def do_all_1D_projection_plots_in_dir(directories,
components_styles_dicts,
output_dir,
region_str=None,
do_ratio=True,
normalise_hists=True,
jet_config_str="",
title=None,
bin_by=None):
"""
Given a set of TDirs, loop over all 2D hists, do projection hists for chosen bins, and plot all TDir contributions on a canvas for comparison.
components_styles_dicts should be a list of style dicts, one for each directory/contribution to a plot.
This should include label for this component.
"""
if len(directories) != len(components_styles_dicts):
raise IndexError("Need same number of style dicts and directories")
list_of_obj = [get_list_of_obj(d) for d in directories]
# check all have same list of plots
if not all(x == list_of_obj[0] for x in list_of_obj):
print("Different number of object in the TDirectorys")
common_list_of_obj = set(list_of_obj[0])
for l in list_of_obj[1:]:
common_list_of_obj = common_list_of_obj & set(l)
common_list_of_obj = sorted(list(common_list_of_obj))
pt_bins = qgc.PT_BINS
lumi = cu.get_lumi_str(do_dijet=False, do_zpj=True)
if "dijet" in region_str.lower():
lumi = cu.get_lumi_str(do_dijet=True, do_zpj=False)
show_integrals = not normalise_hists
for obj_name in common_list_of_obj:
if "flav" in obj_name:
continue
if obj_name in [
'eta_jet1_vs_eta_jet2',
'phi_jet1_vs_pt_jet1',
'phi_jet2_vs_pt_jet1',
# 'reliso_mu1_vs_pt_jet1', 'reliso_mu2_vs_pt_jet1',
# 'dphi_mumu_jet1_vs_pt_jet1',
# 'dphi_mumu_vs_pt_jet1',
'pt_jet1_z_pt_jet2_z_ratio',
# 'met_sig_vs_pt_jet1'
'weight_vs_puHat_genHT_ratio',
'eta_jet_response',
]:
continue
if "genjet_ind_recojet_ind" in obj_name:
continue
print("Doing:", obj_name)
objs = [
d.Get(obj_name).Clone(obj_name + str(uuid1())) for d in directories
]
# Do TH1s separately
if not isinstance(objs[0], (ROOT.TH2F, ROOT.TH2D, ROOT.TH2I)):
logx = obj_name in [
"pt_jet", "pt_jet1", "pt_jet2", "pt_mumu", 'gen_ht',
'pt_jet_response_binning', 'pt_genjet_response_binning',
'pt_jet1_unweighted', 'pt_jet_unweighted'
]
this_title = (("{jet_algo}\n"
"{region_label}\n").format(jet_algo=jet_config_str,
region_label=region_str))
if title is not None:
this_title += "\n%s" % (title)
# if obj_name in ['pt_jet', 'pt_jet1', 'pt_jet2', 'pt_mumu']:
# # small rebinned version
# for obj in objs:
# obj.Rebin(1)
# if normalise_hists:
do_1D_plot(objs,
components_styles_dicts=components_styles_dicts,
do_ratio=do_ratio,
normalise_hists=normalise_hists,
logx=logx,
logy=True,
title=this_title,
mean_rel_error=-1,
lumi=lumi,
show_integrals=show_integrals,
output_filename=os.path.join(
output_dir, obj_name + ".%s" % (OUTPUT_FMT)))
# do a rebinned version for some variables
if obj_name in [
'pt_jet', 'pt_jet1', 'pt_jet2', 'pt_mumu',
'pt_jet_response_binning'
]:
for obj in objs:
if obj_name == 'pt_jet_response_binning':
obj.Rebin(2) # to get gen level pt binning
else:
obj.Rebin(
20) # this is multiplied by the previous rebin
this_title = (("{jet_algo}\n"
"{region_label}\n").format(
jet_algo=jet_config_str,
region_label=region_str))
if title is not None:
this_title += "\n%s" % (title)
do_1D_plot(objs,
components_styles_dicts=components_styles_dicts,
do_ratio=do_ratio,
normalise_hists=normalise_hists,
logx=logx,
logy=True,
title=this_title,
lumi=lumi,
show_integrals=show_integrals,
output_filename=os.path.join(
output_dir,
obj_name + "_rebin.%s" % (OUTPUT_FMT)))
else:
for pt_min, pt_max in pt_bins:
# print(pt_min, pt_max)
rebin = 1
# exceptions...why didn't I pick the same number of bins...
do_not_rebin = any([
"n_jets" in obj_name,
"n_mu" in obj_name,
"met_sig" in obj_name,
# obj_name.startswith('dphi_mumu'),
obj_name.startswith('pt_jet3_frac'),
obj_name.startswith('pt_jet1_jet2_ratio'),
obj_name.startswith('pt_jet1_z_ratio'),
obj_name.startswith('pt_jet2_z_ratio'),
obj_name.startswith('dphi_jet1_z_vs_pt_jet1'),
# obj_name.startswith('m_jj'),
])
if not do_not_rebin:
# if objs[0].GetNbinsX() % 5 == 0 and objs[0].GetNbinsX() >= 100:
# rebin = 5
if objs[0].GetNbinsX() % 4 == 0 and objs[0].GetNbinsX(
) >= 80:
rebin = 2
elif objs[0].GetNbinsX() % 3 == 0 and objs[0].GetNbinsX(
) >= 60:
rebin = 3
if obj_name.startswith("m_jj"):
rebin = 2
if "reliso" in obj_name:
rebin = 1
if "pt_jet1_vs_pt_z" in obj_name:
rebin = 20
if "pt_mu1_vs_pt_z" in obj_name:
rebin = 10
if pt_min > 250:
rebin = 20
if "pt_mu2_vs_pt_z" in obj_name:
rebin = 10
if pt_min > 250:
rebin = 25
hists = [
qgp.get_projection_plot(ob, pt_min, pt_max) for ob in objs
]
for h in hists:
h.Rebin(rebin)
def _title(region_str, start_val, end_val):
pt_var_str = "p_{T}^{jet}"
if bin_by == "ave":
pt_var_str = "#LT p_{T}^{jet} #GT"
elif bin_by == "Z" and "_vs_pt_jet" not in obj_name:
pt_var_str = "p_{T}^{Z}"
s = (("{jet_algo}\n"
"{region_label}\n"
"{bin_edge_low:g} < {pt_str} < {bin_edge_high:g} GeV"
).format(jet_algo=jet_config_str,
region_label=region_str,
pt_str=pt_var_str,
bin_edge_low=start_val,
bin_edge_high=end_val))
if title is not None:
s = "%s\n%s" % (s, title)
return s
xtitle = None
if "eta_jet1_vs_pt" in obj_name and "eta_ordered" in directories[
0].GetName().lower():
xtitle = "y^{forward jet}"
elif "eta_jet2_vs_pt" in obj_name and "eta_ordered" in directories[
0].GetName().lower():
xtitle = "y^{central jet}"
logx = any([
x in obj_name
for x in ['pt_jet_response', 'pt_jet1_vs_pt_z']
])
logy = any([
x in obj_name
for x in ['pt_jet_response', 'pt_jet1_vs_pt_z']
])
do_1D_plot(hists,
components_styles_dicts=components_styles_dicts,
do_ratio=do_ratio,
normalise_hists=normalise_hists,
logx=logx,
logy=logy,
title=_title(region_str, pt_min, pt_max),
xtitle=xtitle,
mean_rel_error=-1,
lumi=lumi,
show_integrals=show_integrals,
output_filename=os.path.join(
output_dir, obj_name + "_pt%dto%d.%s" %
(pt_min, pt_max, OUTPUT_FMT)))
def do_all_1D_projection_plots_in_dir(directories,
output_dir,
components_styles_dicts=None,
draw_opts="NOSTACK HISTE",
do_ratio=True,
normalise_hists=True,
filter_noisy=True,
find_best_purity=False,
signal_mask=None):
"""
Given a set of TDirs, loop over all 2D hists, do projection hists for chosen bins, and plot all TDir contributions on a canvas for comparison.
components_styles_dicts should be a list of style dicts, one for each directory/contribution to a plot.
