-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathRatingDataset.py
315 lines (234 loc) · 12.4 KB
/
RatingDataset.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
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
import scipy.sparse as sp
import numpy as np
from torch.utils.data import Dataset
class TestDataset(Dataset):
def __init__(self, train_dataset, test_data_path):
self.feature_size = train_dataset.feature_size
self.metapath_list = train_dataset.metapath_list
self.features = train_dataset.features
self.type2id = train_dataset.type2id
self.id2type = train_dataset.id2type
self.load_test_ratings(test_data_path)
assert len(self.test_ratings_list) == len(self.test_neg_ratings_list)
self.max_test_item_num = 0
for i in range(len(self.test_ratings_list)):
length = len(self.test_ratings_list[i][1:]) + len(self.test_neg_ratings_list[i])
if length > self.max_test_item_num:
self.max_test_item_num = length
print("test ratings %d, test negative rating lists %d, maximum test item number for one rating %d" % (
len(self.test_ratings_list), len(self.test_neg_ratings_list), self.max_test_item_num))
def load_test_ratings(self, filename):
self.test_ratings_list = []
self.test_neg_ratings_list = []
with open(filename, "r") as input:
for line in input.read().splitlines():
tokens = line.split(",")
# id starts from 0
true_ratings = [int(tokens[0])]
true_ratings.extend([i for i in list(map(int, tokens[1].split(" ")))])
negative_ratings = [i for i in list(map(int, tokens[2].split(" ")))]
self.test_ratings_list.append(true_ratings)
self.test_neg_ratings_list.append(negative_ratings)
def __len__(self):
return len(self.test_ratings_list)
def __getitem__(self, idx):
negative_id_num = len(self.test_neg_ratings_list[idx])
real_test_item_size = len(self.test_ratings_list[idx][1:]) + negative_id_num # self.max_test_item_num
real_test_item_sizes = np.empty(self.max_test_item_num, dtype=int)
real_test_item_sizes.fill(real_test_item_size)
positive_item_indices = np.empty(self.max_test_item_num, dtype=int)
positive_item_indices.fill(negative_id_num)
test_item_ids = [0] * self.max_test_item_num
u = self.test_ratings_list[idx][0]
test_item_ids[0:negative_id_num] = self.test_neg_ratings_list[idx]
test_item_ids[negative_id_num:real_test_item_size] = self.test_ratings_list[idx][1:]
# test_item_ids = tuple(test_item_ids)
user_input = np.zeros(self.max_test_item_num, dtype=int)
item_input = np.zeros(self.max_test_item_num, dtype=int)
metapath_input_list = []
for i in range(len(self.metapath_list)):
# metapath_list[i]: metapath_file, path_dict, max_path_num, hop_num
metapath_input_list.append(
np.zeros((self.max_test_item_num, self.metapath_list[i][2], self.metapath_list[i][3],
self.feature_size), dtype=np.float32))
k = 0
# negative item ids
for i in self.test_neg_ratings_list[idx]:
user_input[k] = u
item_input[k] = i
for metapath_idx in range(len(self.metapath_list)):
if (u, i) in self.metapath_list[metapath_idx][1]:
for p_i in range(len(self.metapath_list[metapath_idx][1][(u, i)])):
for p_j in range(len(self.metapath_list[metapath_idx][1][(u, i)][p_i])):
type_id = self.metapath_list[metapath_idx][1][(u, i)][p_i][p_j][0]
node_id = self.metapath_list[metapath_idx][1][(u, i)][p_i][p_j][1]
node_type = self.id2type[type_id]
metapath_input_list[metapath_idx][k][p_i][p_j] = self.features[node_type][node_id]
k += 1
# positive item ids
for i in self.test_ratings_list[idx][1:]:
user_input[k] = u
item_input[k] = i
for metapath_idx in range(len(self.metapath_list)):
if (u, i) in self.metapath_list[metapath_idx][1]:
for p_i in range(len(self.metapath_list[metapath_idx][1][(u, i)])):
for p_j in range(len(self.metapath_list[metapath_idx][1][(u, i)][p_i])):
type_id = self.metapath_list[metapath_idx][1][(u, i)][p_i][p_j][0]
node_id = self.metapath_list[metapath_idx][1][(u, i)][p_i][p_j][1]
node_type = self.id2type[type_id]
metapath_input_list[metapath_idx][k][p_i][p_j] = self.features[node_type][node_id]
k += 1
# metapath_input_list[i]: metapath_file, path_dict, path_num, hop_num, feature_size
# data = [real_test_item_size, positive_item_indices, test_item_ids, user_input, item_input]
data = [real_test_item_sizes, positive_item_indices, np.array(test_item_ids), user_input, item_input]
data.extend(metapath_input_list)
return tuple(data)
class TrainDataset(Dataset):
def __init__(self, train_data_path, metapath_file_paths, negative_num, feature_file_dict):
self.negative_num = negative_num
self.load_train_ratings(train_data_path)
self.load_feature_as_map(feature_file_dict)
self.load_metapath(metapath_file_paths)
print("max_user_id %d, max_item_id %d, train ratings %d" % (
self.max_user_id, self.max_item_id, self.train_rating_mat.nnz))
def __len__(self):
return self.train_rating_mat.nnz
def __getitem__(self, idx):
u, i = self.user_item_pairs[idx]
