-
Notifications
You must be signed in to change notification settings - Fork 0
/
draw_pic.py
247 lines (228 loc) · 9.84 KB
/
draw_pic.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
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
#!/usr/bin/env python
# -*- encoding: utf-8 -*-
"""
@File : draw_pic.py
@Contact : [email protected]
@Modify Time @Author @Version @Desciption
------------ ------- -------- -----------
2022/11/17 13:51 zxx 1.0 None
"""
# import lib
import matplotlib.pyplot as plt
import matplotlib.cm as cm
import matplotlib.colors as mcolors
from matplotlib.gridspec import GridSpec
import numpy as np
import torch
from scipy.stats import gaussian_kde
import re
import os
def get_data_from_log(models_lst, dataset_name="Epinions"):
data_dict = {}
processed_data = {}
for m in models_lst:
# fn_lst = [n + '.txt' for n in os.listdir(f'./log/{m}/{dataset_name}') if 'txt' not in n]
fn_lst = os.listdir(f'./log/{m}/{dataset_name}')
data_dict[m] = {}
processed_data[m] = {}
data_dict[m]['HR@3'] = []
data_dict[m]['HR@5'] = []
data_dict[m]['HR@10'] = []
data_dict[m]['nDCG@3'] = []
data_dict[m]['nDCG@5'] = []
data_dict[m]['nDCG@10'] = []
for fn in fn_lst:
with open(f'./log/{m}/{dataset_name}/{fn}', 'r', encoding='utf-8') as f:
data_lst = f.readlines()[-15:]
datas = [eval(d) for d in re.compile(r'\d\.\d{4}').findall(' '.join(data_lst))]
# print(datas)
data_dict[m]['HR@3'].append(datas[:6])
data_dict[m]['HR@5'].append(datas[6:12])
data_dict[m]['HR@10'].append(datas[12:18])
data_dict[m]['nDCG@3'].append(datas[18:24])
data_dict[m]['nDCG@5'].append(datas[24:30])
data_dict[m]['nDCG@10'].append(datas[30:])
for metric in ['HR', 'nDCG']:
for k in [3, 5, 10]:
processed_data[m][f'{metric}@{k}'] = {}
data_array = np.array(data_dict[m][f'{metric}@{k}'])
assert data_array.shape == (3, 6)
mean_ = np.mean(data_array, axis=0)
std_ = np.std(data_array, axis=0, ddof=1)
processed_data[m][f'{metric}@{k}'].update({'mean': mean_, 'std': std_})
return processed_data
def draw_grouped_bars(models_lst, processed_data, data_name):
labels = ['total', '0-8', '8-16', '16-32', '32-64', '64-']
model_name = models_lst
# metric_name = ['HR@3', 'nDCG@3', 'HR@5', 'nDCG@5', 'HR@10', 'nDCG@10']
metric_name = ['HR@3', 'HR@5', 'HR@10', 'nDCG@3', 'nDCG@5', 'nDCG@10']
# xlims = [(0.14, 0.205), (0.18, 0.27), (0.28, 0.355), (0.1, 0.165), (0.13, 0.188), (0.15, 0.215)] # epinions
# xlims = [(0.1, 0.185), (0.13, 0.23), (0.17, 0.29), (0.08, 0.155), (0.09, 0.165), (0.11, 0.185)] # ciao
plt.style.use('ggplot')
plt.figure(figsize=(36, 24))
plt.rcParams.update({"font.size": 20})
plt.rc('font', family='Times New Roman')
for i in range(6):
metric = metric_name[i]
plt.subplot(eval(str(23) + str(i + 1)))
# plt.ylim(*xlims[i])
plt.xlim(-0.75, 7)
x = np.arange(len(labels))
width = 0.15
for i in range(5):
