def writefiles(self, listname, **kwargs):
"""Write filenames to text file for fast look up in future."""
writeevts = kwargs.pop('nevts', False) # also write nevents to file
listname = repkey(listname,
ERA=self.era,
GROUP=self.group,
SAMPLE=self.name)
print ">>> Write list to %r..." % (listname)
ensuredir(os.path.dirname(listname))
filenevts = self.getfilenevts(checkfiles=True, **
kwargs) if writeevts else None
treename = kwargs.pop('tree', 'Events')
files = self.getfiles(**kwargs)
with open(listname, 'w+') as lfile:
for infile in files:
if writeevts:
nevts = filenevts.get(infile, -1)
if nevts < 0:
LOG.warning(
"Did not find nevents of %s. Trying again..." %
(infile))
nevts = getnevents(infile, treename)
infile = "%s:%d" % (infile, nevts
) # write $FILENAM(:NEVTS)
lfile.write(infile + '\n')
def writefiles(self,listname,**kwargs):
"""Write filenames to text file for fast look up in future.
If there is more than one DAS dataset path, write lists separately for each path."""
kwargs = kwargs.copy() # do not edit given dictionary
writeevts = kwargs.pop('nevts',False) # also write nevents to file
listname = repkey(listname,ERA=self.era,GROUP=self.group,SAMPLE=self.name)
ensuredir(os.path.dirname(listname))
filenevts = self.getfilenevts(checkfiles=True,**kwargs) if writeevts else None
treename = kwargs.pop('tree','Events') # do not pass to Sample.getfiles
kwargs.pop('ncores') # do not pass to Sample.getfiles
kwargs['refresh'] = False # already got file list in Sample.filenevts
files = self.getfiles(**kwargs) # get right URL
if not files:
LOG.warning("writefiles: Did not find any files!")
def _writefile(ofile,fname,prefix=""):
"""Help function to write individual files."""
if writeevts: # add nevents at end of infile string
nevts = filenevts.get(fname,-1) # retrieve from cache
if nevts<0:
LOG.warning("Did not find nevents of %s. Trying again..."%(fname))
nevts = getnevents(fname,treename) # get nevents from file
fname = "%s:%d"%(fname,nevts) # write $FILENAM(:NEVTS)
ofile.write(prefix+fname+'\n')
paths = self.paths if '$PATH' in listname else [self.paths[0]]
for path in paths:
listname_ = repkey(listname,PATH=path.strip('/').replace('/','__'))
with open(listname_,'w+') as lfile:
if '$PATH' in listname: # write only the file list of this path to this text file
print ">>> Write %s files to list %r..."%(len(self.pathfiles[path]),listname_)
for infile in self.pathfiles[path]:
_writefile(lfile,infile)
elif len(self.paths)<=1: # write file list for the only path
if self.nevents>0:
print ">>> Write %s files to list %r..."%(len(files),listname_)
else:
print ">>> Write %s files (%d events) to list %r..."%(len(files),self.nevents,listname_)
for infile in files:
_writefile(lfile,infile)
else: # divide up list per DAS dataset path
if self.nevents>0:
print ">>> Write %s files to list %r..."%(len(files),listname_)
else:
print ">>> Write %s files (%d events) to list %r..."%(len(files),self.nevents,listname_)
for i, path in enumerate(self.paths):
print ">>> %3s files for %s..."%(len(self.pathfiles[path]),path)
lfile.write("DASPATH=%s\n"%(path)) # write special line to text file, which loadfiles() can parse
for infile in self.pathfiles[path]: # loop over this list (general list is sorted)
LOG.insist(infile in files,"Did not find file %s in general list! %s"%(infile,files))
_writefile(lfile,infile,prefix=" ")
if i+1<len(self.paths): # add extra white line between blocks
lfile.write("\n")
def hadd(self, sources, target, **kwargs):
"""Hadd files. Create intermediate target file if needed."""
target = self.expandpath(target, here=True)
verb = kwargs.get('verb', self.verbosity)
tmpdir = kwargs.get(
'tmpdir',
target.startswith(self.parent) and self.cpurl != '')
htarget = target
if tmpdir:
if not isinstance(tmpdir, str):
tmpdir = self.tmpdir
tmpdir = ensuredir(tmpdir, verb=verb)
htarget = os.path.join(tmpdir, os.path.basename(target))
if isinstance(sources, basestring):
sources = [sources]
source = ""
for i, file in enumerate(sources, 1):
source += self.expandpath(file, url=self.fileurl) + ' '
source = source.strip()
if verb >= 2:
print ">>> %-10s = %r" % ('sources', sources)
print ">>> %-10s = %r" % ('source', source)
print ">>> %-10s = %r" % ('target', target)
print ">>> %-10s = %r" % ('htarget', htarget)
out = self.execute("%s %s %s" % (self.haddcmd, htarget, source),
verb=verb)
if tmpdir:
cpout = self.cp(htarget, target, verb=verb)
rmfile(htarget)
return out
def getconfig(verb=0, refresh=False):
"""Get configuration from JSON file."""
global _cfgdefaults, basedir, CONFIG
if CONFIG and not refresh:
return CONFIG
# SETTING
cfgdir = ensuredir(basedir, "config")
cfgname = os.path.join(cfgdir, "config.json")
cfgdict = _cfgdefaults.copy()
rqdstrs = [
k for k, v in _cfgdefaults.iteritems() if isinstance(v, basestring)
]
rqddicts = [k for k, v in _cfgdefaults.iteritems() if isinstance(v, dict)]
# GET CONFIG
if os.path.isfile(cfgname):
with open(cfgname, 'r') as file:
cfgdict = json.load(file, object_pairs_hook=OrderedDict)
update = False
for key in _cfgdefaults.keys(): # check for missing keys
if key not in cfgdict:
LOG.warning(
"Key '%s' not set in config file %s. Setting to default %r"
% (key, os.path.relpath(cfgname), _cfgdefaults[key]))
cfgdict[key] = _cfgdefaults[key]
update = True
if update:
print ">>> Saving defaults..."
with open(cfgname, 'w') as file:
json.dump(cfgdict, file, indent=2)
else:
LOG.warning(
"Config file '%s' does not exist in %s. Creating with defaults..."
