forked from hajohajo/JetRegression_legacy
-
Notifications
You must be signed in to change notification settings - Fork 0
/
plotting.py
207 lines (167 loc) · 9.65 KB
/
plotting.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
'''
Plotting related code
'''
import seaborn as sns
import matplotlib.pyplot as plt
from scipy.stats import norm
import numpy as np
from matplotlib.transforms import ScaledTranslation
def getBinnedStatistics(variable, binningVariable, binning):
indices = np.digitize(binningVariable, binning, right=True)
indices[indices == 0] = 1
indices = indices - 1
binMean = -np.ones(len(binning))
binStd = np.zeros(len(binning))
for i in np.unique(indices):
if (np.sum(indices == i) < 2):
continue
mean, std = norm.fit(variable[indices == i])
binMean[i] = mean
binStd[i] = std
return binMean, binStd
def plotInclusiveResponse(dataframe, name):
fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(3, 6), constrained_layout=True)
fig.suptitle("Inclusive response")
trans1 = ax.transData + ScaledTranslation(-5 / 72, 0, fig.dpi_scale_trans)
trans2 = ax.transData + ScaledTranslation(+5 / 72, 0, fig.dpi_scale_trans)
colors = [sns.xkcd_rgb["cerulean"], sns.xkcd_rgb["rouge"]]
gluons_l1l2l3 = dataframe.loc[dataframe.isPhysG == 1, "jetPt"] / dataframe.loc[dataframe.isPhysG == 1, "genJetPt"]
uds_l1l2l3 = dataframe.loc[dataframe.isPhysG != 1, "jetPt"] / dataframe.loc[dataframe.isPhysG != 1, "genJetPt"]
gluons_dnn = dataframe.loc[dataframe.isPhysG == 1, "response"]
uds_dnn = dataframe.loc[dataframe.isPhysG != 1, "response"]
mean_g_l1l2l3, std_g_l1l2l3 = norm.fit(gluons_l1l2l3)
mean_uds_l1l2l3, std_uds_l1l2l3 = norm.fit(uds_l1l2l3)
median_g_l1l2l3 = np.median(gluons_l1l2l3)
median_uds_l1l2l3 = np.median(uds_l1l2l3)
mean_g_dnn, std_g_dnn = norm.fit(gluons_dnn)
mean_uds_dnn, std_uds_dnn = norm.fit(uds_dnn)
median_g_dnn = np.median(gluons_dnn)
median_uds_dnn = np.median(uds_dnn)
iqr_g_l1l2l3 = np.subtract(*np.percentile(gluons_l1l2l3, [75, 25]))*0.1
iqr_uds_l1l2l3 = np.subtract(*np.percentile(uds_l1l2l3, [75, 25]))*0.1
iqr_g_dnn = np.subtract(*np.percentile(gluons_dnn, [75, 25]))*0.1
iqr_uds_dnn = np.subtract(*np.percentile(uds_dnn, [75, 25]))*0.1
ax.set_xlim(-1.0, 2.0)
ax.set_ylim(0.975, 1.04)
ax.text(-0.5, 1.041, "60 GeV < p$_T$ < 600 GeV, |$\eta$| < 2.5", fontsize=7)
ax.set_ylabel("Median response")
ax.set_xlabel("Jet class")
plt.errorbar(x=['g', 'uds'], y=[median_g_l1l2l3, median_uds_l1l2l3], fmt='o', label="L1L2L3", color=colors[1], ls='none', ecolor='k', transform=trans1)
plt.errorbar(x=['g', 'uds'], y=[median_g_dnn, median_uds_dnn], yerr=[iqr_g_dnn, iqr_uds_dnn], fmt='o', label="DNN", color=colors[0], ls='none', ecolor='k', transform=trans2)
ax.plot([-1.0, 2.0], [1.0, 1.0], linestyle='--', linewidth=1.5, color='k')
ax.legend()
plt.savefig("./plots/InclusiveResponse.pdf".format(name))
plt.clf()
plt.close()
def plotResidual(dataframe, name):
fig, (ax0, ax1) = plt.subplots(nrows=2, ncols=1, sharex=True, gridspec_kw={'height_ratios': [2, 1]}, figsize=(12, 6))
fig.suptitle("Mean response residuals w.r.t gen p$_{T}$")
binning = np.linspace(60.0, 600.0, 55)
binCenters = binning+(binning[1]-binning[0])/2.0
