def compareMCProfiles(samples, channel, era, tag=""):
"""Compare MC profiles."""
print ">>> compareMCProfiles()"
hname = 'pileup'
htitle = 'MC average'
outdir = ensuredir("plots")
avehist = None
hists = []
if tag and tag[0] != '_':
tag = '_' + tag
if 'pmx' in tag:
htitle += " %s pre-mixing" % ("old" if "old" in tag else "new")
# GET histograms
for sample, fname in samples:
print ">>> %s" % (fname)
file, hist = gethist(fname, hname, retfile=True)
hist.SetName(sample)
hist.SetTitle(sample)
hist.SetDirectory(0)
if avehist == None:
avehist = hist.Clone('average%s' % tag)
avehist.SetTitle(htitle)
avehist.SetDirectory(0)
else:
avehist.Add(hist)
hist.Scale(1. / hist.Integral())
hists.append(hist)
file.Close()
# PLOT
hists = [avehist] + hists
colors = [kBlack] + linecolors
avehist.Scale(1. / avehist.Integral())
pname = "%s/pileup_MC_%s%s" % (outdir, era, tag)
xtitle = "Number of true interactions"
plot = Plot(hists, ratio=True)
plot.draw(xtitle=xtitle,
ytitle="A.U.",
rtitle="MC / Ave.",
textsize=0.032,
rmin=0.45,
rmax=1.55,
denom=2,
colors=colors)
plot.drawlegend('TTR', tsize=0.04, latex=False)
plot.saveas(pname + ".png")
plot.saveas(pname + ".pdf")
plot.close(keep=True)
for hist in hists: # clean memory
if hist == avehist:
continue
if hist.GetDirectory():
gDirectory.Delete(hist.GetName())
else:
hist.Delete()
return avehist
def compareDataMCProfiles(datahists,mchist,era="",minbiases=0.0,tag="",rmin=0.6,rmax=1.4,delete=False):
"""Compare data/MC profiles."""
print ">>> compareDataMCProfiles()"
mctitle = "MC average"
outdir = ensuredir("plots")
if islist(datahists): # multiple datahists
if all(islist(x) for x in datahists): # datahists = [(minbias1,datahist1),...]
minbiases = [m for m,h in datahists]
datahists = [h for m,h in datahists]
else: # is single datahist histogram
minbiases = [minbiases]
datahists = [datahists]
hists = datahists+[mchist]
styles = [kSolid]*len(datahists)+[kDashed]
colors = [kBlack]+linecolors if len(datahists)==1 else linecolors[:len(datahists)]+[kBlack]
if 'pmx' in tag:
width = 0.36
position = 'TCR'
else:
width = 0.39
position = 'TR'
if era and isinstance(era,str) and any(s in era for s in ["Run","VFP"]):
width = max(width,0.26+(len(era)-5)*0.018)
position = 'TCR'
if tag and tag[0]!='_':
tag = '_'+tag
if 'pmx' in tag:
mctitle += " (%s pre-mixing)"%("old" if "old" in tag else "new")
if len(minbiases)==1 and minbiases[0]>0:
tag = "_"+str(minbiases[0]).replace('.','p')
for datahist, minbias in zip(datahists,minbiases):
title = "Data"
if era:
title += " %s"%(era)
if minbias>0:
title += ", %.1f pb"%(minbias)
if 'VFP' in era:
title = title.replace("_"," ").replace("VFP","-VFP")
datahist.SetTitle(title)
datahist.Scale(1./datahist.Integral())
mchist.SetTitle(mctitle)
mchist.Scale(1./mchist.Integral())
xtitle = "Number of interactions"
pname = "%s/pileup_Data-MC_%s%s"%(outdir,era,tag)
plot = Plot(hists,ratio=True)
plot.draw(xtitle=xtitle,ytitle="A.U.",rtitle="Data / MC",
textsize=0.045,rmin=rmin,rmax=rmax,denom=-1,colors=colors,styles=styles)
plot.drawlegend(position,width=width)
plot.saveas(pname+".png")
plot.saveas(pname+".pdf")
plot.close(keep=True)
if delete:
deletehist(datahists) # clean memory
def compareDataMCProfiles(datahist,
mchist,
era,
minbias,
tag="",
rmin=0.75,
rmax=1.25):
"""Compare data/MC profiles."""
