def runPlot(opts):
lepton = opts.lepton
batchMode = opts.batch
inputDir = opts.input_dir
outputDir = opts.output_dir
verbose = opts.verbose
debug = opts.debug
dataset.Dataset.verbose_parsing = True if debug else False
groups = dataset.DatasetGroup.build_groups_from_files_in_dir(opts.samples_dir)
regions = regions_to_plot(opts.include_regions, opts.exclude_regions, opts.regions)
inputDir = outputDir+'/'+lepton+'/histos'
outputDir = outputDir+'/'+lepton+'/plots'
mkdirIfNeeded(outputDir)
histonames = dict((g.name, histonamesOneSample(g.name, variables_to_plot(), regions, leptonSources))
for g in groups)
groups_to_stack = [g.name for g in groups if not g.is_data]
if verbose:
print 'groups being included in the compositions: ',groups_to_stack
for region in regions:
all_histos = dict([(g.name,
rootUtils.fetchObjectsFromFile(os.path.join(inputDir, g.name+'_'+region+'.root'),
histonames[g.name][region],
verbose))
for g in groups])
for v in variables_to_plot():
histos = dict()
for s in leptonSources:
histos[s] = summedHisto(histos=[all_histos[g][v][s] for g in groups_to_stack],
label='')
histos['data'] = all_histos['data'][v]['Unknown']
plotStackedHistos(histos=histos, datakey='data', stackkeys=leptonSources,
outputDir=outputDir+'/'+region, region=region,
colors=fakeu.colorsFillSources(), verbose=verbose)
return
def runPlot(opts):
lepton = opts.lepton
batchMode = opts.batch
inputDir = opts.input_dir
outputDir = opts.output_dir
verbose = opts.verbose
debug = opts.debug
dataset.Dataset.verbose_parsing = True if debug else False
groups = dataset.DatasetGroup.build_groups_from_files_in_dir(
opts.samples_dir)
regions = regions_to_plot(opts.include_regions, opts.exclude_regions,
opts.regions)
inputDir = outputDir + '/' + lepton + '/histos'
outputDir = outputDir + '/' + lepton + '/plots'
mkdirIfNeeded(outputDir)
histonames = dict((g.name,
histonamesOneSample(g.name, variables_to_plot(),
regions, leptonSources))
for g in groups)
groups_to_stack = [g.name for g in groups if not g.is_data]
if verbose:
print 'groups being included in the compositions: ', groups_to_stack
for region in regions:
all_histos = dict([(g.name,
rootUtils.fetchObjectsFromFile(
os.path.join(inputDir,
g.name + '_' + region + '.root'),
histonames[g.name][region], verbose))
for g in groups])
for v in variables_to_plot():
histos = dict()
for s in leptonSources:
histos[s] = summedHisto(
histos=[all_histos[g][v][s] for g in groups_to_stack],
label='')
histos['data'] = all_histos['data'][v]['Unknown']
plotStackedHistos(histos=histos,
datakey='data',
stackkeys=leptonSources,
outputDir=outputDir + '/' + region,
region=region,
colors=fakeu.colorsFillSources(),
verbose=verbose)
return
def main():
parser = optparse.OptionParser(usage=usage)
parser.add_option('-i', '--input-dir', default='./out/fakerate')
parser.add_option('-o', '--output-dir', default='./out/fakerate/efficiencies')
parser.add_option('-l', '--lepton', default='el', help='either el or mu')
parser.add_option('-m', '--mode', help='emu')
parser.add_option('-t', '--tag', help='tag used to select the input files (e.g. Apr_04)')
parser.add_option('-f', '--fill-histos', action='store_true', default=False, help='force fill (default only if needed)')
parser.add_option('-T', '--tight-def', help='on-the-fly tight def, one of defs in fakeUtils.py: fakeu.lepIsTight_std, etc.')