user_input = np.zeros(self.negative_num + 1, dtype=int)
item_input = np.zeros(self.negative_num + 1, dtype=int)
labels = np.zeros(self.negative_num + 1, dtype=np.float32)
counter = 0
user_input[counter] = u
item_input[counter] = i
labels[counter] = 1
# metapath: (metapath_file, path_dict, path_num, hop_num)
metapath_input_list = []
for metapath in self.metapath_list:
# PyTorch uses row-wist representation
metapath_input = np.zeros((self.negative_num + 1, metapath[2], metapath[3], self.feature_size),
dtype=np.float32)
if (u, i) in metapath[1]:
for p_i in range(len(metapath[1][(u, i)])):
for p_j in range(len(metapath[1][(u, i)][p_i])):
type_id = metapath[1][(u, i)][p_i][p_j][0]
node_id = metapath[1][(u, i)][p_i][p_j][1]
node_type = self.id2type[type_id]
metapath_input[counter][p_i][p_j] = self.features[node_type][node_id]
metapath_input_list.append(metapath_input)
for t in range(self.negative_num):
counter += 1
j = np.random.randint(1, self.max_item_id + 1)
while j in self.user_item_map[u]:
j = np.random.randint(1, self.max_item_id + 1)
user_input[counter] = u
item_input[counter] = j
labels[counter] = 0
for list_index in range(len(metapath_input_list)):
if (u, j) in self.metapath_list[list_index][1]:
for p_i in range(len(self.metapath_list[list_index][1][(u, j)])):
for p_j in range(len(self.metapath_list[list_index][1][(u, j)][p_i])):
type_id = self.metapath_list[list_index][1][(u, j)][p_i][p_j][0]
node_id = self.metapath_list[list_index][1][(u, j)][p_i][p_j][1]
node_type = self.id2type[type_id]
metapath_input_list[list_index][counter][p_i][p_j] = self.features[node_type][node_id]
data = [user_input, item_input, labels]
data.extend(metapath_input_list)
return tuple(data)
def load_train_ratings(self, filename):
self.max_user_id, self.max_item_id = 0, 0
with open(filename, "r") as input:
for line in input.read().splitlines():
arr = line.split(" ")
u, i = int(arr[0]), int(arr[1])
self.max_user_id = max(self.max_user_id, u)
self.max_item_id = max(self.max_item_id, i)
# id starts from 1, add one more id 0 for invalid updates
shape = (self.max_user_id + 1, self.max_item_id + 1)
self.train_rating_mat = sp.dok_matrix(shape, dtype=np.float32)
self.user_item_map = {}
self.item_user_map = {}
self.user_item_pairs = []
with open(filename, "r") as input:
for line in input.read().splitlines():
arr = line.split(" ")
user, item = int(arr[0]), int(arr[1])
self.train_rating_mat[user, item] = 1.0
if user not in self.user_item_map:
self.user_item_map[user] = {}
if item not in self.item_user_map:
self.item_user_map[item] = {}
self.user_item_map[user][item] = 1.0
self.item_user_map[item][user] = 1.0
self.user_item_pairs.append([user, item])
def load_feature_as_map(self, feature_file_dict):
self.features = {}
self.node_sizes = {}
self.feature_size = -1
for feature_type, feature_file_path in feature_file_dict.items():
with open(feature_file_path) as input:
count = 0
for line in input.read().splitlines():
line = line.strip()
if line == "":
continue
count += 1
arr = line.split(',')
if self.feature_size == -1:
self.feature_size = len(arr) - 1
else:
assert (self.feature_size == (len(arr) - 1))
self.node_sizes[feature_type] = count
for feature_type, feature_file_path in feature_file_dict.items():
self.features[feature_type] = np.zeros((self.node_sizes[feature_type] + 1, self.feature_size),
dtype=np.float32)
with open(feature_file_path) as input:
for line in input.readlines():
line = line.strip()
if line == "":
continue
arr = line.strip().split(',')
node_id = int(arr[0])
for j in range(len(arr[1:])):
self.features[feature_type][node_id][j] = float(arr[j + 1])
def load_metapath(self, metapath_files):
self.type2id = {}
self.id2type = {}
tmp_type_set = set()
for metapath_file in metapath_files:
with open(metapath_file) as input:
for line in input.read().splitlines():
arr = line.split('\t')
for path in arr[2:]:
nodes = path.split(' ')[0].split('-')
for node in nodes:
tmp_type_set.add(node[0])
for node_type in tmp_type_set:
id = len(self.id2type)
self.type2id[node_type] = id
self.id2type[id] = node_type
print("node type " + str(self.type2id))
self.metapath_list = []
for metapath_file in metapath_files:
path_dict = {}
max_path_num = 0
hop_num = 0
with open(metapath_file) as input:
for line in input.read().splitlines():
arr = line.split('\t')
max_path_num = max(int(arr[1]), max_path_num)
hop_num = len(arr[2].strip().split('-'))
with open(metapath_file) as input:
for line in input.read().splitlines():
arr = line.strip().split('\t')
u, i = arr[0].split(',')
u, i = int(u), int(i)
path_dict[(u, i)] = []
for path in arr[2:]:
tmp = path.split(' ')[0].split('-')
node_list = []
for node in tmp:
index = int(node[1:]) - 1
node_list.append([self.type2id[node[0]], index])
path_dict[(u, i)].append(node_list)
self.metapath_list.append((metapath_file, path_dict, max_path_num, hop_num))