def plot_weights(sample,variation,indir,outdir):
'''
21.10.2020
Just implemented all weights and uncertainties, version v12
Goal: show all AK8 jet mass distributions in a single plot. Will be very busy but just want to check if any particular distribution sticks out
'''
plt.style.use([hep.cms.style.ROOT, {'font.size': 24}])
f, ax = plt.subplots()
hep.cms.label(data=False, paper=False, year=args.year, ax=ax, loc=0)
with open( glob(os.path.join(indir,variation,f'*{sample}*json'))[0] ) as f:
hist = json.load(f)
weight_list = [w.split('hist_weights_')[1].split('_2J')[0] for w in hist if w.startswith('hist_weights_') and w.endswith('2J2WdeltaR_Pass_weights_nominal')]
for w in ['ones','nominal']:
if w in weight_list: weight_list.remove(w)
for w in weight_list:
wn = f'hist_weights_{w}_2J2WdeltaR_Pass_weights_nominal'
hep.histplot(hist[wn]['contents'], hist[wn]['edges'], edges=True, label=w)
plt.xlim(0,1.6)
# Shrink current axis by 20% to make space for legend
#plt.semilogy()
plt.legend()
ax.set_xlabel('Event weights', ha='right', x=1)
binsize = hist[wn]['edges'][1] - hist[wn]['edges'][0]
ax.set_ylabel(f'Entries / {binsize:.2f}', ha='right', y=1)
#plt.show()
for ext in ['.png','.pdf']:
plt.savefig(os.path.join(outdir,f'weights_{args.mask}_{sample}{ext}'), bbox_inches='tight')
def plot_corrections(sample,histName,indir,outdir):
'''
21.10.2020
Just implemented all weights and uncertainties, version v12
Goal: show all AK8 jet mass distributions in a single plot. Will be very busy but just want to check if any particular distribution sticks out
'''
rebin_factor = 5
plt.style.use([hep.cms.style.ROOT, {'font.size': 24}])
mx = { 'sample' : '', 'max' : 0 }
mn = { 'sample' : '', 'max' : 1000 }
with open( glob(os.path.join(indir,'nominal',f'*{sample}*json'))[0] ) as f:
nominal = json.load(f)[histName]
nominal['edges'], nominal['contents'], nominal['contents_w2'] = rebin(nominal['edges'], nominal['contents'], nominal['contents_w2'], rebin_factor)
for variation in [v for v in os.listdir(indir) if v.endswith('Up')]:
f, ax = plt.subplots()
hep.cms.label(data=False, paper=False, year=args.year, ax=ax, loc=0)
variation = variation.replace('Up','')
if variation.startswith('jesTotal'): continue
hep.histplot(nominal['contents'], nominal['edges'], edges=True, label='nominal', color='k')
for v in ['Up','Down']:
with open( glob(os.path.join(indir,variation+v,f'*{sample}*json'))[0] ) as f:
hist = json.load(f)[histName]
hist['edges'], hist['contents'], hist['contents_w2'] = rebin(hist['edges'], hist['contents'], hist['contents_w2'], rebin_factor)
hep.histplot(hist['contents'], hist['edges'], edges=True, label=variation+v, color=('r' if v=='Up' else 'b'), ls='--')
plt.legend(fontsize=14)
plt.xlim(100,160)
ax.set_xlabel('Leading AK8 jet softdrop mass [GeV]', ha='right', x=1)
ax.set_ylabel(f'Events / {rebin_factor} GeV', ha='right', y=1)
#plt.show()
for ext in ['.png','.pdf']:
plt.savefig(os.path.join(outdir,f'{args.variable}_{args.mask}_{sample}_{variation}{ext}'), bbox_inches='tight')
plt.close()
def compareShapes(h1,h2,outFile,xTitle="",yTitle="",yRange=[],xRange=[],log=False,label1="",label2="",labelR="Data/MC",data=False):
hRpf_mj = createRatio(h1,h2)
hRatio = []
hRatioErrors = []
hDataErrors = []
hData = []
hDataCentres = []
for i in range(1,hRpf_mj.GetNbinsX()+1):
hRatio.append(hRpf_mj.GetBinContent(i))
hRatioErrors.append(hRpf_mj.GetBinError(i))
for i in range(1,h2.GetNbinsX()+1):
hData.append(h2.GetBinContent(i))
hDataCentres.append(h2.GetBinCenter(i))
hDataErrors.append(h2.GetBinError(i))
#numpy histograms
h1, h1Edges = hist2array(h1,return_edges=True)
h2, h2Edges = hist2array(h2,return_edges=True)
hRatio = np.asarray(hRatio)
hRatioErrors = np.asarray(hRatioErrors)
centresRatio = (h1Edges[0][:-1] + h1Edges[0][1:])/2.