parser.add_option('-v', '--verbose', action='store_true', default=False)
(options, args) = parser.parse_args()
inputDir = options.input_dir
outputDir = options.output_dir
lepton = options.lepton
mode = options.mode
tag = options.tag
verbose = options.verbose
if not tag : parser.error('tag is a required option')
if lepton not in ['el', 'mu'] : parser.error("invalid lepton '%s'"%lepton)
validModes = ['emu']
if mode not in validModes : parser.error("invalid mode %s"%mode)
tupleStem, treeName = filter(lambda _: _[0]==mode, fakeu.tupleStemsAndNames)[0]
templateInputFilename = "*_%(stem)s_tuple_%(tag)s.root" % {'tag':tag, 'stem':tupleStem}
templateOutputFilename = "%(stem)s_%(l)s_eff.root" % {'stem':tupleStem.replace('tuple','histos'), 'l':lepton}
outputFileName = os.path.join(outputDir, templateOutputFilename)
cacheFileName = outputFileName.replace('.root', '_'+mode+'_cache.root')
doFillHistograms = options.fill_histos or not os.path.exists(cacheFileName)
onthefly_tight_def = eval(options.tight_def) if options.tight_def else None # eval will take care of aborting on typos
optionsToPrint = ['inputDir', 'outputDir', 'mode', 'tag', 'doFillHistograms', 'cacheFileName', 'onthefly_tight_def']
if verbose :
print "working from %s"%os.getcwd()
print "being called as : %s"%' '.join(os.sys.argv)
print "options parsed:\n"+'\n'.join(["%s : %s"%(o, eval(o)) for o in optionsToPrint])
print 'input filenames: ',os.path.join(inputDir, templateInputFilename)
# collect inputs
tupleFilenames = glob.glob(os.path.join(inputDir, templateInputFilename))
samples = setSameGroupForAllData(fastSamplesFromFilenames(tupleFilenames, verbose))
samplesPerGroup = collections.defaultdict(list)
filenamesPerGroup = collections.defaultdict(list)
mkdirIfNeeded(outputDir)
for s, f in zip(samples, tupleFilenames) :
samplesPerGroup[s.group].append(s)
filenamesPerGroup[s.group].append(f)
vars = ['pt', 'pt_eta']
groups = [g for g in samplesPerGroup.keys() if g is not 'higgs']
if lepton=='el' : groups = [g for g in groups if g is not 'heavyflavor']
sourcesThisMode = ['real', 'conv', 'heavy', 'light', 'unknown'] if lepton=='el' else ['real', 'heavy', 'light', 'unknown']
#fill histos
if doFillHistograms :
start_time = time.clock()
num_processed_entries = 0
histosPerGroupPerSource = bookHistosPerSamplePerSource(vars, groups, sourcesThisMode, mode=mode)
for group in groups:
filenames = filenamesPerGroup[group]
sources = histosPerGroupPerSource.keys()
histosThisGroupPerSource = dict((s, histosPerGroupPerSource[s][group]) for s in sources)
histosAnyGroupPerSource = dict((s, histosPerGroupPerSource[s]['anygroup']) for s in sources) if group!='data' else {}
chain = r.TChain(treeName)
[chain.Add(fn) for fn in filenames]
if verbose: print "%s : %d entries"%(group, chain.GetEntries())
is_data = group in ['data']
print 'is_data ',is_data
num_processed_entries += fillHistos(chain=chain,
histosPerSource=histosThisGroupPerSource,
histosPerSourceAnygroup=histosAnyGroupPerSource,
lepton=lepton,
onthefly_tight_def=onthefly_tight_def,
verbose=verbose)
writeHistos(cacheFileName, histosPerGroupPerSource, verbose)
end_time = time.clock()
delta_time = end_time - start_time
one_minute = 60
if verbose:
print ("processed {0:d} entries ".format(num_processed_entries)
+"in "+("{0:d} min ".format(int(delta_time/60)) if delta_time>one_minute else
"{0:.1f} s ".format(delta_time))
+"({0:.1f} kHz)".format(num_processed_entries/delta_time))
# plot histos
histosPerGroupPerSource = fetchHistos(cacheFileName, histoNamesPerSamplePerSource(vars, groups, sourcesThisMode, mode), verbose)
# effs = computeEfficiencies(histosPerGroupPerSource) # still [var][gr][source][l/t]
for v in vars:
varIs1D, varIs2D = v=='pt', v=='pt_eta'
densThisSourceThisVar = dictSum(dict((s, histosPerGroupPerSource[v]['anygroup'][s]['loose']) for s in sourcesThisMode),
{'data' : histosPerGroupPerSource[v]['data']['unknown']['loose']})
numsThisSourceThisVar = dictSum(dict((s, histosPerGroupPerSource[v]['anygroup'][s]['tight']) for s in sourcesThisMode),
{'data' : histosPerGroupPerSource[v]['data']['unknown']['tight']})
if varIs1D:
lT, lX, lY = '#varepsilon(T|L)', 'p_{T} [GeV]', '#varepsilon(T|L)'
cname = 'stack_loose_'+lepton