% (cfgname, cfgdir))
with open(cfgname, 'w') as file:
json.dump(cfgdict, file, indent=2)
# SANITY CHECKS - format of values for required keys
for key in rqdstrs:
assert key in cfgdict, "Required key '%s' not found in the configuration file..." % (
key)
assert isinstance(cfgdict[key],basestring),\
"Required value for '%s' must be a string or unicode! Instead is of type %s: %r"%(key,type(cfgdict[key]),key)
for key in rqddicts:
assert key in cfgdict, "Required key '%s' not found in the configuration file..." % (
key)
assert isinstance(cfgdict[key],dict),\
"Required value for '%s' must be a dictionary! Instead is of type %s: %r"%(key,type(cfgdict[key]),cfgdict[key])
# RETURN
if verb >= 1:
print '-' * 80
print ">>> Reading config JSON file '%s'" % cfgname
for key, value in cfgdict.iteritems():
print ">>> %-13s = %s" % (key, value)
print '-' * 80
CONFIG = Config(cfgdict, cfgname)
return CONFIG
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 main():
nevts = 1000000
predefine = True #and False # initialize histogram before calling filling
sample = 'ZTT' #'Data'
outdir = ensuredir('plots')
filedict = makesamples(nevts, sample=sample, outdir=outdir)
file, tree = filedict[sample]
nevts = tree.GetEntries()
print ">>> Using pseudo data %s..." % (file.GetName())
variables = [
('m_vis', 20, 0, 140),
('m_vis', [
0, 20, 40, 50, 60, 65, 70, 75, 80, 85, 90, 95, 100, 110, 130, 160,
200
]),
('pt_1', 40, 0, 120),
('pt_2', 40, 0, 120),
('pt_1+pt_2', 40, 0, 200),
('eta_1', 30, -3, 3),
('eta_2', 30, -3, 3),
('min(eta_1,eta_2)', 30, -3, 3),
('njets', 10, 0, 10),
]
variables2D = [
('pt_1', 50, 0, 100, 'pt_2', 50, 0, 100),
('pt_1', 50, 0, 100, 'eta_1', 50, -3, 3),
('pt_2', 50, 0, 100, 'eta_2', 50, -3, 3),
('eta_1', 50, -3, 3, 'eta_1', 50, -3, 3),
('pt_1', 50, 0, 100, 'm_vis', 50, 0, 150),
('pt_2', 50, 0, 100, 'm_vis', 50, 0, 150),
]
selections = [
('pt_1>30 && pt_2>30 && abs(eta_1)<2.4 && abs(eta_2)<2.4', "weight"),
]
dtime1 = singledraw(tree,
variables,
selections,
outdir=outdir,
predefine=predefine)
dtime2 = multidraw(tree,
variables,
selections,
outdir=outdir,
predefine=predefine)
dtime3 = multidraw2D(tree,
variables2D,
selections,
outdir=outdir,
predefine=predefine)
file.Close()
print ">>> Result: MultiDraw is %.2f times faster than TTree::Draw for %s events and %s variables!" % (
dtime1 / dtime2, nevts, len(variables))
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 setdefaultconfig(verb=0):
"""Set configuration to default values."""
global _cfgdefaults, basedir, CONFIG
# SETTING
cfgdir = ensuredir(basedir,"config")
cfgname = os.path.join(cfgdir,"config.json")
cfgdict = _cfgdefaults.copy()
if os.path.isfile(cfgname):
LOG.warning("Config file '%s' already exists. Overwriting with defaults..."%(cfgname))
CONFIG = Config(cfgdict,cfgname)
CONFIG.write()
return CONFIG
def hadd(self, sources, target, **kwargs):
"""Hadd files. Create intermediate target file if needed."""
target = self.expandpath(target, here=True)
dryrun = kwargs.get('dry', False)
verb = kwargs.get('verb', self.verbosity)
fileurl = kwargs.get('url', self.fileurl)
maxopen = kwargs.get('maxopenfiles',
0) # maximum number of files opened via -n option
tmpdir = kwargs.get(
'tmpdir',
target.startswith(self.parent) and self.cpurl != '')
htarget = target
if tmpdir: # create temporary dir for hadd target, and copy after
if not isinstance(tmpdir, str):
tmpdir = self.tmpdir
tmpdir = ensuredir(tmpdir, verb=verb)
htarget = os.path.join(tmpdir, os.path.basename(target))
if isinstance(sources, basestring):
sources = [sources]
source = ""
for i, file in enumerate(sources, 1):
fname = os.path.basename(file)
if '$PATH' in file and fileurl and isglob(
fname): # expand glob pattern
parent = os.path.dirname(file)
files = self.getfiles(parent, filter=fname, url=fileurl)
source += ' '.join(files) + ' '
else:
source += self.expandpath(file, url=fileurl) + ' '
source = source.strip()
if verb >= 2:
print ">>> %-10s = %r" % ('sources', sources)
print ">>> %-10s = %r" % ('source', source)
print ">>> %-10s = %r" % ('target', target)
print ">>> %-10s = %r" % ('htarget', htarget)
print ">>> %-10s = %r" % ('maxopen', maxopen)
haddcmd = self.haddcmd
if maxopen >= 1:
haddcmd += " -n %s" % (maxopen)
out = self.execute("%s %s %s" % (haddcmd, htarget, source),
dry=dryrun,
verb=verb)
if tmpdir: # copy hadd target and remove temporary file
cpout = self.cp(htarget, target, dry=dryrun, verb=verb)
