forked from Mukayyyy/del_jtvae
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmain_plot_figure_in_paper.py
executable file
·220 lines (183 loc) · 10.6 KB
/
main_plot_figure_in_paper.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
208
209
210
211
212
213
214
215
216
217
218
219
220
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Sun Aug 16 20:54:36 2020
@author: yifeng
"""
import os
import pandas as pd
from matplotlib import pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
from matplotlib.gridspec import GridSpec
import seaborn as sns
import numpy as np
from utils.plots import plot_counts_multi,plot_props_multi
def plot_dgm_loss_from_diff_expriments(run_dir1, run_dir2, run_dir3, group='epoch', postfix='PCBA'):
df1 = pd.read_csv(os.path.join(run_dir1, 'results/performance/loss_details.csv')) # 1
df2 = pd.read_csv(os.path.join(run_dir2, 'results/performance/loss_details.csv')) # 0.1
df3 = pd.read_csv(os.path.join(run_dir3, 'results/performance/loss_details.csv')) # 0.01
if group=='epoch':
df1 = df1.groupby('epoch').mean()
df2 = df2.groupby('epoch').mean()
df3 = df3.groupby('epoch').mean()
#fig, axs = plt.subplots(5, 1, sharex=True)
fig, axs = plt.subplots(2, 2)
fig.set_size_inches(8,4)
axs[0,0].plot(df1['CE_loss'], color='red', alpha=0.5, label=r'$\beta=1$')
axs[0,0].plot(df2['CE_loss'], color='blue', alpha=0.5, label=r'$\beta=0.1$')
axs[0,0].plot(df3['CE_loss'], color='olive', alpha=0.5, label=r'$\beta=0.01$')
axs[0,0].set_ylabel('CE Loss')
axs[0,0].set_xlim(0,49)
axs[0,0].legend(loc='upper right')
axs[0,1].plot(df1['KL_loss'], color='red', alpha=0.5, label=r'$\beta=1$')
axs[0,1].plot(df2['KL_loss'], color='blue', alpha=0.5, label=r'$\beta=0.1$')
axs[0,1].plot(df3['KL_loss'], color='olive', alpha=0.5, label=r'$\beta=0.01$')
axs[0,1].set_ylabel('KL Loss')
axs[0,1].set_xlim(0,49)
#axs[0,1].legend(loc='upper right')
axs[1,0].plot(df1['MSE_loss'], color='red', alpha=0.5, label=r'$\beta=1$')
axs[1,0].plot(df2['MSE_loss'], color='blue', alpha=0.5, label=r'$\beta=0.1$')
axs[1,0].plot(df3['MSE_loss'], color='olive', alpha=0.5, label=r'$\beta=0.01$')
axs[1,0].set_ylabel('MSE Loss')
axs[1,0].set_x_label('Epoch')
axs[1,0].set_xlim(0,49)
#axs[1,0].legend(loc='upper right')
axs[1,1].plot(df1['loss'], color='red', alpha=0.5, label=r'$\beta=1$')
axs[1,1].plot(df2['loss'], color='blue', alpha=0.5, label=r'$\beta=0.1$')
axs[1,1].plot(df3['loss'], color='olive', alpha=0.5, label=r'$\beta=0.01$')
axs[1,1].set_ylabel('Loss') #r'CE_Loss + $\beta$ MSE_Loss + $\alpha$ KL_loss'
axs[1,1].set_x_label('Epoch')
axs[1,1].set_xlim(0,49)
#axs[1,1].legend(loc='upper right')