This should include label for this component.
filter_noisy aims to remove low stats/large error components that muddle things up
find_best_purity aism to find cut for best purity.
signal_mask should be a list of bools to specify which directories are signal
"""
if len(directories) != len(components_styles_dicts):
raise IndexError("Need same number of style dicts and directories")
list_of_obj = [get_list_of_obj(d) for d in directories]
print(list_of_obj[0])
# check all have same list of plots
if not all(x == list_of_obj[0] for x in list_of_obj):
print("Different number of object in the TDirectorys")
common_list_of_obj = set(list_of_obj[0])
for l in list_of_obj[1:]:
common_list_of_obj = common_list_of_obj & set(l)
pt_bins = [(20, 40), (40, 60), (60, 80), (100, 120), (160, 200),
(260, 300), (500, 600), (1000, 2000)]
for obj_name in common_list_of_obj:
objs = [d.Get(obj_name) for d in directories]
# Ignore TH1s
if not isinstance(objs[0], (ROOT.TH2F, ROOT.TH2D, ROOT.TH2I)):
print(obj_name, "is not a TH2")
continue
if "flav" in obj_name:
continue
if obj_name in [
# 'eta_jet1_vs_eta_jet2',
'phi_jet1_vs_pt_jet1',
'phi_jet2_vs_pt_jet1',
'reliso_mu1_vs_pt_jet1',
'reliso_mu2_vs_pt_jet1',
'dphi_mumu_jet1_vs_pt_jet1',
'dphi_mumu_vs_pt_jet1',
'pt_jet1_z_pt_jet2_z_ratio'
]:
continue
for pt_min, pt_max in pt_bins:
# print(pt_min, pt_max)
rebin = 1
# exceptions...why didn't i pick the same number of bins...
do_not_rebin = any([
"n_jets" in obj_name,
"n_mu" in obj_name,
"met_sig" in obj_name,
obj_name.startswith('dphi_mumu'),
obj_name.startswith('pt_jet3_frac'),
obj_name.startswith('pt_jet1_jet2_ratio'),
obj_name.startswith('pt_jet1_z_ratio'),
obj_name.startswith('pt_jet2_z_ratio'),
obj_name.startswith('dphi_jet1_z_vs_pt_jet1'),
# obj_name.startswith('m_jj'),
])
if not do_not_rebin:
if objs[0].GetNbinsX() % 5 == 0 and objs[0].GetNbinsX() >= 100:
rebin = 5
elif objs[0].GetNbinsX() % 4 == 0 and objs[0].GetNbinsX(
) >= 80:
rebin = 2
elif objs[0].GetNbinsX() % 3 == 0 and objs[0].GetNbinsX(
) >= 60:
rebin = 3
if obj_name.startswith("m_jj"):
rebin = 2
hists = [
qgg.get_projection_plot(ob, pt_min, pt_max) for ob in objs
]
# need the clone in here as unnormalised hists are used for purity/eff scans
contributions = [
Contribution(hist.Clone(hist.GetName() + "Clone"),
normalise_hist=normalise_hists,
rebin_hist=rebin,
**csd)
for hist, csd in zip(hists, components_styles_dicts)
]
if len(contributions) == 0:
continue
# Ignore if all empty objs
total_entries = sum(c.obj.GetEntries() for c in contributions)
if total_entries == 0:
continue
if filter_noisy:
# filter contributions to ensure odd low stat ones don't dominate the scale
# combination of factors to remove noisy samples
mean_errs = []
max_over_mean_errs = []
rel_err_vars = []
for cont in contributions:
obj = cont.obj
errs = [
obj.GetBinError(i) for i in range(obj.GetNbinsX() + 1)
]
if obj.GetEntries() > 0:
mean_errs.append(np.mean(errs))
max_over_mean_errs.append(np.max(errs) / np.mean(errs))
rel_err_vars.append(np.std(errs) / np.mean(errs))
# print("obj", cont.label)
# print('mean errs', np.mean(errs))
# print('rel max err', np.max(errs) / np.mean(errs))
# print('stddev err', np.std(errs))
# print('rel std dev err', np.std(errs) / np.mean(errs))
else:
# Dud values if 0 entries
mean_errs.append(9999999)
max_over_mean_errs.append(9999999)
rel_err_vars.append(9999999)
ref_mean_err = np.median(mean_errs)
ref_rel_err_var = np.median(rel_err_vars)
# print('-'*20)
# print('mean mean err', ref_mean_err)
# print('mean var', ref_rel_err_var)
contributions = [
cont for cont, merr, rev, mom in zip(
contributions, mean_errs, rel_err_vars,
max_over_mean_errs)
if (merr < 2.5 * ref_mean_err) and (
rev < 5 * ref_rel_err_var or mom < 5)
]
purity_text = ROOT.TPaveText(0.6, 0.4, 0.95, 0.6, "NDC")
purity_text.SetFillColor(ROOT.kWhite)
purity_text.SetFillStyle(0)
purity_text.SetLineWidth(0)
purity_text.SetTextAlign(ROOT.kHAlignLeft + ROOT.kVAlignTop)
if find_best_purity:
if not signal_mask:
raise RuntimeError("Need signal_mask to find purity")
if len(signal_mask) != len(directories):
raise RuntimeError(
"signal_mask must be same length as directories")
signal_hists = [
h for h, mask in zip(hists, signal_mask) if mask
]
bkg_hists = [
h for h, mask in zip(hists, signal_mask) if not mask
]
signal_hist = signal_hists[0].Clone("Signal")
if len(signal_hists) > 1:
for h in signal_hists[1:]:
signal_hist.Add(h)
bkg_hist = bkg_hists[0].Clone("Background")
if len(bkg_hists) > 1:
for h in bkg_hists[1:]:
bkg_hist.Add(h)
signal = signal_hist.Integral()
bkg = bkg_hist.Integral()
starting_purity = 1. * signal / (signal + bkg)
starting_purity_x_eff = starting_purity
signal_cumul = signal_hist.GetCumulative()
bkg_cumul = bkg_hist.GetCumulative()
best_purity = starting_purity
best_purity_x_eff = 1 * starting_purity
best_purity_x_eff = 0
purity_cut_value, purity_cut_type, purity_bin_ind = None, None, None
purity_x_eff_cut_value, purity_x_eff_cut_type, purity_x_eff_bin_ind = None, None, None
purity_x_effs_1, cut_vals_1 = [], []
purity_x_effs_2, cut_vals_2 = [], []
puritys_1, effs_1 = [], []
puritys_2, effs_2 = [], []
total_s = signal_hist.Integral()
# try cut lower than X, see if any cut improves upon starting purity
for i in range(1, signal_cumul.GetNbinsX() + 1):
s = signal_cumul.GetBinContent(i)
b = bkg_cumul.GetBinContent(i)
if s == 0 and b == 0:
this_purity = 0
this_purity_x_eff = 0
else:
this_purity = s / (s + b)
if s > total_s * 1.00001:
raise RuntimeError(
"s = %f, total_s = %f, something terribly wrong"
% (s, total_s))
this_eff = s / total_s
this_purity_x_eff = this_eff * this_purity
cut_vals_1.append(signal_cumul.GetBinLowEdge(i + 1))
purity_x_effs_1.append(this_purity_x_eff)
effs_1.append(this_eff)
puritys_1.append(this_purity)
# print("this_purity =", this_purity)
if this_purity > best_purity:
purity_cut_type = "<"
purity_bin_ind = i
purity_cut_value = signal_cumul.GetBinLowEdge(i + 1)
best_purity = this_purity
if this_purity_x_eff > best_purity_x_eff:
best_purity_x_eff = this_purity_x_eff
purity_x_eff_cut_type = "<"
purity_x_eff_bin_ind = i
purity_x_eff_cut_value = signal_cumul.GetBinLowEdge(i +
1)
# try cut below, see if any cut improves upon starting purity
signal_cumul_rev = signal_hist.GetCumulative(False)
bkg_cumul_rev = bkg_hist.GetCumulative(False)
for i in range(1, signal_cumul_rev.GetNbinsX() + 1):
s = signal_cumul_rev.GetBinContent(i)
b = bkg_cumul_rev.GetBinContent(i)
if s == 0 and b == 0:
this_purity = 0
this_eff = 0
else:
this_purity = s / (s + b)
if s > total_s * 1.00001:
raise RuntimeError(
"s = %f, total_s = %f, something terribly wrong"
% (s, total_s))
this_eff = (s / total_s)
this_purity_x_eff = this_eff * this_purity
cut_vals_2.append(signal_cumul_rev.GetBinLowEdge(i +
1))
purity_x_effs_2.append(this_purity_x_eff)
effs_2.append(this_eff)
puritys_2.append(this_purity)
# print("this_purity =", this_purity)
if this_purity > best_purity:
purity_cut_type = ">"
purity_bin_ind = i
purity_cut_value = signal_cumul_rev.GetBinLowEdge(i)
best_purity = this_purity
if this_purity_x_eff > best_purity_x_eff:
best_purity_x_eff = this_purity_x_eff
purity_x_eff_cut_type = ">"
purity_x_eff_bin_ind = i
purity_x_eff_cut_value = signal_cumul_rev.GetBinLowEdge(
i + 1)
c = ROOT.TCanvas("ctmp", "", 600, 600)
c.SetTicks(1, 1)
gr = ROOT.TGraph(len(purity_x_effs_1), array('d', cut_vals_1),
array('d', purity_x_effs_1))
gr.SetMarkerStyle(22)
gr2 = ROOT.TGraph(len(purity_x_effs_2), array('d', cut_vals_2),
array('d', purity_x_effs_2))
gr2.SetMarkerStyle(23)
gr2.SetMarkerColor(ROOT.kBlue)
mg = ROOT.TMultiGraph("mg",
";Cut value;Purity #times Efficiency")
mg.Add(gr)
mg.Add(gr2)
mg.Draw("AP")
c.SaveAs(
os.path.join(
output_dir, obj_name + "pt%dto%d_purity_x_eff.%s" %
(pt_min, pt_max, OUTPUT_FMT)))
gr = ROOT.TGraph(len(puritys_1), array('d', effs_1),
array('d', puritys_1))
gr.SetMarkerStyle(22)
gr2 = ROOT.TGraph(len(puritys_2), array('d', effs_2),
array('d', puritys_2))
gr2.SetMarkerStyle(23)
gr2.SetMarkerColor(ROOT.kBlue)
mg = ROOT.TMultiGraph("mg2", ";efficiency;purity")
mg.Add(gr)
mg.Add(gr2)
mg.Draw("ALP")
c.SaveAs(
os.path.join(
output_dir, obj_name + "pt%dto%d_purity_vs_eff.%s" %
(pt_min, pt_max, OUTPUT_FMT)))
if best_purity == starting_purity:
purity_text.AddText("Best purity (%.3f) with no cut" %
best_purity)
else:
purity_text.AddText("No cut purity = %.3f" %
(starting_purity))
purity_text.AddText(
"Best purity (%.3f) with %s %.3g" %
(best_purity, purity_cut_type, purity_cut_value))
if best_purity_x_eff == starting_purity_x_eff:
purity_text.AddText(
"Best purity * eff (%.3f) with no cut" %
best_purity_x_eff)
else:
purity_text.AddText("No cut purity * eff = %.3f" %
(starting_purity_x_eff))
purity_text.AddText(
"Best purity * eff (%.3f) with cut %s %.3g" %
(best_purity_x_eff, purity_x_eff_cut_type,
purity_x_eff_cut_value))
ylim = None
# if "pt_jet" in obj_name and "ratio" not in obj_name and "frac" not in obj_name:
# ylim = [1E-9, 1E-1]
title = "%d < p_{T}^{jet 1} < %d GeV" % (pt_min, pt_max)
p = Plot(contributions,
what='hist',
ytitle="p.d.f." if normalise_hists else "N",
title=title,
subplot_type="ratio" if do_ratio else None,
subplot_title="#splitline{Ratio wrt}{%s}" %
contributions[0].label,
subplot=contributions[0],
ylim=ylim)
p.legend.SetX1(0.7)
p.legend.SetX2(0.9)
p.plot(draw_opts)
p.main_pad.cd()
if find_best_purity:
purity_text.Draw("SAME")
p.save(
os.path.join(
output_dir,
obj_name + "_pt%dto%d.%s" % (pt_min, pt_max, OUTPUT_FMT)))
def do_plots(root_dir):
# QG variable plots
pt_bins = qgc.PT_BINS[:]
var_list = qgc.COMMON_VARS
radius, pus = cu.get_jet_config_from_dirname(root_dir)
jet_str = "AK%s" % (radius.upper())
dy_tfile, qcd_tfile = None, None
if os.path.isfile(os.path.join(root_dir, qgc.DY_FILENAME)):
dy_tfile = cu.TFileCacher(os.path.join(root_dir, qgc.DY_FILENAME))
if os.path.isfile(os.path.join(root_dir, qgc.QCD_FILENAME)):
qcd_tfile = cu.TFileCacher(os.path.join(root_dir, qgc.QCD_FILENAME))
for gr_append in ["", "_groomed"]:
if gr_append == "_groomed":
print("Doing groomed plots...")