print ">>> compareDataMCProfiles()"
mctitle = "MC average"
outdir = ensuredir("plots")
hists = [datahist, mchist]
colors = [kBlack] + linecolors
width = 0.36 if 'pmx' in tag else 0.26
if isinstance(era, str) and "Run" in era:
width = max(width, 0.26 + (len(era) - 5) * 0.02)
if tag and tag[0] != '_':
tag = '_' + tag
if 'pmx' in tag:
mctitle += " (%s pre-mixing)" % ("old" if "old" in tag else "new")
datahist.SetTitle("Data %s, %.1f pb" % (era, minbias))
mchist.SetTitle(mctitle)
datahist.Scale(1. / datahist.Integral())
mchist.Scale(1. / mchist.Integral())
xtitle = "Number of interactions"
pname = "%s/pileup_Data-MC_%s_%s%s" % (outdir, era, str(minbias).replace(
'.', 'p'), tag)
plot = Plot(hists, ratio=True)
plot.draw(xtitle=xtitle,
ytitle="A.U.",
rtitle="Data / MC",
textsize=0.045,
rmin=rmin,
rmax=rmax,
denom=2,
colors=colors)
plot.drawlegend('TCR', width=width)
plot.saveas(pname + ".png")
plot.saveas(pname + ".pdf")
plot.close(keep=True)
deletehist(datahist) # clean memory
def compare_eras(eras, samplesets, channel, tag="", **kwargs):
"""Test plotting of SampleSet class for data/MC comparison."""
LOG.header("compare_eras", pre=">>>")
if 'mu' in channel:
snames = ['TT', 'DY', 'WJ', 'SingleMuon']
else:
snames = ['TT', 'DY', 'WJ', 'Tau']
snames = kwargs.get('samples', snames) # samples to compare
outdir = kwargs.get('outdir', "plots")
parallel = kwargs.get('parallel', True) #and False
norms = kwargs.get('norm', [True])
entries = kwargs.get('entries', [str(e) for e in eras]) # for legend
exts = kwargs.get('exts', ['png']) # figure file extensions
ensuredir(outdir)
norms = ensurelist(norms)
# SELECTIONS
if 'mutau' in channel:
idiso1 = "iso_1<0.15"
idiso2 = "idDecayModeNewDMs_2 && idDeepTau2017v2p1VSjet_2>=16 && idDeepTau2017v2p1VSe_2>=2 && idDeepTau2017v2p1VSmu_2>=8"
baseline = "q_1*q_2<0 && %s && %s && !lepton_vetoes_notau && metfilter" % (
idiso1, idiso2)
else:
cuts_iso = "idDeepTau2017v2VSjet_1>=16 && idDeepTau2017v2VSjet_2>=16"
antilep = "idDeepTau2017v2VSe_1>=2 && idDeepTau2017v2VSmu_1>=1 && idDeepTau2017v2VSe_2>=2 && idDeepTau2017v2VSmu_2>=1"
baseline = "q_1*q_2<0 && idDecayModeNewDMs_1 && idDecayModeNewDMs_2 && %s && %s && !lepton_vetos_notau && metfilter" % (
antilep, cuts_iso)
selections = [
Sel('baseline', baseline),
Sel('baseline, muon pt > 24 GeV',
baseline + " && pt_1>24",
fname="baseline-ptgt24"),
]
# VARIABLES
variables = [
Var('m_vis', 23, 0, 460),
Var('pt_1', "Leading tau_h pt", 18, 0, 270),
Var('pt_2', "Subleading tau_h pt", 18, 0, 270),
#Var('jpt_1', 18, 0, 270),
#Var('jpt_2', 18, 0, 270),
#Var('met', 20, 0, 300),
Var('rawDeepTau2017v2p1VSe_2',
"rawDeepTau2017v2p1VSe",
30,
0.70,
1,
fname="$VAR_zoom",
logy=True,
pos='L;y=0.85'),
Var('rawDeepTau2017v2p1VSmu_2',
"rawDeepTau2017v2p1VSmu",
20,
0.80,
1,
fname="$VAR_zoom",
logy=True,
logyrange=4,
pos='L;y=0.85'),
Var('rawDeepTau2017v2p1VSjet_2',
"rawDeepTau2017v2p1VSjet",
100,
0.0,
1,
pos='L;y=0.85',
logy=True,
ymargin=2.5),
Var('rawDeepTau2017v2p1VSjet_2',
"rawDeepTau2017v2p1VSjet",
20,
0.80,
1,
fname="$VAR_zoom",
pos='L;y=0.85'),
]
# PLOT
for sname in snames:
#LOG.header(sname)
print ">>> %s" % (sname)
samples = [samplesets[e].get(sname, unique=True) for e in eras]
header = samples[0].title
for selection in selections:
print ">>> %s: %r" % (selection, selection.selection)
hdict = {}
text = "%s: %s" % (channel.replace("tau",
"tau_{h}"), selection.title)
fname = "%s/compare_eras_$VAR_%s_%s%s$TAG" % (
outdir, sname, selection.filename, tag)
for sample in samples:
vars = [v for v in variables if v.data or not sample.isdata]
hists = sample.gethist(vars, selection, parallel=parallel)
for variable, hist in zip(variables, hists):
hdict.setdefault(variable, []).append(hist)
for variable, hists in hdict.iteritems():
for norm in norms:
ntag = '_norm' if norm else "" #_lumi"
plot = Plot(variable, hists, norm=norm)
plot.draw(ratio=True, lstyle=1)
plot.drawlegend(header=header, entries=entries)
plot.drawtext(text)
plot.saveas(fname, ext=['png'], tag=ntag) #,'pdf'
plot.close(keep=True)
deletehist(hists)
print ">>> "
def compare_cuts(sampleset, channel, tag="", outdir="plots"):
"""Compare different selections."""
