def calculate_xsections(normalisation, category, channel):
global variable, met_type, path_to_JSON
# calculate the x-sections
branching_ratio = 0.15
if channel == 'combined':
branching_ratio = branching_ratio * 2
TTJet_xsection = calculate_xsection(normalisation['TTJet_measured'], luminosity, branching_ratio) # L in pb1
TTJet_xsection_unfolded = calculate_xsection(normalisation['TTJet_unfolded'], luminosity, branching_ratio) # L in pb1
MADGRAPH_xsection = calculate_xsection(normalisation['MADGRAPH'], luminosity, branching_ratio) # L in pb1
POWHEG_xsection = calculate_xsection(normalisation['POWHEG'], luminosity, branching_ratio) # L in pb1
MCATNLO_xsection = calculate_xsection(normalisation['MCATNLO'], luminosity, branching_ratio) # L in pb1
matchingdown_xsection = calculate_xsection(normalisation['matchingdown'], luminosity, branching_ratio) # L in pb1
matchingup_xsection = calculate_xsection(normalisation['matchingup'], luminosity, branching_ratio) # L in pb1
scaledown_xsection = calculate_xsection(normalisation['scaledown'], luminosity, branching_ratio) # L in pb1
scaleup_xsection = calculate_xsection(normalisation['scaleup'], luminosity, branching_ratio) # L in pb1
xsection_unfolded = {'TTJet_measured' : TTJet_xsection,
'TTJet_unfolded' : TTJet_xsection_unfolded,
'MADGRAPH': MADGRAPH_xsection,
'POWHEG': POWHEG_xsection,
'MCATNLO': MCATNLO_xsection,
# systematics
'matchingdown': matchingdown_xsection,
'matchingup': matchingup_xsection,
'scaledown': scaledown_xsection,
'scaleup': scaleup_xsection
}
write_data_to_JSON(xsection_unfolded, path_to_JSON + '/xsection_measurement_results' + '/kv' + str(unfoldCfg.SVD_k_value) + '/' + category + '/xsection_' + channel + '_' + met_type + '.txt')
def write_normalised_xsection_measurement(options, measurement, measurement_unfolded, summary = '' ):
'''
Writes the list of normalised measurements and normalised unfolded measurements of the form:
[Central Value, Lower Systemtic, Upper Systematic] to a json. Different combinations of
systematic uncertainty are stored as different json by appending different 'summary'
'''
path_to_JSON=options['path_to_JSON']
method=options['method']
channel=options['channel']
output_file = '{path_to_JSON}/TEST/central/normalised_xsection_{channel}_{method}_with_errors.txt'
output_file = output_file.format(
path_to_JSON = path_to_JSON,
channel = channel,
method = method,
)
if not summary == '':
output_file = output_file.replace( 'with_errors', summary + '_errors' )
output = {'TTJet_measured':measurement, 'TTJet_unfolded': measurement_unfolded}
write_data_to_JSON( output, output_file )
print("Data written to : ", output_file)
def calculate_normalised_xsections(normalisation, category, channel, normalise_to_one=False):
global variable, met_type, path_to_JSON
TTJet_normalised_xsection = calculate_normalised_xsection(normalisation['TTJet_measured'], bin_widths[variable], normalise_to_one)
TTJet_normalised_xsection_unfolded = calculate_normalised_xsection(normalisation['TTJet_unfolded'], bin_widths[variable], normalise_to_one)
MADGRAPH_normalised_xsection = calculate_normalised_xsection(normalisation['MADGRAPH'], bin_widths[variable], normalise_to_one)
POWHEG_normalised_xsection = calculate_normalised_xsection(normalisation['POWHEG'], bin_widths[variable], normalise_to_one)
MCATNLO_normalised_xsection = calculate_normalised_xsection(normalisation['MCATNLO'], bin_widths[variable], normalise_to_one)
matchingdown_normalised_xsection = calculate_normalised_xsection(normalisation['matchingdown'], bin_widths[variable], normalise_to_one)
matchingup_normalised_xsection = calculate_normalised_xsection(normalisation['matchingup'], bin_widths[variable], normalise_to_one)
scaledown_normalised_xsection = calculate_normalised_xsection(normalisation['scaledown'], bin_widths[variable], normalise_to_one)
scaleup_normalised_xsection = calculate_normalised_xsection(normalisation['scaleup'], bin_widths[variable], normalise_to_one)
normalised_xsection = {'TTJet_measured' : TTJet_normalised_xsection,
'TTJet_unfolded' : TTJet_normalised_xsection_unfolded,
'MADGRAPH': MADGRAPH_normalised_xsection,
'POWHEG': POWHEG_normalised_xsection,
'MCATNLO': MCATNLO_normalised_xsection,
# systematics
'matchingdown': matchingdown_normalised_xsection,
'matchingup': matchingup_normalised_xsection,
'scaledown': scaledown_normalised_xsection,
'scaleup': scaleup_normalised_xsection
}
filename = path_to_JSON + '/xsection_measurement_results' + '/kv' + str(unfoldCfg.SVD_k_value) + '/' + category + '/normalised_xsection_' + channel + '_' + met_type + '.txt'
if normalise_to_one:
filename = filename.replace('normalised_xsection', 'normalised_to_one_xsection')
write_data_to_JSON(normalised_xsection, filename)
def write_normalised_xsection_measurement(measurement, measurement_unfolded, channel, summary=''):
global path_to_JSON, met_type
output_file = path_to_JSON + 'central/normalised_xsection_' + channel + '_' + met_type + '_with_errors.txt'
if not summary == '':
output_file = output_file.replace('with_errors', summary + '_errors')
output = {'TTJet_measured':measurement, 'TTJet_unfolded': measurement_unfolded}
write_data_to_JSON(output, output_file)
def save_pulls(pulls, method, channel, tau_value, output_folder):
'''
Saves pull distributions in JSON format
'''
file_template = 'Pull_data_%s_%s_%.4g.txt'
output_file = output_folder + \
file_template % (method, channel, tau_value)
write_data_to_JSON(pulls, output_file)
print('Pulls saved in file: ', output_file)
return output_file
def main():
global config, options
parser = OptionParser()
parser.add_option("-p",
"--path",
dest="path",
default='data/fit_checks/no_merging',
help="set path to JSON files")
parser.add_option('--create_fit_data',
dest="create_fit_data",
action="store_true",
help="create the fit data for testing.")
parser.add_option('--refit',
dest="refit",
action="store_true",
help="Fit again even if the output already exists")
parser.add_option('--test',
dest="test",
action="store_true",
help="Test only: run just one selected sample")
variables = config.histogram_path_templates.keys()
fit_variables = fit_var_inputs
mc_samples = ['TTJet', 'SingleTop', 'QCD', 'V+Jets']
tests = closure_tests
channels = ['electron', 'muon']
COMEnergies = ['7', '8']
(options, _) = parser.parse_args()
print 'Running from path', options.path
if (options.create_fit_data):
create_fit_data(options.path,
variables,
fit_variables,
mc_samples,
COMEnergies=COMEnergies,
channels=channels)
output_file = options.path + '/fit_test_output.txt'
if options.test:
output_file = options.path + '/fit_test_output_test.txt'
if options.refit or not os.path.isfile(output_file) or options.test:
if os.path.isfile(options.path + '/fit_check_data.txt'):
fit_data = read_data_from_JSON(options.path +
'/fit_check_data.txt')
results = run_tests(fit_data, COMEnergies, variables,
fit_variables, mc_samples, channels, tests)
write_data_to_JSON(results, output_file)
else:
print 'Please run bin/prepare_data_for_fit_checks first'
print 'Then run this script with the option --create_fit_data.'
