jobs += [(sample.name, str(njob))]
#Merges subfiles if needed
if args.isTest:
input_file_path = os.getcwd()+'/data/testArea/compareTauID/includeReco/'
else:
input_file_path = os.getcwd()+'/data/compareTauID/includeReco/'
if args.onlyReco:
input_file_path += 'onlyReco/'
merge_files = glob.glob(input_file_path + '*')
for mf in merge_files:
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')
output_file.mkdir(v)
output_file.cd(v)
list_of_hist[sample.output][v].getHist().Write()
output_file.Close()
#If the option to not run over the events again is made, load in the already created histograms here
else:
import glob
hist_list = glob.glob(os.getcwd() + '/data/plotTau/' + reco_or_gen_str +
'/*')
# Merge files if necessary
for n in hist_list:
merge(n)
list_of_hist = {}
# Load in the histograms from the files
for h in hist_list:
sample_name = h.split('/')[-1]
list_of_hist[sample_name] = {}
for v in var.keys():
list_of_hist[sample_name][v] = Histogram(
getObjFromFile(h + '/variables.root',
v + '/' + sample_name + '-' + v))
#
# Plot!
#
if args.runOnCream or args.isTest:
base_path_out = os.path.join(
os.path.expandvars('$CMSSW_BASE'), 'src', 'HNL',
'BackgroundEstimation', 'data', 'testArea', 'Results',
__file__.split('.')[0].rsplit('/', 1)[-1], data_str, args.year,
args.flavor, '-'.join([region_to_select, args.selection]))
else:
base_path_out = os.path.join(
os.path.expandvars('$CMSSW_BASE'), 'src', 'HNL',
'BackgroundEstimation', 'data', 'Results',
__file__.split('.')[0].rsplit('/', 1)[-1], data_str, args.year,
args.flavor, '-'.join([region_to_select, args.selection]))
in_files = glob.glob(os.path.join(base_path_in, '*'))
merge(in_files,
__file__,
jobs, ('sample', 'subJob'),
argParser,
istest=args.isTest)
if args.inData:
os.system("hadd -f " + base_path_in + "/events.root " + base_path_in +
'/*/events.root')
samples_to_plot = ['Data'] if args.inData else sample_manager.getOutputs()
for sample_output in samples_to_plot:
if args.isTest and sample_output != 'DY': continue
if args.inData:
output_dir = makePathTimeStamped(base_path_out)
in_file = base_path_in + '/events.root'
else:
else:
list_of_hist[sample.output][v].write(
output_name + 'variables' + subjobAppendix + '.root',
subdirs=[v],
is_test=args.isTest,
append=True)
#If the option to not run over the events again is made, load in the already created histograms here
else:
import glob
hist_list = glob.glob(os.getcwd() + '/data/plotTau/' + reco_or_gen_str +
'/*')
# Merge files if necessary
merge(hist_list, __file__, jobs, ('sample', 'subJob'), argParser)
list_of_hist = {}
# Load in the histograms from the files
for h in hist_list:
sample_name = h.split('/')[-1]
list_of_hist[sample_name] = {}
for v in var.keys():
list_of_hist[sample_name][v] = Histogram(
getObjFromFile(h + '/variables.root',
v + '/' + sample_name + '-' + v))
closeLogger(log)
#
# Plot!
out_file.cd()
else:
from HNL.Tools.mergeFiles import merge
from HNL.ObjectSelection.tauSelector import getCorrespondingLightLepDiscr
import glob
list_to_merge = [
f for f in glob.glob(
os.path.join(os.getcwd(), 'data',
__file__.split('.')[0].rsplit('/')[-1], args.year,
'*', '*')) if not 'TextResults' in f
]
merge(list_to_merge,
__file__,
jobs, ('sample', 'subJob'),
argParser,
istest=args.isTest)
def sortkey(c):
mass_dir = c.split('/')[-2]
# return int(mass_dir.split('-M')[-1])
return int(mass_dir.split('-m')[-1])
list_of_input = {
'Signal':
sorted(glob.glob(
os.path.join(os.getcwd(), 'data',
__file__.split('.')[0].rsplit('/')[-1], args.year,
'Signal', '*', 'events.root')),
key=sortkey),
for entry in xrange(chain.GetEntries()):
chain.GetEntry(entry)
new_vars[prompt_str][c].event_weight = num_of_entries[
'tot'][c] / num_of_entries[chain.HNL_mass][c]
out_trees[prompt_str][c].Fill()
out_file.cd(prompt_str + '/' + c)
out_trees[prompt_str][c].Write()
merge_base = '/storage_mnt/storage/user/lwezenbe/public/ntuples/HNL/' if not args.isTest else '/user/lwezenbe/private/PhD/Analysis_CMSSW_10_2_22/CMSSW_10_2_22/src/HNL/TMVA/data/testArea/'
signal_mergefiles = merge_base + 'TMVA/' + str(
args.year) + '/' + args.region + '-' + args.selection + '/Signal'
background_mergefiles = merge_base + 'TMVA/' + str(
args.year) + '/' + args.region + '-' + args.selection + '/Background'
merge([signal_mergefiles, background_mergefiles], __file__, jobs,
('sample', 'subJob'), argParser)
combined_dir = lambda sd: merge_base + 'TMVA/' + str(
args.year
) + '/' + args.region + '-' + args.selection + '/Combined/' + sd
makeDirIfNeeded(combined_dir('SingleTree') + '/test')