if not dryrun:
rmfile(htarget)
return out
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 main(args):
eras = args.eras
injson = args.injson
periods = args.periods
rrange = args.rrange
diffs = args.diffs
outdir = args.outdir
verbosity = args.verbosity
for era in eras:
if diffs:
compareJSONs(diffs, verb=verbosity)
elif periods:
periods = cleanPeriods(periods)
jname = injson or getJSON(era)
outdir = ensuredir("json")
for period in periods:
jsonout = filterJSONByRunNumberRange(jname,
era,
period=period,
outdir=outdir,
verb=verbosity)
elif rrange:
jname = injson or getJSON(era)
jsonout = filterJSONByRunNumberRange(jname,
era,
rrange=rrange,
outdir=outdir,
verb=verbosity)
else:
print ">>> Please specifiy a period (-p BCD) or a range of run numbers (-r 272007-278770)"
print ">>> JSON for %r: %s" % (era, jname)
print ">>> datasets: "
datasets = getRuns(era)
for period in sorted(datasets.keys()):
start, end = datasets[period]
print ">>> %s: %s-%s" % (period, start, end)
def testBatch(path,verb=0):
# SETTINGS
verbosity = args.verbosity
dryrun = args.dryrun # prepare job and submit command, but do not submit
ntasks = args.ntasks # only run a few test tasks per job
nchecks = args.nchecks # number of times to check job status
queue = args.queue # queue option for the batch system (job flavor for HTCondor)
time = args.time # maximum time for the batch system
batchopts = args.batchopts # extra options for the batch system
#prompt = args.prompt # ask user confirmation before submitting
outdir = ensuredir("testBatch")
logdir = ensuredir("testBatch/log")
jobname = "testBatch"
tasklist = os.path.join(outdir,"testBatch.txt")
# INITIALIZE
LOG.header("__init__")
#batch = ensuremodule(system,"PicoProducer.batch."+batch)
batch = getbatch(args.batch,verb=verbosity+1)
print ">>> %r"%(batch)
print ">>> %-10s = %s"%('jobname',jobname)
print ">>> %-10s = %s"%('ntasks',ntasks)
print ">>> %-10s = %s"%('nchecks',nchecks)
print ">>> %-10s = %s"%('outdir',outdir)
print ">>> %-10s = %s"%('logdir',logdir)
print ">>> %-10s = %s"%('dryrun',dryrun)
print ">>> %-10s = %s"%('queue',queue)
print ">>> %-10s = %s"%('time',time)
print ">>> %-10s = %s"%('batchopts',batchopts)
print ">>> %-10s = %s"%('verbosity',verbosity)
print ">>> "
# PREPARE JOBS
createtasks(tasklist,ntasks)
# SUBMIT
LOG.header("Submit")
jkwargs = { # key-word arguments for batch.submit
'name': jobname, 'opt': batchopts, 'dry': dryrun,
'short': True, 'queue':queue, 'time':time
}
if batch.system=='HTCondor':
# use specific settings for KIT condor
if 'etp' in platform.node():
script = "python/batch/submit_HTCondor_KIT.sub"
else:
script = "python/batch/submit_HTCondor.sub"
appcmds = ["initialdir=%s"%(outdir),
"mylogfile='log/%s.$(ClusterId).$(ProcId).log'"%(jobname)]
jkwargs.update({ 'app': appcmds })
elif batch.system=='SLURM':
script = "python/batch/submit_SLURM.sh"
logfile = os.path.join(logdir,"%x.%A.%a.log") # $JOBNAME.o$JOBID.$TASKID.log
jkwargs.update({ 'log': logfile, 'array': ntasks })
#elif batch.system=='SGE':
#elif batch.system=='CRAB':
else:
LOG.throw(NotImplementedError,"Submission for batch system '%s' has not been implemented (yet)..."%(batch.system))
jobid = batch.submit(script,tasklist,**jkwargs)
print ">>> jobid: %s"%(jobid)
# CHECK JOBS
LOG.header("Check jobs")
for i in xrange(nchecks):
jobs = batch.jobs(jobid,verb=verbosity-1) # get refreshed job list
#jobs = batch.jobs(verb=verbosity-1) # get refreshed job list
print ">>> job objects: %r"%(jobs)
print ">>> "
#for job in jobs:
# print ">>> Found job %r, status=%r, args=%r"%(job,job.getstatus(),job.args.rstrip())
if i<nchecks-1:
sleep(2)
def main(args):
print ""
analysis = 'ztt_tid'
obsset = args.observables
channels = args.channels
eras = args.eras
tag = ""
bin = 'Tight'
title = "m_{T} < 60 GeV, D_{#zeta} > -25 GeV, |#Delta#eta| < 1.5" #"mutau, DeepTau2017v2p1VSjet"
fname = "$DIR/$ANALYSIS_$OBS_$CHANNEL-$BIN-$ERA$TAG.shapes.root"
pname = "$DIR/$OBS_$CHANNEL-$BIN-$ERA$TAG_$FIT.png"
#outfile = "$TAG/$ANALYSIS_%s_$CHANNEL-$ERA%s.inputs.root"%(obs,tag+outtag)
# PLOT SETTINGS
ZTT = "Z -> #tau#tau" # "Z -> #tau_{#mu}#tau_{h}"
#STYLE.sample_colors['ZTT'] = STYLE.kOrange-4
STYLE.sample_titles.update({
'ZTT': ZTT,
'ZTT_DM0': ZTT + ", h^{#pm}",
'ZTT_DM1': ZTT + ", h^{#pm}#pi^{0}",
'ZTT_DM10': ZTT + ", h^{#pm}h^{#mp}h^{#pm}",
'ZTT_DM11': ZTT + ", h^{#pm}h^{#mp}h^{#pm}#pi^{0}",
'ZJ': "Z -> ll",
})