m = model_name[i]
y = processed_data[m][metric]['mean']
yerr = processed_data[m][metric]['std']
if m == 'FusionLightGCN':
m = 'F-LightGCN'
plt.bar(x - (i - 2) * width, y, width, label=m, yerr=yerr, error_kw=dict(elinewidth=1, capsize=4))
plt.ylabel('Scores', fontsize=34)
plt.xlabel('Groups', fontsize=34)
plt.title(metric, fontsize=50, y=1.03)
plt.xticks(x, labels=labels, fontsize=30)
plt.tick_params(labelsize=30)
plt.legend(loc='upper right', fontsize=24)
# plt.legend(loc=(0.73, 0.35), fontsize=20)
plt.suptitle(data_name, fontsize=80, y=0.98)
plt.subplots_adjust(left=0.08, right=0.95, top=0.9, bottom=0.05, hspace=0.25)
plt.show()
def draw_grouped_bars_plot(models_lst, processed_data, data_name, y_lst):
labels = ['total', '0-8', '8-16', '16-32', '32-64', '64-']
datasets = ['train', 'val', 'test']
markers = ['s', 'd', 'D', 'o', '*']
lines = ['-', '-.', '--', '-.', '-']
model_name = models_lst
metric_name = ['HR@3', 'HR@5', 'HR@10', 'nDCG@3', 'nDCG@5', 'nDCG@10']
plt.style.use('seaborn-paper')
# 获取调色盘
colors = plt.rcParams['axes.prop_cycle'].by_key()['color']
plt.rcParams.update({"font.size": 20})
plt.rc('font', family='Times New Roman')
fig = plt.figure(figsize=(36, 24))
for i in range(6):
metric = metric_name[i]
x = np.arange(len(labels))
width = 0.25
ax1 = fig.add_subplot(eval(str(23) + str(i + 1)))
# 画数据集边数分布图
for i in range(3):
y = np.log10([sum(y_lst[i])] + y_lst[i])
ax1.bar(x - (i - 1) * width, y, width, label=datasets[i], alpha=0.5, color=colors[i])
plt.legend(loc='upper left', fontsize=24)
plt.xticks(x, labels=labels, fontsize=30)
ax1.tick_params(axis='y',labelsize=25) # y轴字体大小设置
ax1.set_ylabel('Num of edges(log)', fontsize=34)
# 画指标折线图
ax2 = ax1.twinx()
for i in range(5):
m = model_name[i]
y = processed_data[m][metric]['mean']
yerr = processed_data[m][metric]['std']
if m == 'FusionLightGCN':
m = 'F-LightGCN'
if i > 2:
color = colors[3]
else:
color = colors[4]
ax2.errorbar(x, y, label=m, yerr=yerr,
linestyle=lines[i], color=color, marker=markers[i],
markersize=15, linewidth=4, capsize=5, capthick=3)
ax2.set_ylabel('Scores', fontsize=34)
ax2.set_xlabel('Groups', fontsize=34)
ax2.tick_params(axis='y', labelsize=25)
plt.title(metric, fontsize=50, y=1.03)
# plt.xticks(x, labels=labels, fontsize=30)
# plt.tick_params(labelsize=30)
plt.legend(loc='upper right', fontsize=24)
# plt.legend(loc=(0.73, 0.35), fontsize=20)
plt.suptitle(data_name, fontsize=80, y=0.98)
plt.subplots_adjust(left=0.08, right=0.95, top=0.9, bottom=0.05, hspace=0.25, wspace=0.25)
plt.show()
def draw_dist(g, dataset_name):
# 获取节点度数
x = (torch.log(g.out_degrees(etype='rate')) / torch.log(torch.tensor(2))).numpy()
y = (torch.log(g.out_degrees(etype='trust')) / torch.log(torch.tensor(2))).numpy()
xy = np.vstack([x, y])
# 拟合二维高斯分布
z = gaussian_kde(xy)(xy)
fig = plt.figure(figsize=(10, 8))