plt.style.use([hep.style.CMS])
f, axs = plt.subplots(2,1, sharex=True, sharey=False,gridspec_kw={'height_ratios': [4, 1],'hspace': 0.05})
axs = axs.flatten()
plt.sca(axs[0])
if(log):
axs[0].set_yscale("log")
hep.histplot(h1,h1Edges[0],stack=False,ax=axs[0],label = label1, histtype="step",color="r")
if(label2=="data_obs"):
print("Plotting data")
plt.errorbar(hDataCentres,hData, yerr=hDataErrors, fmt='o',color="k",label = label2)
else:
hep.histplot(h2,h2Edges[0],stack=False,ax=axs[0],label = label2, histtype="step",color="k")
plt.ylabel(yTitle, horizontalalignment='right', y=1.0)
axs[0].legend()
axs[1].set_xlabel(xTitle)
#axs[0].set_ylabel(yTitle)
axs[1].set_ylabel(labelR)
if(yRange):
axs[0].set_ylim(yRange)
if(xRange):
axs[0].set_xlim(xRange)
if data:
hep.cms.text("WiP",loc=1)
hep.cms.lumitext(text='59.8 $fb^{-1} (13 TeV)$', ax=axs[0], fontname=None, fontsize=None)
else:
hep.cms.lumitext(text='2018', ax=axs[0], fontname=None, fontsize=None)
hep.cms.text("Simulation WiP",loc=1)
plt.legend(loc="best")#loc = (0.4,0.2))
plt.sca(axs[1])#switch to lower pad
axs[1].axhline(y=1.0, xmin=0, xmax=1, color="r")
axs[1].set_ylim([0.0,2.0])
plt.xlabel(xTitle, horizontalalignment='right', x=1.0)
plt.errorbar(centresRatio,hRatio, yerr=hRatioErrors, fmt='o',color="k")
print("Saving {0}".format(outFile))
plt.savefig(outFile)
def plot(var: str, xlabel: str, weight: str,
bins: tuple = (50, 0, 5000)) -> None:
# plot
t0 = time.time()
for dsid in df['mcChannelNumber'].unique():
# per dsid
df_dsid = df.loc[df['mcChannelNumber'] == dsid]
hist = bh.Histogram(bh.axis.Regular(*bins))
hist.fill(df_dsid[var], weight=df_dsid[weight], threads=N_THREADS // 2)
hep.histplot(hist, label=dsid)
# inclusive
hist = bh.Histogram(bh.axis.Regular(*bins))
hist.fill(df[var], weight=df[weight])
hep.histplot(hist, label='Inclusive', color='k')
plt.xlabel(xlabel)
plt.ylabel("Entries")
plt.semilogy()
plt.legend(fontsize=10, ncol=2)
hep.atlas.label(italic=(True, True),
llabel='Internal',
rlabel=r'$W^+\rightarrow\tau\nu$ 13TeV')
plt.savefig(OUT_DIR + 'wplustaunu_' + var + '.png')
plt.show()
print(f"Making {var} plot: {time.time() - t0:.3f}s")
plt.clf()
def main():
rebsmear_file = uproot.open(
rebsmear_path('plot/output/merged_2021-07-01_rebsmear_v2_run/rebsmear_qcd_estimate.root')
)
qcd_file = uproot.open(
rebsmear_path('input/qcd_from_ic/out_MTR_2017.root_qcdDD.root')
)
outdir = 'output/merged_2021-07-01_rebsmear_v2_run'
h_qcd = qcd_file['rebin_QCD_hist_counts']
h_rebsmear = rebsmear_file['rebsmear_qcd_2017']
fig, ax = plt.subplots()
hep.histplot(h_qcd.values, h_qcd.edges, yerr=np.sqrt(h_qcd.variances), binwnorm=1, label='QCD (IC)')
hep.histplot(h_rebsmear.values, h_rebsmear.edges, yerr=np.sqrt(h_rebsmear.variances), binwnorm=1, label='Rebalance & Smear')
ax.set_xlabel(r'$M_{jj} \ (GeV)$')
ax.set_ylabel('Events / GeV')
ax.legend()
ax.set_yscale('log')
ax.set_ylim(1e-4,1e2)
outpath = pjoin(outdir, 'qcd_comparison.pdf')
fig.savefig(outpath)
plt.close(fig)
print(f'File saved: {outpath}')
def scale_and_plot(frame,
new_lumi: float = 1.,
c=None,
linewidth=1,
label=None):
xs = frame['weight_mc'].abs().sum() / len(frame.index)
lumi = frame['weight'].sum() / xs
hist = bh.Histogram(bh.axis.Regular(100,
300,
10000,
transform=bh.axis.transform.log),
storage=bh.storage.Weight())
hist.fill(frame['MC_WZ_dilep_m_born'],
weight=new_lumi * frame['weight'],
threads=6)
# scale cross-section
hist /= hist.axes[0].widths
hep.histplot(hist.view().value,
bins=hist.axes[0].edges,
color=c,
linewidth=linewidth,
label=label)
return xs, lumi
def plot_dist(yvals_f, var, bin, label, outpath, sample):
hists_gen = get_distribution(yvals_f, "gen", bin, var)
# hists_cand = get_distribution(yvals_f, "cand", bin, var)
hists_pred = get_distribution(yvals_f, "pred", bin, var)
plt.figure()
ax = plt.axes()
v1 = mplhep.histplot(
[h[bh.rebin(2)] for h in hists_gen],
stack=True,
label=[class_names[k] for k in [13, 11, 22, 1, 2, 130, 211]],
lw=1)
mplhep.histplot(
[h[bh.rebin(2)] for h in hists_pred],
stack=True,
color=[x.stairs.get_edgecolor() for x in v1],
lw=2,
histtype="errorbar",
)
legend1 = plt.legend(v1, [x.legend_artist.get_label() for x in v1],
loc=(0.60, 0.44),
title="true")
# legend2 = plt.legend(v2, [x.legend_artist.get_label() for x in v1], loc=(0.8, 0.44), title="pred")
plt.gca().add_artist(legend1)
plt.ylabel("Total number of particles / bin")
cms_label(ax)
sample_label(sample, ax)
plt.yscale("log")
plt.ylim(top=1e9)
plt.xlabel(f"PFCandidate {label} [GeV]")
plt.savefig(f"{outpath}/pfcand_{var}.pdf", bbox_inches="tight")
plt.close()
def test_basic():
fig, ax = plt.subplots(figsize=(10, 10))
h = [1, 3, 2]
bins = [0, 1, 2, 3]
hep.histplot(h, bins, yerr=True, label="X")
ax.legend()
return fig
def plot_htmiss_before_and_after(outdir,
infile,
dataset_tag='jetht',
plot_gen=True):
'''Do the actual plotting of distributions.'''