results = read_data_from_JSON(output_file)
plot_results(results)
def get_unfolded_normalisation(TTJet_fit_results, category, channel):
global variable, met_type, path_to_JSON, file_for_unfolding, file_for_powheg, file_for_mcatnlo
global file_for_matchingdown, file_for_matchingup, file_for_scaledown, file_for_scaleup
global ttbar_generator_systematics
files_for_systematics = {
ttbar_theory_systematic_prefix + 'matchingdown':file_for_matchingdown,
ttbar_theory_systematic_prefix + 'matchingup':file_for_matchingup,
ttbar_theory_systematic_prefix + 'scaledown':file_for_scaledown,
ttbar_theory_systematic_prefix + 'scaleup':file_for_scaleup,
}
h_truth, h_measured, h_response = None, None, None
if category in ttbar_generator_systematics and not 'ptreweight' in category:
h_truth, h_measured, h_response = get_unfold_histogram_tuple(files_for_systematics[category], variable, channel, met_type)
elif 'mcatnlo_matrix' in category:
h_truth, h_measured, h_response = get_unfold_histogram_tuple(file_for_mcatnlo, variable, channel, met_type)
else:
h_truth, h_measured, h_response = get_unfold_histogram_tuple(file_for_unfolding, variable, channel, met_type)
MADGRAPH_results = hist_to_value_error_tuplelist(h_truth)
POWHEG_results = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_powheg, variable, channel, met_type)[0])
MCATNLO_results = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_mcatnlo, variable, channel, met_type)[0])
matchingdown_results = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_matchingdown, variable, channel, met_type)[0])
matchingup_results = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_matchingup, variable, channel, met_type)[0])
scaledown_results = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_scaledown, variable, channel, met_type)[0])
scaleup_results = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_scaleup, variable, channel, met_type)[0])
TTJet_fit_results_unfolded = unfold_results(TTJet_fit_results,
category,
channel,
h_truth,
h_measured,
h_response,
'RooUnfoldSvd'
# 'TSVDUnfold'
)
normalisation_unfolded = {
'TTJet_measured' : TTJet_fit_results,
'TTJet_unfolded' : TTJet_fit_results_unfolded,
'MADGRAPH': MADGRAPH_results,
# other generators
'POWHEG': POWHEG_results,
'MCATNLO': MCATNLO_results,
# systematics
'matchingdown': matchingdown_results,
'matchingup': matchingup_results,
'scaledown': scaledown_results,
'scaleup': scaleup_results
}
write_data_to_JSON(normalisation_unfolded, path_to_JSON + '/xsection_measurement_results' + '/kv' + str(unfoldCfg.SVD_k_value) + '/' + category + '/normalisation_' + channel + '_' + met_type + '.txt')
return normalisation_unfolded
def create_fit_data( path, variables, fit_variables, mc_samples, COMEnergies, channels ):
'''
Creates the fit data in path + fit_check_data.txt in the JSON format.
The resulting dictionary is of the form
{centre-of-mass-energy: {
channel : {
variable : {
variable_bin : {
fit_variable : {
templates : { sample : []},
initial_values : { sample : []}
}
}
}
}
}
}
'''
# first step is to read the data
raw_data = {}
# the folder structure is
# path/fit_variable/centre-of-mass-energy/variable/fit_results/central/*
for fit_variable in fit_variables:
raw_data[fit_variable] = {}
for COMEnergy in COMEnergies:
raw_data[fit_variable][COMEnergy] = {}
for variable in variables:
raw_data[fit_variable][COMEnergy][variable] = {}
for channel in channels:
raw_data[fit_variable][COMEnergy][variable][channel] = {}
data_path = path + '/' + fit_variable + '/' + str( COMEnergy ) + 'TeV'
templates = read_fit_templates( data_path, variable,
channel = channel )
initial_values = read_fit_input( data_path,
variable,
channel = channel )
raw_data[fit_variable][COMEnergy][variable][channel]['templates'] = templates
raw_data[fit_variable][COMEnergy][variable][channel]['initial_values'] = initial_values
# put it into the new structure
fit_data = {}
for COMEnergy in COMEnergies:
fit_data[COMEnergy] = {}
for channel in channels:
fit_data[COMEnergy][channel] = {}
for variable in variables:
fit_data[COMEnergy][channel][variable] = {}
for v_bin in variable_bins_ROOT[variable]:
fit_data[COMEnergy][channel][variable][v_bin] = {}
for fit_variable in fit_variables:
fit_data[COMEnergy][channel][variable][v_bin][fit_variable] = {}
fit_data[COMEnergy][channel][variable][v_bin][fit_variable]['templates'] = raw_data[fit_variable][COMEnergy][variable][channel]['templates']
fit_data[COMEnergy][channel][variable][v_bin][fit_variable]['initial_values'] = raw_data[fit_variable][COMEnergy][variable][channel]['initial_values']
write_data_to_JSON( fit_data, path + '/fit_check_data.txt', indent = False )
def save_pulls(pulls, test, method, channel, output_folder, n_toy_mc,
n_toy_data, offset_toy_mc, offset_toy_data):
'''
Saves pull distributions in JSON format
'''
file_template = 'Pulls_%s_%s_%s_toy_MC_%d_to_%d_MC_%d_to_%d_data.txt'
output_file = output_folder + \
file_template % (test, method, channel, offset_toy_mc + 1, n_toy_mc +
offset_toy_mc, offset_toy_data + 1,
n_toy_data + offset_toy_data)
write_data_to_JSON(pulls, output_file)
print('Pulls saved in file: ', output_file)
def write_normalised_xsection_measurement( measurement, measurement_unfolded, channel, summary = '' ):
global path_to_JSON, met_type, k_values
output_file = path_to_JSON + '/' + channel + '/kv' + str( k_values[channel] ) + '/central/normalised_xsection_' + met_type + '_with_errors.txt'
if channel == 'combined':
output_file = output_file.replace( 'kv' + str( k_values[channel] ), '' )
if not summary == '':
output_file = output_file.replace( 'with_errors', summary + '_errors' )
output = {'TTJet_measured':measurement, 'TTJet_unfolded': measurement_unfolded}
write_data_to_JSON( output, output_file )
def save_pulls(pulls, test, method, channel, output_folder, n_toy_mc,
n_toy_data, offset_toy_mc, offset_toy_data):
'''
Saves pull distributions in JSON format
'''
file_template = 'Pulls_%s_%s_%s_toy_MC_%d_to_%d_MC_%d_to_%d_data.txt'
output_file = output_folder + \
file_template % (test, method, channel, offset_toy_mc + 1, n_toy_mc +
offset_toy_mc, offset_toy_data + 1,
n_toy_data + offset_toy_data)
write_data_to_JSON(pulls, output_file)
print('Pulls saved in file: ', output_file)
def calculate_normalised_xsections( normalisation, category, channel, normalise_to_one = False ):
global variable, met_type, path_to_JSON, phase_space
binWidths = None
if phase_space == 'VisiblePS':
binWidths = bin_widths_visiblePS
elif phase_space == 'FullPS':
binWidths = bin_widths
TTJet_normalised_xsection = calculate_normalised_xsection( normalisation['TTJet_measured'], binWidths[variable], normalise_to_one )
TTJet_withoutFakes_normalised_xsection = calculate_normalised_xsection( normalisation['TTJet_measured_withoutFakes'], binWidths[variable], normalise_to_one )
TTJet_normalised_xsection_unfolded = calculate_normalised_xsection( normalisation['TTJet_unfolded'], binWidths[variable], normalise_to_one )
normalised_xsection = {'TTJet_measured' : TTJet_normalised_xsection,
'TTJet_measured_withoutFakes' : TTJet_withoutFakes_normalised_xsection,
'TTJet_unfolded' : TTJet_normalised_xsection_unfolded
}
if category == 'central':
powhegPythia8_normalised_xsection = calculate_normalised_xsection( normalisation['powhegPythia8'], binWidths[variable], normalise_to_one )
amcatnlo_normalised_xsection = calculate_normalised_xsection( normalisation['amcatnlo'], binWidths[variable], normalise_to_one )
powhegHerwig_normalised_xsection = calculate_normalised_xsection( normalisation['powhegHerwig'], binWidths[variable], normalise_to_one )
amcatnloHerwig_normalised_xsection = calculate_normalised_xsection( normalisation['amcatnloHerwig'], binWidths[variable], normalise_to_one )
madgraphMLM_normalised_xsection = calculate_normalised_xsection( normalisation['madgraphMLM'], binWidths[variable], normalise_to_one )
