-
Notifications
You must be signed in to change notification settings - Fork 0
/
prepare_data.py
195 lines (161 loc) · 8.02 KB
/
prepare_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
import os
import re
import json
from json import JSONDecodeError
from glob import glob
from tqdm import tqdm
from typing import List, Union, Dict, Any
from datasets import load_dataset
from datasets.arrow_dataset import Batch
def clean(s):
s = s.replace("\n", "")
s = s.replace("\t", "")
s = re.sub(", +}", ",}", s)
s = re.sub(", +]", ",]", s)
s = s.replace(",}", "}")
s = s.replace(",]", "]")
s = s.replace("'", "\"")
s = s.replace(".,", ",")
return s
def prepare_dialect_dataset(filenames: List[str]):
results = []
for filename in tqdm(filenames):
with open(filename, encoding="utf-8-sig") as f:
s = f.read()
try:
data = json.loads(s)
except JSONDecodeError:
data = json.loads(clean(s))
utterance = data["utterance"]
for u in utterance:
if u["standard_form"] != u["dialect_form"]:
dialect_idx = []
for e in u["eojeolList"]:
if e["isDialect"]:
dialect_idx.append(e["id"])
sample = {
"id": u["id"],
"do": filename.split("/")[1],
"standard": u["standard_form"],
"dialect": u["dialect_form"],
"dialect_idx": dialect_idx,
}
results.append(sample)
return results
NAMES = ['NAME', 'NAEM', 'anem', 'anme', 'mane', 'naem', 'nam', 'nmae', '이름', '고자영', '최미영']
PAT_LIST = [
'&NAEM4&', '&NAME&', '&NAME18&', '&adderess2&', '&adderss11&', '&address&', '&address1&',
'&address10&', '&address11&', '&address12&', '&address13&', '&address14&', '&address15&',
'&address16&', '&address17&', '&address18&', '&address19&', '&address2&', '&address20&',
'&address21&', '&address22&', '&address23&', '&address3&', '&address4&', '&address5&',
'&address6&', '&address7&', '&address8&', '&address9&', '&addressa&', '&adress&', '&anem6&',
'&anme1&', '&anme5&', '&anme6&', '&company2&', '&company3&', '&company_name1&', '&company_name2&',
'&mane1&', '&mane4&', '&mane5&', '&naem1&', '&naem16&', '&naem2&', '&naem6&', '&naem7&',
'&naem9&', '&nam13&', '&nam16e&', '&nam1e&', '&nam3&', '&nam4&', '&nam51&', '&nam7&',
'&namE5&', '&name&', '&name0&', '&name1&', '&name10&', '&name11&', '&name12&', '&name13&',
'&name14&', '&name145&', '&name15&', '&name16&', '&name17&', '&name18&', '&name19&',
'&name2&', '&name20&', '&name21&', '&name22&', '&name23&', '&name24&', '&name25&',
'&name26&', '&name27&', '&name28&', '&name29&', '&name3&', '&name30&', '&name31&',
'&name32&', '&name33&', '&name34&', '&name35&', '&name36&', '&name37&', '&name38&',
'&name39&', '&name4&', '&name40&', '&name41&', '&name42&', '&name43&', '&name44&',
'&name45&', '&name46&', '&name47&', '&name48&', '&name49&', '&name5&', '&name50&',
'&name51&', '&name52&', '&name54&', '&name55&', '&name56&', '&name57&', '&name59&',
'&name6&', '&name60&', '&name61&', '&name62&', '&name63&', '&name64&', '&name65&',
'&name67&', '&name68&', '&name7&', '&name8&', '&name9&', '&names5&', '&nmae2&', '&nmae3&',
'&가자&', '&고자영2&', '&상호명1&', '&상호명2&', '&서연림1&', '&선옥언니&', '&월령&',
'&유튜브&', '&이름1&', '&이름2&', '&이름4&', '&이름5&', '&인가&', '&좌미영2&', '&한림농협&'
]
PAT_MAP = {}
for p in PAT_LIST:
PAT_MAP[p] = "[OTHER]"
if "add" in p:
PAT_MAP[p] = "[ADDRESS]"
for n in NAMES:
if n in p:
PAT_MAP[p] = "[NAME]"
PATTERN1 = re.compile("(\(\(\)\))|(\{\w+\})|[#-]")
PATTERN2 = re.compile("(\(\(\w+\)\))")
PATTERN3 = re.compile("(&\w+&)")
PATTERN4 = re.compile("\((\w+)\)/\((\w+)\)")
def preprocess(examples: Batch) -> Union[Dict, Any]:
ids = examples["id"]
dos = examples["do"]
standard_texts = examples["standard"]
dialect_texts = examples["dialect"]
dialect_idxs = examples["dialect_idx"]