colors = [sns.xkcd_rgb["cerulean"], sns.xkcd_rgb["rouge"]]
binningVariable = dataframe.loc[:, "genJetPt"]
variable_1, std_1 = getBinnedStatistics(dataframe[(dataframe.isPhysG == 1)].loc[:, "response"], binningVariable[(dataframe.isPhysG == 1)], binning)
variable_2, std_2 = getBinnedStatistics(dataframe[(dataframe.isPhysG != 1)].loc[:, "response"], binningVariable[(dataframe.isPhysG != 1)], binning)
binMean = variable_1 - variable_2
binStd = np.sqrt(std_1 ** 2 + std_2 ** 2)
up = np.add(binMean, binStd)
down = np.add(binMean, -binStd)
controlVariable_1, std_1 = getBinnedStatistics(dataframe.loc[dataframe.isPhysG == 1, "jetPt"]/dataframe.loc[dataframe.isPhysG == 1, "genJetPt"], binningVariable[dataframe.isPhysG == 1], binning)
controlVariable_2, std_2 = getBinnedStatistics(dataframe.loc[dataframe.isPhysG != 1, "jetPt"]/dataframe.loc[dataframe.isPhysG != 1, "genJetPt"], binningVariable[dataframe.isPhysG != 1], binning)
controlBinMean = controlVariable_1 - controlVariable_2
controlBinStd = np.sqrt(std_1 ** 2 + std_2 ** 2)
controlUp = np.add(controlBinMean, controlBinStd)
controlDown = np.add(controlBinMean, -controlBinStd)
nonzeroPoints = (binMean > -1)
ax0.fill_between(binCenters[nonzeroPoints], up[nonzeroPoints], down[nonzeroPoints], alpha=0.6, color=colors[0])
ax0.plot(binCenters[nonzeroPoints], up[nonzeroPoints], c=colors[0], linewidth=1.0, alpha=0.8)
ax0.plot(binCenters[nonzeroPoints], down[nonzeroPoints], c=colors[0], linewidth=1.0, alpha=0.8)
ax0.scatter(binCenters[nonzeroPoints], binMean[nonzeroPoints], c=colors[0], label="DNN", linewidth=1.0, edgecolors='k')
# ax0.fill_between(binCenters[nonzeroPoints], controlUp[nonzeroPoints], controlDown[nonzeroPoints], alpha=0.6, color=colors[1])
ax0.plot(binCenters[nonzeroPoints], controlUp[nonzeroPoints], c=colors[1], linewidth=1.5, linestyle='--', alpha=0.8)
ax0.plot(binCenters[nonzeroPoints], controlDown[nonzeroPoints], c=colors[1], linewidth=1.5, linestyle='--', alpha=0.8)
ax0.scatter(binCenters[nonzeroPoints], controlBinMean[nonzeroPoints], c=colors[1], label="L1L2L3", linewidth=1.0, edgecolors='k')
ax0.plot([binCenters[0], binCenters[-1]], [0.0, 0.0], linestyle='--', linewidth=1.5, color='k')
ax0.set_ylim(-0.2, 0.2)
locs = np.linspace(-0.2, 0.2, 21)
ax0.set_yticks(locs)
ax0.set_xlim(binning[0], binning[-1])
ax0.set_ylabel("R$_G$-R$_{UDS}$")
ax1.set_xlabel("gen p$_{T}$")
handles, labels = ax0.get_legend_handles_labels()
ax0.legend(handles[::-1], labels[::-1], loc='upper right')
ax0.spines['top'].set_visible(False)
ax0.spines['right'].set_visible(False)
ax0.spines['bottom'].set_visible(False)
ax0.spines['left'].set_visible(False)
ax0.grid(False, axis='x')
ax0.plot([binning[0], binning[-1]], [1.0, 1.0], linestyle='--', linewidth=1.5, color='k')
ax1.set_xlim(binning[0], binning[-1])
ax1.hist(dataframe.loc[:, "genJetPt"], bins=binning, alpha=1.0, edgecolor="black", linewidth=1.0, color=colors[0])
ax1.set_ylabel("Jets/bin")
plt.tight_layout(pad=2.2)
fig.align_ylabels((ax0, ax1))
plt.savefig("./plots/MeanResidual.pdf".format(name))
plt.clf()
plt.close()
def plotCorrectedPtVsGenPt(dataframe, binningToUse, name):
