def produce_pickle_file(hist_passed_data, hist_total_data, file_name):
output_pickle = open(file_name, 'wb')
dictionary = {}
number_of_pt_bin_edges = len(jet_pt_bins)
data_efficiency = Efficiency(hist_passed_data, hist_total_data)
for i in range(number_of_pt_bin_edges - 1):
lower_edge_pt = jet_pt_bins[i]
upper_edge_pt = jet_pt_bins[i + 1]
pt_bin_range = 'pt_' + str(lower_edge_pt) + '_' + str(upper_edge_pt)
dictionary[pt_bin_range] = {}
data_efficiency_in_bin = data_efficiency.GetEfficiency(i + 1)
data_efficiency_in_bin_error_up = data_efficiency.GetEfficiencyErrorUp(
i + 1)
data_efficiency_in_bin_error_down = data_efficiency.GetEfficiencyErrorLow(
i + 1)
dictionary[pt_bin_range]['data'] = {
'efficiency': data_efficiency_in_bin,
'err_up': data_efficiency_in_bin_error_up,
'err_down': data_efficiency_in_bin_error_down,
}
pickle.dump(dictionary, output_pickle)
def efficiencies_plot(ana, category, discri_var, plot_level, working_points):
vars = {
'pt': VARIABLES['pt'],
'eta': VARIABLES['eta'],
'good_npv': VARIABLES['good_npv'],
'averageintpercrossing': VARIABLES['averageintpercrossing'],
}
canvases = {}
efficiencies = {}
for v in vars.keys():
efficiencies[v] = []
for wp in working_points:
cut = wp.cut if isinstance(wp.cut, str) else '{0} >= {1}'.format(
discri_var, wp.cut)
hist_samples = ana.get_hist_samples_array(vars,
plot_level,
category=category)
hist_samples_cut = ana.get_hist_samples_array(vars,
plot_level,
category=category,
cuts=cut)
for v in vars.keys():
efficiencies[v].append(
Efficiency(hist_samples_cut[v]['tau'],
hist_samples[v]['tau'],
title=wp.name))
for v, effs in efficiencies.items():
canvases[v] = draw_efficiencies(effs, plot_level + '_' + v, category)
return canvases
def get_efficiency(parentDirectory,
denomHistName,
numerHistName,
binsToMerge=None,
numerAdd=None):
sampleNames = [k.name for k in parentDirectory.keys()]
if len(sampleNames) == 0:
raise ValueError('No sample directory found!')
hDenom = getattr(getattr(parentDirectory, sampleNames[0]),
denomHistName).clone()
hNumer = getattr(getattr(parentDirectory, sampleNames[0]),
numerHistName).clone()
if numerAdd:
hNumer.Add(
getattr(getattr(parentDirectory, sampleNames[0]),
numerHistName + numerAdd).clone())
for s in sampleNames[1:]:
hDenom.Add(
getattr(getattr(parentDirectory, s), denomHistName).clone())
hNumer.Add(
getattr(getattr(parentDirectory, s), numerHistName).clone())
if numerAdd:
hNumer.Add(
getattr(getattr(parentDirectory, s),
numerHistName + numerAdd).clone())
if binsToMerge:
hDenom = hDenom.merge_bins(binsToMerge)
hNumer = hNumer.merge_bins(binsToMerge)
return Efficiency(hNumer, hDenom)
def rejection_linear(eff):
htot = asrootpy(eff.GetTotalHistogram()).Clone()
hpass = asrootpy(eff.GetPassedHistogram()).Clone()
hnotpass = htot - hpass
hnotpass.Sumw2()
rej = Efficiency(hnotpass, htot, name='Rejlin_{0}'.format(eff.name), title=eff.title)
return rej
def get_efficiency_array():
eff_array = {}
for var in VARIABLES['plotting']:
hpassed = Hist(var['bins'], var['range'][0], var['range'][1])
htotal = Hist(var['bins'], var['range'][0], var['range'][1])
eff_array[var['name']] = Efficiency(hpassed,
htotal,
title=get_label(var))
return eff_array
def get_efficiency_graphs(d):
efficiency_graphs = []
#for s in sample_names2:
# sample_dir = getattr(d2, s)
# numer_ = getattr(sample_dir, 'lxyDpToEl__match').clone()
# denom_ = getattr(sample_dir, 'lxyDpToEl__total').clone()
# g = Efficiency(numer_, denom_).graph
# efficiency_graphs.append(g)
for s in sample_names:
sample_dir = getattr(d, s)
numer_ = getattr(sample_dir, 'lxyDpToEl__match').clone()
denom_ = getattr(sample_dir, 'lxyDpToEl__total').clone()