lT, lY = 'loose '+lepton+', denominator to #varepsilon(T|L)', '#varepsilon(T|L)'
title = lT+' '+'anysource'+' '+lepton+';'+lX+';'+lY
plotStackedHistosWithData(densThisSourceThisVar,
outputDir, cname, title,
colors=fakeu.colorsFillSources(),
verbose=verbose)
cname = 'stack_tight_'+lepton
lT, lY = 'tight '+lepton+', numerator to #varepsilon(T|L)', '#varepsilon(T|L)'
title = lT+' '+'anysource'+' '+lepton+';'+lX+';'+lY
plotStackedHistosWithData(numsThisSourceThisVar,
outputDir, cname, title,
colors=fakeu.colorsFillSources(),
verbose=verbose)
for s in sourcesThisMode:
for v in vars:
groups = first(histosPerGroupPerSource).keys()
varIs1D, varIs2D = v=='pt', v=='pt_eta'
# effsThisSourceThisVar = dict((g, effs[v][g][s]) for g in groups)
densThisSourceThisVar = dictSum(dict((g, histosPerGroupPerSource[v][g][s]['loose'])
for g in groups if g not in ['anygroup','data']),
{'data' : histosPerGroupPerSource[v]['data']['unknown']['loose']})
numsThisSourceThisVar = dictSum(dict((g, histosPerGroupPerSource[v][g]['unknown']['tight'])
for g in groups if g not in ['anygroup','data']),
{'data' : histosPerGroupPerSource[v]['data']['unknown']['tight']})
if varIs1D:
# cname = 'eff_'+lepton+'_'+s
lT, lX, lY = '#varepsilon(T|L)', 'p_{T} [GeV]', '#varepsilon(T|L)'
# title = lT+' '+s+' '+lepton+';'+lX+';'+lY
# zoomIn = True
# fakeu.plot1dEfficiencies(effsThisSourceThisVar, cname, outputDir, title, zoomIn)
cname = 'stack_loose_'+lepton+'_'+s
lT, lY = 'loose '+lepton+', denominator to #varepsilon(T|L)', '#varepsilon(T|L)'
title = lT+' '+s+' '+lepton+';'+lX+';'+lY
plotStackedHistosWithData(densThisSourceThisVar,
outputDir, cname, title,
colors=SampleUtils.colors,
verbose=verbose)
cname = 'stack_tight_'+lepton+'_'+s
lT, lY = 'tight '+lepton+', numerator to #varepsilon(T|L)', '#varepsilon(T|L)'
title = lT+' '+s+' '+lepton+';'+lX+';'+lY
plotStackedHistosWithData(numsThisSourceThisVar,
outputDir, cname, title,
colors=SampleUtils.colors,
verbose=verbose)
# elif varIs2D:
# cname = 'eff_'+lepton+'_'+s
# lT, lX, lY = '#varepsilon(T|L)', 'p_{T} [GeV]', '#eta'
# title = lT+' '+s+' '+lepton+';'+lX+';'+lY
# fakeu.plot2dEfficiencies(effsThisSourceThisVar, cname, outputDir, title, zoomIn=zoomIn)
# writeHistos(outputFileName, effs, verbose)
if verbose : print "saved scale factors to %s" % outputFileName
colors=SampleUtils.colors,
verbose=verbose)
# elif varIs2D:
# cname = 'eff_'+lepton+'_'+s
# lT, lX, lY = '#varepsilon(T|L)', 'p_{T} [GeV]', '#eta'
# title = lT+' '+s+' '+lepton+';'+lX+';'+lY
# fakeu.plot2dEfficiencies(effsThisSourceThisVar, cname, outputDir, title, zoomIn=zoomIn)
# writeHistos(outputFileName, effs, verbose)
if verbose : print "saved scale factors to %s" % outputFileName
#___________________________________________________
leptonTypes = ['loose', 'tight']
leptonSources = []
colorsFillSources = fakeu.colorsFillSources()
colorsLineSources = fakeu.colorsLineSources()
markersSources = fakeu.markersSources()
enum2source = fakeu.enum2source
def fillHistos(chain, histosPerSource, histosPerSourceAnygroup={}, lepton='', onthefly_tight_def=None, verbose=False):
"""fill the histograms, returns the number of events
processed. histosPerSource is required; histosPerSourceAnygroup is
filled only when provided"""
class Counters: # scope trick (otherwise unavailable within nested func
nLoose, nTight = 0, 0
totWeightLoose, totWeightTight = 0.0, 0.0
def str(self):
counterNames = ['nLoose', 'nTight', 'totWeightLoose', 'totWeightTight']
return ', '.join(["%s : %.1f"%(c, getattr(self, c)) for c in counterNames])
counters = Counters()
histos['data'] = all_histos['data'][v]['Unknown']
plotStackedHistos(histos=histos,
datakey='data',
stackkeys=leptonSources,
outputDir=outputDir + '/' + region,
region=region,
colors=fakeu.colorsFillSources(),
verbose=verbose)
return
#___________________________________________________
allLeptonSources = fakeu.allLeptonSources()
leptonSources = fakeu.leptonSources()
colorsFillSources = fakeu.colorsFillSources()
colorsLineSources = fakeu.colorsLineSources()
markersSources = fakeu.markersSources()
enum2source = fakeu.enum2source
def histoName(sample, var, selection, source):
return 'h_' + sample + '_' + var + '_' + selection + '_' + source
