ntkmassplot.ratiotitle = "Ratio wrt Sum"
ntkmassplot.tag = 'm_{t} = 172.5 GeV'
ntkmassplot.subtag = '%s, %s' %(pairing, seltag)
ntkmassplot.show("ntkscan_%s"%tag, opt.outDir)
ntkmassplot.reset()
#fractions
for tag,sel,seltag in SELECTIONS:
print 70*'-'
print '%-10s: %s %s' % (tag, sel,seltag)
if not('inclusive' in tag): continue
fcor, fwro, funm = {}, {}, {}
for mass,proc in sorted(massfiles.keys()):
# mass = 172.5
ncount={}
for comb in ['tot','cor','wro','unm']:
hist = masshistos[(tag,proc, mass,comb)]
ncount[comb]=hist.Integral()
fcor[mass] = 100.*(ncount['cor']/float(ncount['tot']))
fwro[mass] = 100.*(ncount['wro']/float(ncount['tot']))
funm[mass] = 100.*(ncount['unm']/float(ncount['tot']))
print (' %-6s %5.1f GeV: %7d entries \t'
'(%4.1f%% corr, %4.1f%% wrong, %4.1f%% unmatched)' %
(proc, mass, ncount['tot'],
fcor[mass], fwro[mass], funm[mass]))
oname = os.path.join(opt.outDir, 'fracvsmt_%s'%tag)
plotFracVsTopMass(fcor, fwro, funm, tag, seltag, oname)
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 makeSVMassPlots(histodict):
ratplot = RatioPlot('SVMass')
ratplot.normalized = True
ratplot.ratiorange = (0.5,1.5)
ratplot.legpos = (0.65, 0.60)
ratplot.ratiotitle = 'Data/MC'
ratplot.extratext = '' #Work in Progress'
ratplot.tag = "Inclusive"
# ratplot.tag = "e#mu channel"
ratplot.tagpos = (0.80,0.55)
ratplot.add(histodict[('nominal','SVMass',3)], 'N_{trk}=3 t#bar{t} MC', includeInRatio=False)
ratplot.add(histodict[('data', 'SVMass',3)], 'N_{trk}=3 Data', includeInRatio=True)
ratplot.add(histodict[('nominal','SVMass',4)], 'N_{trk}=4 t#bar{t} MC', includeInRatio=False)
ratplot.add(histodict[('data', 'SVMass',4)], 'N_{trk}=4 Data', includeInRatio=True)
ratplot.add(histodict[('nominal','SVMass',5)], 'N_{trk}=5 t#bar{t} MC', includeInRatio=False)
ratplot.add(histodict[('data', 'SVMass',5)], 'N_{trk}=5 Data', includeInRatio=True)
ratplot.reference = [histodict[('nominal','SVMass',3)]]
ratplot.reference.append(histodict[('nominal','SVMass',4)])
ratplot.reference.append(histodict[('nominal','SVMass',5)])
ratplot.colors = [ROOT.kMagenta+1, ROOT.kMagenta+1,
ROOT.kViolet+2, ROOT.kViolet+2,
ROOT.kAzure+7, ROOT.kAzure+7]
ratplot.drawoptions = ['hist', 'PE', 'hist', 'PE', 'hist', 'PE']
ratplot.markerstyles = [24, 24, 25, 25, 26, 26]
ratplot.markersizes = len(ratplot.histos)*[1.4]
ratplot.show('svmassplot',opt.outDir)
ratplot.saveRatios('svmass_weights',opt.outDir,['SVMass_weight_3','SVMass_weight_4','SVMass_weight_5'])
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
def main(args, opt):
os.system('mkdir -p %s' % opt.outDir)
systfiles = {} # procname -> filename
try:
for fname in os.listdir(os.path.join(args[0], 'syst')):
if not os.path.splitext(fname)[1] == '.root': continue
for syst, _, systfile, _ in SYSTSFROMFILES:
if fname in systfile:
systfiles[syst] = [os.path.join(args[0], 'syst', fname)]
# Get the split nominal files
systfiles['nominal'] = []
for fname in os.listdir(os.path.join(args[0], 'Chunks')):
if not os.path.splitext(fname)[1] == '.root': continue
isdata, procname, splitno = resolveFilename(fname)
if not procname == 'TTJets_MSDecays_172v5': continue
if not splitno: continue # file is split
systfiles['nominal'].append(os.path.join(args[0], 'Chunks', fname))
if len(systfiles['nominal']) < 20:
print "ERROR >>> Missing files for split nominal sample?"