fig, (ax0, ax1) = plt.subplots(nrows=2, ncols=1, sharex=True, gridspec_kw={'height_ratios': [2, 1]}, figsize=(12, 6))
fig.suptitle("Mean response with respect to genPt ({})".format(name))
variableName = "response"
binning = binningToUse
binCenters = binning+(binning[1]-binning[0])/2.0
colors = [sns.xkcd_rgb["cerulean"], sns.xkcd_rgb["rouge"]]
variable = dataframe.loc[:, variableName]
binningVariable = dataframe.loc[:, "genJetPt"]
binMean, binStd = getBinnedStatistics(variable, binningVariable, binning)
up = np.add(binMean, binStd)
down = np.add(binMean, -binStd)
controlBinMean, controlBinStd = getBinnedStatistics(dataframe.loc[:, "jetPt"]/dataframe.loc[:, "genJetPt"], binningVariable, binning)
controlUp = np.add(controlBinMean, controlBinStd)
controlDown = np.add(controlBinMean, -controlBinStd)
nonzeroPoints = (binMean > -1)
ax0.fill_between(binCenters[nonzeroPoints], up[nonzeroPoints], down[nonzeroPoints], alpha=0.6, color=colors[0])
ax0.plot(binCenters[nonzeroPoints], up[nonzeroPoints], c=colors[0], linewidth=1.0, alpha=0.8)
ax0.plot(binCenters[nonzeroPoints], down[nonzeroPoints], c=colors[0], linewidth=1.0, alpha=0.8)
ax0.scatter(binCenters[nonzeroPoints], binMean[nonzeroPoints], c=colors[0], label="DNN", linewidth=1.0, edgecolors='k')
# ax0.fill_between(binCenters[nonzeroPoints], controlUp[nonzeroPoints], controlDown[nonzeroPoints], alpha=0.6, color=colors[1])
ax0.plot(binCenters[nonzeroPoints], controlUp[nonzeroPoints], c=colors[1], linewidth=1.5, linestyle='--', alpha=0.8)
ax0.plot(binCenters[nonzeroPoints], controlDown[nonzeroPoints], c=colors[1], linewidth=1.5, linestyle='--', alpha=0.8)
ax0.scatter(binCenters[nonzeroPoints], controlBinMean[nonzeroPoints], c=colors[1], label="L1L2L3", linewidth=1.0, edgecolors='k')
ax0.plot([binCenters[0], binCenters[-1]], [1.0, 1.0], linestyle='--', linewidth=1.5, color='k')
ax0.set_ylim(0.8, 1.2)
locs = np.linspace(0.8, 1.2, 21)
ax0.set_yticks(locs)
ax0.set_xlim(binning[0], binning[-1])
ax0.set_ylabel("Mean response")
handles, labels = ax0.get_legend_handles_labels()
ax0.legend(handles[::-1], labels[::-1], loc='upper right')
ax0.spines['top'].set_visible(False)
ax0.spines['right'].set_visible(False)
ax0.spines['bottom'].set_visible(False)
ax0.spines['left'].set_visible(False)
ax0.grid(False, axis='x')
ax1.set_xlim(binning[0], binning[-1])
ax1.hist(dataframe.loc[:, "genJetPt"], bins=binning, alpha=1.0, edgecolor="black", linewidth=1.0, color=colors[0])
ax1.set_ylabel("Jets/bin")
plt.tight_layout(pad=2.2)
fig.align_ylabels((ax0, ax1))
plt.savefig("./plots/MeanResponse_{}.pdf".format(name))
plt.clf()
plt.close()
def plotLossHistory(history):
colors = [sns.xkcd_rgb["gold"], sns.xkcd_rgb["grass"]]
loss = history.history["loss"][1:]
valLoss = history.history["val_loss"][1:]
epochs = range(1, len(loss)+2)[1:]
fig = plt.figure(figsize=(8, 6))
plt.plot(epochs, loss, label="Training loss", color=colors[0], linewidth=2.0)
plt.plot(epochs, valLoss, label="Validation loss", color=colors[1], linewidth=2.0)
plt.xlabel("Epoch")
plt.ylabel("Loss function value")
plt.yscale('log')
fig.suptitle("Training and validation loss")
plt.legend()
plt.savefig("./plots/Losses.pdf")
plt.clf()
plt.close()