else:
print("Doing ungroomed plots...")
# GEnJets matched to recojet, reco selection
zpj_dirname = "ZPlusJets_QG%s" % (gr_append)
dj_cen_dirname = "Dijet_QG_central_tighter%s" % (gr_append)
dj_fwd_dirname = "Dijet_QG_forward_tighter%s" % (gr_append)
# Gen Jets from gen selection
zpj_dirname = "ZPlusJets_QG_gen%s" % (gr_append)
dj_cen_dirname = "Dijet_QG_gen_central%s" % (gr_append)
dj_fwd_dirname = "Dijet_QG_gen_forward%s" % (gr_append)
for ang in var_list[:]:
print("...Doing", ang.name)
var_template = "{prefix}%s_vs_pt" % (ang.var
) # {prefix} is for flavour
for pt_ind, (start_val, end_val) in enumerate(pt_bins):
zpj_entries = []
dijet_cen_entries = []
dijet_fwd_entries = []
# Get all plots
lw = 2
msize = 1.1
mc_msize = 1E-3 # small enough to not be seen but not 0 - otherwise the horizontal lines on error bars don't get drawn
mc_msize = msize # small enough to not be seen but not 0 - otherwise the horizontal lines on error bars don't get drawn
flavours_dicts = [
{
"prefix": "q",
"label": "u/d/s"
},
{
"prefix": "g",
"label": "g"
},
{
"prefix": "bc",
"label": "b/c"
},
]
if dy_tfile:
# Z+JETS REGION
marker = cu.Marker(qgc.DY_MARKER)
for flav_dict, color, mark in zip(flavours_dicts,
qgc.DY_COLOURS,
marker.cycle()):
h2d_dyj_mc = dy_tfile.get(
"%s/%s" %
(zpj_dirname, var_template.format(**flav_dict)))
dy_kwargs_mc = dict(line_color=color,
line_width=lw,
fill_color=color,
marker_color=color,
marker_style=mark,
marker_size=mc_msize,
label="%s flavour" %
(flav_dict['label']),
subplot=None)
zpj_entries.append(
(qgp.get_projection_plot(h2d_dyj_mc, start_val,
end_val), dy_kwargs_mc))
if qcd_tfile:
# DIJET CENTRAL REGION
marker = cu.Marker(qgc.QCD_MARKER)
for flav_dict, color, mark in zip(flavours_dicts,
qgc.QCD_COLOURS,
marker.cycle()):
h2d_qcd_cen_mc = qcd_tfile.get(
"%s/%s" %
(dj_cen_dirname, var_template.format(**flav_dict)))
qcd_cen_kwargs_mc = dict(line_color=color,
line_width=lw,
fill_color=color,
marker_color=color,
marker_style=mark,
marker_size=mc_msize,
label="%s flavour" %
(flav_dict['label']),
subplot=None)
dijet_mgpy_hist = qgp.get_projection_plot(
h2d_qcd_cen_mc, start_val, end_val)
dijet_cen_entries.append(
(dijet_mgpy_hist, qcd_cen_kwargs_mc))
# DIJET FORWARD REGION
marker = cu.Marker(qgc.QCD_MARKER)
for flav_dict, color, mark in zip(flavours_dicts,
qgc.QCD_COLOURS,
marker.cycle()):
h2d_qcd_fwd_mc = qcd_tfile.get(
"%s/%s" %
(dj_fwd_dirname, var_template.format(**flav_dict)))
qcd_fwd_kwargs_mc = dict(line_color=color,
line_width=lw,
fill_color=color,
marker_color=color,
marker_style=mark,
marker_size=mc_msize,
label="%s flavour" %
(flav_dict['label']),
subplot=None)
dijet_mgpy_hist = qgp.get_projection_plot(
h2d_qcd_fwd_mc, start_val, end_val)
dijet_fwd_entries.append(
(dijet_mgpy_hist, qcd_fwd_kwargs_mc))
#################
# SETUP PLOTTING
#################
rebin = 2
v_lower = var_template.lower()
if "multiplicity" in v_lower:
rebin = 1
# elif "flavour" in v_lower or "thrust" in v_lower:
# rebin = 1
# elif 'ptd' in v_lower:
# rebin = 4
# elif 'lha_charged' in v_lower:
# rebin = 4
# rebin = 1
xlim = None
if "width" in v_lower or "ptd" in v_lower:
xlim = (0, 1)
elif "thrust" in v_lower:
xlim = (0, 0.5)
# elif "multiplicity" in v_lower and "ak4" in root_dir.lower():
# if end_val <= 150:
# xlim = (0, 50)
# else:
# xlim = (0, 80)
auto_xlim = False
if "multiplicity" in v_lower or "thrust" in v_lower:
auto_xlim = True
ylim = [0, None]
if "flavour" in v_lower:
ylim = (0, 1)
# elif "lha" in v_lower:
# ylim = (0, 5)
plot_dir = os.path.join(
root_dir,
"plots_gen_lambda_flav_comparison%s" % (gr_append))
subplot_title = "Simulation / Data"
subplot_limits = (0, 2)
xlabel = ang.name + " (" + ang.lambda_str + ")"
if gr_append is not "":
xlabel = "Groomed " + ang.name + " (" + ang.lambda_str + ")"
def _title(region_str, start_val, end_val):
pt_var_str = "p_{T}^{jet}"
s = (("{jet_algo}\n"
"{region_label}\n"
"{mc_sample}\n"
"{bin_edge_low:g} < {pt_str} < {bin_edge_high:g} GeV"
).format(jet_algo=jet_str,
region_label=region_str,
mc_sample="MG5+Pythia8",
pt_str=pt_var_str,
bin_edge_low=start_val,
bin_edge_high=end_val))
return s
has_data = False
draw_opt = "NOSTACK HIST E1"
# dj central only
# dont' try to do multiple signal regions per plot, it looks rubbish
if len(dijet_cen_entries) > 0:
qgp.do_comparison_plot(
dijet_cen_entries,
"%s/ptBinned/%s_pt%dto%d_dijet_central.%s" %
(plot_dir, ang.var, start_val, end_val, OUTPUT_FMT),
rebin=rebin,
draw_opt=draw_opt,
title=_title(qgc.Dijet_CEN_LABEL, start_val, end_val),
xtitle=xlabel,
xlim='auto' if auto_xlim else
xlim, # don't use calc_auto_xlim, since do_comparison_plot will rebin it anyway
ylim=ylim,
data_first=has_data,
mean_rel_error=0.4,
subplot_type=None,
lumi=cu.get_lumi_str(
do_dijet=False,
do_zpj=True)) # full lumi as just MC
# dj forward only
if len(dijet_fwd_entries) > 0:
qgp.do_comparison_plot(
dijet_fwd_entries,
"%s/ptBinned/%s_pt%dto%d_dijet_forward.%s" %
(plot_dir, ang.var, start_val, end_val, OUTPUT_FMT),
rebin=rebin,
draw_opt=draw_opt,
title=_title(qgc.Dijet_FWD_LABEL, start_val, end_val),
xtitle=xlabel,
xlim='auto' if auto_xlim else xlim,
ylim=ylim,
data_first=has_data,
mean_rel_error=0.4,
subplot_type=None,
lumi=cu.get_lumi_str(
do_dijet=False,
do_zpj=True)) # full lumi as just MC
# zpj only
if len(zpj_entries) > 0:
if start_val > 149:
# rebin *= 2
rebin += 1
# find nearest divisor
while (zpj_entries[0][0].GetNbinsX() % rebin != 0):
rebin += 1
if "multiplicity" in v_lower:
if start_val > 300:
rebin += 1
# find nearest divisor
while (zpj_entries[0][0].GetNbinsX() % rebin != 0):
rebin += 1
qgp.do_comparison_plot(
zpj_entries,
"%s/ptBinned/%s_pt%dto%d_zpj.%s" %
(plot_dir, ang.var, start_val, end_val, OUTPUT_FMT),
rebin=rebin,
draw_opt=draw_opt,
title=_title(qgc.ZpJ_LABEL, start_val, end_val),
xtitle=xlabel,
xlim=qgp.calc_auto_xlim([d[0] for d in zpj_entries])
if auto_xlim else xlim,
ylim=ylim,
data_first=has_data,
mean_rel_error=0.4,
subplot_type=None,
lumi=cu.get_lumi_str(do_dijet=False, do_zpj=True))
def make_1D_rebin_hists(input_filename, plot_dirname, output_filename):
# QG variable plots
in_f = cu.open_root_file(input_filename)
out_f = ROOT.TFile(output_filename, "RECREATE")
plot_file = open(output_filename.replace(".root", ".plot"), "w")
for ang_ind, ang in enumerate(qgc.COMMON_VARS[:5]):
if ang.var not in qgc.ANGLE_REBIN_DICT:
continue
in_f.cd()
var_prepend = ""
obj_name = "%s%s_vs_pt" % (var_prepend, ang.var)
if var_prepend == "gen" and "multiplicity" in ang.var.lower():
obj_name = obj_name.replace("puppiMultiplicity", "multiplicity")
