def createEfficiencies(output_name, event_categories, var_for_eff):
efficiency = {}
if args.divideByCategory is None:
efficiency['noCategories'] = {}
else:
for c in event_categories:
efficiency[c] = {}
bins_for_eff = var_for_eff['HNLmass'][1] if not args.makePlots else None
if args.compareTriggerCuts is None:
for k in efficiency.keys():
efficiency[k]['regularRun'] = Efficiency('efficiency_'+str(k), var_for_eff['HNLmass'][0], var_for_eff['HNLmass'][2], output_name, bins=bins_for_eff,
subdirs=['efficiency_'+str(k), 'l1_'+str('r')+'_l2_'+str('e')+'_l3_'+str('g')])
elif args.compareTriggerCuts == 'full':
if args.divideByCategory is None:
print "Inconsistent input: This mode is to be used together with divideByCategory. Exiting"
exit(0)
for k in efficiency.keys():
efficiency[k]['full'] = Efficiency('efficiency_'+str(k), var_for_eff['HNLmass'][0], var_for_eff['HNLmass'][2], output_name, bins=bins_for_eff,
subdirs=['efficiency_'+str(k), 'l1_'+str('f')+'_l2_'+str('u')+'_l3_'+str('l')])
# efficiency[k]['full'] = Efficiency('efficiency_'+str(k), var_for_eff['HNLmass'][0], var_for_eff['HNLmass'][2], output_name, bins=bins_for_eff)
else: #'single' or 'cumulative'
if args.divideByCategory is None:
print "Inconsistent input: This mode is to be used together with divideByCategory. Exiting"
exit(0)
for c in event_categories:
for ptcuts in returnCategoryPtCuts(c):
efficiency[c][ptcuts] = Efficiency('efficiency_'+str(c)+'_l1_'+str(ptcuts[0])+'_l2_'+str(ptcuts[1])+'_l3_'+str(ptcuts[2]),
var_for_eff['HNLmass'][0], var_for_eff['HNLmass'][2], output_name, bins=bins_for_eff, subdirs=['efficiency_'+str(c),
'l1_'+str(ptcuts[0])+'_l2_'+str(ptcuts[1])+'_l3_'+str(ptcuts[2])])
return efficiency
for f_name in f_names:
if 'FO' in f_name: continue
sub_files = glob.glob(input_name + '/'+f_name+'.root')
#
# TODO: THIS CODE SEEMS VERY OUTDATED, PLEASE UPDATE
#
print 'plotting', f_name
for f in sub_files:
output_dir = os.path.join(os.getcwd(), 'data', 'Results', 'calcSignalEfficiency', f_name, category_split_str, trigger_str, mass_str, flavor_name)
output_dir = makePathTimeStamped(output_dir)
eff = Efficiency('efficiency_noCategories', None, None, f, subdirs = ['efficiency_noCategories', 'l1_r_l2_e_l3_g'])
h = eff.getEfficiency()
p = Plot(h, 'V_{'+args.flavor+'N} = 0.01', 'efficiency', h.GetXaxis().GetTitle(), h.GetYaxis().GetTitle(), x_log=True, color_palette = 'Black')
p.drawHist(output_dir = output_dir, draw_option = 'EP')
# dict_of_categories = {}
# dict_of_names = {}
# for category in CATEGORIES:
# dict_of_categories[category] = []
# dict_of_names[category] = []
# rf = ROOT.TFile(f)
# if args.triggerTest or args.cumulativeCuts: get_nested = True
if "Results" in mf: merge_files.pop(merge_files.index(mf))
if not args.onlyReco and 'onlyReco' in mf: merge_files.pop(merge_files.index(mf))
script = os.path.expandvars(os.path.join('$CMSSW', 'src', 'HNL', 'ObjectSelection', 'compareTauID.py'))
merge(merge_files, script, jobs, ('sample', 'subJob'), argParser, istest=args.isTest)
list_of_bkgr_eff = {}
list_of_signal_eff = {}
list_of_bkgr_pt = {}
list_of_signal_pt = {}
inputfiles_eff = glob.glob(input_file_path + '*/*root')
samples = {f.split('/')[-2] for f in inputfiles_eff}