g = Efficiency(numer_, denom_).graph
efficiency_graphs.append(g)
return efficiency_graphs
def test_overall_efficiency():
for stat_op in range(0, 8):
Eff = Efficiency(Hist(20, -3, 3), Hist(20, -3, 3))
Eff_1bin = Efficiency(Hist(1, -3, 3), Hist(1, -3, 3))
Eff.SetStatisticOption(stat_op)
Eff_1bin.SetStatisticOption(stat_op)
for i in range(1000):
x = gauss(0, 3.6)
w = uniform(0, 1)
passed = w > 0.5
Eff.Fill(passed, x)
Eff_1bin.Fill(passed, x)
assert_almost_equal(Eff.overall_efficiency(overflow=True)[0],
Eff_1bin.overall_efficiency(overflow=True)[0])
assert_almost_equal(Eff.overall_efficiency(overflow=True)[1],
Eff_1bin.overall_efficiency(overflow=True)[1])
assert_almost_equal(Eff.overall_efficiency(overflow=True)[2],
Eff_1bin.overall_efficiency(overflow=True)[2])
def eff_curve(accept, total, var, weight_field=None, prefix='off'):
"""
Draw the efficiency curve from two record arrays
(selection already applied so computation should be fast over several vars)
"""
if var not in VARIABLES.keys():
log.error('Wrong variable name (see variables.py)')
raise ValueError('Wrong variable name (see variables.py)')
var_info = VARIABLES[var]
hnum = Hist(var_info['bins'], var_info['range'][0], var_info['range'][1])
hden = Hist(var_info['bins'], var_info['range'][0], var_info['range'][1])
log.info('filling')
if prefix is not None:
field = prefix + '_' + var_info['name']
else:
field = var_info['name']
num = accept[field]
den = total[field]
if 'scale' in var_info.keys():
num *= var_info['scale']
den *= var_info['scale']
fill_hist(hnum, accept[prefix + '_' + var], accept[weight_field])
fill_hist(hden, total[prefix + '_' + var], total[weight_field])
eff = Efficiency(hnum,
hden,
name='eff_{0}'.format(var),
title=get_label(VARIABLES[var]))
return eff
def test_efficiency():
# 1D
eff = Efficiency(Hist(10, 0, 1), Hist(10, 0, 1))
eff.Fill(False, 0.1)
eff.Fill(True, 0.8)
assert_equal(len(eff), len(eff.total))
if ROOT_VERSION >= ROOTVersion(53417):
assert eff.graph
assert eff.painted_graph
assert_equal(len(list(eff.efficiencies())), 10)
assert_equal(len(list(eff.efficiencies(overflow=True))), 12)
assert_equal(len(list(eff.errors())), 10)
assert_equal(len(list(eff.errors(overflow=True))), 12)
# 2D
eff = Efficiency(Hist2D(10, 0, 1, 10, 0, 1), Hist2D(10, 0, 1, 10, 0, 1))
eff.Fill(False, 0.1)
eff.Fill(True, 0.8)
assert_equal(len(eff), len(eff.total))
if ROOT_VERSION >= ROOTVersion(53417):
assert eff.histogram
assert eff.painted_histogram
def efficiency_graph(tobject_numerator, tobject_denominator):
"""Compute TEfficiency with proper clopper-pearson intervals"""
_eff = Efficiency(tobject_numerator, tobject_denominator)
return asrootpy(_eff.CreateGraph())
def efficiency(tobject_numerator, tobject_denominator):
"""Compute TEfficiency"""
return Efficiency(tobject_numerator, tobject_denominator)
def test_efficiency():
# 1D
eff = Efficiency(Hist(10, 0, 1), Hist(10, 0, 1))
eff.Fill(False, 0.1)
eff.Fill(True, 0.8)
assert_equal(len(eff), len(eff.total))
if ROOT_VERSION >= ROOTVersion(53417):
assert eff.graph
assert eff.painted_graph
assert_equal(len(list(eff.efficiencies())), 10)
assert_equal(len(list(eff.efficiencies(overflow=True))), 12)
assert_equal(len(list(eff.errors())), 10)
assert_equal(len(list(eff.errors(overflow=True))), 12)
# 2D
eff = Efficiency(Hist2D(10, 0, 1, 10, 0, 1), Hist2D(10, 0, 1, 10, 0, 1))
eff.Fill(False, 0.1)
eff.Fill(True, 0.8)
assert_equal(len(eff), len(eff.total))
if ROOT_VERSION >= ROOTVersion(53417):
assert eff.histogram
assert eff.painted_histogram
def test_overall_efficiency():
for stat_op in range(0, 8):
Eff = Efficiency(Hist(20, -3, 3), Hist(20, -3, 3))
