def region_comparison_plot(tag):
for year in [2017,2018]:
regions = ['1m', '2m', '1e','1m_hlt','2m_hlt']
opts = markers('data')
opts['markersize'] = 5.
opts['fillstyle'] = 'none'
emarker = opts.pop('emarker', '')
fig, ax, rax = fig_ratio()
x, y, yerr = {}, {}, {}
for region in regions:
if region.endswith('e'):
file = f'output/{tag}/table_{region}_EGamma_{year}.txt'
else:
file = f'output/{tag}/table_{region}_SingleMuon_{year}.txt'
x[region], y[region], yerr[region] = get_xy(file)
opts['color'] = colors[region]
ax.errorbar(x[region], y[region], yerr=yerr[region],label=f'{region} region', **opts)
# opts.pop('elinewidth')
if region=='1m':
continue
rax.errorbar(x['1m'], y[region]/y['1m'], yerr[region]/y['1m'], **opts)
# for f in files: plot(f)
outdir = f"./output/{tag}"
# ax.set_ylim(0.9,1)
ax.legend()
ax.set_ylabel("Efficiency")
ax.xaxis.set_major_locator(MultipleLocator(200))
ax.xaxis.set_minor_locator(MultipleLocator(50))
ax.yaxis.set_major_locator(MultipleLocator(0.05))
ax.yaxis.set_minor_locator(MultipleLocator(0.01))
ax.set_ylim(0.9,1.02)
ax.grid(1)
rax.set_ylim(0.9,1.1)
rax.grid(1)
rax.set_xlabel("Recoil or $p_{T}^{miss}$ (GeV)")
rax.set_ylabel(r"Ratio to single-$\mu$")
plt.text(1., 1., r"$\approx$ %.1f fb$^{-1}$ (13 TeV)" % lumi(year),
fontsize=16,
horizontalalignment='right',
verticalalignment='bottom',
transform=ax.transAxes
)
plt.text(0., 1., f'{year}',
fontsize=16,
horizontalalignment='left',
verticalalignment='bottom',
transform=ax.transAxes
)
fig.savefig(pjoin(outdir, f'region_comparison_data_{year}.pdf'))
fig.clear()
plt.close(fig)
def pdf_plot(acc):
outdir = './output/pdfstudy/'
if not os.path.exists(outdir):
os.makedirs(outdir)
datasets = [
'WJetsToLNu_HT_MLM_2017',
'DYJetsToLL_M-50_HT_MLM_2017',
]
for ds in datasets:
fig, ax, rax = fig_ratio()
h = acc['gen_vpt']
h = h.rebin(h.axis('vpt'),
hist.Bin("vpt", r"$p_{T}^{V}$ (GeV)", 10, 0, 2000))
h = merge_extensions(h, acc, reweight_pu=False)
scale_xs_lumi(h)
h = merge_datasets(h)
h = h.project(h.axis('dataset'), ds)
for pdf in h.axis('pdf').identifiers():
if str(pdf) == 'none':
continue
data_err_opts['color'] = colors[str(pdf)]
hist.plot1d(
h.project('pdf', pdf),
# overlay='pdf',
error_opts=data_err_opts,
ax=ax,
overflow='all',
clear=False)
hist.plotratio(
h.project('pdf', pdf),
h.project('pdf', 'none'),
ax=rax,
denom_fill_opts={},
guide_opts={},
unc='num',
overflow='all',
error_opts=data_err_opts,
clear=False,
)
ax.set_ylim(1e-3, 1e8)
rax.set_ylim(0.9, 1.6)
ax.set_yscale('log')
leg = ax.legend()
for i, pdf in enumerate(h.axis('pdf').identifiers()):
if str(pdf) == 'none':
continue
leg.get_texts()[i].set_text(str(pdf))
fig.savefig(pjoin(outdir, f'{ds}.pdf'))
plt.close(fig)
def main():
"""
A script to easily compare template files between different runs.