f = uproot.open(infile)
htmiss_bef = f['htmiss_before']
htmiss_aft = f['htmiss_after']
fig, ax = plt.subplots()
hep.histplot(htmiss_bef.values,
htmiss_bef.edges,
ax=ax,
label='Before rebalancing')
hep.histplot(htmiss_aft.values,
htmiss_aft.edges,
ax=ax,
label='After rebalancing')
ax.set_xlabel(r'$H_T^{miss} \ (GeV)$', fontsize=14)
ax.set_ylabel(r'Counts', fontsize=14)
ax.set_yscale('log')
ax.set_ylim(1e-1, 1e7)
ax.legend(title='Rebalancing')
ax.text(0.,
1.,
f'{tag_to_plottag(dataset_tag)} 2017',
fontsize=14,
ha='left',
va='bottom',
transform=ax.transAxes)
# If we're looking at QCD and plot_gen=True, plot the GEN HTmiss distribution as well
if dataset_tag == 'qcd' and plot_gen:
# Coffea file to take GEN HT-miss distribution from
accpath = './input/qcd_QCD_HT700to1000-mg_new_pmx_2017.coffea'
acc = load(accpath)
distribution = 'gen_htmiss_noweight'
h = acc[distribution].integrate('dataset').integrate(
'region', 'inclusive')
hist.plot1d(h, ax=ax, clear=False)
handles, labels = ax.get_legend_handles_labels()
for handle, label in zip(handles, labels):
if label == 'None':
handle.set_label(r'GEN $H_T^{miss}$')
ax.legend(handles=handles)
outpath = pjoin(outdir, f'htmiss_before_after_reb.pdf')
fig.savefig(outpath)
plt.close(fig)
print(f'File saved: {outpath}')
def plotSidebandsMJY_mpl(h_3d,outFile,xTitle="",yTitle="",yRange=[],xRange=[],log=False):
nBinsX = h_3d.GetNbinsX()
nBinsY = h_3d.GetNbinsY()
nBinsZ = h_3d.GetNbinsZ()
wp = 0.8
point = 0.0
taggerLowBin = h_3d.GetXaxis().FindBin(point)
taggerHighBin = h_3d.GetXaxis().FindBin(wp) - 1
h_mj_fail = h_3d.ProjectionZ("mj_failProjection" ,taggerLowBin,taggerHighBin,1,nBinsY)
h_mj_pass = h_3d.ProjectionZ("mj_passProjection" ,taggerHighBin,nBinsX,1,nBinsY)
h_mj_fail.Scale(1/h_mj_fail.Integral())
h_mj_pass.Scale(1/h_mj_pass.Integral())
h_mj_pass.Rebin(2)
h_mj_fail.Rebin(2)
hRpf_mj = createRatio(h_mj_pass,h_mj_fail)
hRatio = []
hRatioErrors = []
for i in range(1,hRpf_mj.GetNbinsX()+1):
hRatio.append(hRpf_mj.GetBinContent(i))
hRatioErrors.append(hRpf_mj.GetBinError(i))
#numpy histograms
hMJYFail, hMJYFailEdges = hist2array(h_mj_fail,return_edges=True)
hMJYPass, hMJYPassEdges = hist2array(h_mj_pass,return_edges=True)
hRatio = np.asarray(hRatio)
hRatioErrors = np.asarray(hRatioErrors)
centresRatio = (hMJYPassEdges[0][:-1] + hMJYPassEdges[0][1:])/2.