scaledown_normalised_xsection = calculate_normalised_xsection( normalisation['scaledown'], binWidths[variable], normalise_to_one )
scaleup_normalised_xsection = calculate_normalised_xsection( normalisation['scaleup'], binWidths[variable], normalise_to_one )
massdown_normalised_xsection = calculate_normalised_xsection( normalisation['massdown'], binWidths[variable], normalise_to_one )
massup_normalised_xsection = calculate_normalised_xsection( normalisation['massup'], binWidths[variable], normalise_to_one )
normalised_xsection['powhegPythia8'] = powhegPythia8_normalised_xsection
normalised_xsection['amcatnlo'] = amcatnlo_normalised_xsection
normalised_xsection['madgraphMLM' ] = madgraphMLM_normalised_xsection
normalised_xsection['powhegHerwig'] = powhegHerwig_normalised_xsection
normalised_xsection['amcatnloHerwig'] = amcatnloHerwig_normalised_xsection
normalised_xsection['scaledown'] = scaledown_normalised_xsection
normalised_xsection['scaleup'] = scaleup_normalised_xsection
normalised_xsection['massdown'] = massdown_normalised_xsection
normalised_xsection['massup'] = massup_normalised_xsection
file_template = '{path_to_JSON}/{category}/normalised_xsection_{channel}_{method}.txt'
filename = file_template.format(
path_to_JSON = path_to_JSON,
category = category,
channel = channel,
method = method,
)
if normalise_to_one:
filename = filename.replace( 'normalised_xsection', 'normalised_to_one_xsection' )
write_data_to_JSON( normalised_xsection, filename )
def write_fit_results( channel, category, fit_results ):
global variable, met_type, output_path
folder_template = '%(path)s/%(fit_vars)s/%(CoM)dTeV/%(variable)s/fit_results/%(category)s/'
inputs = {
'path':output_path,
'CoM': measurement_config.centre_of_mass_energy,
'variable': variable,
'category': category,
'fit_vars': '_'.join( fit_variables )
}
output_folder = folder_template % inputs
write_data_to_JSON( fit_results, output_folder + 'fit_results_' + channel + '_' + met_type + '.txt' )
def save(self, output_path):
if not self.have_normalisation:
self.calculate_normalisation()
folder_template = '{path}/normalisation/{method}/{CoM}TeV/{variable}/'
folder_template += '{phase_space}/{category}/'
inputs = {
'path': output_path,
'CoM': self.config.centre_of_mass_energy,
'variable': self.variable,
'category': self.category,
'method': self.method_string(),
'phase_space': self.phase_space,
}
output_folder = folder_template.format(**inputs)
file_template = '{type}_{channel}_{met_type}.txt'
inputs = {
'channel': self.channel,
'met_type': self.met_type,
}
write_data_to_JSON(
self.normalisation, output_folder +
file_template.format(type='normalisation', **inputs))
write_data_to_JSON(
self.initial_normalisation, output_folder +
file_template.format(type='initial_normalisation', **inputs))
write_data_to_JSON(
self.templates,
output_folder + file_template.format(type='templates', **inputs))
write_data_to_JSON(
self.auxiliary_info, output_folder +
file_template.format(type='auxiliary_info', **inputs))
return output_folder
def get_unfolded_normalisation(TTJet_fit_results, category, channel):
global variable, met_type, path_to_JSON
h_truth, h_measured, h_response = get_unfold_histogram_tuple(
file_for_unfolding, variable, channel, met_type)
MADGRAPH_results = hist_to_value_error_tuplelist(h_truth)
POWHEG_results = hist_to_value_error_tuplelist(
get_unfold_histogram_tuple(file_for_powheg, variable, channel,
met_type)[0])
MCATNLO_results = hist_to_value_error_tuplelist(
get_unfold_histogram_tuple(file_for_mcatnlo, variable, channel,
met_type)[0])
matchingdown_results = hist_to_value_error_tuplelist(
get_unfold_histogram_tuple(file_for_matchingdown, variable, channel,
met_type)[0])
matchingup_results = hist_to_value_error_tuplelist(
get_unfold_histogram_tuple(file_for_matchingup, variable, channel,
met_type)[0])
scaledown_results = hist_to_value_error_tuplelist(
get_unfold_histogram_tuple(file_for_scaledown, variable, channel,
met_type)[0])
scaleup_results = hist_to_value_error_tuplelist(
get_unfold_histogram_tuple(file_for_scaleup, variable, channel,
met_type)[0])
TTJet_fit_results_unfolded = unfold_results(
TTJet_fit_results, category, channel, h_truth, h_measured, h_response,
'RooUnfoldSvd'
# 'TSVDUnfold'
)
normalisation_unfolded = {
'TTJet_measured': TTJet_fit_results,
'TTJet_unfolded': TTJet_fit_results_unfolded,
'MADGRAPH': MADGRAPH_results,
#other generators
'POWHEG': POWHEG_results,
'MCATNLO': MCATNLO_results,
#systematics
'matchingdown': matchingdown_results,
'matchingup': matchingup_results,
'scaledown': scaledown_results,
'scaleup': scaleup_results
}
write_data_to_JSON(
normalisation_unfolded, path_to_JSON + '/' + variable +
'/xsection_measurement_results' + '/kv' + str(unfoldCfg.SVD_k_value) +
'/' + category + '/normalisation_' + channel + '_' + met_type + '.txt')
return normalisation_unfolded
def save(self, output_path):
if not self.have_normalisation:
self.calculate_normalisation()
folder_template = '{path}/normalisation/{method}/{CoM}TeV/{variable}/'
folder_template += '{phase_space}/{category}/'
inputs = {
'path': output_path,
'CoM': self.config.centre_of_mass_energy,
'variable': self.variable,
'category': self.category,
'method': self.method_string(),
'phase_space': self.phase_space,
}
output_folder = folder_template.format(**inputs)
file_template = '{type}_{channel}_{met_type}.txt'
inputs = {
'channel': self.channel,
'met_type': self.met_type,
}
write_data_to_JSON(self.normalisation,
output_folder + file_template.format(type='normalisation', **inputs))
write_data_to_JSON(self.initial_normalisation,
output_folder + file_template.format(type='initial_normalisation', **inputs))
write_data_to_JSON(self.templates,
output_folder + file_template.format(type='templates', **inputs))
write_data_to_JSON(self.auxiliary_info,
output_folder + file_template.format(type='auxiliary_info', **inputs))
return output_folder
def write_normalised_xsection_measurement( measurement, measurement_unfolded, channel, summary = '' ):
global path_to_JSON, method
output_file = '{path_to_JSON}/central/normalised_xsection_{channel}_{method}_with_errors.txt'
output_file = output_file.format(
path_to_JSON = path_to_JSON,
channel = channel,
method = method,
)
if not summary == '':
output_file = output_file.replace( 'with_errors', summary + '_errors' )
output = {'TTJet_measured':measurement, 'TTJet_unfolded': measurement_unfolded}
write_data_to_JSON( output, output_file )
def write_fit_results(channel, category, fit_results):
global variable, met_type, output_path
folder_template = '%(path)s/%(fit_vars)s/%(CoM)dTeV/%(variable)s/%(category)s/'
inputs = {
'path': output_path,
'CoM': measurement_config.centre_of_mass_energy,
'variable': variable,
'category': category,
'fit_vars': '_'.join(fit_variables)
}
output_folder = folder_template % inputs
write_data_to_JSON(
fit_results,
output_folder + 'normalisation_' + channel + '_' + met_type + '.txt')
def calculate_xsections( normalisation, category, channel ):
global variable, met_type, path_to_JSON
# calculate the x-sections
branching_ratio = 0.15
if channel == 'combined':
branching_ratio = branching_ratio * 2
TTJet_xsection = calculate_xsection( normalisation['TTJet_measured'], luminosity, branching_ratio ) # L in pb1
TTJet_withoutFakes_xsection = calculate_xsection( normalisation['TTJet_measured_withoutFakes'], luminosity, branching_ratio ) # L in pb1
TTJet_xsection_unfolded = calculate_xsection( normalisation['TTJet_unfolded'], luminosity, branching_ratio ) # L in pb1
xsection_unfolded = {'TTJet_measured' : TTJet_xsection,
'TTJet_measured_withoutFakes' : TTJet_withoutFakes_xsection,
'TTJet_unfolded' : TTJet_xsection_unfolded,
}
if category == 'central':
powhegPythia8_xsection = calculate_xsection( normalisation['powhegPythia8'], luminosity, branching_ratio ) # L in pb1
amcatnlo_xsection = calculate_xsection( normalisation['amcatnlo'], luminosity, branching_ratio ) # L in pb1
