-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathdataset.py
231 lines (164 loc) · 9.5 KB
/
dataset.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
# -*- coding: utf-8 -*-
"""
Created on Thu Jun 17 19:32:48 2021
@author: Administrator
"""
import random
import torch
from enum import Enum
from torch.utils.data import Dataset
import numpy as np
import time
class Split(Enum):
TRAIN = 0
VALIDATE = 1
TEST = 2
class PoiDataset(Dataset):
def __init__(self, users, timeSeconds, timeSlots, coords, locs, catgs, catgLayers, sequence_length, split, loc_count):
self.users = users
self.timeSeconds = timeSeconds
self.timeSlots = timeSlots
self.coords = coords
self.locs = locs
self.catgs = catgs
self.catgsLayers = catgLayers
self.sequence_length = sequence_length
self.split = split
self.loc_count = loc_count
train_split = '2012-12-14 12:00:00' # '2012-08-14 12:00:00' # followd by Bi-STDDP
test_split = '2013-01-15 12:00:00' # '2012-11-15 12:00:00'
train_split_timetamp = time.mktime(time.strptime(train_split, "%Y-%m-%d %H:%M:%S"))
test_split_timetamp = time.mktime(time.strptime(test_split, "%Y-%m-%d %H:%M:%S"))
self.new_users = []
self.new_timeSeconds = []
self.new_timeSlots = []
self.new_coords = []
self.new_locs = []
self.new_catgs = []
self.new_catgLayers = []
for i, (user, timeSecond, timeSlot, coord, loc, catg, catgLayer) in enumerate(zip(self.users, self.timeSeconds, self.timeSlots, self.coords, self.locs, self.catgs, self.catgsLayers)):
item_len = len(timeSecond)
train_threshold = np.sum(np.array(timeSecond) < train_split_timetamp) # 计算训练样本数量
test_threshold = np.sum(np.array(timeSecond) < test_split_timetamp) # 计算测试样本数量
train_rate = train_threshold / item_len
# delete users who has less ten samples
'''（1）如果训练样本数量小于10；
（2）或者训练样本占比小于50%；
（3）或者测试样本小于1，则pass'''
if train_threshold <= (sequence_length*2+10) or train_rate < 0.7 or (item_len - test_threshold) <= (sequence_length*2+1):
# print(user+1)
# print(train_rate)
continue
if (split == Split.TRAIN):
self.new_users.append(user)
self.new_timeSeconds.append(timeSecond[:train_threshold])
self.new_timeSlots.append(timeSlot[:train_threshold])
self.new_coords.append(coord[:train_threshold])
self.new_locs.append(loc[:train_threshold])
self.new_catgs.append(catg[:train_threshold])
self.new_catgLayers.append(catgLayer[:train_threshold])
if (split == Split.VALIDATE):
if (test_threshold - train_threshold) <= (sequence_length*2+1):
continue
else:
self.new_users.append(user)
self.new_timeSeconds.append(timeSecond[train_threshold:test_threshold])
self.new_timeSlots.append(timeSlot[train_threshold:test_threshold])
self.new_coords.append(coord[train_threshold:test_threshold])
self.new_locs.append(loc[train_threshold:test_threshold])
self.new_catgs.append(catg[train_threshold:test_threshold])
self.new_catgLayers.append(catgLayer[train_threshold:test_threshold])
if (split == Split.TEST):
if (item_len - test_threshold) <= (sequence_length*2+1):
continue
else:
self.new_users.append(user)
self.new_timeSeconds.append(timeSecond[test_threshold:])
self.new_timeSlots.append(timeSlot[test_threshold:])
self.new_coords.append(coord[test_threshold:])
self.new_locs.append(loc[test_threshold:])
self.new_catgs.append(catg[test_threshold:])
self.new_catgLayers.append(catgLayer[test_threshold:])
self.sequence_users = []
self.sequence_pois_forward = []
self.sequence_pois_backward = []
self.sequence_catgs_forward = []
self.sequence_catgs_backward = []
self.sequence_catgLayers_forward = []
self.sequence_catgLayers_backward = []
self.sequence_timeSeconds_forward = []
self.sequence_timeSeconds_backward = []
self.sequence_timeSlots_forward = []
self.sequence_timeSlots_backward = []
self.sequence_coords_forward = []