Usage: ./compare_templates.py /path/to/first/template_file.root /path/to/second/template_file.root
The script will loop over all templates in the files and create a comparison plot for each of them.
All plots are dumped into a folder for inspection.
"""
args = parse_commandline()
# Based on input locations, derive tag names to identify the files
tag1 = os.path.basename(os.path.dirname(args.fname1))
tag2 = os.path.basename(os.path.dirname(args.fname2))
# filter
regex = re.compile(args.filter)
# Convert to dictionary
h1 = make_dict(args.fname1)
h2 = make_dict(args.fname2)
# Make sure the two files have consistent keys
# assert(h1.keys()==h2.keys())
# Create plot folder
if not os.path.exists(args.outdir):
os.makedirs(args.outdir)
# Do the actual plotting
table = []
for key in tqdm.tqdm(h1.keys()):
if key not in h2:
print("Found missing key ", key)
if not regex.match(str(key)):
continue
fig, ax, rax = fig_ratio()
x = 0.5 * np.sum(h1[key].bins, axis=1)
edges = np.unique(h1[key].bins)
try:
v1 = np.sum(h1[key].allvalues)
v2 = np.sum(h2[key].allvalues)
if v2 == v1:
ratio = 1
elif v1 == 0:
ratio = "-"
else:
ratio = v2 / v1
table.append([key, v1, v2, ratio])
except KeyError:
plt.close(fig)
continue
hep.histplot(
sum_overflow_into_last_bin(h1[key].allvalues),
edges,
ax=ax,
label=f"{tag1}, Integral={np.sum(h1[key].values):.1f}",
color='navy',
)
hep.histplot(sum_overflow_into_last_bin(h2[key].allvalues),
edges,
yerr=np.sqrt(h2[key].variances),
ax=ax,
label=f"{tag2}, Integral={np.sum(h2[key].values):.1f}",
color='crimson',
marker='o',
markersize=5,
histtype='errorbar')
ax.legend()
# Bottom panel: ratio plot
valid = h1[key].values != 0
rax.errorbar(
x[valid],
h2[key].values[valid] / h1[key].values[valid],
np.sqrt(h2[key].variances[valid]) / h1[key].values[valid],
linestyle='none',
marker='o',
color="crimson",
)
# Add indicators for bins where we could not calculate the ratio
if np.any(~valid):
rax.plot(0.5 * np.sum(h1[key].bins, axis=1)[~valid],
np.ones(np.sum(~valid)),
'x',
color="k",
fillstyle="none")
# Aesthetics
ax.set_title(key)
rax.set_ylim(*map(float, args.rlim.split(",")))
rax.set_xlabel("Recoil (GeV)")
rax.set_ylabel("Ratio")
ax.set_ylabel("Events / bin")
ax.set_yscale("log")
try:
ax.set_ylim(
0.5 * min(h2[key].values[h2[key].values > 0]),
1.5 * max(h2[key].values),
)
except ValueError:
plt.close(fig)
continue
rax.grid(linestyle='--')
fig.savefig(pjoin(args.outdir, f"{key}.png"))
plt.close(fig)
print(tabulate(table))
def sf_comparison_plot(tag):
for year in [2017,2018]:
regions = ['1m', '2m', '1m_hlt', '2m_hlt']
opts = markers('data')
opts['markersize'] = 5
opts['fillstyle'] = 'none'
emarker = opts.pop('emarker', '')
fig, ax, rax = fig_ratio()
x, y, yerr = {}, {}, {}