from HNL.Tools.efficiency import Efficiency
from HNL.Tools.helpers import getObjFromFile
for sample in samples:
eff = Efficiency('efficiency_pt', None, None, input_file_path + sample+'/efficiency.root', subdirs = ['deeptauVSjets-none-none-'+args.wp, 'efficiency_pt'])
if 'HNL' in sample:
list_of_signal_eff[sample] = eff.getEfficiency()
list_of_signal_pt[sample] = getObjFromFile(os.path.expandvars('$CMSSW_BASE/src/HNL/Test/data/plotTau/gen/' + sample + '/variables.root'), 'pt/'+sample+'-pt')
list_of_signal_pt[sample].Scale(1./list_of_signal_pt[sample].GetSumOfWeights())
else:
list_of_bkgr_eff[sample] = eff.getEfficiency()
list_of_bkgr_pt[sample] = getObjFromFile(os.path.expandvars('$CMSSW_BASE/src/HNL/Test/data/plotTau/gen/' + sample + '/variables.root'), 'pt/'+sample+'-pt')
list_of_bkgr_pt[sample].Scale(1./list_of_bkgr_pt[sample].GetSumOfWeights())
from HNL.Plotting.plot import Plot
if args.isTest:
output_dir = os.getcwd()+'/data/testArea/Results/compareTauID/includeReco/'
else:
output_dir = os.getcwd()+'/data/Results/compareTauID/includeReco/'
lambda c: c.eta_l,
np.arange(-2.5, 3., 0.5),
('#eta^{#tau}(offline) [GeV]', 'Efficiency')
]
}
list_of_var_hist['efficiency'][algo] = {}
list_of_var_hist['fakerate'][algo] = {}
for v in var_hist.keys():
list_of_var_hist['efficiency'][algo][v] = {}
list_of_var_hist['fakerate'][algo][v] = {}
for wp in algo_wp:
tot_name = output_name + '/' + sample.name + '_efficiency-' + str(
args.flavor) + '_' + args.subJob + '.root'
list_of_var_hist['efficiency'][algo][v][wp] = Efficiency(
'efficiency_' + v + algo, var_hist[v][0], var_hist[v][2],
tot_name, var_hist[v][1])
tot_name = output_name + '/' + sample.name + '_fakerate-' + str(
args.flavor) + '_' + args.subJob + '.root'
list_of_var_hist['fakerate'][algo][v][wp] = Efficiency(
'fakerate_' + v + algo, var_hist[v][0], var_hist[v][2],
tot_name, var_hist[v][1])
#Determine if testrun so it doesn't need to calculate the number of events in the getEventRange
if args.isTest:
event_range = xrange(150)
else:
event_range = sample.getEventRange(int(args.subJob))
from HNL.Tools.helpers import progress
from HNL.ObjectSelection.electronSelector import isBaseElectron
var = ['pt', 'eta']
inputFiles = {'efficiency' : glob.glob(os.getcwd()+'/data/compareLightLeptonId/'+args.signal+'/efficiency-'+str(args.flavor)+'.root'), 'fakerate': glob.glob(os.getcwd()+'/data/compareLightLeptonId/'+args.bkgr+'/fakerate-'+str(args.flavor)+'.root')}
for eff_name in ['efficiency', 'fakerate']:
for f in inputFiles[eff_name]:
print f
rf = TFile(f)
key_names = [k[0] for k in rootFileContent(rf)]
filtered_key_names = {fk.rsplit('-', 1)[0].split('/')[-1] for fk in key_names}
list_of_eff = {}
for fk in filtered_key_names: #algo
list_of_eff[fk] = {}
sample = f.split('/')[-2]
for k in [i.split('-', 1)[1] for i in key_names if fk in i]: #wps
list_of_eff[fk][k] = {}
for v in var:
list_of_eff[fk][k][v] = Efficiency(eff_name+'_'+v, None, None, f, subdirs = [fk+'-'+k, eff_name+'_'+v])
for fk in filtered_key_names: #algo
for v in var:
bkgr_hist = getObjFromFile(f, fk+'-'+k+'/'+eff_name+'_'+v+'/'+eff_name+'_'+v+'_denom')
tmp_list = [list_of_eff[fk][i][v] for i in list_of_eff[fk].keys()]
scale_factor = 0.25 if v == 'pt' else 1.
