-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathdata.py
221 lines (202 loc) · 7.58 KB
/
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
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
from datasets import load_dataset, Dataset
import pandas as pd
import csv
import sys
import os
import logging
import pdb
log = logging.getLogger(__name__)
csv.field_size_limit(sys.maxsize)
def load_news_commentary_de_en():
base_path = os.path.dirname(__file__)
file_path = os.path.join(base_path,"data/news-commentary-v18/news-commentary-v18.de-en.tsv")
with open(file_path, mode="r", encoding="utf-8") as file:
tsv_reader = csv.reader(file, delimiter="\t")
dataset = []
id = 0
for row in tsv_reader:
example = {}
id += 1
try:
example['en'] = row[1]
example['de'] = row[0]
dataset.append(example)
except IndexError:
print(id,": ",row)
# dataset = Dataset.from_dict(dataset)
df = pd.DataFrame(dataset)
dataset = Dataset.from_pandas(df)
dataset = dataset.train_test_split(test_size = 0.1, seed = 2024)
return dataset
def load_news_commentary_de_ja():
base_path = os.path.dirname(__file__)
file_path = os.path.join(base_path,"data/news-commentary-v18/news-commentary-v18.de-ja.tsv")
with open(file_path, mode="r", encoding="utf-8") as file:
tsv_reader = csv.reader(file, delimiter="\t")
dataset = {'de':[],'ja':[]}
id = 0
for row in tsv_reader:
id += 1
try:
dataset['ja'].append(row[1])
dataset['de'].append(row[0])
except IndexError:
print(id,": ",row)
dataset = Dataset.from_dict(dataset)
dataset = dataset.train_test_split(test_size = 0.1, seed = 2024)
return dataset
def load_news_commentary_de_zh():
base_path = os.path.dirname(__file__)
file_path = os.path.join(base_path,"data/news-commentary-v18/news-commentary-v18.de-zh.tsv")
with open(file_path, mode="r", encoding="utf-8") as file:
tsv_reader = csv.reader(file, delimiter="\t")
dataset = {'de':[],'zh':[]}
id = 0
for row in tsv_reader:
id += 1
try:
dataset['zh'].append(row[1])
dataset['de'].append(row[0])
except IndexError:
print(id,": ",row)
dataset = Dataset.from_dict(dataset)
dataset = dataset.train_test_split(test_size = 0.1, seed = 2024)
return dataset
def load_news_commentary_en_ja():
base_path = os.path.dirname(__file__)
file_path = os.path.join(base_path,"data/news-commentary-v18/news-commentary-v18.en-ja.tsv")
with open(file_path, mode="r", encoding="utf-8") as file:
tsv_reader = csv.reader(file, delimiter="\t")
dataset = {'en':[],'ja':[]}
id = 0
for row in tsv_reader:
id += 1
try:
dataset['ja'].append(row[1])
dataset['en'].append(row[0])
except IndexError:
print(id,": ",row)
dataset = Dataset.from_dict(dataset)
dataset = dataset.train_test_split(test_size = 0.1, seed = 2024)
return dataset
def load_news_commentary_en_zh():
base_path = os.path.dirname(__file__)
file_path = os.path.join(base_path,"data/news-commentary-v18/news-commentary-v18.en-zh.tsv")
with open(file_path, mode="r", encoding="utf-8") as file:
tsv_reader = csv.reader(file, delimiter="\t")
dataset = {'en':[],'zh':[]}
id = 0
for row in tsv_reader:
id += 1
try:
dataset['zh'].append(row[1])
dataset['en'].append(row[0])
except IndexError:
print(id,": ",row)
dataset = Dataset.from_dict(dataset)
dataset = dataset.train_test_split(test_size = 0.1, seed = 2024)
return dataset
def load_news_commentary_ja_zh():
base_path = os.path.dirname(__file__)
file_path = os.path.join(base_path,"data/news-commentary-v18/news-commentary-v18.ja-zh.tsv")
with open(file_path, mode="r", encoding="utf-8") as file:
tsv_reader = csv.reader(file, delimiter="\t")
dataset = {'zh':[],'ja':[]}
id = 0
for row in tsv_reader:
id += 1
try:
dataset['ja'].append(row[0])
dataset['zh'].append(row[1])
except IndexError:
print(id,": ",row)
dataset = Dataset.from_dict(dataset)
dataset = dataset.train_test_split(test_size = 0.1, seed = 2024)
return dataset
def load_Tatoeba_en_ja():
base_path = os.path.dirname(__file__)
file_path = os.path.join(base_path,"data/tatoeba/tatoeba_jp_en_sentence.tsv")
with open(file_path, mode="r", encoding="utf-8") as file:
tsv_reader = csv.reader(file, delimiter="\t")
dataset = {'en':[],'ja':[]}
id = 0
for row in tsv_reader:
id += 1
try:
dataset['ja'].append(row[1])
dataset['en'].append(row[3])
except IndexError:
print(id,": ",row)
dataset = Dataset.from_dict(dataset)
dataset = dataset.train_test_split(test_size = 0.1, seed = 2024)
return dataset
def load_Tatoeba_en_zh():
base_path = os.path.dirname(__file__)
file_path = os.path.join(base_path,"data/tatoeba/tatoeba_en_zh_sentence.tsv")
with open(file_path, mode="r", encoding="utf-8") as file:
tsv_reader = csv.reader(file, delimiter="\t")
dataset = {'en':[],'zh':[]}
id = 0
for row in tsv_reader:
id += 1
try:
dataset['en'].append(row[3])
dataset['zh'].append(row[1])
except IndexError:
print(id,": ",row)
dataset = Dataset.from_dict(dataset)
dataset = dataset.train_test_split(test_size = 0.1, seed = 2024)
return dataset
def load_Tatoeba_en_de():
base_path = os.path.dirname(__file__)
file_path = os.path.join(base_path,"data/tatoeba/tatoeba_de_en_sentence.tsv")
with open(file_path, mode="r", encoding="utf-8") as file:
tsv_reader = csv.reader(file, delimiter="\t")
dataset = {'en':[],'de':[]}
id = 0
for row in tsv_reader:
id += 1
try:
dataset['de'].append(row[1])
dataset['en'].append(row[3])
except IndexError:
print(id,": ",row)
dataset = Dataset.from_dict(dataset)
dataset = dataset.train_test_split(test_size = 0.1, seed = 2024)
return dataset
def SeqToSeqEncode(example, input_lang, target_lang, tokenizer, max_length=None):
inputs = tokenizer(
example[input_lang],
return_tensors="pt",
padding=True,
truncation=True,
max_length=max_length,
)
outputs = tokenizer(
example[target_lang],
return_tensors="pt",
padding=True,
truncation=True,
max_length=max_length,
)
results = {
"input_ids": inputs["input_ids"],
"attention_mask": inputs["attention_mask"],
"labels": outputs["input_ids"],
"decoder_attention_mask": outputs["attention_mask"],
}
return results
DATASET_MAP = {
"news_commentary_de_en":load_news_commentary_de_en,
"news_commentary_de_zh":load_news_commentary_de_zh,
"news_commentary_de_ja":load_news_commentary_de_ja,
"news_commentary_en_zh":load_news_commentary_en_zh,
"news_commentary_en_ja":load_news_commentary_en_ja,
"news_commentary_ja_zh":load_news_commentary_ja_zh,
"tatoeba_en_ja":load_Tatoeba_en_ja,
"tatoeba_en_zh":load_Tatoeba_en_zh,
"tatoeba_de_en":load_Tatoeba_en_de
}
if __name__ =="__main__":
dataset = load_Tatoeba_en_ja()
print(dataset)