self.sequence_coords_backward = []
self.sequence_query_timeSeconds = []
self.sequence_query_timeSlots = []
self.sequence_quary_coords = []
self.sequence_result_poi = []
self.sequence_result_catg = []
self.sequence_result_catgLayer = []
for i , (user, timeSecond, timeSlot, coord, loc, catg, catgLayer) in enumerate(zip(self.new_users, self.new_timeSeconds, self.new_timeSlots, self.new_coords, self.new_locs, self.new_catgs, self.new_catgLayers)):
# 样本数量
seq_count = len(loc) - sequence_length * 2
# for every user
for i in range(seq_count):
start_forward = i
end_forward = start_forward + sequence_length
start_backward = end_forward + 1
end_backward = start_backward + sequence_length
# for user
self.sequence_users.append(user)
# for timeSecond
self.sequence_timeSeconds_forward.append(timeSecond[start_forward : end_forward])
self.sequence_timeSeconds_backward.append(timeSecond[start_backward : end_backward][::-1])
# for timeSlot
self.sequence_timeSlots_forward.append(timeSlot[start_forward : end_forward])
self.sequence_timeSlots_backward.append(timeSlot[start_backward : end_backward][::-1])
# for timeSlot
# for coord
self.sequence_coords_forward.append(coord[start_forward : end_forward])
self.sequence_coords_backward.append(coord[start_backward : end_backward][::-1])
# for loc
self.sequence_pois_forward.append(loc[start_forward : end_forward])
self.sequence_pois_backward.append(loc[start_backward : end_backward][::-1])
# for catg
self.sequence_catgs_forward.append(catg[start_forward : end_forward])
self.sequence_catgs_backward.append(catg[start_backward : end_backward][::-1])
# for catgLayer
self.sequence_catgLayers_forward.append(catgLayer[start_forward : end_forward])
self.sequence_catgLayers_backward.append(catgLayer[start_backward : end_backward][::-1])
# for query item
self.sequence_query_timeSeconds.append(timeSecond[end_forward])
self.sequence_query_timeSlots.append(timeSlot[end_forward])
self.sequence_quary_coords.append(coord[end_forward])
# for result
self.sequence_result_poi.append(loc[end_forward])
self.sequence_result_catg.append(catg[end_forward])
self.sequence_result_catgLayer.append(catgLayer[end_forward])
print('用户：',len(self.new_users))
print('长度：',len(self.sequence_users))
self.sample_len = len(self.sequence_result_poi)
def __len__(self):
return self.sample_len
def __getitem__(self, index):
x_user = torch.tensor(self.sequence_users[index])
x_tf = torch.tensor(self.sequence_timeSeconds_forward[index])
x_tb= torch.tensor(self.sequence_timeSeconds_backward[index])
x_tsf = torch.tensor(self.sequence_timeSlots_forward[index])
x_tsb= torch.tensor(self.sequence_timeSlots_backward[index])
x_cof = torch.tensor(self.sequence_coords_forward[index])
x_cob= torch.tensor(self.sequence_coords_backward[index])
x_poi_f = torch.tensor(self.sequence_pois_forward[index])
x_poi_b= torch.tensor(self.sequence_pois_backward[index])
x_catg_f= torch.tensor(self.sequence_catgs_forward[index])
x_catg_b = torch.tensor(self.sequence_catgs_backward[index])
x_catgLayer_f = torch.tensor(self.sequence_catgLayers_forward[index])
x_catgLayer_b = torch.tensor(self.sequence_catgLayers_backward[index])
y_tsecond = torch.tensor(self.sequence_query_timeSeconds[index])
y_tslot = torch.tensor(self.sequence_query_timeSlots[index])
y_coord = torch.tensor(self.sequence_quary_coords[index])
y_poi = torch.tensor(self.sequence_result_poi[index])
y_catg = torch.tensor(self.sequence_result_catg[index])
y_catgLayer = torch.tensor(self.sequence_result_catgLayer[index])
return x_user, x_tf, x_tb, x_tsf, x_tsb, x_cof, x_cob, x_poi_f, x_poi_b, x_catg_f, x_catg_b, x_catgLayer_f, x_catgLayer_b, y_tsecond, y_tslot, y_coord, y_poi, y_catg, y_catgLayer