procs = [
'ZTT',
'ZL',
'ZJ',
'TTT',
'TTJ',
'W',
'ST',
'VV',
'QCD',
'data_obs' #'STT', 'STJ'
]
groups = [
(['^TT*'], 'Top', 'ttbar'), #,STYLE.sample_colors['TT']),
#(['^TT*','ST*'],'Top','ttbar and single top'),
(['W*', 'ZJ', 'VV', 'ST*',
'QCD'], 'EWK', 'Electroweak'), #,STYLE.sample_colors['EWK']),
]
title_dict = {
'mvis': "m_{vis} (GeV)",
'mtau': "m(#tau_{h}) (GeV)",
}
tsize = 0.054
PLOT._lsize = 0.040 # label size
ratio = False
pos = 'x=0.56,y=0.88'
ncol = 1
square = not ratio and False
exts = ['png', 'pdf', 'root', 'C']
if "mtau" in obsset:
procs = ['ZTT_DM0', 'ZTT_DM1', 'ZTT_DM10', 'ZTT_DM11'] + procs[1:]
pos = 'x=0.22,y=0.85'
ncol = 2
# PLOT
for era in eras:
setera(era, extra="")
for channel in channels:
for obs in obsset:
indir = "output/%s" % era
outdir = ensuredir("plots/%s" % era)
xtitle = title_dict.get(obs)
fname_ = repkey(fname,
DIR=indir,
ANALYSIS=analysis,
OBS=obs,
CHANNEL=channel,
BIN=bin,
ERA=era,
TAG=tag)
pname_ = repkey(pname,
DIR=outdir,
ANALYSIS=analysis,
OBS=obs,
CHANNEL=channel,
BIN=bin,
ERA=era,
TAG=tag)
drawpostfit(fname_,
bin,
procs,
pname=pname_,
tag=tag,
group=groups,
title=title,
xtitle=xtitle,
tsize=tsize,
pos=pos,
ncol=ncol,
ratio=ratio,
square=square,
exts=exts)
def main():
eras = args.eras
periods = cleanEras(args.periods)
channel = args.channel
types = args.types
verbosity = args.verbosity
minbiases = [69.2] if periods else [69.2, 80.0, 69.2 * 1.046, 69.2 * 0.954]
for era in args.eras:
year = getyear(era)
mcfilename = "MC_PileUp_%s.root" % (era)
jsondir = os.path.join(datadir, 'json', str(year))
pileup = os.path.join(jsondir, "pileup_latest.txt")
CMSStyle.setCMSEra(year)
if era == '2016':
# https://twiki.cern.ch/twiki/bin/viewauth/CMS/PdmV2017Analysis
# /afs/cern.ch/cms/CAF/CMSCOMM/COMM_DQM/certification/Collisions16/13TeV/ReReco/Final/Cert_271036-284044_13TeV_23Sep2016ReReco_Collisions16_JSON.txt"
# /afs/cern.ch/cms/CAF/CMSCOMM/COMM_DQM/certification/Collisions16/13TeV/Final/Cert_271036-284044_13TeV_PromptReco_Collisions16_JSON.txt
# /afs/cern.ch/cms/CAF/CMSCOMM/COMM_DQM/certification/Collisions16/13TeV/PileUp/pileup_latest.txt
JSON = os.path.join(
jsondir,
"Cert_271036-284044_13TeV_ReReco_07Aug2017_Collisions16_JSON.txt"
)
datasets = {
'B': (272007, 275376),
'C': (275657, 276283),
'D': (276315, 276811),
'E': (276831, 277420),
'F': (277772, 278808),
'G': (278820, 280385),
'H': (280919, 284044),
}
campaign = "Moriond17"
samples = [
(
'TT',
"TT",
),
(
'DY',
"DYJetsToLL_M-10to50",
),
(
'DY',
"DYJetsToLL_M-50",
),
(
'DY',
"DY1JetsToLL_M-50",
),
(
'DY',
"DY2JetsToLL_M-50",
),
(
'DY',
"DY3JetsToLL_M-50",
),
(
'WJ',
"WJetsToLNu",
),
(
'WJ',
"W1JetsToLNu",
),
(
'WJ',
"W2JetsToLNu",
),
(
'WJ',
"W3JetsToLNu",
),
(
'WJ',
"W4JetsToLNu",
),
(
'ST',
"ST_tW_top",
),
(
'ST',
"ST_tW_antitop",
),
(
'ST',
"ST_t-channel_top",
),
(
'ST',
"ST_t-channel_antitop",
),
#( 'ST', "ST_s-channel", ),
(
'VV',
"WW",
),
(
'VV',
"WZ",
),
(
'VV',
"ZZ",
),
]
elif '2017' in era:
# https://twiki.cern.ch/twiki/bin/viewauth/CMS/PdmV2017Analysis
# /afs/cern.ch/cms/CAF/CMSCOMM/COMM_DQM/certification/Collisions16/13TeV/Final/Cert_271036-284044_13TeV_PromptReco_Collisions16_JSON.txt
# /afs/cern.ch/cms/CAF/CMSCOMM/COMM_DQM/certification/Collisions17/13TeV/PileUp/pileup_latest.txt
JSON = os.path.join(
jsondir,
"Cert_294927-306462_13TeV_PromptReco_Collisions17_JSON.txt")
datasets = {
'B': (297020, 299329),
'C': (299337, 302029),
'D': (302030, 303434),
'E': (303435, 304826),
'F': (304911, 306462),
}
samples_bug = []
samples_fix = []
if 'UL' in era:
campaign = "Summer19"
samples_fix = [
#( 'DY', "DYJetsToLL_M-10to50", ),
(
'DY',
"DYJetsToLL_M-50",
),
(
'DY',
"DY1JetsToLL_M-50",
),
(
'DY',
"DY2JetsToLL_M-50",
),
(
'DY',
"DY3JetsToLL_M-50",
),
(
'DY',
"DY4JetsToLL_M-50",
),
#( 'TT', "TTTo2L2Nu", ),
(
'TT',
"TTToHadronic",
),
#( 'TT', "TTToSemiLeptonic", ),
(
'WJ',
"WJetsToLNu",
),
(
'WJ',
"W1JetsToLNu",
),
(
'WJ',
"W2JetsToLNu",
),
(
'WJ',
"W3JetsToLNu",
),
(
'WJ',
"W4JetsToLNu",
),
(
'ST',
"ST_tW_top",
),
(
'ST',
"ST_tW_antitop",
),
(
'ST',
"ST_t-channel_top",
),
(
'ST',
"ST_t-channel_antitop",
),
#( 'ST', "ST_s-channel", ),
#( 'VV', "WW", ),
#( 'VV', "WZ", ),
#( 'VV', "ZZ", ),
]
else:
campaign = "Winter17_V2"
samples_bug = [
(
'DY',
"DYJetsToLL_M-50",
),
(
'WJ',
"W3JetsToLNu",
),
(
'VV',
"WZ",
),
]
samples_fix = [
(
'DY',
"DYJetsToLL_M-10to50",
),
(
'DY',
"DY1JetsToLL_M-50",
),
(
'DY',
"DY2JetsToLL_M-50",
),
(
'DY',