makeDirIfNeeded(combined_dir('TwoTrees') + '/test')
import glob
all_signal_files = glob.glob(merge_base + 'TMVA/' + str(args.year) + '/' +
args.region + '-' + args.selection +
'/Signal/*root')
low_mass = [10, 20, 40, 50, 60, 70, 80]
high_mass = [90, 100, 120, 130, 150, 200, 300, 400, 500, 600, 800]
tau = lambda mass: merge_base + 'TMVA/' + str(
args.year
#
# Argument parser and logging
#
import argparse
argParser = argparse.ArgumentParser(description = "Argument parser")
argParser.add_argument('signal', action='store', default=None, help='Select bkgr')
argParser.add_argument('bkgr', action='store', default=None, help='Select bkgr')
argParser.add_argument('flavor', action='store', default=None, help='0: electrons, 1:muons')
args = argParser.parse_args()
#Merges subfiles if needed
merge_files = glob.glob(os.getcwd()+'/data/compareLightLeptonId/*')
for mf in merge_files:
if "Results" in mf: continue
merge(mf)
input_signal = glob.glob(os.getcwd()+'/data/compareLightLeptonId/'+args.signal+'/*ROC-'+args.flavor+'.root')
bkgr_prefix = os.getcwd()+'/data/compareLightLeptonId/'+args.bkgr
from HNL.Plotting.plot import Plot
output_dir = makePathTimeStamped(os.getcwd()+'/data/Results/compareLightLeptonId/ROC/'+args.signal+'-'+args.bkgr)
curves = []
ordered_f_names = []
for f_path in input_signal:
f_name = f_path.rsplit('/', 1)[1].split('.')[0]
ordered_f_names.append(f_name)
roc_curve = ROC(f_name.split('-')[0], 1., '', f_path, misid_path =bkgr_prefix + '/'+f_name+'.root')
curves.append(roc_curve.returnGraph())
p = Plot(curves, ordered_f_names, args.signal+'_'+args.flavor, 'efficiency', 'misid', y_log=True)
if i == 0 and j == 0: efficiency[c_key][t_key].write(is_test=arg_string)
else: efficiency[c_key][t_key].write(append=True, is_test=arg_string)
closeLogger(log)
else:
import glob
from HNL.Tools.mergeFiles import merge
from HNL.Tools.helpers import makePathTimeStamped
from HNL.Plotting.plottingTools import extraTextFormat, drawLineFormat
cut_str = 'noFilter' if args.baselineCut is None else args.baselineCut
input_files = glob.glob(getOutputBase(args.flavor))
# Merge if needed
merge(input_files, __file__, jobs[args.flavor], ('sample', 'subJob'), argParser, istest=args.isTest, groups_to_merge = ['signalSelection-'+cut_str])
in_file = getOutputBase(args.flavor)+'/signalSelection-'+cut_str+'.root'
efficiency = {}
efficiency[cut_str] = createEfficiencies(in_file, CATEGORIES, {'HNLmass': (None, None, None)})
if args.stackBaselineCuts is not None:
for cut in args.stackBaselineCuts:
efficiency[cut] = createEfficiencies(getOutputBase(args.flavor)+'/signalSelection-'+cut+'.root', CATEGORIES, {'HNLmass': (None, None, None)})
if args.compareTriggerCuts is None:
if args.divideByCategory is not None:
for cut_key in efficiency.keys():
for sk_key in SUPER_CATEGORIES.keys():
from HNL.EventSelection.eventCategorization import CATEGORY_NAMES, ANALYSIS_CATEGORIES
# args.selection, args.region, sample.output, prompt_str
if not args.isTest:
base_path = os.path.join(os.getcwd(), 'data',
__file__.split('.')[0], args.selection,
args.region)
else:
base_path = os.path.join(os.getcwd(), 'data', 'testArea',
__file__.split('.')[0], args.selection,
args.region)
in_files = glob.glob(os.path.join(base_path, '*', '*'))
for f in in_files:
if '.txt' in f or '.root' in f: continue
merge(f)
def mergeValues(values_to_merge, errors_to_merge):
if len(values_to_merge) != len(errors_to_merge):
raise RuntimeError(
'length of both input arrays should be the same. Now they are: len(values) = '
+ str(len(values_to_merge)) + ' and len(errors) = ' +
str(len(errors_to_merge)))
out_val, out_err = (values_to_merge[0], errors_to_merge[0])
if len(values_to_merge) > 1:
for val, err in zip(values_to_merge[1:], errors_to_merge[1:]):
out_val += val
out_err = np.sqrt(out_err**2 + err**2)
return out_val, out_err
#
]
else:
bkgr_list = []
# data
if len(args.includeData) > 0:
data_list = glob.glob(getOutputName('data', year) + '/Data')
else:
data_list = []
print 'check merge'
# Merge files if necessary
mixed_list = signal_list + bkgr_list + data_list
merge(mixed_list,
__file__,
jobs[year], ('sample', 'subJob', 'includeData'),
argParser,
istest=args.isTest,
additionalArgs=[('year', year)])
if args.groupSamples:
background_collection = sample_manager.sample_groups.keys() + [
'non-prompt'
]
else:
background_collection = [b.split('/')[-1] for b in bkgr_list]
#Clean list of hist
tmp_signal_list = []
tmp_bkgr_list = []
for i_s, s in enumerate(signal_list):
sample_name = s.split('/')[-1]