for region in regions:
if '1e' in region:
fnum = f'output/{tag}/table_{region}_EGamma_{year}.txt'
fden = f'output/{tag}/table_{region}_WJetsToLNu-MLM_{year}.txt'
elif '1m' in region:
fnum = f'output/{tag}/table_{region}_SingleMuon_{year}.txt'
fden = f'output/{tag}/table_{region}_WJetsToLNu-MLM_{year}.txt'
elif '2m' in region:
fnum = f'output/{tag}/table_{region}_SingleMuon_{year}.txt'
fden = f'output/{tag}/table_{region}_DYNJetsToLL_M-50-MLM_{year}.txt'
xnum, ynum, yerrnum = get_xy(fnum)
xden, yden, yerrden = get_xy(fden)
x[region] = xnum
y[region] = ynum / yden
yerr[region] = ratio_unc(ynum, yden, yerrnum, yerrden)
opts['color'] = colors[region]
opts['marker'] = region_marker[region]
ax.errorbar(x[region], y[region], yerr=yerr[region],label=f'{region} region', **opts)
# opts.pop('elinewidth')
if region=='1m':
continue
rax.errorbar(x['1m'], y[region]/y['1m'], ratio_unc(y[region],y['1m'],yerr[region],yerr['1m']), **opts)
# for f in files: plot(f)
outdir = f"./output/{tag}"
# ax.set_ylim(0.9,1)
ax.legend()
ax.set_ylabel("Data / MC SF")
ax.xaxis.set_major_locator(MultipleLocator(200))
ax.xaxis.set_minor_locator(MultipleLocator(50))
ax.yaxis.set_major_locator(MultipleLocator(0.05))
ax.yaxis.set_minor_locator(MultipleLocator(0.01))
ax.set_ylim(0.9,1.1)
ax.grid(1)
rax.set_ylim(0.95,1.05)
rax.yaxis.set_major_locator(MultipleLocator(0.05))
rax.yaxis.set_minor_locator(MultipleLocator(0.01))
rax.grid(1)
rax.set_xlabel("Recoil or $p_{T}^{miss}$ (GeV)")
rax.set_ylabel(r"Ratio to single-$\mu$")
plt.text(1., 1., r"$\approx$ %.1f fb$^{-1}$ (13 TeV)" % lumi(year),
fontsize=16,
horizontalalignment='right',
verticalalignment='bottom',
transform=ax.transAxes
)
plt.text(0., 1., f'{year}',
fontsize=16,
horizontalalignment='left',
verticalalignment='bottom',
transform=ax.transAxes
)
fig.savefig(pjoin(outdir, f'sf_comparison_{year}.pdf'))
fig.clear()
plt.close(fig)
def fit(tag, year):
outdir = './output/gamma/fit/'
try:
os.makedirs(outdir)
except FileExistsError:
pass
x, eff, eff_up, eff_down = {}, {}, {}, {}
x['data'], eff['data'], eff_up['data'], eff_down['data'] = load(tag, 'JetHT', year)
x['mc'], eff['mc'], eff_up['mc'], eff_down['mc'] = load(tag, 'GJets_HT_MLM', year)
pars = {}
cross = {}
for key in ['data','mc']:
pars[key], _ = curve_fit(sigmoid, x[key], eff[key], sigma=0.5*(eff_up[key]-eff_down[key]), p0=[1e-3,200,0.1,1])
cross[key] = minimize( lambda x: np.abs(sigmoid(x, *pars[key]) - 0.95),x0=230)
fig, ax, rax = fig_ratio()
xinterp = np.linspace(min(x['data']), max(x['data']), 1000)
handles = []
ax.errorbar(x['data'], eff['data'], 0.5*(eff_up['data']-eff_down['data']),fmt='o',label='Data',color=colors["data"])
ax.errorbar(x['mc'], eff['mc'], 0.5*(eff_up['mc']-eff_down['mc']),fmt='s',label='MC',fillstyle='none',color=colors["mc"])