"DY3JetsToLL_M-50",
),
(
'DY',
"DY4JetsToLL_M-50",
),
(
'TT',
"TTTo2L2Nu",
),
(
'TT',
"TTToHadronic",
),
(
'TT',
"TTToSemiLeptonic",
),
(
'WJ',
"WJetsToLNu",
),
(
'WJ',
"W1JetsToLNu",
),
(
'WJ',
"W2JetsToLNu",
),
(
'WJ',
"W4JetsToLNu",
),
(
'ST',
"ST_tW_top",
),
(
'ST',
"ST_tW_antitop",
),
(
'ST',
"ST_t-channel_top",
),
(
'ST',
"ST_t-channel_antitop",
),
#( 'ST', "ST_s-channel", ),
(
'VV',
"WW",
),
(
'VV',
"ZZ",
),
]
samples = samples_bug + samples_fix
else:
# https://twiki.cern.ch/twiki/bin/viewauth/CMS/PdmV2018Analysis
# /afs/cern.ch/cms/CAF/CMSCOMM/COMM_DQM/certification/Collisions18/13TeV/PromptReco
# /afs/cern.ch/cms/CAF/CMSCOMM/COMM_DQM/certification/Collisions18/13TeV/PileUp/pileup_latest.txt
JSON = os.path.join(
jsondir,
"Cert_314472-325175_13TeV_PromptReco_Collisions18_JSON.txt")
datasets = {
'A': (315252, 316995),
'B': (317080, 319310),
'C': (319337, 320065),
'D': (320673, 325175),
}
campaign = "Autumn18"
samples = [
(
'TT',
"TTTo2L2Nu",
),
(
'TT',
"TTToHadronic",
),
(
'TT',
"TTToSemiLeptonic",
),
(
'DY',
"DYJetsToLL_M-10to50",
),
(
'DY',
"DYJetsToLL_M-50",
),
(
'DY',
"DY1JetsToLL_M-50",
),
(
'DY',
"DY2JetsToLL_M-50",
),
(
'DY',
"DY3JetsToLL_M-50",
),
(
'DY',
"DY4JetsToLL_M-50",
),
#( 'WJ', "WJetsToLNu", ),
(
'WJ',
"W1JetsToLNu",
),
(
'WJ',
"W2JetsToLNu",
),
(
'WJ',
"W3JetsToLNu",
),
(
'WJ',
"W4JetsToLNu",
),
(
'ST',
"ST_tW_top",
),
(
'ST',
"ST_tW_antitop",
),
(
'ST',
"ST_t-channel_top",
),
(
'ST',
"ST_t-channel_antitop",
),
#( 'ST', "ST_s-channel", ),
(
'VV',
"WW",
),
(
'VV',
"WZ",
),
(
'VV',
"ZZ",
),
]
# SAMPLES FILENAMES
fname = "$PICODIR/$SAMPLE_$CHANNEL.root"
if '$PICODIR' in fname:
import TauFW.PicoProducer.tools.config as GLOB
CONFIG = GLOB.getconfig(verb=verbosity)
fname = repkey(fname, PICODIR=CONFIG['picodir'])
for i, (group, sample) in enumerate(samples):
fname = repkey(fname,
ERA=era,
GROUP=group,
SAMPLE=sample,
CHANNEL=channel)
samples[i] = (sample, fname)
if verbosity >= 1:
print ">>> samples = %r" % (samples)
# JSON
jsons = {}
if periods:
outdir = ensuredir("json")
for period in periods:
start, end = getPeriodRunNumbers(period, datasets)
erarun = "Run%d%s" % (era, period)
jsonout = "json/" + re.sub(r"\d{6}-\d{6}", erarun,
JSON.split('/')[-1])
filterJSONByRunNumberRange(JSON,
jsonout,
start,
end,
verb=verbosity)
jsons[erarun] = jsonout
else:
jsons[era] = JSON
# DATA
datahists = {period: [] for period in jsons}
if 'data' in types:
for period, json in jsons.iteritems():
for minbias in minbiases:
filename = "Data_PileUp_%s_%s.root" % (
period, str(minbias).replace('.', 'p'))
datahist = getDataProfile(filename, json, pileup, 100, era,
minbias)
datahists[period].append((minbias, datahist))
elif args.plot:
for era in jsons:
for minbias in minbiases:
filename = "Data_PileUp_%s_%s.root" % (
era, str(minbias).replace('.', 'p'))
file, hist = gethist(filename, 'pileup', retfile=True)
if not file or not hist: continue
hist.SetDirectory(0)
file.Close()
datahists[era].append((minbias, hist))
# MC
if 'mc' in types:
mcfilename = "MC_PileUp_%s.root" % (era)
#mcfilename = "MC_PileUp_%s_%s.root"%(era,campaign)
getMCProfile(mcfilename, samples, channel, era)
if args.plot:
mchist = compareMCProfiles(samples, channel, era)
for era in jsons:
for minbias, datahist in datahists[era]:
compareDataMCProfiles(datahist, mchist, era, minbias)
deletehist(mchist) # clean memory
if era == '2017': # also check new/old pmx separately
mcfilename_bug = mcfilename.replace(".root", "_old_pmx.root")
mcfilename_fix = mcfilename.replace(".root", "_new_pmx.root")
getMCProfile(mcfilename_bug, samples_bug, channel, era)
getMCProfile(mcfilename_fix, samples_fix, channel, era)
if args.plot:
mchist_bug = compareMCProfiles(samples_bug,
channel,
era,
tag="old_pmx")
mchist_fix = compareMCProfiles(samples_fix,
channel,
era,
tag="new_pmx")
for era in jsons:
for minbias, datahist in datahists[era]:
compareDataMCProfiles(datahist,
mchist_bug,
era,
minbias,
tag="old_pmx")
compareDataMCProfiles(datahist,
mchist_fix,
era,
minbias,
tag="new_pmx")
# FLAT
if 'flat' in types:
filename = "MC_PileUp_%d_FlatPU0to75.root" % era
hist_flat = getFlatProfile(filename, 75)
for era in jsons:
for minbias, datahist in datahists[era]:
compareDataMCProfiles(datahist,
hist_flat,
era,
minbias,
tag="FlatPU0to75",
rmin=0.0,
rmax=3.1)
def createinputs(fname, sampleset, observables, bins, **kwargs):
"""Create histogram inputs in ROOT file for datacards.