new_examples = {"id": [], "do": [], "standard": [], "dialect": [], "dialect_idx": []}
iterator = zip(ids, dos, standard_texts, dialect_texts, dialect_idxs)
for _id, do, standard_text, dialect_text, dialect_idx in iterator:
# remove (()), {\w+} patterns
standard_text = re.sub(PATTERN1, "", standard_text).replace(" ", " ")
dialect_text = re.sub(PATTERN1, "", dialect_text).replace(" ", " ")
# remove \n, \t patterns
standard_text = re.sub("[\t\n]", " ", standard_text).replace(" ", " ")
dialect_text = re.sub("[\t\n]", " ", dialect_text).replace(" ", " ")
# remove sample which has ((\w+)) patterns
if PATTERN2.findall(standard_text) + PATTERN2.findall(dialect_text):
continue
# $\w+$ pattern mapping
for k in PATTERN3.findall(standard_text):
standard_text = standard_text.replace(k, PAT_MAP.get(k, "[OHTER]"))
for k in PATTERN3.findall(dialect_text):
dialect_text = dialect_text.replace(k, PAT_MAP.get(k, "[OHTER]"))
# (\w+)/(\w+)
standard_text = re.sub(PATTERN4, r"\2", standard_text)
dialect_text = re.sub(PATTERN4, r"\1", dialect_text)
new_examples["id"].append(_id)
new_examples["do"].append(do)
new_examples["standard"].append(standard_text)
new_examples["dialect"].append(dialect_text)
new_examples["dialect_idx"].append(dialect_idx)
return new_examples
def prepare_for_style_classification(examples: Batch) -> Union[Dict, Any]:
ids = examples["id"]
dos = examples["do"]
standard_texts = examples["standard"]
dialect_texts = examples["dialect"]
new_examples = {"id": [], "do": [], "text": [], "label": []}
iterator = zip(ids, dos, standard_texts, dialect_texts)
for _id, do, standard_text, dialect_text in iterator:
new_examples["id"].extend([_id, _id])
new_examples["do"].extend([do, do])
new_examples["text"].extend([standard_text, dialect_text])
new_examples["label"].extend([0, 1])
return new_examples
def prepare_for_style_transfer(examples: Batch) -> Union[Dict, Any]:
ids = examples["id"]
dos = examples["do"]
standard_texts = examples["standard"]
dialect_texts = examples["dialect"]
new_examples = {"id": [], "source": [], "target": [], "src_lang": [], "tgt_lang": []}
iterator = zip(ids, dos, standard_texts, dialect_texts)
for _id, do, standard_text, dialect_text in iterator:
new_examples["id"].extend([_id, _id])
new_examples["source"].extend([standard_text, dialect_text])
new_examples["target"].extend([dialect_text, standard_text])
new_examples["src_lang"].extend(["standard", do])
new_examples["tgt_lang"].extend([do, "standard"])
return new_examples
if __name__ == "__main__":
if not os.path.isfile("data/train_dialect.json"):
train_files = glob("data/*/train/*.json")
train_samples = prepare_dialect_dataset(train_files)
json.dump({"data": train_samples}, open("data/train_dialect.json", "w"))
if not os.path.isfile("data/valid_dialect.json"):
valid_files = glob("data/*/valid/*.json")
valid_samples = prepare_dialect_dataset(valid_files)
json.dump({"data": valid_samples}, open("data/valid_dialect.json", "w"))
data_files = {"train": "data/train_dialect.json", "valid": "data/valid_dialect.json"}
dialect = load_dataset("json", data_files=data_files, field="data")
# Data preprocessing
dialect_dataset = dialect.map(function=preprocess, batched=True, batch_size=1000)
dialect_dataset_for_sc = dialect_dataset.map(
function=prepare_for_style_classification, batched=True, batch_size=1000,
remove_columns=dialect_dataset.column_names["train"],
)
dialect_dataset_for_st = dialect_dataset.map(
function=prepare_for_style_transfer, batched=True, batch_size=1000,
remove_columns=dialect_dataset.column_names["train"],
)
dialect_dataset_for_sc.save_to_disk("data/style_classification")
dialect_dataset_for_st.save_to_disk("data/style_transfer")