plt.style.use([hep.style.CMS])
f, axs = plt.subplots(2,1, sharex=True, sharey=False,gridspec_kw={'height_ratios': [4, 1],'hspace': 0.05})
axs = axs.flatten()
plt.sca(axs[0])
if(log):
axs[0].set_yscale("log")
hep.histplot(hMJYPass,hMJYPassEdges[0],stack=False,ax=axs[0],label = "Pass", histtype="step",color="r")
hep.histplot(hMJYFail,hMJYFailEdges[0],stack=False,ax=axs[0],label = "Fail", histtype="step",color="b")
axs[0].legend()
axs[1].set_xlabel(xTitle)
axs[0].set_ylabel(yTitle)
axs[1].set_ylabel("Data/MC")
plt.ylabel(yTitle, horizontalalignment='right', x=1.0)
if(yRange):
axs[0].set_ylim(yRange)
if(xRange):
axs[0].set_xlim(xRange)
hep.cms.lumitext(text='59.8 $fb^{-1} (13 TeV)$', ax=axs[0], fontname=None, fontsize=None)
hep.cms.text("Simulation WiP",loc=1)
plt.legend(loc="best")#loc = (0.4,0.2))
plt.sca(axs[1])#switch to lower pad
axs[1].axhline(y=1.0, xmin=0, xmax=1, color="r")
axs[1].set_ylim([0.0,2.0])
plt.xlabel(xTitle, horizontalalignment='right', x=1.0)
plt.errorbar(centresRatio,hRatio, yerr=hRatioErrors, fmt='o',color="k")
print("Saving {0}".format(outFile))
plt.savefig(outFile)
def test_log():
fig, axs = plt.subplots(2, 2, figsize=(10, 10))
for ax in axs[0]:
hep.histplot([1, 2, 3, 2], range(5), ax=ax)
plt.semilogy()
for ax in axs[1]:
hep.histplot([1, 2, 3, 2], range(5), ax=ax, edges=False)
plt.semilogy()
return fig
def effCutflow(inFile,
sample,
outFile,
xTitle="",
yTitle="",
label="",
yRange=[],
xRange=[],
log=False):
tempFile = r.TFile.Open(inFile)
hTemp = tempFile.Get("{0}_cutflow".format(sample))
hist, edges = hist2array(hTemp, return_edges=True)
#print(sample)
#print(hist)
for i in range(1, 5): #tagger regions only pnet and dak8 TT/LL
hist[-i] = hist[-i] / hist[0]
for i in range(5, len(hist)): #skip tagging regions
hist[-i] = hist[-i] / hist[0]
hist[0] = 1.
#print(hist)
plt.style.use([hep.style.CMS])
f, ax = plt.subplots()
ax.locator_params(nbins=12, axis='x')
hep.histplot(hist,
edges[0],
ax=ax,
label=label,
histtype="step",
color='black')
if (log):
ax.set_yscale("log")
ax.legend()
ax.set_xlabel(xTitle)
ax.set_ylabel(yTitle)
hep.cms.lumitext(text='2018', ax=None, fontname=None, fontsize=None)
hep.cms.text("Simulation WiP", loc=1)
plt.legend(loc="best") #loc = (0.4,0.2))
axisTicks = ax.get_xticks().tolist()
axisTicks = [
0, "No cuts", "Triggers", "Jets definition", "Preselection",
"H-jet $m_{SD}$", "H,Y pT cut", "ParticlNet TT", "ParticlNet LL",
"DeepAK8 TT", "DeepAK8 LL"
]
ax.set_xticklabels(axisTicks, rotation=45, fontsize=14)
if (yRange):
ax.set_ylim(yRange)
if (xRange):
ax.set_xlim(xRange)
print("Saving {0}".format(outFile))
plt.savefig(outFile)
plt.clf()
def plot_unc(indir, histName, uncName, outdir):
sample = 'ttHTobb'
if uncName=='msd':
hist = {
'nominal' : load_mc(indir, 'msd_nanoAOD', histName, sample),
'up' : load_mc(indir, 'msd_nom', histName, sample),
'down' : load_mc(indir, 'msd_raw', histName, sample)
}
else:
hist = {
'nominal' : load_mc(indir, 'msd_nom', histName, sample),
'up' : load_mc(indir, f'{uncName}Up', histName, sample),
'down' : load_mc(indir, f'{uncName}Down', histName, sample)
}
color = {
'nominal' : 'k',
'up' : 'r',
'down' : 'b'
}
ls = {
'nominal' : '-',
'up' : '--',
'down' : '-.'
