forked from THU-KEG/EAkit
-
Notifications
You must be signed in to change notification settings - Fork 0
/
load_data.py
150 lines (131 loc) · 6.3 KB
/
load_data.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
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
File name: load_data.py
Author: locke
Date created: 2020/3/25 下午7:00
"""
import time
import numpy as np
class AlignmentData:
def __init__(self, data_dir="data/DBP15K/ja_en", rate=0.3, share=False, swap=False, val=0.0, with_r=False):
t_ = time.time()
self.rate = rate
self.val = val
self.ins2id_dict, self.id2ins_dict, [self.kg1_ins_ids, self.kg2_ins_ids] = self.load_dict(data_dir + "/ent_ids_", file_num=2)
self.rel2id_dict, self.id2rel_dict, [self.kg1_rel_ids, self.kg2_rel_ids] = self.load_dict(data_dir + "/rel_ids_", file_num=2)
self.ins_num = len(self.ins2id_dict)
self.rel_num = len(self.rel2id_dict)
self.triple_idx = self.load_triples(data_dir + "/triples_", file_num=2)
self.ill_idx = self.load_triples(data_dir + "/ill_ent_ids", file_num=1)
np.random.shuffle(self.ill_idx)
self.ill_train_idx, self.ill_val_idx, self.ill_test_idx = np.array(self.ill_idx[:int(len(self.ill_idx) // 1 * rate)], dtype=np.int32), np.array(self.ill_idx[int(len(self.ill_idx) // 1 * rate) : int(len(self.ill_idx) // 1 * (rate+val))], dtype=np.int32), np.array(self.ill_idx[int(len(self.ill_idx) // 1 * (rate+val)):], dtype=np.int32)
self.ins_G_edges_idx, self.ins_G_values_idx, self.r_ij_idx = self.gen_sparse_graph_from_triples(self.triple_idx, self.ins_num, with_r)
assert (share != swap or (share == False and swap == False))
if share:
self.triple_idx = self.share(self.triple_idx, self.ill_train_idx) # 1 -> 2:base
self.kg1_ins_ids = (self.kg1_ins_ids - set(self.ill_train_idx[:, 0])) | set(self.ill_train_idx[:, 1])
self.ill_train_idx = []
if swap:
self.triple_idx = self.swap(self.triple_idx, self.ill_train_idx)
self.labeled_alignment = set()
self.boot_triple_idx = []
self.boot_pair_dix = []
self.init_time = time.time() - t_
def load_triples(self, data_dir, file_num=2):
if file_num == 2:
file_names = [data_dir + str(i) for i in range(1, 3)]
else:
file_names = [data_dir]
triple = []
for file_name in file_names:
with open(file_name, "r", encoding="utf-8") as f:
data = f.read().strip().split("\n")
data = [tuple(map(int, i.split("\t"))) for i in data]
triple += data
np.random.shuffle(triple)
return triple
def load_dict(self, data_dir, file_num=2):
if file_num == 2:
file_names = [data_dir + str(i) for i in range(1, 3)]
else:
file_names = [data_dir]
what2id, id2what, ids = {}, {}, []
for file_name in file_names:
with open(file_name, "r", encoding="utf-8") as f:
data = f.read().strip().split("\n")
data = [i.split("\t") for i in data]
what2id = {**what2id, **dict([[i[1], int(i[0])] for i in data])}
id2what = {**id2what, **dict([[int(i[0]), i[1]] for i in data])}
ids.append(set([int(i[0]) for i in data]))
return what2id, id2what, ids
def gen_sparse_graph_from_triples(self, triples, ins_num, with_r=False):
edge_dict = {}
for (h, r, t) in triples:
if h != t:
if (h, t) not in edge_dict:
edge_dict[(h, t)] = []
edge_dict[(t, h)] = []
edge_dict[(h, t)].append(r)
edge_dict[(t, h)].append(-r)
if with_r:
edges = [[h, t] for (h, t) in edge_dict for r in edge_dict[(h, t)]]
values = [1 for (h, t) in edge_dict for r in edge_dict[(h, t)]]
r_ij = [abs(r) for (h, t) in edge_dict for r in edge_dict[(h, t)]]
edges = np.array(edges, dtype=np.int32)
values = np.array(values, dtype=np.float32)
r_ij = np.array(r_ij, dtype=np.float32)
return edges, values, r_ij
else:
edges = [[h, t] for (h, t) in edge_dict]
values = [1 for (h, t) in edge_dict]
# add self-loop
edges += [[e, e] for e in range(ins_num)]
values += [1 for e in range(ins_num)]
edges = np.array(edges, dtype=np.int32)
values = np.array(values, dtype=np.float32)
return edges, values, None
def share(self, triples, ill):
from_1_to_2_dict = dict(ill)
new_triples = []
for (h, r, t) in triples:
if h in from_1_to_2_dict:
h = from_1_to_2_dict[h]
if t in from_1_to_2_dict:
t = from_1_to_2_dict[t]
new_triples.append((h, r, t))
new_triples = list(set(new_triples))
return new_triples
def swap(self, triples, ill):
from_1_to_2_dict = dict(ill)
from_2_to_1_dict = dict(ill[:, ::-1])
new_triples = []
for (h, r, t) in triples:
new_triples.append((h, r, t))
if h in from_1_to_2_dict:
new_triples.append((from_1_to_2_dict[h], r, t))
if t in from_1_to_2_dict:
new_triples.append((h, r, from_1_to_2_dict[t]))
if h in from_2_to_1_dict:
new_triples.append((from_2_to_1_dict[h], r, t))
if t in from_2_to_1_dict:
new_triples.append((h, r, from_2_to_1_dict[t]))
new_triples = list(set(new_triples))
return new_triples
def __repr__(self):
return self.__class__.__name__ + " dataset summary:" + \
"\n\tins_num: " + str(self.ins_num) + \
"\n\trel_num: " + str(self.rel_num) + \
"\n\ttriple_idx: " + str(len(self.triple_idx)) + \
"\n\trate: " + str(self.rate) + "\tval: " + str(self.val) + \
"\n\till_idx(train/test/val): " + str(len(self.ill_idx)) + " = " + str(len(self.ill_train_idx)) + " + " + str(len(self.ill_test_idx)) + " + " + str(len(self.ill_val_idx)) + \
"\n\tins_G_edges_idx: " + str(len(self.ins_G_edges_idx)) + \
"\n\t----------------------------- init_time: " + str(round(self.init_time, 3)) + "s"
if __name__ == '__main__':
# TEST
d = AlignmentData(share=False, swap=False)
print(d)
d = AlignmentData(share=True, swap=False)
print(d)
d = AlignmentData(share=False, swap=True)
print(d)