def histonamesOneSample(sample_name, variables, selections, sources):
"dict of histogram names with keys [sel][var][source]"
hn = histoName
return dict([(se,
dict([(v,
dict([(so, hn(sample_name, v, se, so))
def main():
parser = optparse.OptionParser(usage=usage)
parser.add_option('-i', '--input-dir', default='./out/fakerate')
parser.add_option('-o', '--output-dir', default='./out/tight_variables_plots', help='dir for plots')
parser.add_option('-l', '--lepton', default='el', help='either el or mu')
parser.add_option('-r', '--region', help='one of the regions for which we saved the fake ntuples')
parser.add_option('-t', '--tag', help='tag used to select the input files (e.g. Apr_04)')
parser.add_option('-f', '--fill-histos', action='store_true', default=False, help='force fill (default only if needed)')
parser.add_option('-v', '--verbose', action='store_true', default=False)
(options, args) = parser.parse_args()
inputDir = options.input_dir
outputDir = options.output_dir
lepton = options.lepton
region = options.region
tag = options.tag
verbose = options.verbose
if not tag : parser.error('tag is a required option')
if lepton not in ['el', 'mu'] : parser.error("invalid lepton '%s'"%lepton)
filestems, treenames = utils.verticalSlice(fakeu.tupleStemsAndNames)
regions = filestems
assert region in regions,"invalid region '%s', must be one of %s"%(region, str(regions))
templateInputFilename = "*_%(region)s_tuple_%(tag)s.root" % {'tag':tag, 'region':region}
templateOutputFilename = "%(region)s_%(l)s_tight_plots.root" % {'region':region, 'l':lepton}
treeName = treenames[regions.index(region)]
outputDir = outputDir+'/'+region+'/'+lepton # split the output in subdirectories, so we don't overwrite things
mkdirIfNeeded(outputDir)
outputFileName = os.path.join(outputDir, templateOutputFilename)
cacheFileName = outputFileName.replace('.root', '_'+region+'_cache.root')
doFillHistograms = options.fill_histos or not os.path.exists(cacheFileName)
optionsToPrint = ['inputDir', 'outputDir', 'region', 'tag', 'doFillHistograms']
if verbose :
print "working from %s"%os.getcwd()
print "being called as : %s"%' '.join(os.sys.argv)
print "options:\n"+'\n'.join(["%s : %s"%(o, eval(o)) for o in optionsToPrint])
# collect inputs
if verbose : print 'input files ',os.path.join(inputDir, templateInputFilename)
tupleFilenames = glob.glob(os.path.join(inputDir, templateInputFilename))
samples = setSameGroupForAllData(fastSamplesFromFilenames(tupleFilenames, verbose))
if not samples : samples = [guessSampleFromFilename(f) for f in tupleFilenames] # if the fast guess didn't work, try the slow one
samplesPerGroup = collections.defaultdict(list)
filenamesPerGroup = collections.defaultdict(list)
for s, f in zip(samples, tupleFilenames) :
samplesPerGroup[s.group].append(s)
filenamesPerGroup[s.group].append(f)
vars = ['pt','eta','d0sig','z0SinTheta','etCone','ptCone','etConeCorr','ptConeCorr']
vars += ['relEtConeStd', 'relPtConeStd', 'relEtConeMod', 'relPtConeMod']
groups = samplesPerGroup.keys()
sources = leptonSources
#fill histos
if doFillHistograms :
lepLabel = "(probe %s)"%lepton
histosPerGroup = bookHistosPerGroup(vars, groups, lepLabel=lepLabel)
histosPerSource = bookHistosPerSource(vars, sources, lepLabel=lepLabel)
for group in groups:
isData = isDataSample(group)
filenames = filenamesPerGroup[group]
histosThisGroup = histosPerGroup[group]
chain = r.TChain(treeName)
[chain.Add(fn) for fn in filenames]
print "%s : %d entries"%(group, chain.GetEntries())
fillHistos(chain, histosThisGroup, histosPerSource, isData, lepton, group, verbose)
writeHistos(cacheFileName, {'perGroup':histosPerGroup, 'perSource':histosPerSource}, verbose)
# compute scale factors
histosPerGroup = fetchHistos(cacheFileName, histoNames(vars, groups), verbose)
histosPerSource = fetchHistos(cacheFileName, histoNames(vars, sources), verbose)
plotStackedHistos(histosPerGroup, outputDir+'/by_group', region, colors=SampleUtils.colors, verbose=verbose)
plotStackedHistos(histosPerSource, outputDir+'/by_source', region, colors=fakeu.colorsFillSources(), verbose=verbose)
plotIsoComparison(histosPerSource, outputDir+'/', region, lepton, verbose)