return -1
except IndexError:
print "Please provide a valid input directory"
exit(-1)
hname_to_keys = {} # hname -> (tag, syst, comb)
tasklist = {} # treefile -> tasklist
for fsyst in systfiles.keys():
if not fsyst in tasklist: tasklist[fsyst] = []
for tag, sel, _ in SELECTIONS:
if fsyst == 'nominal':
for syst, _, weight, combs in SYSTSFROMWEIGHTS:
tasks = makeSystTask(tag,
sel,
syst,
hname_to_keys,
weight=weight,
combs=combs)
tasklist[fsyst] += tasks
tasks = []
for comb, combsel in COMBINATIONS.iteritems():
for var, nbins, xmin, xmax, titlex in CONTROLVARS:
hname = "%s_%s_%s" % (var, comb, tag)
finalsel = "%s*(%s&&%s)" % (COMMONWEIGHT, sel, combsel)
tasks.append(
(hname, var, finalsel, nbins, xmin, xmax, titlex))
hname_to_keys[hname] = (tag, var, comb)
tasklist[fsyst] += tasks
tasks = []
for name, nus in [('nu', 1), ('nonu', 0), ('nuunm', -1)]:
hname = "SVLMass_%s_%s_%s" % ('tot', tag, name)
finalsel = "%s*(%s&&BHadNeutrino==%d)" % (COMMONWEIGHT,
sel, nus)
tasks.append((hname, 'SVLMass', finalsel, NBINS, XMIN,
XMAX, MASSXAXISTITLE))
hname_to_keys[hname] = (tag, name, 'tot')
tasklist[fsyst] += tasks
else:
tasks = makeSystTask(tag, sel, fsyst, hname_to_keys)
tasklist[fsyst] += tasks
if not opt.cache:
# print ' Will process the following tasks:'
# for filename,tasks in sorted(tasklist.iteritems()):
# print filename
# for task in tasks:
# print task
# raw_input("Press any key to continue...")
runTasks(systfiles, tasklist, opt, 'syst_histos')
systhistos = {} # (tag, syst, comb) -> histo
systhistos = gatherHistosFromFiles(
tasklist, systfiles, os.path.join(opt.outDir, 'syst_histos'),
hname_to_keys)
cachefile = open(".svlsysthistos.pck", 'w')
pickle.dump(systhistos, cachefile, pickle.HIGHEST_PROTOCOL)
cachefile.close()
# print "Wrote syst histos to cache file"
# raw_input("press key")
cachefile = open(".svlsysthistos.pck", 'r')
systhistos = pickle.load(cachefile)
print '>>> Read syst histos from cache (.svlsysthistos.pck)'
cachefile.close()
ROOT.gStyle.SetOptTitle(0)
ROOT.gStyle.SetOptStat(0)
ROOT.gROOT.SetBatch(1)
for var, _, _, _, _ in CONTROLVARS:
for sel, tag in [
#('inclusive', 'Fully Inclusive'),
#('inclusive_mrank1', 'Mass ranked, leading p_{T}'),
#('inclusive_mrank1dr', 'Mass ranked, #DeltaR<2, leading p_{T}'),
#('inclusive_drrank1dr', '#DeltaR ranked, #DeltaR<2, leading p_{T}'),
('inclusive_optmrank', 'Optimized mass rank')
]:
try:
makeControlPlot(systhistos, var, sel, tag, opt)
except KeyError:
print 'control plots for %s selection not found' % sel
for tag, _, _ in SELECTIONS:
if not 'inclusive' in tag: continue
print "... processing %s" % tag
# Make plot of mass with and without neutrino:
# for comb in COMBINATIONS.keys():
# plot = RatioPlot('neutrino_%s'%tag)
# plot.rebin = 2
# plot.add(systhistos[(tag,'nonu', 'tot')], 'Without neutrino')
# plot.add(systhistos[(tag,'nu', 'tot')], 'With neutrino')
# plot.add(systhistos[(tag,'nuunm','tot')], 'Unmatched')
# plot.reference = systhistos[(tag,'nominal','tot')]
# plot.tag = "Mass shape with and without neutrinos"
# plot.subtag = SELNAMES[tag] + COMBNAMES['tot']
# plot.ratiotitle = 'Ratio wrt Total'
# plot.ratiorange = (0.7, 1.3)
# plot.colors = [ROOT.kBlue-3, ROOT.kRed-4, ROOT.kOrange-3]
# plot.show("neutrino_%s_%s"%(tag,'tot'),
# os.path.join(opt.outDir, 'syst_plots'))
# plot.reset()
for name, title, systs, colors, comb in SYSTPLOTS:
print name, title, systs, colors, comb
plot = RatioPlot('%s_%s' % (name, comb))
plot.rebin = 2
for syst in systs:
try:
plot.add(systhistos[(tag, syst, comb)], SYSTNAMES[syst])
except:
print 'failed to add', (tag, syst, comb), syst
plot.tag = title
subtag = SELNAMES[tag] + COMBNAMES[comb]
plot.subtag = subtag
plot.ratiotitle = 'Ratio wrt %s' % SYSTNAMES[systs[0]]
plot.ratiorange = (0.85, 1.15)
plot.colors = colors
filename = "%s_%s" % (name, tag)
if comb != 'tot': filename += '_%s' % comb
plot.show(filename, os.path.join(opt.outDir, 'syst_plots'))
plot.reset()
# Make top pt plot with both correct and wrong
plot = RatioPlot('toppt_paper_cor_wro')
plot.canvassize = (600, 600)
plot.tag = 'Top quark p_{T} mismodeling'
plot.rebin = 2
plot.subtag = 'Inclusive channels'
plot.tagpos = (0.92, 0.85)
plot.subtagpos = (0.92, 0.78)
plot.titlex = 'm_{svl} [GeV]'
plot.ratiotitle = '1 / Nominal'
# plot.ratiorange = (0.85, 1.15)
plot.ratiorange = (0.92, 1.08)
plot.legpos = (0.55, 0.38)
plot.ratioydivisions = 405
plot.colors = [
ROOT.kGreen + 2, ROOT.kGreen - 6, ROOT.kRed + 2, ROOT.kRed - 6
]
plot.add(systhistos[(tag, 'nominal', 'cor')],
'Nominal (correct)',
includeInRatio=False)
plot.add(systhistos[(tag, 'toppt', 'cor')],
'p_{T} weighted (correct)',
includeInRatio=True)
plot.add(systhistos[(tag, 'nominal', 'wro')],
'Nominal (wrong)',
includeInRatio=False)
plot.add(systhistos[(tag, 'toppt', 'wro')],
'p_{T} weighted (wrong)',
includeInRatio=True)
plot.reference = [
systhistos[(tag, 'nominal', 'cor')],
systhistos[(tag, 'nominal', 'wro')]
]
plot.show('toppt_cor_wro_forpaper_%s' % tag,
os.path.join(opt.outDir, 'syst_plots'))
plot.reset()
# Make b fragmentation plot for paper
plot = RatioPlot('bfrag_paper')
plot.canvassize = (600, 600)
plot.tag = 'b fragmentation'
plot.titlex = 'm_{svl} [GeV]'
plot.rebin = 2
plot.subtag = 'Inclusive channels'
plot.tagpos = (0.92, 0.85)
plot.subtagpos = (0.92, 0.78)
plot.ratiotitle = '1 / Z2* #it{r}_{b} LEP'
# plot.ratiorange = (0.85, 1.15)
plot.ratiorange = (0.92, 1.08)
plot.legpos = (0.65, 0.20)
plot.ratioydivisions = 405
plot.colors = [
ROOT.kMagenta, ROOT.kMagenta + 2, ROOT.kMagenta - 9,
ROOT.kAzure + 7
]
plot.add(systhistos[(tag, 'nominal', 'tot')],
'Z2* #it{r}_{b} LEP',
includeInRatio=False)
plot.add(systhistos[(tag, 'bfragdn', 'tot')],
'Z2* #it{r}_{b} LEP soft')
plot.add(systhistos[(tag, 'bfragup', 'tot')],
'Z2* #it{r}_{b} LEP hard')
plot.add(systhistos[(tag, 'bfragz2s', 'tot')], 'Z2* nominal')
plot.reference = [systhistos[(tag, 'nominal', 'tot')]]
plot.show('bfrag_paper_%s' % tag, os.path.join(opt.outDir,
'syst_plots'))
plot.reset()