fname: filename pattern of ROOT file
sampleset: SampleSet object
observables: list of Variables objects
bins: list of Selection objects
"""
#LOG.header("createinputs")
outdir = kwargs.get('outdir', "")
tag = kwargs.get('tag', "") # file tag
htag = kwargs.get('htag', "") # hist tag for systematic
filters = kwargs.get('filter',
None) # only create histograms for these processes
vetoes = kwargs.get('veto', None) # veto these processes
parallel = kwargs.get('parallel', True) # MultiDraw histograms in parallel
recreate = kwargs.get('recreate', False) # recreate ROOT file
replaceweight = kwargs.get('replaceweight', None) # replace weight
extraweight = kwargs.get('weight', "") # extraweight
shiftQCD = kwargs.get('shiftQCD', 0) # e.g 0.30 for 30%
verbosity = kwargs.get('verb', 0)
option = 'RECREATE' if recreate else 'UPDATE'
method = 'QCD_OSSS' if filters == None or 'QCD' in filters else None
method = kwargs.get('method', method)
# FILE LOGISTICS: prepare file and directories
files = {}
ensuredir(outdir)
fname = os.path.join(outdir, fname)
for obs in observables:
obsname = obs.filename
ftag = tag + obs.tag
fname_ = repkey(fname, OBS=obsname, TAG=tag)
file = TFile.Open(fname_, option)
if recreate:
print ">>> created file %s" % (fname_)
for selection in bins:
if not obs.plotfor(selection): continue
obs.changecontext(selection)
ensureTDirectory(file, selection.filename, cd=True, verb=verbosity)
if recreate:
string = joincuts(selection.selection, obs.cut)
TNamed("selection", string).Write(
) # write exact selection string to ROOT file for the record / debugging
#TNamed("weight",sampleset.weight).Write()
LOG.verb(
"%s selection %r: %r" % (obsname, selection.name, string),
verbosity, 1)
files[obs] = file
# GET HISTS
for selection in bins:
bin = selection.filename # bin name
print ">>>\n>>> " + color(
" %s " % (bin), 'magenta', bold=True, ul=True)
if htag:
print ">>> systematic uncertainty: %s" % (color(
htag.lstrip('_'), 'grey'))
if recreate or verbosity >= 1:
print ">>> %r" % (selection.selection)
hists = sampleset.gethists(observables,
selection,
method=method,
split=True,
parallel=parallel,
filter=filters,
veto=vetoes)
# SAVE HIST
ljust = 4 + max(11, len(htag)) # extra space
TAB = LOG.table("%10.1f %10d %-18s %s")
TAB.printheader('events', 'entries', 'variable',
'process'.ljust(ljust))
for obs, hist in hists.iterhists():
name = lreplace(hist.GetName(), obs.filename).strip(
'_') # histname = $VAR_$NAME (see Sample.gethist)
if not name.endswith(htag):
name += htag # HIST = $PROCESS_$SYSTEMATIC
name = repkey(name, BIN=bin)
drawopt = 'E1' if 'data' in name else 'EHIST'
lcolor = kBlack if any(
s in name
for s in ['data', 'ST', 'VV']) else hist.GetFillColor()
hist.SetOption(drawopt)
hist.SetLineColor(lcolor)
hist.SetFillStyle(0) # no fill in ROOT file
hist.SetName(name)
hist.GetXaxis().SetTitle(obs.title)
for i, yval in enumerate(hist):
if yval < 0:
print ">>> replace bin %d (%.3f<0) of %r" % (
i, yval, hist.GetName())
hist.SetBinContent(i, 0)
files[obs].cd(bin) # $FILE:$BIN/$PROCESS_$SYSTEMATC
hist.Write(name, TH1.kOverwrite)
TAB.printrow(hist.GetSumOfWeights(), hist.GetEntries(),
obs.printbins(), name)
deletehist(hist) # clean memory
# CLOSE
for obs, file in files.iteritems():
file.Close()
def plotpanels(ratio, **kwargs):
# SETTING
width = kwargs.get('width', 800)
height = kwargs.get('height', 750)
lmargin = kwargs.get('lmargin', 1.)
outdir = ensuredir("plots/")
fname = 'multipad_%sx%s_%s' % (width, height, str(ratio))
fname = outdir + fname.replace('.', 'p') + '.png'
# CANVAS
canvas = TCanvas('canvas', 'canvas', 100, 100, width, height)
canvas.SetFillColor(0)
canvas.SetBorderMode(0)
canvas.SetFrameBorderMode(0)
canvas.SetMargin(0.0, 0.0, 0.0, 0.0) # LRBT
canvas.Divide(2)
pad1 = canvas.GetPad(1)
pad2 = canvas.GetPad(2)
pad1.SetPad('pad1', 'pad1', 0.0, ratio, 1.0, 1.0)
pad1.SetMargin(0.14 * lmargin, 0.04, 0.11, 0.08)
pad1.SetFillColor(0)
#pad1.SetFillStyle(4000) # transparant (for pdf)
#pad1.SetFillStyle(0)
#pad1.SetBorderMode(0)
pad1.Draw()
pad2.SetPad('pad2', 'pad2', 0.0, 0.0, 1.0, ratio)
pad2.SetMargin(0.14 * lmargin, 0.04, 0.30, 0.10)
pad2.SetFillColor(0) #pad2.SetFillColorAlpha(0,0.0)
#pad2.SetFillStyle(4000) # transparant (for pdf)
#pad2.SetBorderMode(0)
pad2.Draw()
# SIZES
# https://root.cern.ch/doc/master/classTPad.html
TH, TW = pad1.GetWh() * pad1.GetHNDC(), pad1.GetWw() * pad1.GetWNDC()
BH, BW = pad2.GetWh() * pad2.GetHNDC(), pad2.GetWw() * pad2.GetWNDC()