filename='./fig_dgm_loss_details_'+group+'_'+postfix+'.pdf'
fig.savefig(filename,bbox_inches='tight')
def plot_del_loss_from_diff_expriments(run_dir1, run_dir2, run_dir3, run_dir4, group='epoch', postfix='PCBA'):
df1 = pd.read_csv(os.path.join(run_dir1, 'results/performance/loss_details.csv')) # 0.1
df2 = pd.read_csv(os.path.join(run_dir2, 'results/performance/loss_details.csv')) # 0.1 -> 0.4
df3 = pd.read_csv(os.path.join(run_dir3, 'results/performance/loss_details.csv')) # 0.01
df4 = pd.read_csv(os.path.join(run_dir4, 'results/performance/loss_details.csv')) # 0.01 -> 0.4
if group=='epoch':
df1 = df1.groupby('epoch').mean()
df2 = df2.groupby('epoch').mean()
df3 = df3.groupby('epoch').mean()
df4 = df4.groupby('epoch').mean()
#fig, axs = plt.subplots(5, 1, sharex=True)
fig, axs = plt.subplots(3, 2)
fig.set_size_inches(10,8)
axs[0,0].plot(df1['CE_loss'], color='red', alpha=0.5, label=r'$\beta=0.1, \alpha=1$')
axs[0,0].plot(df2['CE_loss'], color='orange', alpha=0.5, label=r'$\beta=0.1 \rightarrow 0.4, \alpha=1 \rightarrow 4$')
axs[0,0].plot(df3['CE_loss'], color='blue', alpha=0.5, label=r'$\beta=0.01, \alpha=1$')
axs[0,0].plot(df4['CE_loss'], color='steelblue', alpha=0.5, label=r'$\beta=0.01 \rightarrow 0.4, \alpha=1 \rightarrow 4$')
axs[0,0].set_ylabel('CE Loss')
#axs[0,0].set_xlim(0,49)
axs[0,0].legend(loc='upper right')
axs[0,1].plot(df1['KL_loss'], color='red', alpha=0.5, label=r'$\beta=0.1, \alpha=1$')
axs[0,1].plot(df2['KL_loss'], color='orange', alpha=0.5, label=r'$\beta=0.1 \rightarrow 0.4, \alpha=1 \rightarrow 4$')
axs[0,1].plot(df3['KL_loss'], color='blue', alpha=0.5, label=r'$\beta=0.01, \alpha=1$')
axs[0,1].plot(df4['KL_loss'], color='steelblue', alpha=0.5, label=r'$\beta=0.01 \rightarrow 0.4, \alpha=1 \rightarrow 4$')
axs[0,1].set_ylabel('KL Loss')
#axs[0,1].set_xlim(0,49)
#axs[0,1].legend(loc='upper right')
axs[1,0].plot(df1['MSE_loss'], color='red', alpha=0.5, label=r'$\beta=0.1, \alpha=1$')
axs[1,0].plot(df2['MSE_loss'], color='orange', alpha=0.5, label=r'$\beta=0.1 \rightarrow 0.4, \alpha=1 \rightarrow 4$')
axs[1,0].plot(df3['MSE_loss'], color='blue', alpha=0.5, label=r'$\beta=0.01, \alpha=1$')
axs[1,0].plot(df4['MSE_loss'], color='steelblue', alpha=0.5, label=r'$\beta=0.01 \rightarrow 0.4, \alpha=1 \rightarrow 4$')
axs[1,0].set_ylabel('MSE Loss')
#axs[1,0].set_xlim(0,49)
#axs[1,0].legend(loc='upper right')
axs[1,1].plot(df1['loss'], color='red', alpha=0.5, label=r'$\beta=0.1, \alpha=1$')
axs[1,1].plot(df2['loss'], color='orange', alpha=0.5, label=r'$\beta=0.1 \rightarrow 0.4, \alpha=1 \rightarrow 4$')
axs[1,1].plot(df3['loss'], color='blue', alpha=0.5, label=r'$\beta=0.01, \alpha=1$')