}
rebin_factor = 5
plt.style.use([hep.cms.style.ROOT, {'font.size': 24}])
f, ax = plt.subplots()
hep.cms.label(data=False, paper=False, year=args.year, ax=ax, loc=0)
for hn,h in hist.items():
h['edges'], h['contents'], h['contents_w2'] = rebin(h['edges'], h['contents'], h['contents_w2'], rebin_factor)
if uncName=='msd':
if hn=='nominal': label = 'nanoAOD'
elif hn=='up': label = 'nom'
elif hn=='down': label = 'raw'
else:
if hn=='nominal': label = hn
else: label = f'{uncName} {hn}'
hep.histplot(h['contents'], h['edges'], edges=True, label=label,ls=ls[hn],color=color[hn])
#print(uncName, hn, h['contents'][-3:])
plt.xlim(90,170)
#plt.xlim(0,1000)
#ax.set_ylim(ymin=.5e-1)
#plt.semilogy()
plt.legend()
#import pdb
#pdb.set_trace()
ax.set_xlabel('Leading AK8 jet softdrop mass [GeV]', ha='right', x=1)
ax.set_ylabel(f'Events / {rebin_factor} GeV', ha='right', y=1)
for ext in ['.png','.pdf']:
plt.savefig(os.path.join(outdir,f'{sample}_{args.variable}_{args.mask}_{uncName}{ext}'))
def plot_projections(data_counts, model_prediction, data_bin_edges, model_prediction_grid,
fit_params=None, close_image=False, save_fig=True, output_folder='.', image_name=None):
fig, axs = plt.subplots(1, 2, figsize=(20,7))
data_counts_x = data_counts['x']
data_counts_y = data_counts['y']
#
model_prediction_x = model_prediction['x']['model']
# model_prediction_x_sig = model_prediction['x']['sig']
# model_prediction_x_bkgr = model_prediction['x']['bkgr']
#
model_prediction_y = model_prediction['y']['model']
# model_prediction_y_sig = model_prediction['y']['sig']
# model_prediction_y_bkgr = model_prediction['y']['bkgr']
##
axs[0].plot(model_prediction_grid, model_prediction_x, label="Model", linewidth=5)
# axs[0].plot(model_prediction_grid, model_prediction_x_sig, label="Signal", linewidth=5)
# axs[0].plot(model_prediction_grid, model_prediction_x_bkgr, label="Background", linewidth=5)
mplhep.histplot(data_counts_x, data_bin_edges, yerr=True, color='black', histtype='errorbar',
markersize=17, capsize=2.5,
markeredgewidth=1.5, zorder=1,
elinewidth=1.5, ax=axs[0]
)
axs[0].set_title('X projection')
if fit_params:
assert 'x' in fit_params
assert 'mu' in fit_params['x'] and 'sigma' in fit_params['x'] and 'fr' in fit_params['x']
mu_patch = mpatches.Patch(color='none', label=f"mu = {fit_params['x']['mu']:.2f}")
sigma_patch = mpatches.Patch(color='none', label=f"sigma = {fit_params['x']['sigma']:.2f}")
fr_patch = mpatches.Patch(color='none', label=f"fr = {fit_params['x']['fr']:.4f}")
N_patch = mpatches.Patch(color='none', label=f"N = {fit_params['x']['N']:.0f}")
axs[0].legend(handles=[mu_patch, sigma_patch, fr_patch, N_patch])
##
axs[1].plot(model_prediction_grid, model_prediction_y, label="Model", linewidth=5)
# axs[1].plot(model_prediction_grid, model_prediction_y_sig, label="Signal", linewidth=5)
# axs[1].plot(model_prediction_grid, model_prediction_y_bkgr, label="Background", linewidth=5)
mplhep.histplot(data_counts_y, data_bin_edges, yerr=True, color='black', histtype='errorbar',
markersize=17, capsize=2.5,
markeredgewidth=1.5, zorder=1,
elinewidth=1.5, ax=axs[1]
)
axs[1].set_title('Y projection')
if fit_params:
assert 'y' in fit_params
assert 'mu' in fit_params['y'] and 'sigma' in fit_params['y'] and 'fr' in fit_params['y']
mu_patch = mpatches.Patch(color='none', label=f"mu = {fit_params['y']['mu']:.2f}")
sigma_patch = mpatches.Patch(color='none', label=f"sigma = {fit_params['y']['sigma']:.2f}")
fr_patch = mpatches.Patch(color='none', label=f"fr = {fit_params['y']['fr']:.4f}")
N_patch = mpatches.Patch(color='none', label=f"N = {fit_params['y']['N']:.0f}")
axs[1].legend(handles=[mu_patch, sigma_patch, fr_patch, N_patch])
if save_fig:
fig.savefig(f"{output_folder}/{image_name}_fitted.png")
if close_image:
plt.close(fig)
def create_plot1d(hist, save_name, log=False):
import matplotlib.pyplot as plt
import mplhep as hep