h2d = cu.get_from_tfile(in_f, "%s/%s" % (plot_dirname, obj_name))
for pt_ind, (start_val, end_val) in enumerate(qgc.PT_BINS):
hist = qgp.get_projection_plot(h2d, start_val, end_val,
'y') # y cos lambda is on x axis
this_rebins = qgc.ANGLE_REBIN_DICT[ang.var]
# new_name = "%s_Pt%sto%d" % (ang.var, start_val, end_val)
yoda_name = create_yoda_hist_name(plot_dirname, ang.var,
(start_val, end_val))
rebin_hist = hist.Rebin(
len(this_rebins) - 1, yoda_name, array('d', this_rebins))
# Normalise
if rebin_hist.Integral() > 0:
rebin_hist.Scale(1. / rebin_hist.Integral())
# Divide bin contents by bin width
for bin_ind in range(1, rebin_hist.GetNbinsX() + 1):
val = rebin_hist.GetBinContent(bin_ind)
width = rebin_hist.GetBinWidth(bin_ind)
rebin_hist.SetBinContent(bin_ind, val / width)
out_f.WriteTObject(rebin_hist) # saves faffing with cd()
# create yoda plot info
xlabel = "%s $(%s)$" % (ang.name, ang.lambda_str.replace(
"#", "\\"))
title_str = "$%d < p_{T}^{jet} < %d$ GeV" % (start_val, end_val)
xmax = 1
if "multiplicity" in ang.var.lower():
xmax = 50
this_plot_info = YODA_PLOT_TEMPLATE.format(plotname=yoda_name,
xlabel=xlabel,
title=title_str,
ylabel="p.d.f",
logx=0,
logy=0,
xmax=xmax,
xmin=0)
plot_file.write(this_plot_info)
# Add in pt hist
h_pt = cu.get_from_tfile(in_f, "Dijet_tighter/pt_jet_response_binning")
pt_name = "d01-x20-y01"
h_pt.SetName(pt_name)
this_plot_info = YODA_PLOT_TEMPLATE.format(
plotname=pt_name,
xlabel="$\langle p_{T}^{j}\\rangle$ GeV",
title="",
ylabel="N",
logy=1,
logx=1,
xmin=10,
xmax=1000)
plot_file.write(this_plot_info)
out_f.WriteTObject(h_pt)
out_f.Close()
plot_file.close()
dijet_cen_nominal_vs_syst_sf_entries = []
dijet_fwd_no_vs_nominal_sf_entries = []
dijet_fwd_nominal_vs_syst_sf_entries = []
# Get all plots
lw = 2
msize = 1.1
data_line_width = lw
####################
# Z+JETS REGION
####################
zpj_obj_name = "%s/%s" % (zpj_dirname, v)
# NO SF
zpj_no_sf_hist = qgp.get_projection_plot(
DY_NO_SF_TFILE[zpj_obj_name], start_val, end_val)
dy_kwargs_no_sf = dict(line_color=qgc.SINGLE_MU_COLOUR,
line_width=data_line_width,
fill_color=qgc.SINGLE_MU_COLOUR,
marker_color=qgc.SINGLE_MU_COLOUR,
marker_style=cu.Marker.get(
qgc.DY_MARKER),
marker_size=0,
label="No TRK SF")
zpj_no_vs_nominal_sf_entries.append(
(zpj_no_sf_hist, dy_kwargs_no_sf))
# NOMINAL TK SF
zpj_nominal_sf_hist = qgp.get_projection_plot(
DY_NOMINAL_SF_TFILE[zpj_obj_name], start_val, end_val)
dy_kwargs_nominal_sf = dict(line_color=qgc.DY_COLOUR,
def do_wrong_plots(root_dir,
var_prepend="",
plot_dir="wrong_flavs",
zpj_dirname="ZPlusJets_QG",
dj_dirname="Dijet_QG",
pt_bins=None,
title=""):
"""Plot all the sample/selection/flavour combinations to check distributions indep of sample"""
pt_bins = pt_bins or qgc.THEORY_PT_BINS
plot_vars = [
'jet_LHA', 'jet_pTD', 'jet_width', 'jet_thrust', 'jet_multiplicity'
]
if "puppi" in root_dir.lower():
plot_vars.append('jet_puppiMultiplicity')
for v in plot_vars:
v = "%s%s_vs_pt" % (var_prepend, v)
h2d_dyj_chs = grab_obj(os.path.join(root_dir, qgc.DY_FILENAME),
"%s/q%s" % (zpj_dirname, v))
h2d_dyj_wrong_chs = grab_obj(os.path.join(root_dir, qgc.DY_FILENAME),
"%s/g%s" % (zpj_dirname, v))
h2d_dyj_qcd_chs = grab_obj(os.path.join(root_dir, qgc.DY_FILENAME),
"%s/q%s" % (dj_dirname, v))
h2d_dyj_qcd_wrong_chs = grab_obj(
os.path.join(root_dir, qgc.DY_FILENAME),
"%s/g%s" % (dj_dirname, v))
h2d_qcd_chs = grab_obj(os.path.join(root_dir, qgc.QCD_FILENAME),
"%s/g%s" % (dj_dirname, v))
h2d_qcd_wrong_chs = grab_obj(os.path.join(root_dir, qgc.QCD_FILENAME),
"%s/q%s" % (dj_dirname, v))
lw = 1
dy_kwargs_chs = dict(line_color=qgc.DY_COLOUR,
fill_color=qgc.DY_COLOUR,
label=qgc.DY_ZpJ_QFLAV_LABEL,
line_width=lw,
marker_style=20,
marker_color=qgc.DY_COLOUR)
dy_kwargs_wrong_chs = dict(line_color=qgc.DY_COLOUR + 4,
fill_color=qgc.DY_COLOUR + 4,
label=qgc.DY_ZpJ_GFLAV_LABEL,
line_width=lw,
line_style=1,
marker_style=21,
marker_color=qgc.DY_COLOUR + 4)
dy_kwargs_qcd_chs = dict(line_color=ROOT.kGreen + 2,
fill_color=ROOT.kGreen + 2,
label=qgc.DY_Dijet_QFLAV_LABEL,
line_width=lw,
line_style=1,
marker_style=22,
marker_color=ROOT.kGreen + 2)
dy_kwargs_qcd_wrong_chs = dict(line_color=ROOT.kOrange - 1,
fill_color=ROOT.kOrange - 1,
label=qgc.DY_Dijet_GFLAV_LABEL,
line_width=lw,
line_style=1,
marker_style=23,
marker_color=ROOT.kOrange - 1)
qcd_kwargs_chs = dict(line_color=qgc.QCD_COLOUR,
fill_color=qgc.QCD_COLOUR,
label=qgc.QCD_Dijet_GFLAV_LABEL,
line_width=lw,
marker_style=25,
marker_color=qgc.QCD_COLOUR)
qcd_kwargs_wrong_chs = dict(line_color=ROOT.kRed,
fill_color=ROOT.kRed,
label=qgc.QCD_Dijet_QFLAV_LABEL,
line_width=lw,
line_style=1,
marker_style=24,
marker_color=ROOT.kRed)
rebin = 2
xlim = None
if "thrust" in v:
rebin = 1
xlim = (0, 0.5)
if "multiplicity" in v.lower():
rebin = 2
xlim = (0, 100)
for (start_val, end_val) in pt_bins:
entries = [
(qgg.get_projection_plot(h2d_dyj_chs, start_val,
end_val), dy_kwargs_chs),
(qgg.get_projection_plot(h2d_dyj_wrong_chs, start_val,
end_val), dy_kwargs_wrong_chs),
# (qgg.get_projection_plot(h2d_dyj_qcd_chs, start_val, end_val), dy_kwargs_qcd_chs),
# (qgg.get_projection_plot(h2d_dyj_qcd_wrong_chs, start_val, end_val), dy_kwargs_qcd_wrong_chs),
(qgg.get_projection_plot(h2d_qcd_wrong_chs, start_val,
end_val), qcd_kwargs_wrong_chs),
(qgg.get_projection_plot(h2d_qcd_chs, start_val,
end_val), qcd_kwargs_chs),
]
this_title = "%d < p_{T}^{jet} < %d GeV" % (start_val, end_val)
if title != "":
this_title += ", %s" % title
qgg.do_comparison_plot(
entries,
"%s/%s/%s_pt%dto%d_flavMatched.%s" %
(root_dir, plot_dir, v, start_val, end_val, OUTPUT_FMT),
rebin=rebin,
title=this_title,
xlim=xlim)
def do_plots(root_dir, title):
# QG variable plots
pt_bins = qgc.PT_BINS[:]
print(pt_bins)
var_list = qgc.COMMON_VARS
var_prepend = ""
radius, pus = cu.get_jet_config_from_dirname(root_dir)
jet_str = "AK%s PF %s" % (radius.upper(), pus.upper())
if radius == "8":
pt_bins = qgc.PT_BINS[2:]
print(pt_bins)
single_mu_tfile = cu.TFileCacher(os.path.join(root_dir, qgc.SINGLE_MU_FILENAME))
dy_tfile = cu.TFileCacher(os.path.join(root_dir, qgc.DY_FILENAME))
dy_hpp_tfile = cu.TFileCacher(os.path.join(root_dir, qgc.DY_HERWIG_FILENAME))
jetht_zb_tfile = cu.TFileCacher(os.path.join(root_dir, qgc.JETHT_ZB_FILENAME))
qcd_tfile = cu.TFileCacher(os.path.join(root_dir, qgc.QCD_FILENAME))
qcd_hpp_tfile = cu.TFileCacher(os.path.join(root_dir, qgc.QCD_HERWIG_FILENAME))
for gr_append in ["", "_groomed"]:
if gr_append == "_groomed":
print("Doing groomed plots...")