axs[1,1].plot(df4['loss'], color='steelblue', alpha=0.5, label=r'$\beta=0.01 \rightarrow 0.4, \alpha=1 \rightarrow 4$')
axs[1,1].set_ylabel('Loss') #r'CE_Loss + $\beta$ MSE_Loss + $\alpha$ KL_loss'
#axs[1,1].set_xlim(0,49)
#axs[1,1].legend(loc='upper right')
axs[2,0].plot(df1['beta'], color='red', alpha=0.5, label=r'$\beta=0.1, \alpha=1$')
axs[2,0].plot(df2['beta'], color='orange', alpha=0.5, label=r'$\beta=0.1 \rightarrow 0.4, \alpha=1 \rightarrow 4$')
axs[2,0].plot(df3['beta'], color='blue', alpha=0.5, label=r'$\beta=0.01, \alpha=1$')
axs[2,0].plot(df4['beta'], color='steelblue', alpha=0.5, label=r'$\beta=0.01 \rightarrow 0.4, \alpha=1 \rightarrow 4$')
axs[2,0].set_ylabel(r'$\beta$')
axs[2,0].set_x_label('Epoch')
#axs[2,0].legend(loc='upper right')
axs[2,1].plot(df1['alpha'], color='red', alpha=0.5, label=r'$\beta=0.1, \alpha=1$')
axs[2,1].plot(df2['alpha'], color='orange', alpha=0.5, label=r'$\beta=0.1 \rightarrow 0.4, \alpha=1 \rightarrow 4$')
axs[2,1].plot(df3['alpha'], color='blue', alpha=0.5, label=r'$\beta=0.01, \alpha=1$')
axs[2,1].plot(df4['alpha'], color='steelblue', alpha=0.5, label=r'$\beta=0.01 \rightarrow 0.4, \alpha=1 \rightarrow 4$')
axs[2,1].set_ylabel(r'$\alpha$')
axs[2,1].set_x_label('Epoch')
#axs[2,1].legend(loc='upper right')
filename='./fig_del_loss_details_'+group+'_'+postfix+'.pdf'
fig.savefig(filename,bbox_inches='tight')
def plot_figures_compare_dgm(run_dir1, run_dir2, run_dir3, DEL=False, filename='generated_from_random_aggregated'):
dataset_name = "ZINC" if "ZINC" in run_dir1 else "PCBA"
if DEL:
suffix='_del'
df1 = pd.read_csv(os.path.join(run_dir1, 'results/samples'+suffix+'/'+filename+'.csv'))
df0 = df1[df1.who==dataset_name]
df1 = df1[df1.who=='DEL(1)']
df1.who = r'$\beta$=1'
df2 = pd.read_csv(os.path.join(run_dir2, 'results/samples'+suffix+'/'+filename+'.csv'))
df2 = df2[df2.who=='DEL(1)']
df2.who = r'$\beta$=0.1'
df3 = pd.read_csv(os.path.join(run_dir3, 'results/samples'+suffix+'/'+filename+'.csv'))
df3 = df3[df3.who=='DEL(1)']
df3.who = r'$\beta$=0.01'
df = pd.concat( [df0, df1, df2, df3], ignore_index=True)
run_dir = './'
plot_counts_multi(df, dataset_name, dataset_name+'_compare_dgm', dirsave=os.path.join(run_dir, 'results_for_publication'))
plot_props_multi(df, dataset_name, dataset_name+'_compare_dgm', dirsave=os.path.join(run_dir, 'results_for_publication'))
def plot_MOBO(run_dir):
filename = run_dir + '/results/bo/hvs_random.csv'
sobol = np.loadtxt(filename, dtype=float, delimiter=',')
filename = run_dir + '/results/bo/hvs_qehvi.csv'
qehvi = np.loadtxt(filename, dtype=float, delimiter=',')
filename = run_dir + '/results/bo/hvs_qparego.csv'