plt.style.use(hep.style.CMS)
# plot
ax = plt.gca()
plt.errorbar(
hist.axes[0].centers,
hist.view(),
np.sqrt(hist.view()),
fmt='.',
color='blue',
)
hep.histplot(hist, ax=ax, color='blue')
if log:
ax.set_yscale('log')
else:
ax.ticklabel_format(axis='y',
style='sci',
scilimits=(0, 3),
useMathText=True)
ax.set_xlabel(hist.axes[0].metadata, loc='right')
ax.set_ylabel("Counts", loc='top')
# compute mean and std:
mean = (hist.view() * hist.axes[0].centers).sum() / hist.sum()
std = np.sqrt(
(hist.view() * ((hist.axes[0].centers - mean)**2)).sum() / hist.sum())
annotation = f"Total {hist.sum()}" \
+ "\n" + f"Mean: {round(mean,2)}" \
+ "\n" + f"Std: {round(std,2)}"
ax.annotate(annotation,
xy=(0.85, 0.85),
xycoords='axes fraction',
fontsize="small",
ha='center',
annotation_clip=False,
bbox=dict(boxstyle='round', fc='None'))
ax.set_xlim(hist.axes[0].edges[0],
hist.axes[0].edges[-1] + hist.axes[0].widths[-1])
fig = ax.get_figure()
fig.savefig(save_name)
ax.clear()
fig.clear()
def plot_fit_result(models, data, p_params, obs):
plt_name = "mtw"
lower, upper = obs.limits
h_bin_width = hist_dicts[plt_name]["bin_width"]
h_num_bins = hist_dicts[plt_name]["numbins"]
h_xmin = hist_dicts[plt_name]["xmin"]
h_xmax = hist_dicts[plt_name]["xmax"]
h_xlabel = hist_dicts[plt_name]["xlabel"]
plt_label = "$W \\rightarrow l\\nu$"
bins = [h_xmin + x * h_bin_width for x in range(h_num_bins + 1)]
bin_centers = [
h_xmin + h_bin_width / 2 + x * h_bin_width for x in range(h_num_bins)
]
data_x, _ = np.histogram(data.values, bins=bins)
data_sum = data_x.sum()
plot_scale = data_sum * obs.area() / h_num_bins
plt.clf()
plt.style.use(hep.style.ATLAS)
plt.axes([0.1, 0.30, 0.85, 0.65])
main_axes = plt.gca()
hep.histplot(main_axes.hist(data, bins=bins, log=False, facecolor="none"),
color="black",
yerr=True,
histtype="errorbar",
label="data")
main_axes.set_xlim(h_xmin, h_xmax)
main_axes.xaxis.set_minor_locator(AutoMinorLocator())
main_axes.set_xlabel(h_xlabel)
main_axes.set_title("W Transverse Mass Fit")
main_axes.set_ylabel("Events/4 GeV")
main_axes.ticklabel_format(axis='y', style='sci', scilimits=[-2, 2])
x_plot = np.linspace(lower[-1][0], upper[0][0], num=1000)
for model_name, model in models.items():
if model.is_extended:
main_axes.plot(x_plot,
model.ext_pdf(x_plot) * obs.area() / h_num_bins,
label=model_name)
else:
main_axes.plot(x_plot,
model.pdf(x_plot) * plot_scale,
label=model_name)
main_axes.legend(title=plt_label, loc="best")
plt.savefig(f"../Results/fit_plot_{plt_name}_Complex.pdf")
plt.close()
def test_inputs_bh():
import boost_histogram as bh
hist2d = bh.Histogram(bh.axis.Regular(10, 0.0, 1.0),
bh.axis.Regular(10, 0, 1))
hist2d.fill(np.random.normal(0.5, 0.2, 1000),
np.random.normal(0.5, 0.2, 1000))
fig, axs = plt.subplots(1, 2, figsize=(14, 5))
hep.histplot(hist2d.project(0), ax=axs[0])
hep.hist2dplot(hist2d, labels=True, cbar=False, ax=axs[1])
return fig
def plot1d(arrays, labels, bins, title, x_label, log_x):
plt.style.use(hep.style.ROOT)
fig, ax = plt.subplots()
if log_x:
ax.set_xscale('log')
#hep.cms.label(loc=0)
for array, label in zip(arrays, labels):
h, bins = np.histogram(array, bins, density=True)
print(h, bins)
hep.histplot(h, bins, label=label)
ax.set_xlabel(x_label)
ax.legend()
plt.savefig(f"tagger_efficiency/{title}.pdf")
def test_inputs_uproot():
import uproot4
from skhep_testdata import data_path
fname = data_path("uproot-hepdata-example.root")
f = uproot4.open(fname)
fig, axs = plt.subplots(1, 2, figsize=(14, 5))
TH1, TH2 = f["hpx"], f["hpxpy"]
hep.histplot(TH1, ax=axs[0])
hep.hist2dplot(TH2, ax=axs[1], cbar=False)
return fig
def test_histplot_stack():
np.random.seed(0)
h, bins = np.histogram(np.random.normal(10, 3, 400), bins=10)
fig, axs = plt.subplots(2, 2, sharex=True, sharey=True, figsize=(10, 10))
axs = axs.flatten()
axs[0].set_title("Default", fontsize=18)
hep.histplot([h, 1.5 * h], bins, stack=True, ax=axs[0])