# 刻度的数量
grid_num = 12
plt.style.use('seaborn-paper')
plt.rc('font', family='Times New Roman')
# 用来放置散点图和边缘分布图
gs = GridSpec(4, 4)
ax_joint = fig.add_subplot(gs[1:4, 0:3])
ax_marg_x = fig.add_subplot(gs[0, 0:3])
ax_marg_y = fig.add_subplot(gs[1:4, 3])
# 绘制散点图
ax_joint.scatter(x, y, c=z, marker='o', cmap='Spectral')
ax_joint.grid()
# 用来加上标
sub_map = str.maketrans('0123456789', '⁰¹²³⁴⁵⁶⁷⁸⁹')
ax_joint.set_xticks(range(grid_num))
ax_joint.set_xticklabels(['2' + str(i).translate(sub_map) for i in range(grid_num)], fontsize=15)
ax_joint.set_yticks(range(grid_num))
ax_joint.set_yticklabels(['2' + str(i).translate(sub_map) for i in range(grid_num)], fontsize=15)
# 控制刻度最大最小值
ax_joint.set_xlim(left=0., right=grid_num - 1)
ax_joint.set_ylim(bottom=0., top=grid_num - 1)
# x边缘分布
n, bins, patches = ax_marg_x.hist(x, bins=range(grid_num), density=True, alpha=0.5, align='left')
# 拟合
marg_x_p = gaussian_kde(x)(bins)
ax_marg_x.plot(bins, marg_x_p, '--') # 绘制y的曲线
# y边缘分布
n, bins, patches = ax_marg_y.hist(y, orientation="horizontal", density=True, alpha=0.5, align='left')
# 拟合
marg_y_p = gaussian_kde(y)(bins)
ax_marg_y.plot(marg_y_p, bins, '--') # 绘制y的曲线
# 设置边缘分布刻度
ax_marg_x.set_xticks(range(grid_num))
ax_marg_x.set_xticklabels(range(grid_num), fontsize=15)
ax_marg_x.set_xlim(left=0., right=grid_num - 1)
ax_marg_y.set_yticks(range(grid_num))
ax_marg_y.set_yticklabels(range(grid_num), fontsize=15)
ax_marg_y.set_ylim(bottom=0., top=grid_num - 1)
# 不显示边缘分布刻度标签
plt.setp(ax_marg_x.get_xticklabels(), visible=False)
plt.setp(ax_marg_y.get_yticklabels(), visible=False)
# Set labels on joint
ax_joint.set_xlabel('Rate Degree', fontsize=20)
ax_joint.set_ylabel('Trust Degree', fontsize=20)
# Set labels on marginals
ax_marg_y.set_xlabel('Rate Degree Density', fontsize=10)
ax_marg_x.set_ylabel('Trust Degree Density', fontsize=10)
norm = mcolors.Normalize(vmin=np.min(z), vmax=np.max(z))
# 前面是值到颜色的映射，后面表示在那个子图的位置放
fig.colorbar(cm.ScalarMappable(norm=norm, cmap='Spectral'))
fig.suptitle(dataset_name, fontsize=40)
plt.show()
if __name__ == '__main__':
print(plt.style.available)
models_lst = ['TrustSVD', 'DiffnetPP', 'FusionLightGCN', 'LightGCN', 'MF'] # 列表中越靠右，bar越靠左，legend越靠下
data_name = 'Ciao'
processed_data = get_data_from_log(models_lst, data_name)
y_lst = {
'Yelp': [
[26855, 34889, 31901, 17547, 8873],
[4757, 4271, 3946, 2174, 1104],
[4751, 6125, 4796, 2428, 1163],
],
'Flickr': [
[9106, 17915, 33129, 44235, 132868],
[1763, 2211, 4123, 5518, 16594],
[1778, 3101, 4916, 6038, 17073],
],
'Epinions': [
[25350, 45149, 68707, 66700, 86678],
[4935, 5543, 8532, 8315, 10823],
[5038, 7811, 10210, 9140, 11219],
],
'Ciao': [
[6064, 14428, 25586, 27983, 52351],
[1069, 1778, 3183, 3483, 6540],
[1103, 2514, 3816, 3829, 6752],
]
}
draw_grouped_bars_plot(models_lst, processed_data, data_name, y_lst[data_name])