LOG.header("compare_eras")
# SELECTIONS & VARIABLES
if 'tautau' in channel:
cuts_iso = "idDeepTau2017v2VSjet_1>=16 && idDeepTau2017v2VSjet_2>=16"
antilep = "idDeepTau2017v2VSe_1>=2 && idDeepTau2017v2VSmu_1>=1 && idDeepTau2017v2VSe_2>=2 && idDeepTau2017v2VSmu_2>=1"
baseline = "q_1*q_2<0 && idDecayModeNewDMs_1 && idDecayModeNewDMs_2 && %s && %s && !lepton_vetos_noTau && metfilter" % (
antilep, cuts_iso)
else:
cuts_iso = "idDeepTau2017v2VSjet_2>=16"
antilep = "idDeepTau2017v2VSe_2>=2 && idDeepTau2017v2VSmu_2>=8"
baseline = "q_1*q_2<0 && idMedium_1 && pfRelIso04_all_1<0.15 && idDecayModeNewDMs_2 && %s && %s && !lepton_vetos && metfilter" % (
antilep, cuts_iso)
selections = [
(
'baseline-medium-tight-muon',
'Baseline, Medium vs. Tight muon ID',
Sel('Medium', baseline),
Sel('Tight', baseline.replace('idMedium_1', 'idTight_1')),
),
(
'0b-medium-tight-muon',
'0b, Medium vs. Tight muon ID',
Sel(
'Medium', baseline +
" && pt_1>50 && pt_2>50 && njets50>=1 && m_vis>100 && nbtag50_loose==0"
),
Sel(
'Tight',
baseline.replace(
'idMedium_1', 'idTight_1' +
" && pt_1>50 && pt_2>50 && njets50>=1 && m_vis>100 && nbtag50_loose==0"
)),
),
(
'geq1b-medium-tight-muon',
'geq1b, Medium vs. Tight muon ID',
Sel(
'Medium', baseline +
" && pt_1>50 && pt_2>50 && njets50>=1 && m_vis>100 && nbtag50_loose>=1"
),
Sel(
'Tight',
baseline.replace(
'idMedium_1', 'idTight_1' +
" && pt_1>50 && pt_2>50 && njets50>=1 && m_vis>100 && nbtag50_loose>=1"
)),
),
]
variables = [
Var('m_vis', 23, 0, 460),
Var('pt_1', "Leading tau_h pt", 18, 0, 270),
Var('pt_2', "Subleading tau_h pt", 18, 0, 270),
Var('jpt_1', 18, 0, 270),
Var('jpt_2', 18, 0, 270),
Var('met', 20, 0, 300),
Var('pt_1+pt_2+jpt_1', 30, 0, 600, fname='ST', cut="jpt_1>50"),
]
# SAMPLESETS
sampleset = makesamples(channel, year)
snames = [
'SingleMuon',
'TT',
] #'DY','WJ']
# PLOT
outdir = ensuredir(outdir)
parallel = True and False
for sname in snames:
if 'Tau' and 'tautau' not in channel: continue
if 'SingleMuon' and 'mutau' not in channel: continue
LOG.header(sname)
sample = sampleset.get(sname, unique=True)
header = samples[0].title
for setname, settitle, selection1, selection2 in selections:
hdict = {}
text = "%s: %s" % (channel.replace("tau", "tau_{h}"), settitle)
fname = "%s/compareSels_$VAR_%s_%s%s$TAG" % (outdir, sname,
setname, tag)
for sample in samples:
vars = [v for v in variables if v.data or not sample.isdata]
hists = sample.gethist(vars, selection, parallel=parallel)
for variable, hist in zip(variables, hists):
hdict.setdefault(variable, []).append(hist)
#entries = [str(y) for y in eras] # for legend
for variable, hists in hdict.iteritems():
for norm in [False, True]:
#print norm, hists
ntag = '_norm' if norm else "_lumi"
plot = Plot(variable, hists, norm=norm)
plot.draw(ratio=True, lstyle=1)
plot.drawlegend(header=header, entries=entries)
plot.drawtext(text)
plot.saveas(fname, ext=['png'], tag=ntag) #,'pdf'
plot.close(keep=True)
deletehist(hists)
def comparevars(file, variable, processes, systag, **kwargs):
"""Compare up/down variations of input histograms from ROOT file.
file: TFile or TDirectory object
variable: Variables object
processes: list of strings (name of processes)
systag: string of systematic (file must contain up/down variation)
e.g.
comparevars(file,variable,['ZTT','TTT'],'TES')
"""
text = kwargs.get('text', None)
pname = kwargs.get('pname', "stack.png")
tag = kwargs.get('tag', "")
groups = kwargs.get('group', []) # e.g. [(['^TT','ST'],'Top')]
dname = kwargs.get('dname', None) # directory ('bin') name
pname = kwargs.get('save', "plot_$PROC$SYST$TAG.png") # save as image file
wname = kwargs.get('write', "plot_$PROC$SYST$TAG") # write to file
processes = ensurelist(processes)
uptag = systag + "Up"
downtag = systag + "Down"
tdir = ensureTDirectory(file, dname, cd=True) if dname else file
for process in processes:
hists = []
skip = False
for var in [uptag, "", downtag]:
hname = process + var
hist = gethist(tdir, hname)
if not hist:
skip = True
break
hists.append(hist)
if skip:
LOG.warning("Skipping %r variation for %r..." % (systag, process))
continue
plot = Plot(variable, hists)
plot.draw(ratio=2, lstyle=1)
plot.drawlegend()
if text:
plot.drawtext(text)
if pname:
pname_ = repkey(pname, PROC=process, TAG=tag)
plot.saveas(pname_, ext=['png'])
if wname:
wname_ = repkey(wname, PROC=process, TAG=tag)
plot.canvas.Write(wname_, TH1.kOverwrite)
plot.close()
def compare(fnames, variables, **kwargs):
# SETTING
tnames = kwargs.get('tree', [])
entries = kwargs.get('entries', []) # entries of files
outdir = kwargs.get('outdir', "compare")
tag = kwargs.get('tag', "")
cut = kwargs.get('cut', "")
norm = kwargs.get('norm', False)
ensuredir(outdir)
if isinstance(tnames, str):
tnames = [tnames] * len(fnames)
else:
while len(tnames) < len(fnames):
tnames.append(tnames[-1])
if norm:
tag += "_norm" # normalize each histogram
# GET FILES & TREES
files = []
trees = []
for fname, tname in zip(fnames, tnames):
file = TFile.Open(fname, 'READ')
tree = file.Get(tname)
files.append(file)
trees.append(tree)
for variable in variables[:]:
if not hasattr(tree, variable.name):
LOG.warning(
"compare: tree %s:%s does not contain branch %r! Skipping..."