ax.plot(xinterp, sigmoid(xinterp, *pars['data']), label='Data fit',color=colors["data"],zorder=-1)
ax.plot(xinterp, sigmoid(xinterp, *pars['mc']), label='MC fit',color=colors["mc"],zorder=-1,linestyle='--')
ax.set_ylim(0.,1.1)
ax.set_xlim(100,1100)
ax.legend()
ax.text(350,.4,'f(x) = c + (d-c) / (1 + exp(-a * (x-b)))')
ax.text(
300,
0.1,
'\n'.join([
f"a = {pars['data'][0]:.3f} / GeV",
f"b = {pars['data'][1]:.2f} GeV",
f"c = {pars['data'][2]:.3f}",
f"d = {pars['data'][3]:.3f}"
]),
color=colors['data']
)
ax.text(
600,
0.1,
'\n'.join([
f"a = {pars['mc'][0]:.3f} / GeV",
f"b = {pars['mc'][1]:.2f} GeV",
f"c = {pars['mc'][2]:.3f}",
f"d = {pars['mc'][3]:.3f}"
]),
color=colors['mc']
)
ax.text(700,0.8,
"\n".join([
f"Data > 95% @ {cross['data'].x[0]:.0f} GeV",
f"MC > 95% @ {cross['mc'].x[0]:.0f} GeV",
])
)
ax.text(1., 1., r"%.1f fb$^{-1}$ (13 TeV)" % lumi(year),
fontsize=16,
horizontalalignment='right',
verticalalignment='bottom',
transform=ax.transAxes
)
rax.set_ylim(0.95,1.1)
rax.grid()
# rax.plot([cross['data'].x[0],cross['data'].x[0]], [0.8,1.05],color='k',)
# ax.plot([cross['data'].x[0],cross['data'].x[0]], [0.,1.05],color='k',linestyle='--')
# ax.plot([cross['mc'].x[0],cross['mc'].x[0]], [0.9,1.05],color='r')
# ax.plot([215,215],[0.9,1.05])
rax.errorbar(x["data"], eff["data"] / eff["mc"], 0.5*(eff_up["data"] - eff_down["data"]) / eff["mc"],fmt='o',label='Data / MC',color=colors['data'])
rxinterp = np.linspace(cross['data'].x[0], max(xinterp),1000)
rax.plot(rxinterp, sigmoid(rxinterp, *pars['data']) / sigmoid(rxinterp, *pars['mc']),label=f"Data / MC fit ratio, plateau at {100*(sigmoid(rxinterp, *pars['data']) / sigmoid(rxinterp, *pars['mc']))[-1]:.1f} %",color='k')
rax.plot(rxinterp, 0.99*sigmoid(rxinterp, *pars['data']) / sigmoid(rxinterp, *pars['mc']), label='1% uncertainty on fit',linestyle='--',color='gray')
rax.plot(rxinterp, 1.01*sigmoid(rxinterp, *pars['data']) / sigmoid(rxinterp, *pars['mc']),linestyle='--',color='gray')
# rxinterp2 = np.linspace(min(xinterp),cross['data'].x[0], 1000)
# rax.plot(rxinterp2, sigmoid(rxinterp2, *pars['data']) / sigmoid(rxinterp2, *pars['mc']),color='k',linestyle=':')
rax.legend()
rax.set_ylabel("Ratio")
ax.set_ylabel("Trigger efficiency")
ax.set_xlabel("Photon $p_{T}$ (GeV)")
rax.set_xlabel("Photon $p_{T}$ (GeV)")
ax.figure.savefig(pjoin(outdir, f'fit_{tag}_{year}.pdf'))
ax.figure.clf()
def data_mc_comparison_plot(tag,
ymin=0,
ymax=1.1,
distribution='recoil',
jeteta_config=None,
output_format='pdf'):
if 'gamma' in tag:
regions = [
'g_HLT_PFHT1050', 'g_HLT_PFHT590', 'g_HLT_PFHT680',
'g_HLT_PFHT780', 'g_HLT_PFHT890'
]
elif 'recoil' in tag:
regions = ['1m', '2m']
else:
regions = ['1m', '2m', '1e', '2m_hlt']