powhegHerwig_xsection = calculate_xsection( normalisation['powhegHerwig'], luminosity, branching_ratio ) # L in pb1
amcatnloHerwig_xsection = calculate_xsection( normalisation['amcatnloHerwig'], luminosity, branching_ratio ) # L in pb1
madgraphMLM_xsection = calculate_xsection( normalisation['madgraphMLM'], luminosity, branching_ratio )
scaledown_xsection = calculate_xsection( normalisation['scaledown'], luminosity, branching_ratio ) # L in pb1
scaleup_xsection = calculate_xsection( normalisation['scaleup'], luminosity, branching_ratio ) # L in pb1
massdown_xsection = calculate_xsection( normalisation['massdown'], luminosity, branching_ratio ) # L in pb1
massup_xsection = calculate_xsection( normalisation['massup'], luminosity, branching_ratio ) # L in pb1
xsection_unfolded['powhegPythia8'] = powhegPythia8_xsection
xsection_unfolded['amcatnlo'] = amcatnlo_xsection
xsection_unfolded['madgraphMLM'] = madgraphMLM_xsection
xsection_unfolded['powhegHerwig'] = powhegHerwig_xsection
xsection_unfolded['amcatnloHerwig'] = amcatnloHerwig_xsection
xsection_unfolded['scaledown'] = scaledown_xsection
xsection_unfolded['scaleup'] = scaleup_xsection
xsection_unfolded['massdown'] = massdown_xsection
xsection_unfolded['massup'] = massup_xsection
file_template = '{path_to_JSON}/{category}/xsection_{channel}_{method}.txt'
filename = file_template.format(
path_to_JSON = path_to_JSON,
category = category,
channel = channel,
method = method,
)
write_data_to_JSON( xsection_unfolded, filename )
def calculate_normalised_xsections(normalisation,
category,
channel,
normalise_to_one=False):
global variable, met_type, path_to_JSON
TTJet_normalised_xsection = calculate_normalised_xsection(
normalisation['TTJet_measured'], bin_widths[variable],
normalise_to_one)
TTJet_normalised_xsection_unfolded = calculate_normalised_xsection(
normalisation['TTJet_unfolded'], bin_widths[variable],
normalise_to_one)
MADGRAPH_normalised_xsection = calculate_normalised_xsection(
normalisation['MADGRAPH'], bin_widths[variable], normalise_to_one)
POWHEG_normalised_xsection = calculate_normalised_xsection(
normalisation['POWHEG'], bin_widths[variable], normalise_to_one)
MCATNLO_normalised_xsection = calculate_normalised_xsection(
normalisation['MCATNLO'], bin_widths[variable], normalise_to_one)
matchingdown_normalised_xsection = calculate_normalised_xsection(
normalisation['matchingdown'], bin_widths[variable], normalise_to_one)
matchingup_normalised_xsection = calculate_normalised_xsection(
normalisation['matchingup'], bin_widths[variable], normalise_to_one)
scaledown_normalised_xsection = calculate_normalised_xsection(
normalisation['scaledown'], bin_widths[variable], normalise_to_one)
scaleup_normalised_xsection = calculate_normalised_xsection(
normalisation['scaleup'], bin_widths[variable], normalise_to_one)
normalised_xsection = {
'TTJet_measured': TTJet_normalised_xsection,
'TTJet_unfolded': TTJet_normalised_xsection_unfolded,
'MADGRAPH': MADGRAPH_normalised_xsection,
'POWHEG': POWHEG_normalised_xsection,
'MCATNLO': MCATNLO_normalised_xsection,
#systematics
'matchingdown': matchingdown_normalised_xsection,
'matchingup': matchingup_normalised_xsection,
'scaledown': scaledown_normalised_xsection,
'scaleup': scaleup_normalised_xsection
}
filename = path_to_JSON + '/' + variable + '/xsection_measurement_results' + '/kv' + str(
unfoldCfg.SVD_k_value
) + '/' + category + '/normalised_xsection_' + channel + '_' + met_type + '.txt'
if normalise_to_one:
filename = filename.replace('normalised_xsection',
'normalised_to_one_xsection')
write_data_to_JSON(normalised_xsection, filename)
def main():
global config, options
parser = OptionParser()
parser.add_option( "-p", "--path", dest = "path", default = 'data/fit_checks/no_merging',
help = "set path to JSON files" )
parser.add_option( '--create_fit_data', dest = "create_fit_data", action = "store_true",
help = "create the fit data for testing." )
parser.add_option( '--refit', dest = "refit", action = "store_true",
help = "Fit again even if the output already exists" )
parser.add_option( '--test', dest = "test", action = "store_true",
help = "Test only: run just one selected sample" )
variables = config.histogram_path_templates.keys()
fit_variables = fit_var_inputs
mc_samples = ['TTJet', 'SingleTop', 'QCD', 'V+Jets']
tests = closure_tests
channels = ['electron', 'muon']
COMEnergies = ['7', '8']
( options, _ ) = parser.parse_args()
print 'Running from path', options.path
if ( options.create_fit_data ):
create_fit_data( options.path, variables, fit_variables, mc_samples,
COMEnergies = COMEnergies, channels = channels )
output_file = options.path + '/fit_test_output.txt'
if options.test:
output_file = options.path + '/fit_test_output_test.txt'
if options.refit or not os.path.isfile( output_file ) or options.test:
if os.path.isfile( options.path + '/fit_check_data.txt' ):
fit_data = read_data_from_JSON( options.path + '/fit_check_data.txt' )
results = run_tests( fit_data,
COMEnergies,
variables,
fit_variables,
mc_samples,
channels,
tests )
write_data_to_JSON(results, output_file )
else:
print 'Please run bin/prepare_data_for_fit_checks first'
print 'Then run this script with the option --create_fit_data.'
results = read_data_from_JSON(output_file)
plot_results( results )
def calculate_xsections( normalisation, category, channel, k_value = None ):
global variable, met_type, path_to_JSON
# calculate the x-sections
branching_ratio = 0.15
if channel == 'combined':
branching_ratio = branching_ratio * 2
TTJet_xsection = calculate_xsection( normalisation['TTJet_measured'], luminosity, branching_ratio ) # L in pb1
TTJet_xsection_unfolded = calculate_xsection( normalisation['TTJet_unfolded'], luminosity, branching_ratio ) # L in pb1
MADGRAPH_xsection = calculate_xsection( normalisation['MADGRAPH'], luminosity, branching_ratio ) # L in pb1
MADGRAPH_ptreweight_xsection = calculate_xsection( normalisation['MADGRAPH_ptreweight'], luminosity, branching_ratio ) # L in pb1
POWHEG_PYTHIA_xsection = calculate_xsection( normalisation['POWHEG_PYTHIA'], luminosity, branching_ratio ) # L in pb1
POWHEG_HERWIG_xsection = calculate_xsection( normalisation['POWHEG_HERWIG'], luminosity, branching_ratio ) # L in pb1
MCATNLO_xsection = None
if centre_of_mass == 8:
MCATNLO_xsection = calculate_xsection( normalisation['MCATNLO'], luminosity, branching_ratio ) # L in pb1
matchingdown_xsection = calculate_xsection( normalisation['matchingdown'], luminosity, branching_ratio ) # L in pb1
matchingup_xsection = calculate_xsection( normalisation['matchingup'], luminosity, branching_ratio ) # L in pb1
scaledown_xsection = calculate_xsection( normalisation['scaledown'], luminosity, branching_ratio ) # L in pb1
scaleup_xsection = calculate_xsection( normalisation['scaleup'], luminosity, branching_ratio ) # L in pb1
xsection_unfolded = {'TTJet_measured' : TTJet_xsection,
'TTJet_unfolded' : TTJet_xsection_unfolded,
'MADGRAPH': MADGRAPH_xsection,
'MADGRAPH_ptreweight': MADGRAPH_ptreweight_xsection,
'POWHEG_PYTHIA': POWHEG_PYTHIA_xsection,
'POWHEG_HERWIG': POWHEG_HERWIG_xsection,
# systematics
'matchingdown': matchingdown_xsection,
'matchingup': matchingup_xsection,
'scaledown': scaledown_xsection,
'scaleup': scaleup_xsection
}
if centre_of_mass == 8:
xsection_unfolded['MCATNLO'] = MCATNLO_xsection
if k_value:
filename = path_to_JSON + '/xsection_measurement_results/%s/kv%d/%s/xsection_%s.txt' % ( channel, k_value, category, met_type )
elif not channel == 'combined':
raise ValueError( 'Invalid k-value for variable %s, channel %s, category %s.' % ( variable, channel, category ) )
else:
filename = path_to_JSON + '/xsection_measurement_results/%s/%s/xsection_%s.txt' % ( channel, category, met_type )
write_data_to_JSON( xsection_unfolded, filename )
def write_normalised_xsection_measurement(measurement,
measurement_unfolded,