else:
print("Doing ungroomed plots...")
zpj_dirname = "ZPlusJets_QG%s" % (gr_append)
dj_cen_dirname = "Dijet_QG_central_tighter%s" % (gr_append)
dj_fwd_dirname = "Dijet_QG_forward_tighter%s" % (gr_append)
for ang in var_list[:]:
print("...Doing", ang.name)
v = "%s%s_vs_pt" % (var_prepend, ang.var)
zpj_2d_entries = []
dijet_cen_2d_entries = []
dijet_fwd_2d_entries = []
zpj_1d_entries = []
dijet_cen_1d_entries = []
dijet_fwd_1d_entries = []
for pt_ind, (start_val, end_val) in enumerate(pt_bins):
entries = []
zpj_entries = []
dijet_cen_entries = []
dijet_fwd_entries = []
# Get all plots
lw = 2
msize = 1.1
data_line_width = lw
mc_msize = 1E-3 # small enough to not be seen but not 0 - otherwise the horizontal lines on error bars don't get drawn
mgpy_label = "MG5+Pythia8"
hpp_label = "Herwig++"
####################
# Z+JETS REGION
####################
# SINGLE MU DATA
if single_mu_tfile:
h2d_dyj_data = single_mu_tfile.get("%s/%s" % (zpj_dirname, v))
dy_kwargs_data = dict(line_color=qgc.SINGLE_MU_COLOUR, line_width=data_line_width, fill_color=qgc.SINGLE_MU_COLOUR,
marker_color=qgc.SINGLE_MU_COLOUR, marker_style=cu.Marker.get(qgc.DY_MARKER), marker_size=msize*0.7,
label="Data")
zpj_data_hist = qgp.get_projection_plot(h2d_dyj_data, start_val, end_val)
entries.append((zpj_data_hist, dy_kwargs_data))
zpj_entries.append((zpj_data_hist, dy_kwargs_data))
if pt_ind == 0:
zpj_2d_entries.append((h2d_dyj_data, dy_kwargs_data))
# PYTHIA DY MC
if dy_tfile:
h2d_dyj_mc = dy_tfile.get("%s/%s" % (zpj_dirname, v))
dy_kwargs_mc = dict(line_color=qgc.DY_COLOUR, line_width=lw, fill_color=qgc.DY_COLOUR,
marker_color=qgc.DY_COLOUR, marker_style=cu.Marker.get(qgc.DY_MARKER), marker_size=mc_msize,
label=mgpy_label,
subplot=zpj_data_hist)
entries.append((qgp.get_projection_plot(h2d_dyj_mc, start_val, end_val), dy_kwargs_mc))
zpj_entries.append((qgp.get_projection_plot(h2d_dyj_mc, start_val, end_val), dy_kwargs_mc))
if pt_ind == 0:
zpj_2d_entries.append((h2d_dyj_mc, dy_kwargs_mc))
# HERWIG++ DY
if dy_hpp_tfile:
h2d_dyj_mc_hpp = dy_hpp_tfile.get("%s/%s" % (zpj_dirname, v))
col_hpp = qgc.DY_COLOURS[2]
dy_kwargs_mc_hpp = dict(line_color=col_hpp, line_width=lw, fill_color=col_hpp,
marker_color=col_hpp, marker_style=cu.Marker.get(qgc.DY_MARKER), marker_size=mc_msize,
label=hpp_label,
subplot=zpj_data_hist)
entries.append((qgp.get_projection_plot(h2d_dyj_mc_hpp, start_val, end_val), dy_kwargs_mc_hpp))
zpj_entries.append((qgp.get_projection_plot(h2d_dyj_mc_hpp, start_val, end_val), dy_kwargs_mc_hpp))
if pt_ind == 0:
zpj_2d_entries.append((h2d_dyj_mc_hpp, dy_kwargs_mc_hpp))
# MG+HERWIG++ DY
# if end_val < 151:
# h2d_dyj_mc3 = get("%s/%s" % (zpj_dirname, v))
# col4 = qgc.DY_COLOURS[3]
# dy_kwargs_mc3 = dict(line_color=col4, line_width=lw, fill_color=col4,
# marker_color=col4, marker_style=cu.Marker.get(qgc.DY_MARKER), marker_size=0,
# label="MG+Herwig++",
# subplot=zpj_data_hist)
# entries.append((qgp.get_projection_plot(h2d_dyj_mc3, start_val, end_val), dy_kwargs_mc3))
# zpj_entries.append((qgp.get_projection_plot(h2d_dyj_mc3, start_val, end_val), dy_kwargs_mc3))
# if pt_ind == 0:
# zpj_2d_entries.append((h2d_dyj_mc3, dy_kwargs_mc3))
# else:
# zpj_entries.append(None)
####################
# DIJET CENTRAL REGION
####################
# JETHT/ZEROBIAS DATA
if jetht_zb_tfile:
h2d_qcd_cen_data = jetht_zb_tfile.get("%s/%s" % (dj_cen_dirname, v)) # use already merged jetht+zb
qcd_cen_kwargs_data = dict(line_color=qgc.JETHT_COLOUR, line_width=data_line_width, fill_color=qgc.JETHT_COLOUR,
marker_color=qgc.JETHT_COLOUR, marker_style=cu.Marker.get(qgc.QCD_MARKER), marker_size=msize,
label="Data")
dijet_cen_data_hist = qgp.get_projection_plot(h2d_qcd_cen_data, start_val, end_val)
entries.append((dijet_cen_data_hist, qcd_cen_kwargs_data))
dijet_cen_entries.append((dijet_cen_data_hist, qcd_cen_kwargs_data))
if pt_ind == 0:
dijet_cen_2d_entries.append((h2d_qcd_cen_data, qcd_cen_kwargs_data))
# MG+PYTHIA QCD MC
if qcd_tfile:
h2d_qcd_cen_mc = qcd_tfile.get("%s/%s" % (dj_cen_dirname, v))
qcd_cen_kwargs_mc = dict(line_color=qgc.QCD_COLOUR, line_width=lw, fill_color=qgc.QCD_COLOUR,
marker_color=qgc.QCD_COLOUR, marker_style=cu.Marker.get(qgc.QCD_MARKER), marker_size=mc_msize,
label=mgpy_label,
subplot=dijet_cen_data_hist)
dijet_mgpy_hist = qgp.get_projection_plot(h2d_qcd_cen_mc, start_val, end_val)
entries.append((dijet_mgpy_hist, qcd_cen_kwargs_mc))
dijet_cen_entries.append((dijet_mgpy_hist, qcd_cen_kwargs_mc))
if pt_ind == 0:
dijet_cen_2d_entries.append((h2d_qcd_cen_mc, qcd_cen_kwargs_mc))
# PYTHIA ONLY
# if qcd_py_tfile:
# col = qgc.QCD_COLOURS[2]
# h2d_qcd_cen_mc2 = qcd_py_tfile.get("%s/%s" % (dj_cen_dirname, v))
# qcd_cen_kwargs_mc2 = dict(line_color=col, line_width=lw, fill_color=col,
# marker_color=col, marker_style=cu.Marker.get(qgc.QCD_MARKER), marker_size=mc_msize,
# label="Pythia8",
# subplot=dijet_cen_data_hist)
# entries.append((qgp.get_projection_plot(h2d_qcd_cen_mc2, start_val, end_val), qcd_cen_kwargs_mc2))
# dijet_cen_entries.append((qgp.get_projection_plot(h2d_qcd_cen_mc2, start_val, end_val), qcd_cen_kwargs_mc2))
# if pt_ind == 0:
# dijet_cen_2d_entries.append((h2d_qcd_cen_mc2, qcd_cen_kwargs_mc2))
# HERWIG++ QCD
if qcd_hpp_tfile:
h2d_qcd_cen_mc_hpp = qcd_hpp_tfile.get("%s/%s" % (dj_cen_dirname, v))
qcd_cen_kwargs_mc_hpp = dict(line_color=qgc.HERWIGPP_QCD_COLOUR, line_width=lw, fill_color=qgc.HERWIGPP_QCD_COLOUR,
marker_color=qgc.HERWIGPP_QCD_COLOUR, marker_style=cu.Marker.get(qgc.QCD_MARKER), marker_size=mc_msize,
label=hpp_label,
subplot=dijet_cen_data_hist)
dijet_hpp_hist = qgp.get_projection_plot(h2d_qcd_cen_mc_hpp, start_val, end_val)
entries.append((dijet_hpp_hist, qcd_cen_kwargs_mc_hpp))
dijet_cen_entries.append((dijet_hpp_hist, qcd_cen_kwargs_mc_hpp))
if pt_ind == 0:
dijet_cen_2d_entries.append((h2d_qcd_cen_mc_hpp, qcd_cen_kwargs_mc_hpp))
####################
# DIJET FORWARD REGION
####################
# JETHT/ZEROBIAS DATA
if jetht_zb_tfile:
h2d_qcd_fwd_data = jetht_zb_tfile.get("%s/%s" % (dj_fwd_dirname, v)) # use already merged jetht+zb
qcd_fwd_kwargs_data = dict(line_color=qgc.JETHT_COLOUR, line_width=data_line_width, fill_color=qgc.JETHT_COLOUR,
marker_color=qgc.JETHT_COLOUR, marker_style=cu.Marker.get('triangleDown'), marker_size=msize,
label="Data")
dijet_fwd_data_hist = qgp.get_projection_plot(h2d_qcd_fwd_data, start_val, end_val)
entries.append((dijet_fwd_data_hist, qcd_fwd_kwargs_data))
dijet_fwd_entries.append((dijet_fwd_data_hist, qcd_fwd_kwargs_data))
if pt_ind == 0:
dijet_fwd_2d_entries.append((h2d_qcd_fwd_data, qcd_fwd_kwargs_data))
# MG+PYTHIA QCD MC
if qcd_tfile:
h2d_qcd_fwd_mc = qcd_tfile.get("%s/%s" % (dj_fwd_dirname, v))
qcd_fwd_kwargs_mc = dict(line_color=qgc.QCD_COLOUR, line_width=lw, fill_color=qgc.QCD_COLOUR,
marker_color=qgc.QCD_COLOUR, marker_style=cu.Marker.get(qgc.QCD_MARKER), marker_size=mc_msize,
label=mgpy_label,
subplot=dijet_fwd_data_hist)
dijet_mgpy_hist = qgp.get_projection_plot(h2d_qcd_fwd_mc, start_val, end_val)
entries.append((dijet_mgpy_hist, qcd_fwd_kwargs_mc))
dijet_fwd_entries.append((dijet_mgpy_hist, qcd_fwd_kwargs_mc))
if pt_ind == 0:
dijet_fwd_2d_entries.append((h2d_qcd_fwd_mc, qcd_fwd_kwargs_mc))
# PYTHIA ONLY
# if qcd_py_tfile:
# col = qgc.QCD_COLOURS[2]
# h2d_qcd_fwd_mc2 = qcd_py_tfile.get("%s/%s" % (dj_fwd_dirname, v))
# qcd_fwd_kwargs_mc2 = dict(line_color=col, line_width=lw, fill_color=col,
# marker_color=col, marker_style=cu.Marker.get(qgc.QCD_MARKER), marker_size=mc_msize,
# label="Pythia8",
# subplot=dijet_fwd_data_hist)
# entries.append((qgp.get_projection_plot(h2d_qcd_fwd_mc2, start_val, end_val), qcd_fwd_kwargs_mc2))
# dijet_fwd_entries.append((qgp.get_projection_plot(h2d_qcd_fwd_mc2, start_val, end_val), qcd_fwd_kwargs_mc2))
# if pt_ind == 0:
# dijet_fwd_2d_entries.append((h2d_qcd_fwd_mc2, qcd_fwd_kwargs_mc2))
# HERWIG++ QCD
if qcd_hpp_tfile:
h2d_qcd_fwd_mc_hpp = qcd_hpp_tfile.get("%s/%s" % (dj_fwd_dirname, v))
qcd_fwd_kwargs_mc_hpp = dict(line_color=qgc.HERWIGPP_QCD_COLOUR, line_width=lw, fill_color=qgc.HERWIGPP_QCD_COLOUR,
marker_color=qgc.HERWIGPP_QCD_COLOUR, marker_style=cu.Marker.get(qgc.QCD_MARKER), marker_size=mc_msize,
label=hpp_label,
subplot=dijet_fwd_data_hist)
dijet_hpp_hist = qgp.get_projection_plot(h2d_qcd_fwd_mc_hpp, start_val, end_val)
entries.append((dijet_hpp_hist, qcd_fwd_kwargs_mc_hpp))
dijet_fwd_entries.append((dijet_hpp_hist, qcd_fwd_kwargs_mc_hpp))
if pt_ind == 0:
dijet_fwd_2d_entries.append((h2d_qcd_fwd_mc_hpp, qcd_fwd_kwargs_mc_hpp))
zpj_1d_entries.append(zpj_entries)
dijet_cen_1d_entries.append(dijet_cen_entries)
dijet_fwd_1d_entries.append(dijet_fwd_entries)
#################
# SETUP PLOTTING
#################
rebin = 2
v_lower = v.lower()
if "multiplicity" in v_lower:
rebin = 1
# elif "flavour" in v_lower or "thrust" in v_lower:
# rebin = 1
# elif 'ptd' in v_lower:
# rebin = 4
# elif 'lha_charged' in v_lower:
# rebin = 4
# rebin = 1
xlim = None
if "width" in v_lower or "ptd" in v_lower:
xlim = (0, 1)
elif "thrust" in v_lower:
xlim = (0, 0.5)
# elif "multiplicity" in v_lower and "ak4" in root_dir.lower():
# if end_val <= 150:
# xlim = (0, 50)
# else:
# xlim = (0, 80)
auto_xlim = False
if "multiplicity" in v_lower or "thrust" in v_lower:
auto_xlim = True
ylim = [0, None]
if "flavour" in v_lower:
ylim = (0, 1)
# elif "lha" in v_lower:
# ylim = (0, 5)
plot_dir = os.path.join(root_dir, "plots_lambda_mc_vs_data%s" % (gr_append))
subplot_title = "Simulation / Data"
subplot_limits = (0, 2)
xlabel = ang.name + " (" + ang.lambda_str + ")"
if gr_append is not "":
xlabel = "Groomed " + ang.name + " (" + ang.lambda_str + ")"
def _title(region_str, start_val, end_val):
pt_var_str = "p_{T}^{jet}"
s = (("{jet_algo}\n"
"{region_label}\n"
"{bin_edge_low:g} < {pt_str} < {bin_edge_high:g} GeV")
.format(
jet_algo=jet_str,
region_label=region_str,
pt_str=pt_var_str,
bin_edge_low=start_val,
bin_edge_high=end_val))
if title is not None:
s = "%s\n%s" % (s, title)
return s
has_data = True
draw_opt = "NOSTACK HIST E1"
# dj central only
# dont' try to do multiple signal regions per plot, it looks rubbish
if len(dijet_cen_entries) > 0:
qgp.do_comparison_plot(dijet_cen_entries,
"%s/ptBinned/%s_pt%dto%d_dijet_central.%s" % (plot_dir, v, start_val, end_val, OUTPUT_FMT),
rebin=rebin,
draw_opt=draw_opt,
title=_title(qgc.Dijet_CEN_LABEL, start_val, end_val),
xtitle=xlabel,
xlim='auto' if auto_xlim else xlim, # don't use calc_auto_xlim, since do_comparison_plot will rebin it anyway
ylim=ylim,
data_first=has_data,
mean_rel_error=0.4,
subplot_type='ratio',
subplot_title=subplot_title,
subplot_limits=subplot_limits,
lumi=cu.get_lumi_str(do_dijet=True, do_zpj=False))
# dj forward only
if len(dijet_fwd_entries) > 0:
qgp.do_comparison_plot(dijet_fwd_entries,
"%s/ptBinned/%s_pt%dto%d_dijet_forward.%s" % (plot_dir, v, start_val, end_val, OUTPUT_FMT),
rebin=rebin,
draw_opt=draw_opt,
title=_title(qgc.Dijet_FWD_LABEL, start_val, end_val),
xtitle=xlabel,
xlim='auto' if auto_xlim else xlim,
ylim=ylim,
data_first=has_data,
mean_rel_error=0.4,
subplot_type='ratio',
subplot_title=subplot_title,
subplot_limits=subplot_limits,
lumi=cu.get_lumi_str(do_dijet=True, do_zpj=False))
# zpj only
if len(zpj_entries) > 0:
if start_val > 149:
# rebin *= 2
rebin += 1
# find nearest divisor
while (zpj_entries[0][0].GetNbinsX() % rebin != 0):
rebin += 1
if "multiplicity" in v_lower:
if start_val > 300:
rebin += 1
# find nearest divisor
while (zpj_entries[0][0].GetNbinsX() % rebin != 0):
rebin += 1
qgp.do_comparison_plot(zpj_entries,
"%s/ptBinned/%s_pt%dto%d_zpj.%s" % (plot_dir, v, start_val, end_val, OUTPUT_FMT),
rebin=rebin,
draw_opt=draw_opt,
title=_title(qgc.ZpJ_LABEL, start_val, end_val),
xtitle=xlabel,
xlim=qgp.calc_auto_xlim([d[0] for d in zpj_entries]) if auto_xlim else xlim,
ylim=ylim,
data_first=has_data,
mean_rel_error=0.4,
subplot_type='ratio',
subplot_title=subplot_title,
subplot_limits=subplot_limits,
lumi=cu.get_lumi_str(do_dijet=False, do_zpj=True))
