-
Notifications
You must be signed in to change notification settings - Fork 0
/
DataLoader.py
235 lines (194 loc) · 7.38 KB
/
DataLoader.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
222
223
224
225
226
227
228
229
230
231
232
233
234
235
import json
import re
import subprocess
import os
import nltk
nltk.download("omw-1.4")
nltk.download("punkt")
nltk.download("stopwords")
nltk.download("wordnet")
from nltk.stem import SnowballStemmer, WordNetLemmatizer
from nltk.tokenize import word_tokenize
import numpy as np
import pandas as pd
from sklearn.preprocessing import LabelEncoder
class DataLoader:
def __init__(self, file="corpus-webis-tldr-17.json", verbose=True):
self.file = file
self.verbose = verbose
self.get_line_count()
def get_line_count(self):
p = subprocess.Popen(["wc", "-l", self.file],
stdout=subprocess.PIPE,
stderr=subprocess.PIPE)
result, err = p.communicate()
if p.returncode != 0:
raise IOError(err)
self.line_count = int(result.strip().split()[0])
def load_subreddits(self, subreddits=None, save=False, limit=float("inf")):
X = []
y = []
count = 0
with open("corpus-webis-tldr-17.json", "r") as data:
for i, line in enumerate(data):
if self.verbose and i % 10000 == 0:
print(f"Loading {i}/{self.line_count}...", end="\r")
sample = json.loads(line)
if subreddits is not None and sample["subreddit"] not in subreddits:
continue
content = sample["content"].replace("\n", " ").replace("\r", "")
X.append(content)
y.append(sample["subreddit"])
count += 1
if count >= limit:
break
if self.verbose:
print(f'{" " * 50}', end="\r")
print(f"Loaded {self.line_count}/{self.line_count}")
le = LabelEncoder()
y = le.fit_transform(y)
if subreddits is None:
subreddits = ["all"]
if save:
with open(f'corpus-{"-".join(subreddits)}.json', "w") as f:
json.dump({"X": X, "y": y, "target_names": le.classes_}, f)
return X, y
def preprocess_bow(self, X):
stopwords = nltk.corpus.stopwords.words("english")
stemmer = SnowballStemmer("english")
lemmatizer = WordNetLemmatizer()
for i, _ in enumerate(X):
if self.verbose and i % 1000 == 0:
print(f"Preprocessing {i}/{len(X)}...", end="\r")
text = re.sub(r"[^\w\s]", "", X[i].lower().strip())
tokens = text.split()
# tokens = word_tokenize(X[i])
tokens = [t for t in tokens if t not in stopwords]
tokens = [stemmer.stem(token) for token in tokens]
tokens = [lemmatizer.lemmatize(token) for token in tokens]
X[i] = " ".join(tokens)
if self.verbose:
print(f'{" " * 50}', end="\r")
print(f"Preprocessed {len(X)}/{len(X)}")
return X
def export_for_eda(self, X, y, max_samples=float('inf')):
if max_samples:
path = f"eda_nlp/data/reddit_{max_samples}.txt"
else:
path = "eda_nlp/data/reddit.txt"
with open(path, "w") as f:
for i, (text, label) in enumerate(zip(X, y)):
if i >= max_samples:
break
f.write(f"{label}\t{text}\n")
def export_for_eda_dir(self, X, y, max_samples):
directory = f"eda_nlp/data/reddit_{max_samples}"
df = pd.DataFrame(data={"X": X, "Y": y})
if (max_samples * 10 > len(df.index)):
max_samples = int(len(df.index) / 10)
print(max_samples)
df = df.sample(n=max_samples * 10, ignore_index=True)
for i in range(10):
X = df.loc[(i * max_samples):(i + 1) * max_samples - 1, "X"].values
y = df.loc[(i * max_samples):(i + 1) * max_samples - 1, "Y"].values
with open(directory + "/{:d}.txt".format(i), "w") as f:
for (text, label) in zip(X, y):
f.write(f"{label}\t{text}\n")
def import_unaltered_reddit(self, size=None):
if size:
path = f"eda_nlp/data/reddit_{size}.txt"
else:
path = "eda_nlp/data/reddit.txt"
X = []
y = []
with open(path, "r") as f:
for line in f:
tab = line.index("\t")
label = line[:tab]
text = line[(tab + 1):]
X.append(text)
y.append(int(label))
return X, y
def import_from_mt(self, size):
path = f"eda_nlp/data/mt_reddit_{size}.txt"
X = []
y = []
with open(path, "r") as f:
for line in f:
tab = line.index("\t")
label = line[:tab]
text = line[(tab + 1):]
X.append(text)
y.append(int(label))
return X, y
def import_unaltered_reddit_dir(self, size):
directory = f"eda_nlp/data/reddit_{size}"
p = re.compile('[0-9].txt')
for filename in os.listdir(directory):
if (p.match(filename)):
path = os.path.join(directory, filename)
with open(path, "r") as f:
X = []
y = []
for line in f:
tab = line.index("\t")
label = line[:tab]
text = line[(tab + 1):]
X.append(text)
y.append(int(label))
yield X, y
def import_gpt_label_reddit(self):
path = "eda_nlp/data/reddit.txt"
X = []
y = []
with open(path, "r") as f:
for line in f:
tab = line.index("\t")
label = line[:tab]
text = line[(tab + 1):]
if int(label) == 1:
text = "from lol:" + text
else:
text = "from animal advice: " + text
X.append(text)
y.append(int(label))
return X, y
def import_from_eda(self, size=None):
if size:
path = f"eda_nlp/data/eda_reddit_{size}.txt"
else:
path = "eda_nlp/data/eda_reddit.txt"
X = []
y = []
with open(path, "r") as f:
for line in f:
tab = line.index("\t")
label = line[:tab]
text = line[(tab + 1):]
X.append(text)
y.append(int(label))
return X, y
def import_from_eda_dir(self, size):
directory = f"eda_nlp/data/reddit_{size}"
p = re.compile('eda_[0-9].txt')
for filename in os.listdir(directory):
if (p.match(filename)):
path = os.path.join(directory, filename)
with open(path, "r") as f:
X = []
y = []
for line in f:
tab = line.index("\t")
label = line[:tab]
text = line[(tab + 1):]
X.append(text)
y.append(int(label))
yield X, y
if __name__ == '__main__':
dl = DataLoader()
X, y = dl.load_subreddits(subreddits=['leagueoflegends', 'AdviceAnimals'])
dl.export_for_eda_dir(X, y, 50)
dl.export_for_eda_dir(X, y, 100)
dl.export_for_eda_dir(X, y, 500)
dl.export_for_eda_dir(X, y, 1000)
# dl.export_for_eda_dir(X, y, 5000)