TR = TH / TW
BR = BH / BW
TM = max(TH, TW)
BM = max(BH, BW)
Tm = min(TH, TW)
Bm = min(BH, BW)
print ">>> %6s: %8s %8s %8s %8s %8s %8s %8s" % (
'pad', 'Wh', 'HNDC', 'Wh*HNDC', 'Ww', 'WNDC', 'Ww*WNDC',
'Wh*HNDC/Ww*WNDC')
print ">>> %6s: %8.1f %8.3f %8.1f %8.1f %8.3f %8.1f" % (
'canvas', canvas.GetWh(), canvas.GetHNDC(),
canvas.GetWh() * canvas.GetHNDC(), canvas.GetWw(), canvas.GetWNDC(),
canvas.GetWw() * canvas.GetWNDC())
print ">>> %6s: %8.1f %8.3f %8.1f %8.1f %8.3f %8.1f %10.3f" % (
'pad1', pad1.GetWh(), pad1.GetHNDC(), pad1.GetWh() * pad1.GetHNDC(),
pad1.GetWw(), pad1.GetWNDC(), pad1.GetWw() * pad1.GetWNDC(), TR)
print ">>> %6s: %8.1f %8.3f %8.1f %8.1f %8.3f %8.1f %10.3f" % (
'pad2', pad2.GetWh(), pad2.GetHNDC(), pad2.GetWh() * pad2.GetHNDC(),
pad2.GetWw(), pad2.GetWNDC(), pad2.GetWw() * pad2.GetWNDC(), BR)
#scale = 1.0
#scale = 1./ratio
#scale = (1.-ratio)/ratio
#scale = (1.-ratio)/ratio if canvas.GetWh()<canvas.GetWw() else ratio/(1.-ratio)
#scale = TR/BR
scale = Tm / Bm
#scale = float(H)/W if W>H else 1 # float(W)/H
oscale = 1. / scale
tsize = 0.05
bsize = tsize * scale
toffset = 1.2
boffset = toffset * oscale
#boffset = 1.0+(toffset-1)*oscale
print ">>> 1/r=%.4f, (1-r)/r=%.4f, HNDC1/HNDC2=%.4f, WNDC1/WNDC2=%.4f, TR/BR=%.4f, TM/BM=%.4f, Tm/Bm=%.4f" % (
1. / ratio, (1. - ratio) / ratio, pad1.GetHNDC() / pad2.GetHNDC(),
pad1.GetWNDC() / pad2.GetWNDC(), TR / BR, TM / BM, Tm / Bm)
print ">>> tsize=%.4f, bsize=%.4f, scale=%.4f" % (tsize, bsize, scale)
print ">>> toffset=%.4f, boffset=%.4f, scale=%.4f" % (toffset, boffset,
oscale)
# TOP HIST
canvas.cd(1)
thist = TH1D('top', 'top', 10, 0, 100)
thist.GetXaxis().SetTitle("X title")
thist.GetYaxis().SetTitle("Y title")
print ">>> thist.GetXaxis().GetTitleOffset()=%.4f" % (
thist.GetXaxis().GetTitleOffset())
print ">>> thist.GetYaxis().GetTitleOffset()=%.4f" % (
thist.GetYaxis().GetTitleOffset())
thist.GetXaxis().SetTitleSize(tsize)
thist.GetYaxis().SetTitleSize(tsize)
#thist.GetXaxis().SetTitleOffset(toffset)
thist.GetYaxis().SetTitleOffset(toffset)
thist.GetXaxis().SetLabelSize(tsize)
thist.GetYaxis().SetLabelSize(tsize)
thist.Draw()
# BOTTOM HIST
canvas.cd(2)
bhist = thist.Clone('bottom')
bhist.SetTitle('bottom')
bhist.GetXaxis().SetTitleSize(bsize)
bhist.GetYaxis().SetTitleSize(bsize)
#thist.GetXaxis().SetTitleOffset(boffset)
bhist.GetYaxis().SetTitleOffset(boffset)
bhist.GetXaxis().SetLabelSize(bsize)
bhist.GetYaxis().SetLabelSize(bsize)
bhist.Draw()
# FINISH
canvas.SaveAs(fname)
canvas.Close()
print ">>> "
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 filterJSONByRunNumberRange(jsoninname,
era="",
period=None,
rrange=None,
start=None,
end=None,
outdir='json',
verb=0):
"""Split a given JSON file by start and end run number."""
outdir = ensuredir(outdir)
if period:
datasets = getRuns(era)
start, end = getPeriodRunNumbers(period, datasets)
eraname = "Run%s%s" % (era, period)
elif rrange:
match = range_rexp.match(rrange)
if not match:
raise IOERROR(
"Could not find range pattern in %r. Format should be e.g. 123456-1234567"
% (rrange))
start, end = int(match.group(1)), int(match.group(2))
eraname = "%s-%s" % (start, end)
elif start > 0 and end > 0:
eraname = "%s-%s" % (start, end)
else:
raise IOError(
"filterJSONByRunNumberRange: Please set either period (e.g. period='BCD'), "
"or the range (e.g. rrange='272007-278770'), or start and end (start=272007, end=278770)"
)
jsonoutname = os.path.join(
outdir, range_rexp.sub(eraname, os.path.basename(jsoninname)))
print ">>> filterJSONByRunNumberRange: %r for period=%r, range=%s-%s" % (
eraname, period, start, end)
# READ JSON IN
with open(jsoninname, 'r') as jsonin:
data = json.load(jsonin)
# FILTER run number range
nkeep = 0
ndrop = 0
for element in sorted(data.keys()):
if element.isdigit():
runnumber = int(element)
if runnumber < start or runnumber > end:
ndrop += 1
if verb >= 2: print ">>> dropping %s" % runnumber
del data[element]
else:
nkeep += 1
if verb >= 2: print ">>> keeping %s" % runnumber
else:
print "Warning! filterJSONByRunNumberRange: element is not an integer (run number): '%s'" % element
# WRITE JSON OUT
with open(jsonoutname, 'w') as jsonout:
data = json.dump(data, jsonout, sort_keys=True)
# SUMMARY
print ">>> output: %r" % (jsonoutname)
print ">>> saved %s / %s run numbers" % (nkeep, nkeep + ndrop)
return jsonoutname
def drawpostfit(fname, bin, procs, **kwargs):
"""Plot pre- and post-fit plots PostFitShapesFromWorkspace."""