qparego = np.loadtxt(filename, dtype=float, delimiter=',')
fig = plt.figure(figsize=[4,3])
ax = fig.add_subplot(111)
ax.plot(sobol, label='Sobol', color='olive', alpha=0.5)
ax.plot(qparego, label='qParEGO', color='blue', alpha=0.5)
ax.plot(qehvi, label='qEHVI', color='red', alpha=0.5)
ax.set_xlabel('Batch')
ax.set_ylabel('Hypervolume')
ax.legend(loc='upper left')
filename = run_dir + '/results/bo/fig_hvs.pdf'
fig.savefig(filename,bbox_inches='tight')
# DGM loss
run_dir1 = './RUNS/2020-08-15@10:39:34-yifeng-PCBA' # 1
run_dir2 = './RUNS/2020-08-12@23:59:17-yifeng-PCBA' # 2020-08-11@16:54:54-yifeng-PCBA 0.1
run_dir3 = './RUNS/2020-08-12@20:41:16-yifeng-PCBA' # 0.01
plot_dgm_loss_from_diff_expriments(run_dir1, run_dir2, run_dir3, group='epoch', postfix='PCBA')
run_dir1 = './RUNS/2020-08-15@19:58:55-yifeng-ZINC'
run_dir2 = './RUNS/2020-08-14@03:20:13-yifeng-ZINC'
run_dir3 = './RUNS/2020-08-13@22:56:46-yifeng-ZINC'
plot_dgm_loss_from_diff_expriments(run_dir1, run_dir2, run_dir3, group='epoch', postfix='ZINC')
# DEL loss
run_dir1 = './RUNS/2020-08-11@16:54:54-yifeng-PCBA' # 0.1
run_dir2 = './RUNS/2020-08-12@23:59:17-yifeng-PCBA' # 0.1 to 0.4
run_dir3 = './RUNS/2020-08-12@16:52:46-yifeng-PCBA' # 0.01
run_dir4 = './RUNS/2020-08-12@20:41:16-yifeng-PCBA' # 0.01 to 0.4
plot_del_loss_from_diff_expriments(run_dir1, run_dir2, run_dir3, run_dir4, group='epoch', postfix='PCBA')
run_dir1 = './RUNS/2020-08-13@08:17:08-yifeng-ZINC'
run_dir2 = './RUNS/2020-08-14@03:20:13-yifeng-ZINC'
run_dir3 = './RUNS/2020-08-13@18:51:18-yifeng-ZINC'
run_dir4 = './RUNS/2020-08-13@22:56:46-yifeng-ZINC'
plot_del_loss_from_diff_expriments(run_dir1, run_dir2, run_dir3, run_dir4, group='epoch', postfix='ZINC')
# DGM properties
run_dir1 = './RUNS/2020-08-15@10:39:34-yifeng-PCBA' # 1
run_dir2 = './RUNS/2020-08-12@23:59:17-yifeng-PCBA' # 2020-08-11@16:54:54-yifeng-PCBA 0.1
run_dir3 = './RUNS/2020-08-12@20:41:16-yifeng-PCBA' # 0.01
plot_figures_compare_dgm(run_dir1, run_dir2, run_dir3, DEL=True, filename='generated_from_random_aggregated')
run_dir1 = './RUNS/2020-08-15@19:58:55-yifeng-ZINC'
run_dir2 = './RUNS/2020-08-14@03:20:13-yifeng-ZINC'
run_dir3 = './RUNS/2020-08-13@22:56:46-yifeng-ZINC'
plot_figures_compare_dgm(run_dir1, run_dir2, run_dir3, DEL=True, filename='generated_from_random_aggregated')
# MOBO curves
run_dir = './RUNS/2020-08-12@20:41:16-yifeng-PCBA' # 0.01
plot_MOBO(run_dir)
run_dir = './RUNS/2020-08-12@23:59:17-yifeng-PCBA' # 0.1
plot_MOBO(run_dir)
run_dir = './RUNS/2020-08-14@03:20:13-yifeng-ZINC' # 0.1
plot_MOBO(run_dir)
run_dir = './RUNS/2020-08-13@22:56:46-yifeng-ZINC' # 0.01
plot_MOBO(run_dir)