# for tag,sel,seltag in SELECTIONS:
# print 70*'-'
# print '%-10s: %s' % (tag, sel)
# fcor, fwro, funm = {}, {}, {}
# for mass in sorted(massfiles.keys()):
# # mass = 172.5
# hists = masshistos[(tag, mass)]
# n_tot, n_cor, n_wro, n_unm = (x.GetEntries() for x in hists)
# fcor[mass] = 100.*(n_cor/float(n_tot))
# fwro[mass] = 100.*(n_wro/float(n_tot))
# funm[mass] = 100.*(n_unm/float(n_tot))
# print (' %5.1f GeV: %7d entries \t'
# '(%4.1f%% corr, %4.1f%% wrong, %4.1f%% unmatched)' %
# (mass, n_tot, fcor[mass], fwro[mass], funm[mass]))
# oname = os.path.join(opt.outDir, 'fracvsmt_%s'%tag)
# plotFracVsTopMass(fcor, fwro, funm, tag, seltag, oname)
# print 112*'-'
# print 'Estimated systematics (from a crude chi2 fit)'
# print '%20s | %-15s | %-15s | %-15s | %-15s | %-15s' % (
# 'selection', 'bfrag', 'scale',
# 'toppt', 'matching', 'uecr')
# for tag,_,_ in SELECTIONS:
# sys.stdout.write("%20s | " % tag)
# for syst in ['bfrag', 'scale', 'toppt', 'matching', 'uecr']:
# err, chi2 = systematics[(tag,syst)]
# sys.stdout.write('%4.1f (%4.1f GeV)' % (chi2*1e5, err))
# # sys.stdout.write('%4.1f (%4.1f GeV)' % (chi2, err))
# sys.stdout.write(' | ')
# sys.stdout.write('\n')
# print 112*'-'
return 0
def main(args, opt):
os.system('mkdir -p %s'%opt.outDir)
ROOT.gStyle.SetOptTitle(0)
ROOT.gStyle.SetOptStat(0)
ROOT.gROOT.SetBatch(1)
# Construct histograms
peInputFile = ROOT.TFile.Open(args[0], 'READ')
masshistos = {}
for mass in ALLMASSES+['data']:
masshist = None
masstag = mass
if type(mass) == float:
masstag = 'nominal_'+str(mass).replace('.','v')
for trk,_ in NTRKBINS:
ihist = peInputFile.Get('%s/SVLMass_inclusive_optmrank_%s_%d'%(masstag,masstag,trk))
try:
masshist.Add(ihist)
except AttributeError:
try:
masshist = ihist.Clone(masstag)
except ReferenceError:
print "Histogram not found: %s/SVLMass_inclusive_optmrank_%s_%d"%(masstag,masstag,trk)
continue
masshistos[mass] = masshist
ratplot = RatioPlot('ratioplot')
ratplot.normalized = False
ratplot.ratiotitle = "Ratio wrt 172.5 GeV"
ratplot.tag = 'All channels combined'
ratplot.extratext = 'Preliminary'
ratplot.rebin = 1
ratplot.legpos = (0.55, 0.15)
ratplot.ratiorange = (0.5, 1.5)
ratplot.reference = [masshistos[172.5]]
# Construct dummy data for now
dummy = masshistos[172.5].Clone('dummy_data')
dummy.Reset('ICE')
dummy.FillRandom(masshistos[173.5], ROOT.gRandom.Poisson(masshistos[172.5].Integral()))
ratplot.add(dummy,'Pseudo Data (@173.5 GeV)')
####################################################
# REAL DATA:
# ratplot.add(masshistos['data'], 'Data')
####################################################
for mass in MASSESTOPLOT:
legentry = 'MC (m_{top} = %5.1f GeV)' % mass
try:
histo = masshistos[mass]
ratplot.add(histo, legentry, includeInRatio=(mass != 172.5))
except KeyError: pass
ratplot.colors = [ROOT.kBlack, ROOT.kAzure-4, ROOT.kAzure+1, ROOT.kAzure-6]
# ratplot.colors = [ROOT.kBlack, ROOT.kOrange-4, ROOT.kOrange-3, ROOT.kOrange+9]
ratplot.drawoptions = ['PE', 'hist', 'hist', 'hist']
ratplot.markerstyles = [20,1,1,1]
ratplot.markersizes = [1.5,1,1,1]
ratplot.show("massscan_paper_optmrank", opt.outDir)
return 0
def makeMassPlot(histos, outname, tag='', subtag=''):
ratplot = RatioPlot('ratioplot')
ratplot.normalized = False
ratplot.ratiotitle = "Ratio wrt 172.5 GeV"
ratplot.extratext = ''
ratplot.tag = tag
ratplot.subtag = subtag
ratplot.legpos = (0.65, 0.62)
ratplot.ratiorange = (0.7, 1.3)