# Do overall summary plots across all pt bins
# ------------------------------------------------------------------
ylim_mean = None
if "width" in v_lower or "ptd" in v_lower:
ylim_mean = (0, 0.4)
elif"thrust" in v_lower:
ylim_mean = (0, 0.5)
elif "multiplicity" in v_lower and "ak4" in root_dir.lower():
ylim_mean = (0, 100)
ylim_mean = (0, 80)
if end_val < 150:
ylim_mean = (0, 50)
ylim_mean=None
ylim_rms=None
# Setup variable names for MPL
marker = ""
if "_" in ang.name or "^" in ang.name:
marker = "$"
var_label = marker + ang.name + marker + " ($%s$)" % ang.lambda_str
if gr_append != "":
var_label = "Groomed " + marker + ang.name + marker + " ($%s$)" % ang.lambda_str
# construct a dataframe from the hists
# df = construct_dataframe_from_hists(dijet_cen_entries=dijet_cen_1d_entries,
# dijet_fwd_entries=dijet_fwd_entries,
# zpj_entries=zpj_entries)
# pt_low = 88 if "8" in radius else 50
pt_low = pt_bins[0][0]
qgp.do_mean_rms_summary_plot(dijet_cen_1d_entries,
pt_bins,
"%s/ptBinned/%s_box_dijet_cen_mpl.%s" % (plot_dir, v, OUTPUT_FMT),
var_label=var_label,
xlim=(pt_low, 4000),
ylim_mean=ylim_mean,
ylim_rms=ylim_rms,
region_title="%s jets in %s" % (jet_str, qgc.Dijet_CEN_LABEL.lower()))
qgp.do_mean_rms_summary_plot(dijet_fwd_1d_entries,
pt_bins,
"%s/ptBinned/%s_box_dijet_fwd_mpl.%s" % (plot_dir, v, OUTPUT_FMT),
var_label=var_label,
xlim=(pt_low, 4000),
ylim_mean=ylim_mean,
ylim_rms=ylim_rms,
region_title="%s jets in %s" % (jet_str, qgc.Dijet_FWD_LABEL.lower()))
# zpj_1d_entries[i][j] is the jth sample in the ith pt bin
qgp.do_mean_rms_summary_plot(zpj_1d_entries,
pt_bins,
"%s/ptBinned/%s_box_zpj_mpl.%s" % (plot_dir, v, OUTPUT_FMT),
var_label=var_label,
xlim=(pt_low, 800),
ylim_mean=ylim_mean,
ylim_rms=ylim_rms,
region_title="%s jets in %s" % (jet_str, qgc.ZpJ_LABEL))
def do_plots(root_dir):
# QG variable plots
pt_bins = qgc.PT_BINS[:]
var_list = qgc.COMMON_VARS
var_prepend = ""
radius, pus = cu.get_jet_config_from_dirname(root_dir)
jet_str = "AK%s PF %s" % (radius.upper(), pus.upper())
for gr_append in ["", "_groomed"]:
if gr_append == "_groomed":
print("Doing groomed plots...")
else:
print("Doing ungroomed plots...")
# do a pt plot just for sanity
if gr_append == "":
make_1d_plot(root_dir, "ZPlusJets_gen/pt_jet1", jet_str, gen_only=True)
make_1d_plot(root_dir, "ZPlusJets_gen/pt_mumu", jet_str, gen_only=True)
make_1d_plot(root_dir, "ZPlusJets/pt_jet1", jet_str)
make_1d_plot(root_dir, "ZPlusJets/pt_mumu", jet_str)
continue
zpj_dirname = "ZPlusJets_QG%s" % (gr_append)
for ang in var_list[:]:
print("...Doing", ang.name)
v = "%s%s_vs_pt" % (var_prepend, ang.var)
zpj_2d_entries = []
zpj_1d_entries = []
for pt_ind, (start_val, end_val) in enumerate(pt_bins):
entries = []
zpj_entries = []
# Get all plots
lw = 2
msize = 1.1
data_line_width = 0
####################
# Z+JETS REGION
####################
# SINGLE MU DATA
h2d_dyj_data = grab_obj(os.path.join(root_dir, qgc.SINGLE_MU_FILENAME), "%s/%s" % (zpj_dirname, v))
dy_kwargs_data = dict(line_color=qgc.SINGLE_MU_COLOUR, line_width=data_line_width, fill_color=qgc.SINGLE_MU_COLOUR,
marker_color=qgc.SINGLE_MU_COLOUR, marker_style=cu.Marker.get(qgc.DY_MARKER), marker_size=msize,
label="Data")
zpj_data_hist = qgp.get_projection_plot(h2d_dyj_data, start_val, end_val)
entries.append((zpj_data_hist, dy_kwargs_data))
zpj_entries.append((zpj_data_hist, dy_kwargs_data))
if pt_ind == 0:
zpj_2d_entries.append((h2d_dyj_data, dy_kwargs_data))
# PYTHIA DY MC WITH k FACTORS
h2d_dyj_mc = grab_obj(os.path.join(root_dir, qgc.DY_FILENAME), "%s/%s" % (zpj_dirname, v))
dy_kwargs_mc = dict(line_color=qgc.DY_COLOUR, line_width=lw, fill_color=qgc.DY_COLOUR,
marker_color=qgc.DY_COLOUR, marker_style=cu.Marker.get(qgc.DY_MARKER), marker_size=0,
label="MG+PY8 (w/ k-factor)",
subplot=zpj_data_hist)
entries.append((qgp.get_projection_plot(h2d_dyj_mc, start_val, end_val), dy_kwargs_mc))
zpj_entries.append((qgp.get_projection_plot(h2d_dyj_mc, start_val, end_val), dy_kwargs_mc))
if pt_ind == 0:
zpj_2d_entries.append((h2d_dyj_mc, dy_kwargs_mc))
# WIHTOUT K FACTORS
h2d_dyj_mc_noWeight = grab_obj(os.path.join(root_dir, "uhh2.AnalysisModuleRunner.MC.MC_DYJetsToLL_noZReweight.root"), "%s/%s" % (zpj_dirname, v))
dy_kwargs_mc_noWeight = dict(line_color=qgc.DY_COLOURS[0], line_width=lw, fill_color=qgc.DY_COLOURS[0], line_style=2,
marker_color=qgc.DY_COLOURS[0], marker_style=cu.Marker.get(qgc.DY_MARKER), marker_size=0,
label="MG+PY8 (w/out k-factor)",
subplot=zpj_data_hist)
entries.append((qgp.get_projection_plot(h2d_dyj_mc_noWeight, start_val, end_val), dy_kwargs_mc_noWeight))
zpj_entries.append((qgp.get_projection_plot(h2d_dyj_mc_noWeight, start_val, end_val), dy_kwargs_mc_noWeight))
if pt_ind == 0:
zpj_2d_entries.append((h2d_dyj_mc_noWeight, dy_kwargs_mc_noWeight))
# HERWIG++ DY
# if end_val < 255:
# h2d_dyj_mc2 = grab_obj(os.path.join(root_dir, qgc.DY_HERWIG_FILENAME), "%s/%s" % (zpj_dirname, v))
# col3 = qgc.DY_COLOURS[2]
# dy_kwargs_mc2 = dict(line_color=col3, line_width=lw, fill_color=col3,
# marker_color=col3, marker_style=cu.Marker.get(qgc.DY_MARKER), marker_size=0,
# label="H++",
# subplot=zpj_data_hist)
# entries.append((qgp.get_projection_plot(h2d_dyj_mc2, start_val, end_val), dy_kwargs_mc2))
# zpj_entries.append((qgp.get_projection_plot(h2d_dyj_mc2, start_val, end_val), dy_kwargs_mc2))
# if pt_ind == 0:
# zpj_2d_entries.append((h2d_dyj_mc2, dy_kwargs_mc2))
# else:
# zpj_entries.append(None)
# MG+HERWIG++ DY
# if end_val < 151:
# h2d_dyj_mc3 = grab_obj(os.path.join(root_dir, qgc.DY_MG_HERWIG_FILENAME), "%s/%s" % (zpj_dirname, v))
# col4 = qgc.DY_COLOURS[3]
# dy_kwargs_mc3 = dict(line_color=col4, line_width=lw, fill_color=col4,
# marker_color=col4, marker_style=cu.Marker.get(qgc.DY_MARKER), marker_size=0,
# label="MG+H++",
# subplot=zpj_data_hist)
# entries.append((qgp.get_projection_plot(h2d_dyj_mc3, start_val, end_val), dy_kwargs_mc3))
# zpj_entries.append((qgp.get_projection_plot(h2d_dyj_mc3, start_val, end_val), dy_kwargs_mc3))
# if pt_ind == 0:
# zpj_2d_entries.append((h2d_dyj_mc3, dy_kwargs_mc3))
# else:
# zpj_entries.append(None)
zpj_1d_entries.append(zpj_entries)