axs[1].set_title("Plot No Edges", fontsize=18)
hep.histplot([h, 1.5 * h], bins, edges=False, stack=True, ax=axs[1])
axs[2].set_title("Plot Errorbars", fontsize=18)
hep.histplot([h, 1.5 * h],
bins,
yerr=[np.sqrt(h), np.sqrt(h)],
stack=True,
ax=axs[2])
axs[3].set_title("Filled Histogram", fontsize=18)
hep.histplot([1.5 * h, h], bins, histtype="fill", stack=True, ax=axs[3])
fig.subplots_adjust(hspace=0.1, wspace=0.1)
return fig
def test_histplot_multiple():
np.random.seed(0)
h, bins = np.histogram(np.random.normal(10, 3, 400), bins=10)
fig, axs = plt.subplots(2, 2, sharex=True, sharey=True, figsize=(10, 10))
axs = axs.flatten()
axs[0].set_title("Default Overlay", fontsize=18)
hep.histplot([h, 1.5 * h], bins, ax=axs[0])
axs[1].set_title("Default Overlay w/ Errorbars", fontsize=18)
hep.histplot([h, 1.5 * h],
bins,
yerr=[np.sqrt(h), np.sqrt(1.5 * h)],
ax=axs[1])
axs[2].set_title("Automatic Errorbars", fontsize=18)
hep.histplot([h, 1.5 * h], bins, yerr=True, ax=axs[2])
axs[3].set_title("With Labels", fontsize=18)
hep.histplot([h, 1.5 * h],
bins,
yerr=True,
ax=axs[3],
label=["First", "Second"])
axs[3].legend(fontsize=16, prop={"family": "Tex Gyre Heros"})
fig.subplots_adjust(hspace=0.1, wspace=0.1)
return fig
def test_histplot_kwargs():
np.random.seed(0)
h, bins = np.histogram(np.random.normal(10, 3, 1000), bins=10)
fig, axs = plt.subplots(2, 2, sharex=True, sharey=True, figsize=(10, 10))
axs = axs.flatten()
hep.histplot(
[h * 2, h * 1, h * 0.5],
bins,
label=["1", "2", "3"],
stack=True,
histtype="step",
linestyle="--",
color=["green", "black", (1, 0, 0, 0.4)],
ax=axs[0],
)
axs[0].legend()
hep.histplot(
[h, h, h],
bins,
label=["1", "2", "3"],
stack=True,
histtype="step",
linestyle=["--", ":"],
color=(1, 0, 0, 0.8),
ax=axs[1],
)
axs[1].legend()
hep.histplot(
[h, h, h],
bins,
label=["1", "2", "3"],
histtype="step",
binwnorm=[0.5, 3, 6],
linestyle=["--", ":"],
color=(1, 0, 0, 0.8),
ax=axs[2],
)
axs[2].legend()
hep.histplot(
[h, h, h],
bins,
label=["1", "2", "3"],
histtype="fill",
binwnorm=[0.5, 3, 6],
linestyle=["--", ":"],
color=["green", "darkorange", "red"],
alpha=[0.4, 0.7, 0.2],
ax=axs[3],
)
axs[3].legend()
fig.subplots_adjust(hspace=0.1, wspace=0.1)
return fig
def plot_pdf(title):
plt.figure()
plt.title(title)
y = model.pdf(x).numpy()
y_gauss = (gauss.pdf(x) * frac).numpy()
y_exp = (exponential.pdf(x) * (1 - frac)).numpy()
plt.plot(x, y * plot_scaling, label="Sum - Model")
plt.plot(x, y_gauss * plot_scaling, label="Gauss - Signal")
plt.plot(x, y_exp * plot_scaling, label="Exp - Background")
mplhep.histplot(
np.histogram(data_np, bins=n_bins),
yerr=True,
color="black",
histtype="errorbar",
)
plt.ylabel("Counts")
plt.xlabel("obs: $B_{mass}$")
def plotWeights(model, nBins=20):
plt.figure(figsize=(8, 6))
for ilayer in range(1, len(model.layers)):
if len(model.layers[ilayer].get_weights()) > 0:
label = model.layers[ilayer].name
data = np.histogram(model.layers[ilayer].get_weights()[0])
print(ilayer, label, 'unique weights',
len(np.unique(model.layers[ilayer].get_weights()[0])))
hep.histplot(data[0], data[1], label=label)
else:
print(ilayer, 'no weights')
plt.xlabel('weights')
plt.ylabel('Entries')
plt.yscale('log')
plt.legend()
plt.savefig("%s_weights.pdf" % model.name)
plt.clf()
def plot_pdf(title):
plt.figure()
plt.title(title)
y = model.pdf(x).numpy()
y_gauss = (gauss.pdf(x) * model_unbinned.params["frac_0"]).numpy()
y_exp = (exponential.pdf(x) * model_unbinned.params["frac_1"]).numpy()
plt.plot(x, y * plot_scaling, label="Sum - Model")
plt.plot(x, y_gauss * plot_scaling, label="Gauss - Signal")
plt.plot(x, y_exp * plot_scaling, label="Exp - Background")
# mplhep.histplot(np.histogram(data_np, bins=n_bins), yerr=True, color='black', histtype='errorbar')
mplhep.histplot(data.to_hist(),
yerr=True,
color="black",
histtype="errorbar")
plt.ylabel("Counts")
plt.xlabel("obs: $B_{mass}$")
