-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathloader.py
164 lines (131 loc) · 5.26 KB
/
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
162
163
from torch.utils.data import DataLoader
from dataset.dataset import MonoLingualData, ParallelData, collate_fn, paired_collate_fn
from dataset.tokenization import BertTokenizer
import codecs
import os
from logging import getLogger
import pickle
logger = getLogger()
def load_mono_data(params, vocab):
mono_data_list = ['simp_train', 'simp_dev', 'comp_train', 'comp_dev']
mono_data_path = []
mono_data = {'encdec':{}, 'otf':{}}
mono_data_path.append(params.simp_train_path)
mono_data_path.append(params.simp_dev_path)
mono_data_path.append(params.comp_train_path)
mono_data_path.append(params.comp_dev_path)
for path, name in zip(mono_data_path, mono_data_list):
assert os.path.isfile(path), path
logger.info("Loading data from %s ..." % path)
with codecs.open(path) as f:
read_file = f.readlines()
raw_data = [sent.strip() for sent in read_file]
data_mode = None
if 'comp' in name:
logger.info("Loading data from %s ..." % params.comp_frequent_list)
frequent_list = load_frequent_list(params.comp_frequent_list)
logger.info("Loading data from %s ..." % params.comp_ppdb_rules)
ppdb_rules = load_ppdb_rules(params.comp_ppdb_rules)
data_mode = 'comp'
elif 'simp' in name:
frequent_list = load_frequent_list(params.simp_frequent_list)
logger.info("Loading data from %s ..." % params.simp_ppdb_rules)
ppdb_rules = load_ppdb_rules(params.simp_ppdb_rules)
data_mode = 'simp'
else:
frequent_list, ppdb_rules = None, None
loader = DataLoader(
dataset=MonoLingualData(
params=params,
mono_data=raw_data,
word2index=vocab,
max_len=params.len_max_seq,
frequent_word_list=frequent_list,
ppdb_rules=ppdb_rules,
data_mode=data_mode,
train_mode='autoencoder'
),
batch_size=params.batch_size,
shuffle=True,
collate_fn=paired_collate_fn,
)
otf_loader = DataLoader(
dataset=MonoLingualData(
params=params,
mono_data=raw_data,
word2index=vocab,
max_len=params.len_max_seq,
frequent_word_list=frequent_list,
ppdb_rules=ppdb_rules,
data_mode=data_mode,
train_mode='otf'
),
batch_size=params.batch_size,
shuffle=True,
collate_fn=collate_fn,
)
mono_data['encdec'][name] = loader
mono_data['otf'][name] = otf_loader
return mono_data
def load_frequent_list(path):
frequent_word_list = set()
with codecs.open(path) as f:
for w in f.readlines():
if w not in frequent_word_list:
frequent_word_list.add(w)
return frequent_word_list
def load_ppdb_rules(path):
rule_files = codecs.open(path, mode='rb')
ppdb_rules = pickle.load(rule_files)
return ppdb_rules
def load_parallel_data(params, vocab):
para_data_list = ['dev', 'test']
para_data_path = []
para_data = {}
para_data_path.append(params.para_dev_path)
para_data_path.append(params.para_test_path)
if params.supervised_rate > 0:
para_data_list.append('train')
para_data_path.append(params.para_train_path)
tokenizer = BertTokenizer(params.vocab_path)
for path, name in zip(para_data_path, para_data_list):
assert os.path.isfile(path)
logger.info("Loading data from %s ..." % path)
with codecs.open(path) as f:
read_filev = f.readlines()
comp_sents = []
simp_sents = []
for i in read_filev:
line = i.strip().split('|')
comp_sents.append(tokenizer.tokenize(line[0]))
simp_sents.append(tokenizer.tokenize(line[1]))
if name == 'train':
batch_size = params.batch_size
else:
batch_size = 1
loader = DataLoader(
dataset=ParallelData(simp_sent=simp_sents, complex_sent=comp_sents, word2index=vocab, max_len=params.len_max_seq),
batch_size=batch_size,
shuffle=False,
collate_fn=paired_collate_fn,
)
para_data[name] = loader
return para_data
def load_vocab(params):
vocab_path = params.vocab_path
logger.info(vocab_path)
assert os.path.isfile(vocab_path)
vocab = []
with codecs.open(vocab_path) as f:
read_file = f.readlines()
for i in read_file:
vocab.append(i.strip())
return dict(zip(vocab, range(len(vocab)))), dict(zip(range(len(vocab)), vocab))
def load_data(params):
data = dict()
word2index, index2word = load_vocab(params)
data['index2word'] = index2word
data['word2index'] = word2index
data['mono'] = load_mono_data(params, word2index)
data['para'] = load_parallel_data(params, word2index)
return data