#################
# SETUP PLOTTING
#################
rebin = 2
v_lower = v.lower()
if "multiplicity" in v_lower:
rebin = 2
elif "flavour" in v_lower or "thrust" in v_lower:
rebin = 1
elif 'ptd' in v_lower:
rebin = 2
xlim = None
if "width" in v_lower or "ptd" in v_lower:
xlim = (0, 1)
elif"thrust" in v_lower:
xlim = (0, 0.5)
elif "multiplicity" in v_lower and "ak4" in root_dir.lower():
if end_val <= 150:
xlim = (0, 50)
else:
xlim = (0, 80)
ylim = None
if "flavour" in v_lower:
ylim = (0, 1)
# elif "lha" in v_lower:
# ylim = (0, 5)
plot_dir = os.path.join(root_dir, "plots_dy_vs_qcd_mc_vs_data%s_z_reweight" % (gr_append))
subplot_title = "MC / Data"
subplot_limits = (0.5, 1.5)
xlabel = ang.name + " (" + ang.lambda_str + ")"
if gr_append is not "":
xlabel = "Groomed " + ang.name + " (" + ang.lambda_str + ")"
# zpj only
qgp.do_comparison_plot(zpj_entries,
"%s/ptBinned/%s_pt%dto%d_zpj.%s" % (plot_dir, v, start_val, end_val, OUTPUT_FMT),
rebin=rebin,
title="%d < p_{T}^{jet} < %d GeV\n%s\n%s" % (start_val, end_val, jet_str, qgc.ZpJ_LABEL),
xtitle=xlabel,
xlim=xlim,
ylim=ylim,
subplot_type='ratio',
subplot_title=subplot_title,
subplot_limits=subplot_limits)
# Do overall summary plots across all pt bins
# ------------------------------------------------------------------
ylim_mean = None
if "width" in v_lower or "ptd" in v_lower:
ylim_mean = (0, 0.4)
elif"thrust" in v_lower:
ylim_mean = (0, 0.5)
elif "multiplicity" in v_lower and "ak4" in root_dir.lower():
ylim_mean = (0, 100)
ylim_mean = (0, 80)
if end_val < 150:
ylim_mean = (0, 50)
ylim_mean=None
ylim_rms=None
# Setup variable names for MPL
marker = ""
if "_" in ang.name or "^" in ang.name:
marker = "$"
var_label = marker + ang.name + marker + " ($%s$)" % ang.lambda_str
if gr_append != "":
var_label = "Groomed " + marker + ang.name + marker + " ($%s$)" % ang.lambda_str
# zpj_1d_entries[i][j] is the jth sample in the ith pt bin
qgp.do_mean_rms_summary_plot(zpj_1d_entries[:],
pt_bins[:],
"%s/ptBinned/%s_box_zpj_mpl.%s" % (plot_dir, v, OUTPUT_FMT),
var_label=var_label,
xlim=(50, 614),
ylim_mean=ylim_mean,
ylim_rms=ylim_rms,
region_title="%s jets in %s" % (jet_str, qgc.ZpJ_LABEL))
def do_gen_reco_comparison_plots(root_dir,
var_list=None,
gen_var_prepend="gen",
reco_var_prepend="",
plot_dir="plot_reco_gen",
zpj_reco_dirname=qgc.ZPJ_RECOJET_RDIR,
dj_reco_dirname=qgc.DJ_RECOJET_RDIR,
zpj_gen_dirname=qgc.ZPJ_GENJET_RDIR,
dj_gen_dirname=qgc.DJ_GENJET_RDIR,
pt_bins=qgc.THEORY_PT_BINS,
subplot_type=None):
var_list = var_list or qgc.COMMON_VARS
for ang in var_list:
v_reco = "%s%s_vs_pt" % (reco_var_prepend, ang.var)
v_gen = "%s%s_vs_pt" % (gen_var_prepend, ang.var)
h2d_dyj_reco = grab_obj(os.path.join(root_dir, qgc.DY_FILENAME),
"%s/%s" % (zpj_reco_dirname, v_reco))
h2d_qcd_reco = grab_obj(os.path.join(root_dir, qgc.QCD_FILENAME),
"%s/%s" % (dj_reco_dirname, v_reco))
h2d_dyj_gen = grab_obj(os.path.join(root_dir, qgc.DY_FILENAME),
"%s/%s" % (zpj_gen_dirname, v_gen))
h2d_qcd_gen = grab_obj(os.path.join(root_dir, qgc.QCD_FILENAME),
"%s/%s" % (dj_gen_dirname, v_gen))
if "flavour" not in v_reco:
h2d_dyj_reco_q = grab_obj(os.path.join(root_dir, qgc.DY_FILENAME),
"%s/q%s" % (zpj_reco_dirname, v_reco))
h2d_qcd_reco_g = grab_obj(os.path.join(root_dir, qgc.QCD_FILENAME),
"%s/g%s" % (dj_reco_dirname, v_reco))
h2d_dyj_gen_q = grab_obj(os.path.join(root_dir, qgc.DY_FILENAME),
"%s/q%s" % (zpj_gen_dirname, v_gen))
h2d_qcd_gen_g = grab_obj(os.path.join(root_dir, qgc.QCD_FILENAME),
"%s/g%s" % (dj_gen_dirname, v_gen))
for (start_val, end_val) in pt_bins:
lw = 2
dy_reco_kwargs = dict(line_color=qgc.DY_COLOUR,
fill_color=qgc.DY_COLOUR,
label=qgc.DY_ZpJ_LABEL + " [RecoJet]",
line_width=lw)
qcd_reco_kwargs = dict(line_color=qgc.QCD_COLOUR,
fill_color=qgc.QCD_COLOUR,
label=qgc.QCD_Dijet_LABEL + " [RecoJet]",
line_width=lw)
dy_gen_kwargs = dict(line_color=qgc.DY_COLOUR,
fill_color=qgc.DY_COLOUR,
label=qgc.DY_ZpJ_LABEL + " [GenJet]",
line_width=lw,
line_style=2)
qcd_gen_kwargs = dict(line_color=qgc.QCD_COLOUR,
fill_color=qgc.QCD_COLOUR,
label=qgc.QCD_Dijet_LABEL + " [GenJet]",
line_width=lw,
line_style=2)
entries = [(qgg.get_projection_plot(h2d_dyj_reco, start_val,
end_val), dy_reco_kwargs),
(qgg.get_projection_plot(h2d_qcd_reco, start_val,
end_val), qcd_reco_kwargs),
(qgg.get_projection_plot(h2d_dyj_gen, start_val,
end_val), dy_gen_kwargs),
(qgg.get_projection_plot(h2d_qcd_gen, start_val,
end_val), qcd_gen_kwargs)]
rebin = 2
xlim = None
ylim = None
if "flavour" in v_reco:
rebin = 1
ylim = (0, 1)
if "thrust" in v_reco:
xlim = (0, 0.5)
qgg.do_comparison_plot(
entries,
"%s/%s/%s_pt%dto%d.%s" %
(root_dir, plot_dir, ang.var, start_val, end_val, OUTPUT_FMT),
rebin=rebin,
title="%d < p_{T}^{jet} < %d GeV" % (start_val, end_val),
xtitle=ang.name + " (" + ang.lambda_str + ")",
xlim=xlim,
ylim=ylim,
subplot_type=subplot_type)
# Do flavour-tagged comparison
if "flavour" in v_reco:
continue
dy_reco_kwargs['label'] = qgc.DY_ZpJ_QFLAV_LABEL + " [RecoJet]"
qcd_reco_kwargs['label'] = qgc.QCD_Dijet_GFLAV_LABEL + " [RecoJet]"
dy_gen_kwargs['label'] = qgc.DY_ZpJ_QFLAV_LABEL + " [GenJet]"
qcd_gen_kwargs['label'] = qgc.QCD_Dijet_GFLAV_LABEL + " [GenJet]"
entries = [(qgg.get_projection_plot(h2d_dyj_reco_q, start_val,
end_val), dy_reco_kwargs),
(qgg.get_projection_plot(h2d_qcd_reco_g, start_val,
end_val), qcd_reco_kwargs),
(qgg.get_projection_plot(h2d_dyj_gen_q, start_val,
end_val), dy_gen_kwargs),
(qgg.get_projection_plot(h2d_qcd_gen_g, start_val,
end_val), qcd_gen_kwargs)]
qgg.do_comparison_plot(
entries,
"%s/%s/%s_flavMatched_pt%dto%d.%s" %
(root_dir, plot_dir, ang.var, start_val, end_val, OUTPUT_FMT),
rebin=rebin,
title="%d < p_{T}^{jet} < %d GeV" % (start_val, end_val),
xtitle=ang.name + " (" + ang.lambda_str + ")",
xlim=xlim,
ylim=ylim,
subplot_type=subplot_type)