-
Notifications
You must be signed in to change notification settings - Fork 42
/
Copy pathdata_loader.py
166 lines (130 loc) · 6 KB
/
data_loader.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
import json
import random
from tqdm import tqdm
import logging
def train_generate_simple(dataset, batch_size, few, symbol2id):
logging.info('LOADING TRAINING DATA')
train_tasks = json.load(open(dataset + '/train_tasks.json'))
logging.info('LOADING CANDIDATES')
rel2candidates = json.load(open(dataset + '/rel2candidates.json'))
task_pool = list(train_tasks.keys())
num_tasks = len(task_pool)
rel_idx = 0
while True:
if rel_idx % num_tasks == 0:
random.shuffle(task_pool)
query = task_pool[rel_idx % num_tasks]
rel_idx += 1
candidates = rel2candidates[query]
train_and_test = train_tasks[query]
random.shuffle(train_and_test)
support_triples = train_and_test[:few]
support_pairs = [[symbol2id[triple[0]], symbol2id[triple[2]]] for triple in support_triples]
all_test_triples = train_and_test[few:]
if len(all_test_triples) < batch_size:
query_triples = [random.choice(all_test_triples) for _ in range(batch_size)]
else:
query_triples = random.sample(all_test_triples, batch_size)
query_pairs = [[symbol2id[triple[0]], symbol2id[triple[2]]] for triple in query_triples]
false_pairs = []
for triple in query_triples:
e_h = triple[0]
e_t = triple[2]
while True:
noise = random.choice(candidates)
if noise != e_t:
break
false_pairs.append([symbol2id[e_h], symbol2id[noise]])
yield support_pairs, query_pairs, false_pairs
def train_generate(dataset, batch_size, few, symbol2id, ent2id, e1rel_e2):
logging.info('LOADING TRAINING DATA')
train_tasks = json.load(open(dataset + '/train_tasks.json'))
logging.info('LOADING CANDIDATES')
rel2candidates = json.load(open(dataset + '/rel2candidates.json'))
task_pool = list(train_tasks.keys())
num_tasks = len(task_pool)
rel_idx = 0
while True:
if rel_idx % num_tasks == 0:
random.shuffle(task_pool)
query = task_pool[rel_idx % num_tasks]
rel_idx += 1
candidates = rel2candidates[query]
if len(candidates) <= 20:
# print 'not enough candidates'
continue
train_and_test = train_tasks[query]
random.shuffle(train_and_test)
support_triples = train_and_test[:few]
support_pairs = [[symbol2id[triple[0]], symbol2id[triple[2]]] for triple in support_triples]
support_left = [ent2id[triple[0]] for triple in support_triples]
support_right = [ent2id[triple[2]] for triple in support_triples]
all_test_triples = train_and_test[few:]
if len(all_test_triples) == 0:
continue
if len(all_test_triples) < batch_size:
query_triples = [random.choice(all_test_triples) for _ in range(batch_size)]
else:
query_triples = random.sample(all_test_triples, batch_size)
query_pairs = [[symbol2id[triple[0]], symbol2id[triple[2]]] for triple in query_triples]
query_left = [ent2id[triple[0]] for triple in query_triples]
query_right = [ent2id[triple[2]] for triple in query_triples]
false_pairs = []
false_left = []
false_right = []
for triple in query_triples:
e_h = triple[0]
rel = triple[1]
e_t = triple[2]
while True:
noise = random.choice(candidates)
if (noise not in e1rel_e2[e_h+rel]) and noise != e_t:
break
false_pairs.append([symbol2id[e_h], symbol2id[noise]])
false_left.append(ent2id[e_h])
false_right.append(ent2id[noise])
yield support_pairs, query_pairs, false_pairs, support_left, support_right, query_left, query_right, false_left, false_right
def train_generate_(dataset, batch_size, few, symbol2id, ent2id, e1rel_e2, num_neg=1):
logging.info('LOADING TRAINING DATA')
train_tasks = json.load(open(dataset + '/train_tasks.json'))
logging.info('LOADING CANDIDATES')
rel2candidates = json.load(open(dataset + '/rel2candidates.json'))
task_pool = list(train_tasks.keys())
num_tasks = len(task_pool)
rel_idx = 0
while True:
if rel_idx % num_tasks == 0:
random.shuffle(task_pool)
query = task_pool[rel_idx % num_tasks]
rel_idx += 1
candidates = rel2candidates[query]
train_and_test = train_tasks[query]
random.shuffle(train_and_test)
support_triples = train_and_test[:few]
support_pairs = [[symbol2id[triple[0]], symbol2id[triple[2]]] for triple in support_triples]
support_left = [ent2id[triple[0]] for triple in support_triples]
support_right = [ent2id[triple[2]] for triple in support_triples]
all_test_triples = train_and_test[few:]
if len(all_test_triples) < batch_size:
query_triples = [random.choice(all_test_triples) for _ in range(batch_size)]
else:
query_triples = random.sample(all_test_triples, batch_size)
query_pairs = [[symbol2id[triple[0]], symbol2id[triple[2]]] for triple in query_triples]
query_left = [ent2id[triple[0]] for triple in query_triples]
query_right = [ent2id[triple[2]] for triple in query_triples]
labels = [1] * len(query_triples)
# sample negtive samples for every true triples
for triple in query_triples:
e_h = triple[0]
e_t = triple[2]
if e_t in candidates: candidates.remove(e_t)
if len(candidates) >=num_neg:
noises = random.sample(candidates, num_neg)
else:
noises = candidates
for noise in noises:
query_pairs.append([symbol2id[e_h], symbol2id[noise]])
query_left.append(ent2id[e_h])
query_right.append(ent2id[noise])
labels.append(0)
yield support_pairs, query_pairs, support_left, support_right, query_left, query_right, labels