Eff_1bin = Efficiency(Hist(1, -3, 3), Hist(1, -3, 3))
Eff.SetStatisticOption(stat_op)
Eff_1bin.SetStatisticOption(stat_op)
for i in xrange(1000):
x = gauss(0, 3.6)
w = uniform(0, 1)
passed = w > 0.5
Eff.Fill(passed, x)
Eff_1bin.Fill(passed, x)
assert_almost_equal(
Eff.overall_efficiency(overflow=True)[0],
Eff_1bin.overall_efficiency(overflow=True)[0])
assert_almost_equal(
Eff.overall_efficiency(overflow=True)[1],
Eff_1bin.overall_efficiency(overflow=True)[1])
assert_almost_equal(
Eff.overall_efficiency(overflow=True)[2],
Eff_1bin.overall_efficiency(overflow=True)[2])
eff_ntuple.Draw("nPVs",
"recoPt > 40 && l1gMatch && max(l1gPt, l1gRegionEt) > 25",
hist=uct_pass_vs_pu)
eff_ntuple.Draw("nPVs",
"recoPt > 40 && l1gMatch && max(l1gPt, l1gRegionEt) > 25"
" && l1gJetPt/max(l1gPt, l1gRegionEt) - 1 < 0.5",
hist=uct_iso_pass_vs_pu)
eff_ntuple.Draw("nPVs",
"recoPt > 40 && l1Match && l1Pt > 44",
hist=l1_pass_vs_pu)
canvas = Canvas(800, 800)
l1_efficiency = Efficiency(l1_pass_vs_pu,
total_vs_pu).decorate(linecolor='red',
linewidth=2,
markerstyle=20,
markercolor='red')
uct_efficiency = Efficiency(uct_pass_vs_pu,
total_vs_pu).decorate(linecolor='blue',
linewidth=2,
markerstyle=20,
markercolor='blue')
uct_iso_efficiency = Efficiency(uct_iso_pass_vs_pu,
total_vs_pu).decorate(linecolor='green',
linewidth=2,
markerstyle=20,
markercolor='green')
def make_efficiency_plots_pls(names, outname, colors):
if len(names) < 2:
raise ValueError("I need 2 or more histograms!")
if len(names) != len(colors):
raise ValueError("I need a color for each histogram!")
# First
n = len(names)
i = 1
hname = names[0]
denom = histograms[hname]
hname = names[i]
numer = histograms[hname]
hname = outname
eff = Efficiency(numer, denom, name=hname)
eff.SetStatisticOption(0) # kFCP
eff.SetConfidenceLevel(0.682689492137) # one sigma
eff.linecolor = colors[i]
eff.linewidth = 2
eff.markercolor = colors[i]
eff.markerstyle = 1
#
frame = eff.GetCopyTotalHisto().Clone(hname + "_frame")
frame.Reset()
frame.SetMinimum(0)
frame.SetMaximum(1.2)
frame.GetYaxis().SetTitle("#varepsilon")
frame.SetStats(0)
frame.Draw()
tline = TLine()
tline.SetLineColor(1)
xmin, xmax = frame.GetXaxis().GetXmin(), frame.GetXaxis().GetXmax()
tline.DrawLine(xmin, 1.0, xmax, 1.0)
#
eff.Draw("same p")
keepalive(gPad.func(), eff)
gPad.Print(options.outdir + hname + ".png")
# Second and on
for i in xrange(2, n):
hname = names[0]
denom = histograms[hname]
hname = names[i]
numer = histograms[hname]
hname = outname
eff = Efficiency(numer, denom, name=hname)
eff.SetStatisticOption(0) # kFCP
eff.SetConfidenceLevel(0.682689492137) # one sigma
eff.linecolor = colors[i]
eff.linewidth = 2
eff.markercolor = colors[i]
eff.markerstyle = 1
eff.Draw("same p")
keepalive(gPad.func(), eff)
gPad.Print(options.outdir + hname + ".png")
return
for in_q, q in zip(in_quantiles, quantiles):
print "..", in_q, q
h1a.linecolor = options.palette[i]
h1a.linewidth = 2
h1a.Draw("hist")
gPad.Print(options.outdir + hname + ".png")
# Make efficiency vs gen pT
k = "denom"
hname = "eff_vs_genpt_l1pt20_%s" % k
denom = histograms[hname]
k = "numer"
hname = "eff_vs_genpt_l1pt20_%s" % k
numer = histograms[hname]
hname = "eff_vs_genpt_l1pt20"
eff = Efficiency(numer, denom, name=hname)
eff.SetStatisticOption(0) # kFCP
eff.SetConfidenceLevel(0.682689492137) # one sigma
eff.linecolor = 'gray'
eff.linewidth = 2
eff.markercolor = 'gray'
eff.markerstyle = 1
#
frame = eff.GetCopyTotalHisto().Clone(hname + "_frame")
frame.Reset()
frame.SetMinimum(0)
frame.SetMaximum(1.2)
frame.GetYaxis().SetTitle("#varepsilon")
frame.SetStats(0)
frame.Draw()
tline = TLine()