channel,
summary=''):
global path_to_JSON, method
output_file = '{path_to_JSON}/central/normalised_xsection_{channel}_{method}_with_errors.txt'
output_file = output_file.format(
path_to_JSON=path_to_JSON,
channel=channel,
method=method,
)
if not summary == '':
output_file = output_file.replace('with_errors', summary + '_errors')
output = {
'TTJet_measured': measurement,
'TTJet_unfolded': measurement_unfolded
}
write_data_to_JSON(output, output_file)
def write_fit_results_and_initial_values(channel, category, fit_results, initial_values, templates):
global variable, met_type, output_path
output_folder = output_path + '/' + str(measurement_config.centre_of_mass) + 'TeV/' + variable + '/fit_results/' + category + '/'
write_data_to_JSON(fit_results, output_folder + 'fit_results_' + channel + '_' + met_type + '.txt')
write_data_to_JSON(initial_values, output_folder + 'initial_values_' + channel + '_' + met_type + '.txt')
write_data_to_JSON(templates, output_folder + 'templates_' + channel + '_' + met_type + '.txt')
def calculate_normalised_xsections( normalisation, category, channel, k_value = None, normalise_to_one = False ):
global variable, met_type, path_to_JSON
TTJet_normalised_xsection = calculate_normalised_xsection( normalisation['TTJet_measured'], bin_widths[variable], normalise_to_one )
TTJet_normalised_xsection_unfolded = calculate_normalised_xsection( normalisation['TTJet_unfolded'], bin_widths[variable], normalise_to_one )
MADGRAPH_normalised_xsection = calculate_normalised_xsection( normalisation['MADGRAPH'], bin_widths[variable], normalise_to_one )
MADGRAPH_ptreweight_normalised_xsection = calculate_normalised_xsection( normalisation['MADGRAPH_ptreweight'], bin_widths[variable], normalise_to_one )
POWHEG_PYTHIA_normalised_xsection = calculate_normalised_xsection( normalisation['POWHEG_PYTHIA'], bin_widths[variable], normalise_to_one )
POWHEG_HERWIG_normalised_xsection = calculate_normalised_xsection( normalisation['POWHEG_HERWIG'], bin_widths[variable], normalise_to_one )
MCATNLO_normalised_xsection = None
if centre_of_mass == 8:
MCATNLO_normalised_xsection = calculate_normalised_xsection( normalisation['MCATNLO'], bin_widths[variable], normalise_to_one )
matchingdown_normalised_xsection = calculate_normalised_xsection( normalisation['matchingdown'], bin_widths[variable], normalise_to_one )
matchingup_normalised_xsection = calculate_normalised_xsection( normalisation['matchingup'], bin_widths[variable], normalise_to_one )
scaledown_normalised_xsection = calculate_normalised_xsection( normalisation['scaledown'], bin_widths[variable], normalise_to_one )
scaleup_normalised_xsection = calculate_normalised_xsection( normalisation['scaleup'], bin_widths[variable], normalise_to_one )
normalised_xsection = {'TTJet_measured' : TTJet_normalised_xsection,
'TTJet_unfolded' : TTJet_normalised_xsection_unfolded,
'MADGRAPH': MADGRAPH_normalised_xsection,
'MADGRAPH_ptreweight': MADGRAPH_ptreweight_normalised_xsection,
'POWHEG_PYTHIA': POWHEG_PYTHIA_normalised_xsection,
'POWHEG_HERWIG': POWHEG_HERWIG_normalised_xsection,
# systematics
'matchingdown': matchingdown_normalised_xsection,
'matchingup': matchingup_normalised_xsection,
'scaledown': scaledown_normalised_xsection,
'scaleup': scaleup_normalised_xsection
}
if centre_of_mass == 8:
normalised_xsection['MCATNLO'] = MCATNLO_normalised_xsection
if not channel == 'combined':
filename = path_to_JSON + '/xsection_measurement_results/%s/kv%d/%s/normalised_xsection_%s.txt' % ( channel, k_value, category, met_type )
else:
filename = path_to_JSON + '/xsection_measurement_results/%s/%s/normalised_xsection_%s.txt' % ( channel, category, met_type )
if normalise_to_one:
filename = filename.replace( 'normalised_xsection', 'normalised_to_one_xsection' )
write_data_to_JSON( normalised_xsection, filename )
def calculate_xsections(normalisation, category, channel):
global variable, met_type, path_to_JSON
# calculate the x-sections
TTJet_xsection = calculate_xsection(normalisation['TTJet_measured'],
luminosity, 0.15) # L in pb1
TTJet_xsection_unfolded = calculate_xsection(
normalisation['TTJet_unfolded'], luminosity, 0.15) # L in pb1
MADGRAPH_xsection = calculate_xsection(normalisation['MADGRAPH'],
luminosity, 0.15) # L in pb1
POWHEG_xsection = calculate_xsection(normalisation['POWHEG'], luminosity,
0.15) # L in pb1
MCATNLO_xsection = calculate_xsection(normalisation['MCATNLO'], luminosity,
0.15) # L in pb1
matchingdown_xsection = calculate_xsection(normalisation['matchingdown'],
luminosity, 0.15) # L in pb1
matchingup_xsection = calculate_xsection(normalisation['matchingup'],
luminosity, 0.15) # L in pb1
scaledown_xsection = calculate_xsection(normalisation['scaledown'],
luminosity, 0.15) # L in pb1
scaleup_xsection = calculate_xsection(normalisation['scaleup'], luminosity,
0.15) # L in pb1
xsection_unfolded = {
'TTJet_measured': TTJet_xsection,
'TTJet_unfolded': TTJet_xsection_unfolded,
'MADGRAPH': MADGRAPH_xsection,
'POWHEG': POWHEG_xsection,
'MCATNLO': MCATNLO_xsection,
#systematics
'matchingdown': matchingdown_xsection,
'matchingup': matchingup_xsection,
'scaledown': scaledown_xsection,
'scaleup': scaleup_xsection
}
write_data_to_JSON(
xsection_unfolded, path_to_JSON + '/' + variable +
'/xsection_measurement_results' + '/kv' + str(unfoldCfg.SVD_k_value) +
'/' + category + '/xsection_' + channel + '_' + met_type + '.txt')
def get_unfolded_normalisation(TTJet_fit_results, category, channel):
global variable, met_type, path_to_JSON
h_truth, h_measured, h_response = get_unfold_histogram_tuple(file_for_unfolding, variable, channel, met_type)
MADGRAPH_results = hist_to_value_error_tuplelist(h_truth)
POWHEG_results = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_powheg, variable, channel, met_type)[0])
MCATNLO_results = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_mcatnlo, variable, channel, met_type)[0])
matchingdown_results = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_matchingdown, variable, channel, met_type)[0])
matchingup_results = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_matchingup, variable, channel, met_type)[0])
scaledown_results = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_scaledown, variable, channel, met_type)[0])
scaleup_results = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_scaleup, variable, channel, met_type)[0])
TTJet_fit_results_unfolded = unfold_results(TTJet_fit_results,
category,
channel,
h_truth,
h_measured,
h_response,
'RooUnfoldSvd'
# 'TSVDUnfold'
)
normalisation_unfolded = {
'TTJet_measured' : TTJet_fit_results,
'TTJet_unfolded' : TTJet_fit_results_unfolded,
'MADGRAPH': MADGRAPH_results,
#other generators
'POWHEG': POWHEG_results,
'MCATNLO': MCATNLO_results,
#systematics
'matchingdown': matchingdown_results,
'matchingup': matchingup_results,
'scaledown': scaledown_results,
'scaleup': scaleup_results
}
write_data_to_JSON(normalisation_unfolded, path_to_JSON + '/' + variable + '/xsection_measurement_results' + '/kv' + str(unfoldCfg.SVD_k_value) + '/' + category + '/normalisation_' + channel + '_' + met_type + '.txt')
return normalisation_unfolded
luminosity = 5050 # L in pb1
branching_ratio = 0.15 # to be corrected for
normalisation_electron_unfolded = read_data_from_JSON(csm.normalisation_electron_unfolded_file)
normalisation_muon_unfolded = read_data_from_JSON(csm.normalisation_muon_unfolded_file)
normalisation_combined_unfolded = read_data_from_JSON(csm.normalisation_combined_unfolded_file)
xsection_electron_unfolded = get_cross_sections(normalisation_electron_unfolded, luminosity, branching_ratio)
xsection_muon_unfolded = get_cross_sections(normalisation_muon_unfolded, luminosity, branching_ratio)
xsection_combined_unfolded = get_cross_sections(normalisation_combined_unfolded, luminosity, branching_ratio*2)
normalised_to_one_xsection_electron = get_normalised_xsection(normalisation_electron_unfolded, True)