opts = markers('data')
# opts['markersize'] = 5
# opts['fillstyle'] = 'none'
emarker = opts.pop('emarker', '')
outdir = f"./output/{tag}"
outpath = pjoin(outdir, f'trig_sf.root')
try:
outfile = uproot.recreate(outpath)
except OSError:
outfile = uproot.update(outpath)
for year in [2017, 2018]:
for region in regions:
fig, ax, rax = fig_ratio()
if '1e' in region:
fnum = f'output/{tag}/table_{region}_met_EGamma_{year}.txt'
fden = f'output/{tag}/table_{region}_met_WJetsToLNu_HT_MLM_{year}.txt'
xlabel = "$p_{T}^{miss}$ (GeV)"
elif '1m' in region:
fnum = f'output/{tag}/table_{region}_recoil_SingleMuon_{year}{"_"+jeteta_config if jeteta_config else ""}.txt'
fden = f'output/{tag}/table_{region}_recoil_WJetsToLNu_HT_MLM_{year}{"_"+jeteta_config if jeteta_config else ""}.txt'
xlabel = "Recoil (GeV)"
elif '2m' in region:
fnum = f'output/{tag}/table_{region}_recoil_SingleMuon_{year}{"_"+jeteta_config if jeteta_config else ""}.txt'
fden = f'output/{tag}/table_{region}_recoil_VDYJetsToLL_M-50_HT_MLM_{year}{"_"+jeteta_config if jeteta_config else ""}.txt'
xlabel = "Recoil (GeV)"
elif 'g_' in region:
fnum = f'output/{tag}/table_{region}_photon_pt0_JetHT_{year}.txt'
fden = f'output/{tag}/table_{region}_photon_pt0_GJets_HT_MLM_{year}.txt'
xlabel = "Photon $p_{T}$ (GeV)"
if not os.path.exists(fnum):
print(f"File not found {fnum}")
continue
if not os.path.exists(fden):
print(f"File not found {fden}")
continue
xnum, xedgnum, ynum, yerrnum = get_xy(fnum)
xden, xedgden, yden, yerrden = get_xy(fden)
xsf = xnum
ysf = ynum / yden
ysferr = ratio_unc(ynum, yden, yerrnum, yerrden)
opts['color'] = 'k'
ax.errorbar(xnum,
ynum,
yerr=yerrnum,
label=f'Data, {region} region',
**opts)
opts['color'] = 'r'
ax.errorbar(xden,
yden,
yerr=yerrden,
label=f'MC, {region} region',
**opts)
# rax.plot([0,1000],[0.98,0.98],color='blue')
# rax.plot([0,1000],[0.99,0.99],color='blue',linestyle='--')
if 'g_' in region:
ax.plot([215, 215], [0.9, 1.1], color='blue')
rax.plot([215, 215], [0.95, 1.05], color='blue')
elif distribution == 'recoil':
ax.plot([250, 250], [0.0, 1.1], color='blue')
rax.plot([250, 250], [0.95, 1.05], color='blue')
else:
ax.plot([250, 250], [0.9, 1.1], color='blue')
rax.plot([250, 250], [0.95, 1.05], color='blue')
opts['color'] = 'k'
rax.errorbar(xsf, ysf, ysferr, **opts)
ax.legend()
ax.set_ylabel("Efficiency")
ax.xaxis.set_major_locator(MultipleLocator(200))
ax.xaxis.set_minor_locator(MultipleLocator(50))
ax.set_ylim(ymin, ymax)
ax.grid(1)
if distribution == 'mjj':
ax.yaxis.set_major_locator(MultipleLocator(0.05))
ax.yaxis.set_minor_locator(MultipleLocator(0.01))
elif distribution == 'recoil':
ax.yaxis.set_major_locator(MultipleLocator(0.1))
ax.yaxis.set_minor_locator(MultipleLocator(0.05))