plt.legend()
def shape_in_bins(self, shape_of="dimuon_mass", nbins=4):
for model_name in self.models.keys():
score_name = f"{model_name}_score"
for cls in self.df["class_name"].dropna().unique():
df = self.df[self.df["class_name"] == cls]
score = df[score_name]
fig = plt.figure()
fig, ax = plt.subplots()
for i in range(nbins):
cut_lo = score.quantile(i / nbins)
cut_hi = score.quantile((i + 1) / nbins)
cut = (score > cut_lo) & (score < cut_hi)
if shape_of == "max_abs_eta":
data = df.loc[cut].apply(max_abs_eta, axis=1)
else:
data = df.loc[cut, shape_of]
if shape_of == "dimuon_mass":
data_range = (115, 135)
data_bins = 80
else:
data_range = (data.min(), data.max())
data_bins = 25
weights = (df.lumi_wgt).loc[cut].values
# weights = (df.lumi_wgt * df.mc_wgt).loc[cut].values
hist = np.histogram(
data.values,
bins=data_bins,
range=data_range,
weights=weights,
density=True,
)
hep.histplot(
hist[0],
hist[1],
histtype="step",
label=f"{score_name} bin #{i}",
)
ax.legend(prop={"size": "x-small"})
ax.set_xlabel(shape_of)
ax.set_yscale("log")
ax.set_ylim(0.0001, 1)
out_name = f"{self.out_path}/shapes_{score_name}_{cls}_{shape_of}.png"
fig.savefig(out_name)
def test_bh_to_root_with_weights_1d():
bh_hist = bh.Histogram(bh.axis.Regular(20, -1, 1),
storage=bh.storage.Weight())
bh_hist.fill(data, weight=2)
hep.histplot(bh_hist.view().value,
bins=bh_hist.axes[0].edges,
yerr=np.sqrt(bh_hist.view().variance))
plt.savefig(out_dir + 'bh_to_root_bh_hist_1d.png')
plt.clf()
root_hist = bh_to_TH1(bh_hist, 'test', 'test;x;entries')
c1 = ROOT.TCanvas()
c1.cd()
root_hist.Draw("histE")
c1.Print(out_dir + 'bh_to_root_root_hist_1d.png')
for i, binval in enumerate(bh_hist.view(flow=True)):
assert binval.value == root_hist.GetBinContent(i)
assert np.sqrt(binval.variance) == root_hist.GetBinError(i)
def test_root_to_bh_with_weights_1d():
root_hist = ROOT.TH1F('test', 'test;x;entries', 20, -1, 1)
rnp.fill_hist(root_hist, data, weights=np.full(data.shape[0], 2)) # fill
c3 = ROOT.TCanvas()
c3.cd()
root_hist.Draw("histE")
c3.Print(out_dir + 'root_to_bh_root_hist_1d.png')
bh_hist = TH1_to_bh(root_hist)
hep.histplot(bh_hist.view().value,
bins=bh_hist.axes[0].edges,
yerr=np.sqrt(bh_hist.view().variance))
plt.savefig(out_dir + 'root_to_bh_hist_1d.png')
plt.clf()
for i, binval in enumerate(bh_hist.view(flow=True)):
assert binval.value == root_hist.GetBinContent(i)
assert np.sqrt(binval.variance) == root_hist.GetBinError(i)
def plot_shape(y, ynew, yerr, yerrnew, name):
plt.figure()
hep.histplot(y,
range(0,
len(y) + 1),
yerr=yerr,
histtype='step',
label='before')
hep.histplot(ynew,
range(0,
len(ynew) + 1),
yerr=yerrnew,
histtype='step',
label='after',
ls='--')
plt.legend(title=name)
plt.ylim(bottom=0)
plt.savefig('{name}.png'.format(name=name))
plt.close()
def plot_component(dfs, component):
print("###==========####")
print("Started plotting")
plot_label = "$W \\rightarrow l\\nu$"
hist = hist_dicts['mtw']
h_bin_width = hist["bin_width"]
h_num_bins = hist["numbins"]
h_xmin = hist["xmin"]
h_xmax = hist["xmax"]
h_xlabel = hist["xlabel"]
x_var = hist["xvariable"]
h_title = hist["title"]
bins = [h_xmin + x * h_bin_width for x in range(h_num_bins + 1)]
bin_centers = [
h_xmin + h_bin_width / 2 + x * h_bin_width for x in range(h_num_bins)
]
data_x, _ = np.histogram(dfs[component][x_var].values, bins=bins)
plt.clf()
plt.style.use(hep.style.ATLAS)
_ = plt.figure(figsize=(9.5, 9))
plt.axes([0.1, 0.30, 0.85, 0.65])
plt.yscale("linear")
main_axes = plt.gca()
main_axes.set_title(h_title)
hep.histplot(main_axes.hist(data[component][x_var],
bins=bins,
log=False,
facecolor="none"),
color="black",
yerr=True,
histtype="errorbar",
label='Данные')
main_axes.set_xlim(h_xmin * 0.9, h_xmax * 1.1)
main_axes.xaxis.set_minor_locator(AutoMinorLocator())
main_axes.set_ylabel(f"События/{h_bin_width} ГэВ")
plt.savefig(f"../FitResults_8TeV/{component}_mtw.jpg")