normalised_to_one_xsection_muon = get_normalised_xsection(normalisation_muon_unfolded, True)
normalised_to_one_xsection_combined = get_normalised_xsection(normalisation_combined_unfolded, True)
normalised_xsection_electron = get_normalised_xsection(normalisation_electron_unfolded, False)
normalised_xsection_muon = get_normalised_xsection(normalisation_muon_unfolded, False)
normalised_xsection_combined = get_normalised_xsection(normalisation_combined_unfolded, False)
#write results
write_data_to_JSON(xsection_electron_unfolded, csm.xsection_electron_unfolded_file)
write_data_to_JSON(xsection_muon_unfolded, csm.xsection_muon_unfolded_file)
write_data_to_JSON(xsection_combined_unfolded, csm.xsection_combined_unfolded_file)
write_data_to_JSON(normalised_to_one_xsection_electron, csm.normalised_to_one_xsection_electron_unfolded_file)
write_data_to_JSON(normalised_to_one_xsection_muon, csm.normalised_to_one_xsection_muon_unfolded_file)
write_data_to_JSON(normalised_to_one_xsection_combined, csm.normalised_to_one_xsection_combined_unfolded_file)
write_data_to_JSON(normalised_xsection_electron, csm.normalised_xsection_electron_unfolded_file)
write_data_to_JSON(normalised_xsection_muon, csm.normalised_xsection_muon_unfolded_file)
write_data_to_JSON(normalised_xsection_combined, csm.normalised_xsection_combined_unfolded_file)
# ad-hoc switch for PFMET -> patMETsPFlow
if met_type == 'PFMET':
met_type = 'patMETsPFlow'
filename = ''
# get unfolded normalisation
unfolded_normalisation_electron = get_unfolded_normalisation( TTJet_fit_results_electron, category, 'electron', k_value_electron )
unfolded_normalisation_muon = get_unfolded_normalisation( TTJet_fit_results_muon, category, 'muon', k_value_muon )
if combine_before_unfolding:
unfolded_normalisation_combined = get_unfolded_normalisation( TTJet_fit_results_combined, category, 'combined', k_value_combined )
else:
unfolded_normalisation_combined = combine_complex_results( unfolded_normalisation_electron, unfolded_normalisation_muon )
filename = path_to_JSON + '/xsection_measurement_results/electron/kv%d/%s/normalisation_%s.txt' % ( k_value_electron, category, met_type )
write_data_to_JSON( unfolded_normalisation_electron, filename )
filename = path_to_JSON + '/xsection_measurement_results/muon/kv%d/%s/normalisation_%s.txt' % ( k_value_muon, category, met_type )
write_data_to_JSON( unfolded_normalisation_muon, filename )
filename = path_to_JSON + '/xsection_measurement_results/combined/%s/normalisation_%s.txt' % ( category, met_type )
write_data_to_JSON( unfolded_normalisation_combined, filename )
if measurement_config.include_higgs:
# now the same for the Higgs
Higgs_fit_results_electron = fit_results_electron['Higgs']
Higgs_fit_results_muon = fit_results_muon['Higgs']
Higgs_fit_results_combined = fit_results_combined['Higgs']
unfolded_normalisation_electron_higgs = get_unfolded_normalisation( fit_results_electron['Higgs'], category, 'electron', k_value_electron )
unfolded_normalisation_muon_higgs = get_unfolded_normalisation( fit_results_muon['Higgs'], category, 'muon', k_value_muon )
if combine_before_unfolding:
unfolded_normalisation_combined_higgs = get_unfolded_normalisation( fit_results_combined['Higgs'], category, 'combined', k_value_combined )
else:
def main():
'''
Step 1: Get the 2D histogram for every sample (channel and/or centre of mass energy)
Step 2: Change the size of the first bin until it fulfils the minimal criteria
Step 3: Check if it is true for all other histograms. If not back to step 2
Step 4: Repeat step 2 & 3 until no mo bins can be created
'''
parser = OptionParser()
parser.add_option( '-v', dest = "visiblePhaseSpace", action = "store_true",
help = "Consider visible phase space or not" )
parser.add_option( '-c', dest = "combined", action = "store_true",
help = "Combine channels" )
( options, _ ) = parser.parse_args()
p_min = 0.6
s_min = 0.6
# we also want the statistical error to be larger than 5%
# this translates (error -= 1/sqrt(N)) to (1/0.05)^2 = 400
n_min = 200
n_min_lepton = 500
# n_min = 200 # N = 200 -> 7.1 % stat error
bin_choices = {}
variables = bin_edges.keys()
for variable in variables:
print('--- Doing variable',variable)
variableToUse = variable
if 'Rap' in variable:
variableToUse = 'abs_%s' % variable
histogram_information = get_histograms( variableToUse, options )
if variable == 'HT':
best_binning, histogram_information = get_best_binning( histogram_information , p_min, s_min, n_min, minimum_bin_width[variable], x_min=100. )
elif variable == 'ST':
best_binning, histogram_information = get_best_binning( histogram_information , p_min, s_min, n_min, minimum_bin_width[variable], x_min=123. )
elif variable == 'NJets':
best_binning, histogram_information = get_best_binning( histogram_information , p_min, s_min, n_min, minimum_bin_width[variable], x_min=3.5 )
elif variable == 'lepton_pt':
best_binning, histogram_information = get_best_binning( histogram_information , p_min, s_min, n_min_lepton, minimum_bin_width[variable], x_min=23. )
elif variable == 'abs_lepton_eta':
best_binning, histogram_information = get_best_binning( histogram_information , p_min, s_min, n_min_lepton, minimum_bin_width[variable] )
else:
best_binning, histogram_information = get_best_binning( histogram_information , p_min, s_min, n_min, minimum_bin_width[variable], )
if 'Rap' in variable:
for b in list(best_binning):
if b != 0.0:
best_binning.append(-1.0*b)
best_binning.sort()
# Make last bin smaller if huge
# Won't change final results
if len(best_binning) >= 4:
lastBinWidth = best_binning[-1] - best_binning[-2]
penultimateBinWidth = best_binning[-2] - best_binning[-3]
if lastBinWidth / penultimateBinWidth > 5:
newLastBinWidth = penultimateBinWidth * 5
best_binning[-1] = best_binning[-2] + newLastBinWidth
if variable == 'abs_lepton_eta':
best_binning = [ round(i,2) for i in best_binning ]
elif variable != 'NJets' :
best_binning = [ round(i) for i in best_binning ]
bin_choices[variable] = best_binning
print('The best binning for', variable, 'is:')
print('bin edges =', best_binning)
print('N_bins =', len( best_binning ) - 1)
print('The corresponding purities and stabilities are:')
for info in histogram_information:
# print_latex_table(info, variable, best_binning)
outputInfo = {}
outputInfo['p_i'] = info['p_i']
outputInfo['s_i'] = info['s_i']
outputInfo['N'] = info['N']
outputJsonFile = 'unfolding/13TeV/binningInfo_%s_%s_FullPS.txt' % ( variable, info['channel'] )
if options.visiblePhaseSpace:
outputJsonFile = 'unfolding/13TeV/binningInfo_%s_%s_VisiblePS.txt' % ( variable, info['channel'] )
write_data_to_JSON( outputInfo, outputJsonFile )
for key in outputInfo:
print (key,outputInfo[key])
print('-' * 120)
print('=' * 120)
print('For config/variable_binning.py')
print('=' * 120)
for variable in bin_choices:
print('\''+variable+'\' : '+str(bin_choices[variable])+',')
unfolded_normalisation_muon = {}
# Electron channel
unfolded_normalisation_electron = get_unfolded_normalisation(
TTJet_fit_results_electron,
category,
'electron',
tau_value_electron,
visiblePS=visiblePS)
filename = file_template.format(
path_to_JSON=path_to_JSON,
category=category,
channel='electron',
method=method,
)
write_data_to_JSON(unfolded_normalisation_electron, filename)
# measure xsection
calculate_xsections(unfolded_normalisation_electron, category,
'electron')
calculate_normalised_xsections(unfolded_normalisation_electron,
category, 'electron')
calculate_normalised_xsections(unfolded_normalisation_electron,
category, 'electron', True)
# Muon channel
unfolded_normalisation_muon = get_unfolded_normalisation(
TTJet_fit_results_muon,
category,
'muon',
tau_value_muon,
visiblePS=visiblePS)
def create_fit_data(path, variables, fit_variables, mc_samples, COMEnergies,
channels):
'''
Creates the fit data in path + fit_check_data.txt in the JSON format.
The resulting dictionary is of the form
{centre-of-mass-energy: {
channel : {
variable : {
variable_bin : {
fit_variable : {
templates : { sample : []},
initial_values : { sample : []}
}
}
}
}
}
}
'''
# first step is to read the data
raw_data = {}
# the folder structure is
# path/fit_variable/centre-of-mass-energy/variable/fit_results/central/*
for fit_variable in fit_variables:
raw_data[fit_variable] = {}
for COMEnergy in COMEnergies:
raw_data[fit_variable][COMEnergy] = {}
for variable in variables:
raw_data[fit_variable][COMEnergy][variable] = {}
for channel in channels:
raw_data[fit_variable][COMEnergy][variable][channel] = {}
data_path = path + '/' + fit_variable + '/' + str(
COMEnergy) + 'TeV'
templates = read_fit_templates(data_path,
variable,
channel=channel)
initial_values = read_initial_normalisation(
data_path, variable, channel=channel)
raw_data[fit_variable][COMEnergy][variable][channel][
'templates'] = templates
raw_data[fit_variable][COMEnergy][variable][channel][
'initial_values'] = initial_values
# put it into the new structure
fit_data = {}
for COMEnergy in COMEnergies:
fit_data[COMEnergy] = {}
for channel in channels:
fit_data[COMEnergy][channel] = {}
for variable in variables:
fit_data[COMEnergy][channel][variable] = {}
for v_bin in variable_bins_ROOT[variable]:
fit_data[COMEnergy][channel][variable][v_bin] = {}
for fit_variable in fit_variables:
fit_data[COMEnergy][channel][variable][v_bin][
fit_variable] = {}
fit_data[COMEnergy][channel][variable][v_bin][
fit_variable]['templates'] = raw_data[
fit_variable][COMEnergy][variable][channel][
'templates']
fit_data[COMEnergy][channel][variable][v_bin][
fit_variable]['initial_values'] = raw_data[
fit_variable][COMEnergy][variable][channel][
'initial_values']
write_data_to_JSON(fit_data, path + '/fit_check_data.txt', indent=False)
def main():
"""
Step 1: Get the 2D histogram for every sample (channel and/or centre of mass energy)
Step 2: Change the size of the first bin until it fulfils the minimal criteria
Step 3: Check if it is true for all other histograms. If not back to step 2
Step 4: Repeat step 2 & 3 until no mo bins can be created
"""
parser = OptionParser()
parser.add_option("-v", dest="visiblePhaseSpace", action="store_true", help="Consider visible phase space or not")
parser.add_option("-c", dest="combined", action="store_true", help="Combine channels")
(options, _) = parser.parse_args()
measurement_config = XSectionConfig(13)
p_min = 0.6 # 0.5 for MET
s_min = 0.6
# we also want the statistical error to be larger than 5%
# this translates (error -= 1/sqrt(N)) to (1/0.05)^2 = 400
n_min = 500
n_min_lepton = 500
# n_min = 200 # N = 200 -> 7.1 % stat error
bin_choices = {}
# variables = bin_edges_full.keys()
variables = measurement_config.variables
for variable in variables:
global var
var = variable
print("--- Doing variable", variable)
variableToUse = variable
if "Rap" in variable:
variableToUse = "abs_%s" % variable
histogram_information = get_histograms(variableToUse, options)
if variable == "HT":
best_binning, histogram_information = get_best_binning(
histogram_information, p_min, s_min, n_min, minimum_bin_width[variable], x_min=100.0
)
elif variable == "ST":
best_binning, histogram_information = get_best_binning(
histogram_information, p_min, s_min, n_min, minimum_bin_width[variable], x_min=123.0
)
elif variable == "MET":
best_binning, histogram_information = get_best_binning(
histogram_information, 0.5, 0.5, n_min, minimum_bin_width[variable]
)
elif variable == "NJets":
best_binning, histogram_information = get_best_binning(
histogram_information, p_min, s_min, n_min, minimum_bin_width[variable], x_min=3.5
)
elif variable == "lepton_pt":
best_binning, histogram_information = get_best_binning(
histogram_information, p_min, s_min, n_min_lepton, minimum_bin_width[variable], x_min=23.0
)
elif variable == "abs_lepton_eta":
best_binning, histogram_information = get_best_binning(
histogram_information, p_min, s_min, n_min_lepton, minimum_bin_width[variable]
)
else:
best_binning, histogram_information = get_best_binning(
histogram_information, p_min, s_min, n_min, minimum_bin_width[variable]
)
if "Rap" in variable:
for b in list(best_binning):
if b != 0.0:
best_binning.append(-1.0 * b)
best_binning.sort()
# Make last bin smaller if huge
# Won't change final results
if len(best_binning) >= 4:
lastBinWidth = best_binning[-1] - best_binning[-2]
penultimateBinWidth = best_binning[-2] - best_binning[-3]
if lastBinWidth / penultimateBinWidth > 5:
newLastBinWidth = penultimateBinWidth * 5
best_binning[-1] = best_binning[-2] + newLastBinWidth
if variable == "abs_lepton_eta":
best_binning = [round(i, 2) for i in best_binning]
elif variable != "NJets":
best_binning = [round(i) for i in best_binning]
bin_choices[variable] = best_binning
print("The best binning for", variable, "is:")
print("bin edges =", best_binning)
print("N_bins =", len(best_binning) - 1)
print("The corresponding purities and stabilities are:")
for info in histogram_information:
# print_latex_table(info, variable, best_binning)
outputInfo = {}
outputInfo["p_i"] = info["p_i"]
outputInfo["s_i"] = info["s_i"]
outputInfo["N"] = info["N"]
outputJsonFile = "unfolding/13TeV/binningInfo_%s_%s_FullPS.txt" % (variable, info["channel"])
if options.visiblePhaseSpace:
outputJsonFile = "unfolding/13TeV/binningInfo_%s_%s_VisiblePS.txt" % (variable, info["channel"])
write_data_to_JSON(outputInfo, outputJsonFile)
for key in outputInfo:
print(key, outputInfo[key])
print("-" * 120)
print("=" * 120)
print("For config/variable_binning.py")
print("=" * 120)
for variable in bin_choices:
print("'" + variable + "' : " + str(bin_choices[variable]) + ",")
def save_pulls( pulls, test, method, channel ):
global use_N_toy, offset_toy_mc, offset_toy_data
file_template = 'Pulls_%s_%s_%s_toy_MC_%d_to_%d_MC_%d_to_%d_data.txt'
output_file = output_folder + file_template % ( test, method, channel, offset_toy_mc + 1, use_N_toy + offset_toy_mc, offset_toy_data + 1, use_N_toy + offset_toy_data )
write_data_to_JSON( pulls, output_file )
print 'Pulls saved in file: ', output_file
def write_fit_results_and_initial_values(fit_results_electron, fit_results_muon, initial_values_electron, initial_values_muon):
write_data_to_JSON(fit_results_electron, 'data/fit_results_electron.txt')
write_data_to_JSON(fit_results_muon, 'data/fit_results_muon.txt')
write_data_to_JSON(initial_values_electron, 'data/initial_values_electron.txt')
write_data_to_JSON(initial_values_muon, 'data/initial_values_muon.txt')
def calculate_xsections(normalisation, category, channel):
global variable, met_type, path_to_JSON
# calculate the x-sections
branching_ratio = 0.15
if channel == 'combined':
branching_ratio = branching_ratio * 2
TTJet_xsection = calculate_xsection(normalisation['TTJet_measured'],
luminosity,
branching_ratio) # L in pb1
TTJet_withoutFakes_xsection = calculate_xsection(
normalisation['TTJet_measured_withoutFakes'], luminosity,
branching_ratio) # L in pb1
TTJet_xsection_unfolded = calculate_xsection(
normalisation['TTJet_unfolded'], luminosity,
branching_ratio) # L in pb1
xsection_unfolded = {
'TTJet_measured': TTJet_xsection,
'TTJet_measured_withoutFakes': TTJet_withoutFakes_xsection,
'TTJet_unfolded': TTJet_xsection_unfolded,
}
if category == 'central':
powhegPythia8_xsection = calculate_xsection(
normalisation['powhegPythia8'], luminosity,
branching_ratio) # L in pb1
amcatnlo_xsection = calculate_xsection(normalisation['amcatnlo'],
luminosity,
branching_ratio) # L in pb1
amcatnlo_HERWIG_xsection = calculate_xsection(
normalisation['amcatnlo_HERWIG'], luminosity,
branching_ratio) # L in pb1
scaledown_xsection = calculate_xsection(normalisation['scaledown'],
luminosity,
branching_ratio) # L in pb1
scaleup_xsection = calculate_xsection(normalisation['scaleup'],
luminosity,
branching_ratio) # L in pb1
massdown_xsection = calculate_xsection(normalisation['massdown'],
luminosity,
branching_ratio) # L in pb1
massup_xsection = calculate_xsection(normalisation['massup'],
luminosity,
branching_ratio) # L in pb1
madgraphMLM_xsection = calculate_xsection(normalisation['madgraphMLM'],
luminosity, branching_ratio)
xsection_unfolded['powhegPythia8'] = powhegPythia8_xsection
xsection_unfolded['amcatnlo'] = amcatnlo_xsection
xsection_unfolded['madgraphMLM'] = madgraphMLM_xsection
xsection_unfolded['amcatnlo_HERWIG'] = amcatnlo_HERWIG_xsection
xsection_unfolded['scaledown'] = scaledown_xsection
xsection_unfolded['scaleup'] = scaleup_xsection
xsection_unfolded['massdown'] = massdown_xsection
xsection_unfolded['massup'] = massup_xsection
file_template = '{path_to_JSON}/{category}/xsection_{channel}_{method}.txt'
filename = file_template.format(
path_to_JSON=path_to_JSON,
category=category,
channel=channel,
method=method,
)
write_data_to_JSON(xsection_unfolded, filename)
def save_pulls(pulls, test, method, channel):
global use_N_toy, skip_N_toy
file_template = "Pulls_%s_%s_%s_toy_MC_%d_to_%d.txt"
output_file = output_folder + file_template % (test, method, channel, skip_N_toy + 1, use_N_toy + skip_N_toy)
write_data_to_JSON(pulls, output_file)
def toJSON(self, JSON_file):
output = self.toDict()
filename = JSON_file.split('/')[-1]
directory = JSON_file.replace(filename, '')
fu.make_folder_if_not_exists(directory)
fu.write_data_to_JSON(output, JSON_file)
def calculate_normalised_xsections(normalisation,
category,
channel,
normalise_to_one=False):
global variable, met_type, path_to_JSON, phase_space
binWidths = None
if phase_space == 'VisiblePS':
binWidths = bin_widths_visiblePS
elif phase_space == 'FullPS':
binWidths = bin_widths
TTJet_normalised_xsection = calculate_normalised_xsection(
normalisation['TTJet_measured'], binWidths[variable], normalise_to_one)
TTJet_withoutFakes_normalised_xsection = calculate_normalised_xsection(
normalisation['TTJet_measured_withoutFakes'], binWidths[variable],
normalise_to_one)
TTJet_normalised_xsection_unfolded = calculate_normalised_xsection(
normalisation['TTJet_unfolded'], binWidths[variable], normalise_to_one)
normalised_xsection = {
'TTJet_measured': TTJet_normalised_xsection,
'TTJet_measured_withoutFakes': TTJet_withoutFakes_normalised_xsection,
'TTJet_unfolded': TTJet_normalised_xsection_unfolded
}
if category == 'central':
powhegPythia8_normalised_xsection = calculate_normalised_xsection(
normalisation['powhegPythia8'], binWidths[variable],
normalise_to_one)
amcatnlo_normalised_xsection = calculate_normalised_xsection(
normalisation['amcatnlo'], binWidths[variable], normalise_to_one)
amcatnlo_HERWIG_normalised_xsection = calculate_normalised_xsection(
normalisation['amcatnlo_HERWIG'], binWidths[variable],
normalise_to_one)
scaledown_normalised_xsection = calculate_normalised_xsection(
normalisation['scaledown'], binWidths[variable], normalise_to_one)
scaleup_normalised_xsection = calculate_normalised_xsection(
normalisation['scaleup'], binWidths[variable], normalise_to_one)
massdown_normalised_xsection = calculate_normalised_xsection(
normalisation['massdown'], binWidths[variable], normalise_to_one)
massup_normalised_xsection = calculate_normalised_xsection(
normalisation['massup'], binWidths[variable], normalise_to_one)
madgraphMLM_normalised_xsection = calculate_normalised_xsection(
normalisation['madgraphMLM'], binWidths[variable],
normalise_to_one)
normalised_xsection[
'powhegPythia8'] = powhegPythia8_normalised_xsection
normalised_xsection['amcatnlo'] = amcatnlo_normalised_xsection
normalised_xsection['madgraphMLM'] = madgraphMLM_normalised_xsection
normalised_xsection[
'amcatnlo_HERWIG'] = amcatnlo_HERWIG_normalised_xsection
normalised_xsection['scaledown'] = scaledown_normalised_xsection
normalised_xsection['scaleup'] = scaleup_normalised_xsection
normalised_xsection['massdown'] = massdown_normalised_xsection
normalised_xsection['massup'] = massup_normalised_xsection
file_template = '{path_to_JSON}/{category}/normalised_xsection_{channel}_{method}.txt'
filename = file_template.format(
path_to_JSON=path_to_JSON,
category=category,
channel=channel,
method=method,
)
if normalise_to_one:
filename = filename.replace('normalised_xsection',
'normalised_to_one_xsection')
write_data_to_JSON(normalised_xsection, filename)
POWHEG_results_electron = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_powheg, 'electron')[1])
PYTHIA_results_electron = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_pythia, 'electron')[1])
MCATNLO_results_electron = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_mcatnlo, 'electron')[1])
matchingdown_results_electron = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_matchingdown, 'electron')[1])
matchingup_results_electron = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_matchingup, 'electron')[1])
scaledown_results_electron = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_scaledown, 'electron')[1])
scaleup_results_electron = hist_to_value_error_tuplelist(get_unfold_histogram_tuple(file_for_scaleup, 'electron')[1])
TTJet_fit_results_electron_unfolded = unfold_results(TTJet_fit_results_electron,
h_truth,
h_measured,
h_response,
'RooUnfoldSvd')
write_data_to_JSON(TTJet_fit_results_electron, 'data/TTJet_fit_results_electron.txt')
normalisation_electron_unfolded = {
'TTJet' : TTJet_fit_results_electron_unfolded,
'MADGRAPH': MADGRAPH_results_electron,
#other generators
'POWHEG': POWHEG_results_electron,
'PYTHIA': PYTHIA_results_electron,
'MCATNLO': MCATNLO_results_electron,
#systematics
'matchingdown': matchingdown_results_electron,
'matchingup': matchingup_results_electron,
'scaledown': scaledown_results_electron,
'scaleup': scaleup_results_electron
}
write_data_to_JSON(normalisation_electron_unfolded, 'data/normalisation_electron_unfolded.txt')
def write_fit_results_and_initial_values(channel, category, fit_results, initial_values, templates):
global variable, met_type
write_data_to_JSON(fit_results, 'data/' + variable + '/fit_results/' + category + '/fit_results_' + channel + '_' + met_type + '.txt')
write_data_to_JSON(initial_values, 'data/' + variable + '/fit_results/' + category + '/initial_values_' + channel + '_' + met_type + '.txt')
write_data_to_JSON(templates, 'data/' + variable + '/fit_results/' + category + '/templates_' + channel + '_' + met_type + '.txt')
muon_file = path_to_JSON + '/fit_results/' + ttbar_theory_systematic_prefix + 'mcatnlo' + '/fit_results_muon_' + met_type + '.txt'
TTJet_fit_results_electron = read_data_from_JSON(electron_file)['TTJet']
TTJet_fit_results_muon = read_data_from_JSON(muon_file)['TTJet']
# change back to original MET type for the unfolding
met_type = translate_options[options.metType]
# ad-hoc switch for PFMET -> patMETsPFlow
if met_type == 'PFMET':
met_type = 'patMETsPFlow'
# get unfolded normalisation
unfolded_normalisation_electron = get_unfolded_normalisation(TTJet_fit_results_electron, category, 'electron')
unfolded_normalisation_muon = get_unfolded_normalisation(TTJet_fit_results_muon, category, 'muon')
unfolded_normalisation_combined = combine_complex_results(unfolded_normalisation_electron, unfolded_normalisation_muon)
write_data_to_JSON(unfolded_normalisation_combined,
path_to_JSON + '/xsection_measurement_results' + '/kv' + str(unfoldCfg.SVD_k_value) + '/' + category + '/normalisation_combined_' + met_type + '.txt')
# measure xsection
calculate_xsections(unfolded_normalisation_electron, category, 'electron')
calculate_xsections(unfolded_normalisation_muon, category, 'muon')
calculate_xsections(unfolded_normalisation_combined, category, 'combined')
calculate_normalised_xsections(unfolded_normalisation_electron, category, 'electron')
calculate_normalised_xsections(unfolded_normalisation_muon, category, 'muon')
calculate_normalised_xsections(unfolded_normalisation_combined, category, 'combined')
normalise_to_one = True
calculate_normalised_xsections(unfolded_normalisation_electron, category, 'electron', normalise_to_one)
calculate_normalised_xsections(unfolded_normalisation_muon, category, 'muon', normalise_to_one)
calculate_normalised_xsections(unfolded_normalisation_combined, category, 'combined', normalise_to_one)
file_template = '{path_to_JSON}/{category}/unfolded_normalisation_{channel}_{method}.txt'
filename = ''
# # get unfolded normalisation
unfolded_normalisation_electron = {}
unfolded_normalisation_muon = {}
# Electron channel
unfolded_normalisation_electron = get_unfolded_normalisation( TTJet_fit_results_electron, category, 'electron', tau_value_electron, visiblePS = visiblePS )
filename = file_template.format(
path_to_JSON = path_to_JSON,
category = category,
channel = 'electron',
method = method,
)
write_data_to_JSON( unfolded_normalisation_electron, filename )
# measure xsection
calculate_xsections( unfolded_normalisation_electron, category, 'electron' )
calculate_normalised_xsections( unfolded_normalisation_electron, category, 'electron' )
calculate_normalised_xsections( unfolded_normalisation_electron, category, 'electron' , True )
# Muon channel
unfolded_normalisation_muon = get_unfolded_normalisation( TTJet_fit_results_muon, category, 'muon', tau_value_muon, visiblePS = visiblePS )
filename = file_template.format(
path_to_JSON = path_to_JSON,
category = category,
channel = 'muon',
method = method,
)
write_data_to_JSON( unfolded_normalisation_muon, filename )
# measure xsection
def save_pulls(pulls, test, method, channel):
global use_N_toy, skip_N_toy
file_template = 'Pulls_%s_%s_%s_toy_MC_%d_to_%d.txt'
output_file = output_folder + file_template % (
test, method, channel, skip_N_toy + 1, use_N_toy + skip_N_toy)
write_data_to_JSON(pulls, output_file)