print '>>>\n>>> drawpostfit("%s","%s")' % (fname, bin)
outdir = kwargs.get('outdir', "")
pname = kwargs.get('pname',
"$FIT.png") # replace $FIT = 'prefit', 'postfit'
ratio = kwargs.get('ratio', True)
tag = kwargs.get('tag', "")
xtitle = kwargs.get('xtitle', None)
title = kwargs.get('title', None)
text = kwargs.get('text', "")
tsize = kwargs.get('tsize', 0.050)
xmin = kwargs.get('xmin', None)
xmax = kwargs.get('xmax', None)
ymargin = kwargs.get('ymargin', 1.22)
groups = kwargs.get('group', [])
position = kwargs.get('pos', None) # legend position
ncol = kwargs.get('ncol', None) # legend columns
square = kwargs.get('square', False)
era = kwargs.get('era', "")
exts = kwargs.get('exts', ['pdf', 'png']) # figure extension
ymax = None
fits = ['prefit', 'postfit']
file = ensureTFile(fname, 'READ')
if outdir:
ensuredir(outdir)
if era:
setera(era)
# DRAW PRE-/POST-FIT
for fit in fits:
fitdirname = "%s_%s" % (bin, fit)
dir = file.Get(fitdirname)
if not dir:
LOG.warning('drawpostfit: Did not find dir "%s"' % (fitdirname),
pre=" ")
return
obshist = None
exphists = []
# GET HIST
for proc in procs: #reversed(samples):
hname = "%s/%s" % (fitdirname, proc)
hist = file.Get(hname)
if not hist:
LOG.warning(
'drawpostfit: Could not find "%s" template in directory "%s_%s"'
% (proc, bin, fit),
pre=" ")
continue
if 'data_obs' in proc:
obshist = hist
hist.SetLineColor(1)
ymax = hist.GetMaximum() * ymargin
else:
exphists.append(hist)
if proc in STYLE.sample_titles:
hist.SetTitle(STYLE.sample_titles[proc])
if proc in STYLE.sample_colors:
hist.SetFillStyle(1001)
hist.SetFillColor(STYLE.sample_colors[proc])
if len(exphists) == 0:
LOG.warning(
'drawpostfit: Could not find any templates in directory "%s"' %
(bin),
pre=" ")
continue
if not obshist:
LOG.warning(
'drawpostfit: Could not find a data template in directory "%s"'
% (bin),
pre=" ")
continue
for groupargs in groups:
grouphists(exphists, *groupargs, replace=True)
# PLOT
xtitle = (xtitle or exphists[0].GetXaxis().GetTitle()
) #.replace('[GeV]','(GeV)')
xmax = xmax or exphists[0].GetXaxis().GetXmax()
xmin = xmin or exphists[0].GetXaxis().GetXmin()
errtitle = "Pre-fit stat. + syst. unc." if fit == 'prefit' else "Post-fit unc."
pname_ = repkey(pname, FIT=fit, ERA=era)
rmin, rmax = (0.28, 1.52)
plot = Stack(xtitle, obshist, exphists)
plot.draw(xmin=xmin,
xmax=xmax,
ymax=ymax,
square=square,
ratio=ratio,
rmin=rmin,
rmax=rmax,
staterror=True,
errtitle=errtitle)
plot.drawlegend(position, tsize=tsize, text=text, ncol=ncol)
if title:
plot.drawtext(title, bold=False)
plot.saveas(pname_, outdir=outdir, ext=exts)
plot.close()
file.Close()
def plotinputs(fname, varprocs, observables, bins, **kwargs):
"""Plot histogram inputs from ROOT file for datacards, and write to ROOT file.
fname: filename pattern of ROOT file
varprocs: dictionary for systematic variation to list of processes,
e.g. { 'Nom': ['ZTT','TTT','W','QCD','data_obs'],
'TESUp': ['ZTT','TTT'], 'TESDown': ['ZTT','TTT'] }
observables: list of Variables objects
bins: list of Selection objects
"""
#LOG.header("plotinputs")
tag = kwargs.get('tag', "")
pname = kwargs.get('pname', "$OBS_$BIN$TAG.png")
outdir = kwargs.get('outdir', 'plots')
text = kwargs.get('text', "$BIN")
groups = kwargs.get('group',
[]) # add processes together into one histogram
verbosity = kwargs.get('verb', 0)
ensuredir(outdir)
print ">>>\n>>> " + color(" plotting... ", 'magenta', bold=True, ul=True)
for obs in observables:
obsname = obs.filename
ftag = tag + obs.tag
fname_ = repkey(fname, OBS=obsname, TAG=ftag)
file = ensureTFile(fname_, 'UPDATE')
for set, procs in varprocs.iteritems(
): # loop over processes with variation
if set == 'Nom':
systag = "" # no systematics tag for nominal
procs_ = procs[:]
else:
systag = '_' + set # systematics tag for variation, e.g. '_TESUp'
procs_ = [
(p + systag if p in procs else p) for p in varprocs['Nom']
] # add tag to varied processes
for selection in bins:
if not obs.plotfor(selection): continue
obs.changecontext(selection)
bin = selection.filename
text_ = repkey(
text, BIN=selection.title) # extra text in plot corner
tdir = ensureTDirectory(file, bin,
cd=True) # directory with histograms
if set == 'Nom':
gStyle.Write(
'style', TH1.kOverwrite
) # write current TStyle object to reproduce plots
# STACKS
pname_ = repkey(pname, OBS=obsname, BIN=bin,
TAG=ftag + systag) # image file name
wname = "stack" + systag # name in ROOT file
stackinputs(tdir,
obs,
procs_,
group=groups,
save=pname_,
write=wname,
text=text_)
# VARIATIONS
if 'Down' in set:
systag_ = systag.replace(
'Down', '') # e.g.'_TES' without 'Up' or 'Down' suffix
pname_ = repkey(pname,
OBS=obsname,
BIN=bin,
TAG=ftag + "_$PROC" +
systag) # image file name
wname = "plot_$PROC" + systag # name in ROOT file
comparevars(tdir,
obs,
procs,
systag_,
save=pname_,
write=wname,
text=text_)
file.Close()
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