bkgr_hist.Scale(scale_factor*tmp_list[0].getEfficiency().GetSumOfWeights()/bkgr_hist.GetSumOfWeights())
p = Plot([efficiency.getEfficiency() for efficiency in tmp_list], [i+' '+fk for i in list_of_eff[fk].keys()]+['lepton distribution'], eff_name+'_'+args.flavor+'_'+v, bkgr_hist = bkgr_hist)
p.drawHist(output_dir = os.getcwd()+'/data/Results/compareLightLeptonId/'+fk+'/var/'+sample, draw_option = 'Hist')
#graph = roc_curve.return_graph()
#import ROOT
#c = ROOT.TCanvas('x', 'x')
'l3-2D' : (lambda c : (c.l_pt[2], abs(c.l_eta[2])), (np.arange(0., 100., 15.), np.arange(0., 3., .5)), ('p_{T}(l3) [GeV]', '|#eta|(subleading)'))}
if chain.is_signal:
var['HNLmass'] = (lambda c : c.HNLmass, np.array(mass_range), ('m_{N} [GeV]', 'Efficiency'))
from HNL.Tools.efficiency import Efficiency
from HNL.EventSelection.eventCategorization import returnCategoryTriggers
from HNL.Triggers.triggerSelection import applyCustomTriggers
eff = {}
for c in categories:
for v in {k for k in var.keys()}:
for t in category_map[c]:
if args.separateTriggers is None or args.separateTriggers == 'full':
name = str(c)+ '_' +v + '_' + t
if v != 'HNLmass':
eff[(c, v, t)] = Efficiency(name, var[v][0], var[v][2], getOutputName(v), var[v][1], subdirs=['efficiency_'+str(c), v, 'allTriggers'])
else:
eff[(c, v, t)] = Efficiency(name, var[v][0], var[v][2], getOutputName(v), var[v][1], subdirs=['efficiency_'+str(c), v, 'allTriggers'])
makeDirIfNeeded(getOutputName(v))
else:
for i, trigger in enumerate(category_triggers(chain, c)):
name = str(c)+ '_' +v + '_' + t +'_'+str(i)
if v != 'HNLmass':
eff[(c, v, t, i)] = Efficiency(name, var[v][0], var[v][2], getOutputName(v), var[v][1], subdirs=['efficiency_'+str(c), v, str(returnCategoryTriggerNames(c)[i])])
else:
eff[(c, v, t, i)] = Efficiency(name, var[v][0], var[v][2], getOutputName(v), var[v][1], subdirs=['efficiency_'+str(c), v, str(returnCategoryTriggerNames(c)[i])])
makeDirIfNeeded(getOutputName(v))
#
# Set event range
#
extra_text = [extraTextFormat(CATEGORY_TEX_NAMES[cat], ypos=0.83)
] if cat is not None else None
var = [
k[0].split('/')[-1]
for k in rootFileContent(rf,
starting_dir='efficiency_' + str(cat))
]
for v in var:
if args.separateTriggers is None or args.separateTriggers == 'full':
trigger_list = ['allTriggers']
else:
trigger_list = returnCategoryTriggerNames(cat)
eff_lists[cat][v] = [
Efficiency('_'.join([str(cat), v, 'integral',
str(i)]),
None,
None,
sub_f,
subdirs=['efficiency_' + str(cat), v, t])
for i, t in enumerate(trigger_list)
]
rf.Close()
if not args.ignoreCategories:
for trigger_cat_name in sorted(TRIGGER_CATEGORIES.keys()):
to_merge = [
eff_lists[c] for c in TRIGGER_CATEGORIES[trigger_cat_name]
]
eff_lists[trigger_cat_name] = mergeEfficiencies(
to_merge, trigger_cat_name)
for v in eff_lists[trigger_cat_name].keys():
for k in glob.glob(ip + '/' + algo + '/*')
if not 'all' in k
}
]
if args.includeReco == 'noIso' and d == 'iso':
rf = TFile(ip + '/' + default_iso_algo + '/None/efficiency.root')
var = [k[0].split('/')[1] for k in rootFileContent(rf)]
rf.Close()
for v in var:
list_of_eff = []
var_hist = []
eff_signal = Efficiency(
'efficiency',
None,
None,
ip + '/' + default_iso_algo + '/None/efficiency.root',
subdirs=[v, default_iso_algo + '-None'])
list_of_eff.append(eff_signal.getEfficiency())
var_hist_signal = eff_signal.getDenominator()
var_hist_signal.Scale(1. / var_hist_signal.GetSumOfWeights())
var_hist.append(var_hist_signal)
if args.overlaySignalAndBackground:
eff_bkgr = Efficiency(
'efficiency',
None,
None,
input_paths['background'][d][0] + '/' +
default_iso_algo + '/None/efficiency.root',
subdirs=[v, default_iso_algo + '-None'])
var_hist_bkgr = eff_bkgr.getDenominator()
ec = EventCategory(chain)
from HNL.Tools.efficiency import Efficiency
from HNL.EventSelection.oldEventCategorization import returnCategoryPtCuts
efficiency = {}
if not args.divideByCategory:
efficiency['noCategories'] = {}
else:
for c in ec.categories:
efficiency[c] = {}
if args.compareTriggerCuts is None:
for k in efficiency.keys():
efficiency[k]['regularRun'] = Efficiency('efficiency_'+str(k), var['HNLmass'][0], var['HNLmass'][2], output_name, bins=var['HNLmass'][1])
elif args.compareTriggerCuts == 'full':
if not args.divideByCategory:
print "Inconsistent input: This mode is to be used together with divideByCategory. Exiting"
exit(0)
for k in efficiency.keys():
efficiency[k]['full'] = Efficiency('efficiency_'+str(k), var['HNLmass'][0], var['HNLmass'][2], output_name, bins=var['HNLmass'][1])
else: #'single' or 'cumulative'
if not args.divideByCategory:
print "Inconsistent input: This mode is to be used together with divideByCategory. Exiting"
exit(0)
for c in ec.categories:
for ptcuts in returnCategoryPtCuts(c):
efficiency[c][ptcuts] = Efficiency('efficiency_'+str(c)+'_l1_'+str(ptcuts[0])+'_l2_'+str(ptcuts[1])+'_l3_'+str(ptcuts[2]), var['HNLmass'][0], var['HNLmass'][2], output_name, bins=var['HNLmass'][1])
#
for eff_or_fake in ['efficiency', 'fakerate']:
list_of_var_hist[eff_or_fake][discr] = {}
for algo in algos[discr]:
list_of_var_hist[eff_or_fake][discr][algo] = {}
for v in var_hist.keys():
list_of_var_hist[eff_or_fake][discr][algo][v] = {}
if discr != 'iso' or args.includeReco == 'noIso' or not args.makeCombinations:
for wp in algos[discr][algo]:
out_path = output_name(
discr, algo, wp
) + '/' + sample.name + '_' + eff_or_fake + subjobAppendix + '.root'
list_of_var_hist[eff_or_fake][discr][algo][v][
wp] = Efficiency('-'.join(
[eff_or_fake, v, algo,
str(wp)]),
var_hist[v][0],
var_hist[v][2],
out_path,
var_hist[v][1],
subdirs=[v, algo + '-' + str(wp)])
else:
for wp in algos[discr][algo]:
out_path = output_name(
discr, algo, wp
) + '/' + sample.name + '_' + eff_or_fake + subjobAppendix + '.root'
list_of_var_hist[eff_or_fake][discr][algo][v][wp] = {}
for ele_wp in algos['ele'][linkIsoToLep('ele', algo)]:
list_of_var_hist[eff_or_fake][discr][algo][v][wp][
ele_wp] = {}
for mu_wp in algos['mu'][linkIsoToLep('mu', algo)]:
list_of_var_hist[eff_or_fake][discr][algo][v][
wp][ele_wp][mu_wp] = Efficiency(
if chain.is_signal:
var['HNLmass'] = (lambda c: c.HNLmass, np.array(mass_range),
('m_{N} [GeV]', 'Efficiency'))
from HNL.Tools.efficiency import Efficiency
from HNL.EventSelection.eventCategorization import returnCategoryTriggers
from HNL.Triggers.triggerSelection import applyCustomTriggers
eff = {}
for c in categories:
for v in {k for k in var.keys()}:
for t in category_map[c]:
if args.separateTriggers is None or args.separateTriggers == 'full':
name = str(c) + '_' + v + '_' + t
if v != 'HNLmass':
eff[(c, v, t)] = Efficiency(name, var[v][0], var[v][2],
getOutputName(v), var[v][1])
else:
eff[(c, v, t)] = Efficiency(name, var[v][0], var[v][2],
getOutputName(v), var[v][1])
makeDirIfNeeded(getOutputName(v))
else:
for i, trigger in enumerate(category_triggers(chain, c)):
name = str(c) + '_' + v + '_' + t + '_' + str(i)
if v != 'HNLmass':
eff[(c, v, t, i)] = Efficiency(name, var[v][0],
var[v][2],
getOutputName(v),
var[v][1])
else:
eff[(c, v, t, i)] = Efficiency(name, var[v][0],
var[v][2],
('#eta^{#tau}(offline) [GeV]', 'Efficiency')
]
}
list_of_var_hist['efficiency'][algo] = {}
list_of_var_hist['fakerate'][algo] = {}
for v in var_hist.keys():
list_of_var_hist['efficiency'][algo][v] = {}
list_of_var_hist['fakerate'][algo][v] = {}
for wp in algo_wp:
tot_name = output_name + '/' + sample.name + '_efficiency-' + str(
args.flavor) + '_' + args.subJob + '.root'
list_of_var_hist['efficiency'][algo][v][wp] = Efficiency(
'efficiency_' + v + algo,
var_hist[v][0],
var_hist[v][2],
tot_name,
var_hist[v][1],
subdirs=[algo + '-' + wp, 'efficiency_' + v])
tot_name = output_name + '/' + sample.name + '_fakerate-' + str(
args.flavor) + '_' + args.subJob + '.root'
list_of_var_hist['fakerate'][algo][v][wp] = Efficiency(
'fakerate_' + v + algo,
var_hist[v][0],
var_hist[v][2],
tot_name,
var_hist[v][1],
subdirs=[algo + '-' + wp, 'fakerate_' + v])
#Determine if testrun so it doesn't need to calculate the number of events in the getEventRange
if args.isTest:
