-
Notifications
You must be signed in to change notification settings - Fork 1
/
dataset.py
140 lines (113 loc) · 5.36 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
import os
import pandas as pd
from torch.utils.data import Dataset
class MatchingDataset(Dataset):
def __init__(self, path, tableA, tableB):
assert os.path.isfile(path), "{} is not a file".format(path)
self.data = pd.read_csv(path)
self.tableA = tableA
self.tableB = tableB
self.ltable_id_list = [(i, 'l') for i in self.data["ltable_id"].unique().tolist()]
self.rtable_id_list = [(i, 'r') for i in self.data["rtable_id"].unique().tolist()]
self.center_id_list = self.ltable_id_list + self.rtable_id_list
self.examples = [self._make_example(*i) for i in self.center_id_list]
def __getitem__(self, i):
return self.examples[i]
def __len__(self):
try:
return len(self.examples)
except TypeError:
return 2**32
def __iter__(self):
for x in self.examples:
yield x
def _make_example(self, center_id, type='l'):
if type == 'l':
neighbor = self.data[self.data["ltable_id"] == center_id]
neighbor_ids = neighbor["rtable_id"].values.tolist()
center_example = self.tableA[self.tableA["id"] == center_id].values.tolist()[0]
neighbor_examples = [self.tableB[self.tableB["id"] == i].values.tolist()[0] for i in neighbor_ids]
neighbor_masks = [1] * len(neighbor_ids)
elif type == 'r':
neighbor = self.data[self.data["rtable_id"] == center_id]
neighbor_ids = neighbor["ltable_id"].values.tolist()
center_example = self.tableB[self.tableB["id"] == center_id].values.tolist()[0]
neighbor_examples = [self.tableA[self.tableA["id"] == i].values.tolist()[0] for i in neighbor_ids]
neighbor_masks = [1] * len(neighbor_ids)
else:
raise NotImplementedError
labels = neighbor["label"].values.tolist()
example = {
"type": type,
"center": center_example,
"neighbors": neighbor_examples,
"neighbors_mask": neighbor_masks,
"labels": labels,
}
return example
class MergedMatchingDataset(Dataset):
def __init__(self, path, tableA, tableB, other_path=None):
if other_path and not isinstance(other_path, list):
other_path = [other_path]
assert os.path.isfile(path), "{} is not a file".format(path)
for p in other_path:
if p:
assert os.path.isfile(p), "{} is not a file".format(p)
self.data = pd.read_csv(path)
if other_path:
self.other_data = pd.read_csv(other_path[0])
for p in other_path[1:]:
if p:
self.other_data = pd.concat([self.other_data, pd.read_csv(p)], axis=0)
else:
self.other_data = None
self.tableA = tableA
self.tableB = tableB
self.ltable_id_list = [(i, 'l') for i in self.data["ltable_id"].unique().tolist()]
self.rtable_id_list = [(i, 'r') for i in self.data["rtable_id"].unique().tolist()]
self.center_id_list = self.ltable_id_list + self.rtable_id_list
self.examples = [self._make_example(*i) for i in self.center_id_list]
def __getitem__(self, i):
return self.examples[i]
def __len__(self):
try:
return len(self.examples)
except TypeError:
return 2**32
def __iter__(self):
for x in self.examples:
yield x
def _make_example(self, center_id, type='l'):
if type == 'l':
self_neighbor = self.data[self.data["ltable_id"] == center_id]
self_neighbor_ids = self_neighbor["rtable_id"].values.tolist()
other_neighbor = self.other_data[self.other_data["ltable_id"] == center_id]
other_neighbor_ids = other_neighbor["rtable_id"].values.tolist()
neighbor_ids = self_neighbor_ids + other_neighbor_ids
center_example = self.tableA[self.tableA["id"] == center_id].values.tolist()[0]
neighbor_examples = [self.tableB[self.tableB["id"] == i].values.tolist()[0] for i in neighbor_ids]
# 0 for not use label, 1 for use label
neighbor_masks = [1] * len(self_neighbor_ids) + [0] * len(other_neighbor_ids)
elif type == 'r':
self_neighbor = self.data[self.data["rtable_id"] == center_id]
self_neighbor_ids = self_neighbor["ltable_id"].values.tolist()
other_neighbor = self.other_data[self.other_data["rtable_id"] == center_id]
other_neighbor_ids = other_neighbor["ltable_id"].values.tolist()
neighbor_ids = self_neighbor_ids + other_neighbor_ids
center_example = self.tableB[self.tableB["id"] == center_id].values.tolist()[0]
neighbor_examples = [self.tableA[self.tableA["id"] == i].values.tolist()[0] for i in neighbor_ids]
# 0 for not use label, 1 for use label
neighbor_masks = [1] * len(self_neighbor_ids) + [0] * len(other_neighbor_ids)
else:
raise NotImplementedError
labels = self_neighbor["label"].values.tolist() + other_neighbor["label"].values.tolist()
example = {
"type": type,
"center": center_example,
"neighbors": neighbor_examples,
"neighbors_mask": neighbor_masks,
"labels": labels
}
return example
def collate_fn(batch):
return batch