ratplot.titlex = 'm_{svl} [GeV]'
ratplot.xlabelsize = 28
ratplot.xtitlesize = 34
ratplot.ylabelsize = 28
ratplot.ytitlesize = 34
ratplot.ytitleoffset = 1.8
ratplot.xtitleoffset = 3
ratplot.legendtextsize = 28
ratplot.legendrowsize = 0.08
ratplot.reference = [histos[172.5]]
ratplot.titley = "Combinations / %3.1f GeV" %histos['data'].GetXaxis().GetBinWidth(1)
####################################################
# Construct dummy data for now
# dummy = histos[172.5].Clone('dummy_data')
# dummy.Reset('ICE')
# dummy.FillRandom(histos[173.5], ROOT.gRandom.Poisson(histos[172.5].Integral()))
# ratplot.add(dummy,'Pseudo Data (@173.5 GeV)')
####################################################
####################################################
# REAL DATA:
ratplot.add(histos['data'], 'Data')
####################################################
# histos[MASSESTOPLOT[0]].SetLineStyle(2)
# histos[MASSESTOPLOT[-1]].SetLineStyle(3)
for mass in MASSESTOPLOT:
legentry = 'm_{t} = %5.1f GeV' % mass
try:
histo = histos[mass]
ratplot.add(histo, legentry, includeInRatio=(mass != 172.5))
except KeyError: pass
# ratplot.colors = [ROOT.kBlack, ROOT.kAzure-4, ROOT.kAzure+1, ROOT.kAzure-6]
# ratplot.colors = [ROOT.kBlack, ROOT.kAzure-2, ROOT.kGray, ROOT.kOrange+9]
ratplot.colors = [ROOT.kBlack, ROOT.kSpring-8, ROOT.kGray+1, ROOT.kOrange+9]
# ratplot.colors = [ROOT.kBlack, ROOT.kOrange-4, ROOT.kOrange-3, ROOT.kOrange+9]
ratplot.drawoptions = ['PE', 'hist', 'hist', 'hist']
ratplot.markerstyles = [20,1,1,1]
ratplot.markersizes = [1.5,1,1,1]
ratplot.ratioydivisions = 405
ratplot.show("massscan_paper_%s"%outname, opt.outDir)
def makeSVMassPlots(histodict):
ratplot = RatioPlot('SVMass')
ratplot.normalized = True
ratplot.ratiorange = (0.5, 1.5)
ratplot.legpos = (0.65, 0.60)
ratplot.ratiotitle = 'Data/MC'
ratplot.extratext = '' #Work in Progress'
ratplot.tag = "Inclusive"
# ratplot.tag = "e#mu channel"
ratplot.tagpos = (0.80, 0.55)
ratplot.add(histodict[('nominal', 'SVMass', 3)],
'N_{trk}=3 t#bar{t} MC',
includeInRatio=False)
ratplot.add(histodict[('data', 'SVMass', 3)],
'N_{trk}=3 Data',
includeInRatio=True)
ratplot.add(histodict[('nominal', 'SVMass', 4)],
'N_{trk}=4 t#bar{t} MC',
includeInRatio=False)
ratplot.add(histodict[('data', 'SVMass', 4)],
'N_{trk}=4 Data',
includeInRatio=True)
ratplot.add(histodict[('nominal', 'SVMass', 5)],
'N_{trk}=5 t#bar{t} MC',
includeInRatio=False)
ratplot.add(histodict[('data', 'SVMass', 5)],
'N_{trk}=5 Data',
includeInRatio=True)
ratplot.reference = [histodict[('nominal', 'SVMass', 3)]]
ratplot.reference.append(histodict[('nominal', 'SVMass', 4)])
ratplot.reference.append(histodict[('nominal', 'SVMass', 5)])
ratplot.colors = [
ROOT.kMagenta + 1, ROOT.kMagenta + 1, ROOT.kViolet + 2,
ROOT.kViolet + 2, ROOT.kAzure + 7, ROOT.kAzure + 7
]
ratplot.drawoptions = ['hist', 'PE', 'hist', 'PE', 'hist', 'PE']
ratplot.markerstyles = [24, 24, 25, 25, 26, 26]
ratplot.markersizes = len(ratplot.histos) * [1.4]
ratplot.show('svmassplot', opt.outDir)
ratplot.saveRatios(
'svmass_weights', opt.outDir,
['SVMass_weight_3', 'SVMass_weight_4', 'SVMass_weight_5'])