% (fname, tname, variable.name))
variables.remove(variable)
while len(entries) < len(trees):
i = len(entries)
entry = "%s %d" % (tnames[i], i + 1)
entries.append(entry)
# PLOT
for variable in variables:
fname = "%s/%s%s.png" % (outdir, variable, tag)
rrange = 0.5
header = "Compare" # legend header
text = "" # corner text
ratio = True #and False
logy = True and False
grid = True #and False
staterr = True #and False # add uncertainty band to first histogram
lstyle = 1 # solid lines
LOG.header(fname)
# GET HISOGRAMS
hists = []
for i, (tree, htitle) in enumerate(zip(trees, entries)):
hname = "%s_%d" % (variable.filename, i)
#hist = variable.gethist(hname,htitle)
#dcmd = variable.drawcmd(hname)
hist = variable.draw(
tree, cut, name=hname,
title=htitle) # create and fill hist from tree
evts = hist.GetEntries()
hist.SetTitle("%s (%d)" % (htitle, evts))
hists.append(hist)
print ">>> %r: entries=%d, integral=%s, mean=%#.6g, s.d.=%#.6g" % (
htitle, evts, hist.Integral(), hist.GetMean(),
hist.GetStdDev())
# DRAW PLOT
plot = Plot(variable, hists, norm=norm)
plot.draw(ratio=True,
logy=logy,
ratiorange=rrange,
lstyle=lstyle,
grid=grid,
staterr=staterr)
plot.drawlegend(header=header)
plot.drawtext(text)
plot.saveas(fname)
plot.close()
print
def compare(fnames, varsets, **kwargs):
# SETTING
tname = kwargs.get('tree', 'Events')
entries = kwargs.get('entries', []) # entries of files
outdir = kwargs.get('outdir', "compare")
tag = kwargs.get('tag', "")
cut = kwargs.get('cut', "")
norm = kwargs.get('norm', False)
ensuredir(outdir)
if norm:
tag += "_norm" # normalize each histogram
tree = TChain(tname)
# GET FILES & TREES
for fname in fnames:
tree.Add(fname)
# PLOT
for varset in varsets:
fname = "%s/%s%s.png" % (outdir, '-'.join(v.filename
for v in varset), tag)
rrange = 0.5
header = "Compare" # legend header
text = "" # corner text
ratio = True #and False
logy = True and False
grid = True #and False
staterr = True #and False # add uncertainty band to first histogram
lstyle = 1 # solid lines
xvar = varset[0]
LOG.header(fname)
# GET HISOGRAMS
hists = []
for i, variable in enumerate(varset):
hname = "%s_%d" % (variable.filename, i)
htitle = variable.name
#hist = variable.gethist(hname,htitle)
#dcmd = variable.drawcmd(hname)
hist = variable.draw(
tree, cut, name=hname,
title=htitle) # create and fill hist from tree
evts = hist.GetEntries()
hist.SetTitle(htitle) #"%s (%d)"%(htitle,evts)
hists.append(hist)
print ">>> %r: entries=%d, integral=%s, mean=%#.6g, s.d.=%#.6g" % (
htitle, evts, hist.Integral(), hist.GetMean(),
hist.GetStdDev())
# DRAW PLOT
plot = Plot(xvar, hists, norm=norm)
plot.draw(ratio=True,
logy=logy,
ratiorange=rrange,
lstyle=lstyle,
grid=grid,
staterr=staterr)
plot.drawlegend(header=header, latex=False)
plot.drawtext(text)
plot.saveas(fname)
plot.close()
print
def plothist(xtitle,hists,ratio=False,logy=False,norm=False):
# SETTING
outdir = ensuredir("plots/")
fname = outdir+"testPlot"
if ratio:
fname += "_ratio"
if logy:
fname += "_logy"
if norm:
fname += "_norm" # normalize each histogram
rrange = 0.5
width = 0.2 # legend width
position = 'topright' # legend position
header = "Gaussians" # legend header
text = "#mu#tau_{h}" # corner text
grid = True #and False
staterr = True and False # add uncertainty band to first histogram
lstyle = 1 # solid lines
# PLOT
LOG.header(fname)
plot = Plot(xtitle,hists,norm=norm)
plot.draw(ratio=ratio,logy=logy,ratiorange=rrange,lstyle=lstyle,grid=grid,staterr=staterr)
plot.drawlegend(position,header=header,width=width)
plot.drawtext(text)
plot.saveas(fname+".png")
plot.saveas(fname+".pdf")
#plot.saveas(fname+".C")
#plot.saveas(fname+".png",fname+".C")
#plot.saveas(fname,ext=['png','pdf'])
plot.close()
print
def plothist(xtitle,hists,ratio=False,logy=False,norm=False,**kwargs):
# SETTING
outdir = ensuredir("plots/")
fname = outdir+"testLegend"
rrange = 0.5
header = "Gaussians" # legend header
text = "#mu#tau_{h}" # corner text
grid = True and False
staterr = True and False # add uncertainty band to first histogram
lstyle = 1 # solid lines
panel = 1 # 1 = main (top) panel, 2 = ratio (bottom) panel
hists = [h.Clone("%s_clone%d"%(h.GetName(),i)) for i, h in enumerate(hists)]
if kwargs.get('pos',False):
fname += "_"+kwargs['pos'].replace(';','')
header = kwargs['pos']
if ratio:
fname += "_ratio"
# PLOT
LOG.header(fname)
plot = Plot(xtitle,hists,norm=norm)
plot.draw(ratio=ratio,logy=logy,ratiorange=rrange,lstyle=lstyle,grid=grid,staterr=staterr)
plot.drawlegend(border=True,header=header,panel=panel,**kwargs)
plot.drawtext(text)
plot.saveas(fname+".png")
#plot.saveas(fname+".pdf")
plot.close()
print