rax.set_xlabel(xlabel)
rax.set_ylabel("Data / MC SF")
rax.set_ylim(0.95, 1.05)
rax.yaxis.set_major_locator(MultipleLocator(0.05))
rax.yaxis.set_minor_locator(MultipleLocator(0.01))
rax.grid(1)
plt.text(1.,
1.,
r"$\approx$ %.1f fb$^{-1}$ (13 TeV)" %
lumi_by_region(region, year),
fontsize=16,
horizontalalignment='right',
verticalalignment='bottom',
transform=ax.transAxes)
plt.text(1.,
0.95,
f'{jeteta_config if jeteta_config else ""}',
fontsize=12,
horizontalalignment='right',
verticalalignment='bottom',
transform=ax.transAxes)
plt.text(0.,
1.,
f'{year}',
fontsize=16,
horizontalalignment='left',
verticalalignment='bottom',
transform=ax.transAxes)
fig.savefig(
pjoin(
outdir,
f'data_mc_comparison_{region}_{year}{"_"+jeteta_config if jeteta_config else ""}.{output_format}'
))
fig.clear()
plt.close(fig)
vals = np.array(sorted(list(set(list(xedgnum.flatten())))))
ysf[np.isnan(ysf) | np.isinf(np.abs(ysf))] = 1
outfile[f'{tag}_{region}_{year}'] = (ysf, vals)
def extract_yields_in_cr(acc, distribution, region='cr_vbf_qcd_rs', year=2017):
'''Calculate the data - (nonQCD MC) in the QCD CR.'''
acc.load(distribution)
h = acc[distribution]
h = merge_extensions(h, acc)
scale_xs_lumi(h)
h = merge_datasets(h)
if distribution in BINNINGS.keys():
new_ax = BINNINGS[distribution]
h = h.rebin(new_ax.name, new_ax)
h = h.integrate('region', region)
data = f'MET_{year}'
mc = re.compile(
f'(ZJetsToNuNu.*|EW.*|Top_FXFX.*|Diboson.*|DYJetsToLL_M-50_HT_MLM.*|WJetsToLNu.*HT.*).*{year}'
)
fig, ax, rax = fig_ratio()
data_err_opts = {
'linestyle': 'none',
'marker': '.',
'markersize': 10.,
'color': 'k',
'elinewidth': 1,
}
hist.plot1d(h[data],
ax=ax,
overlay='dataset',
binwnorm=1,
error_opts=data_err_opts)
hist.plot1d(h[mc],
ax=ax,
overlay='dataset',
binwnorm=1,
stack=True,
clear=False)
ax.set_yscale('log')
ax.set_ylim(1e-4, 1e6)
ax.set_ylabel('Events / GeV')
ax.yaxis.set_ticks_position('both')
handles, labels = ax.get_legend_handles_labels()
for handle, label in zip(handles, labels):
for regex, new_label in PRETTY_LEGEND_LABELS.items():
if re.match(regex, label):
handle.set_label(new_label)
ax.legend(title='VBF QCD CR', handles=handles, ncol=2)
# Calculate data - MC
h_data = h[data].integrate('dataset')
h_mc = h[mc].integrate('dataset')
h_mc.scale(-1)
h_data.add(h_mc)
# Plot data - MC on the bottom pad
hist.plot1d(h_data, ax=rax, binwnorm=1)
rax.set_ylabel('(Data - MC) / GeV')
rax.set_ylim(1e-3, 1e1)
rax.set_yscale('log')
rax.get_legend().remove()
rax.yaxis.set_ticks_position('both')
outdir = './output/qcd_cr'
try:
os.makedirs(outdir)
except FileExistsError:
pass
outpath = pjoin(outdir, f'qcd_cr_{distribution}.pdf')
fig.savefig(outpath)
plt.close(fig)
print(f'File saved: {outpath}')
# Return the QCD yield
return h_data