-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathsubtitle_preprocessor.py
195 lines (169 loc) · 7.05 KB
/
subtitle_preprocessor.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
"""
Running basic pre-processing for the .srt subtitle files
"""
import re
import os
import pysrt
import glob
from joblib import Parallel, delayed
from tqdm import tqdm
from json_utils import load_annotation_list, save_jsonl
from configs.preprocess_configs import DOWNLOAD_ROOT, SUBTITLE_ROOT
def convert_sub_time_to_seconds(sub_time):
"""sub_time is a SubRipTime object defined by pysrt"""
return 60 * sub_time.minutes + sub_time.seconds + 0.001 * sub_time.milliseconds
def clean_single_sub_sentence(sub_sentence):
"""sub_sentence: str, """
sub_sentence = sub_sentence.replace("\n", " ")
sub_sentence = sub_sentence.replace("(", " ")
sub_sentence = sub_sentence.replace(")", " ")
sub_sentence = sub_sentence.replace(":", " : ")
sub_sentence = re.sub(r"\s{2,}", " ",
sub_sentence) # 这里是正则表达式，单空格替换双空格 https://www.runoob.com/python/python-reg-expressions.html
return sub_sentence
def split_multi_lines(cur_sub):
"""
return: subtitle list that split all multi-line items
"""
start_time = []
end_time = []
duration = cur_sub.duration
text_length = len(cur_sub.text)
text = cur_sub.text.split('\n')
start_time.append(cur_sub.start)
for i in range(len(text) - 1):
middle_time = start_time[i].__add__(duration.__mul__(len(text[i]) / text_length))
start_time.append(middle_time)
end_time.append(middle_time)
end_time.append(cur_sub.end)
splited_item = dict(
text_list=text,
start_time_list=start_time,
end_time_list=end_time
)
return splited_item
def preprocess_subtitles_single_video(video_path, save_path, segment_list):
"""
return: A python dict, the keys are the video names, the entries are lists,
each contains all the text from a .srt file
sub_times are the start time of the sentences.
"""
video_id = os.path.basename(video_path).split('.')[0]
subs = pysrt.open(video_path, encoding="iso-8859-1") # 打开字幕.srt文件到subs
if len(subs) == 0:
subs = pysrt.open(video_path)
# 转成单行
sub_single_line = []
for cur_sub in subs:
# 对读取到subs中的每一个srt item进行处理
text = cur_sub.text
if text == '':
continue
if '\n' in text:
splited_item = split_multi_lines(cur_sub)
text_list = splited_item['text_list']
start_time_list = splited_item['start_time_list']
end_time_list = splited_item['end_time_list']
for sentence_index in range(len(text_list)):
sub_single_line.append(dict(
text=text_list[sentence_index],
start=start_time_list[sentence_index],
end=end_time_list[sentence_index]
))
else:
sub_single_line.append(dict(
text=cur_sub.text,
start=cur_sub.start,
end=cur_sub.end
))
# 去重
sub_data = []
prev = sub_single_line[0]
for sentence_index in range(1, len(sub_single_line)):
cur_sub = sub_single_line[sentence_index]
if cur_sub['text'] != prev['text']:
sub_data.append(dict(
text=clean_single_sub_sentence(prev['text']),
start=convert_sub_time_to_seconds(prev['start']),
end=convert_sub_time_to_seconds(cur_sub['start'])
))
prev = cur_sub
if sentence_index == len(sub_single_line) - 1:
sub_data.append(dict(
text=clean_single_sub_sentence(prev['text']),
start=convert_sub_time_to_seconds(prev['start']),
end=convert_sub_time_to_seconds(cur_sub['end'])
))
# 划分入segment，并在segment分界点拆开
split_point_list = [segment_list[0][0]] + [segment_list[i][1] for i in range(len(segment_list))]
split_point_index = 0
sentence_index = 0
while sentence_index < len(sub_data) and split_point_index < len(split_point_list):
split_point = split_point_list[split_point_index]
if sub_data[sentence_index]['start'] < split_point < sub_data[sentence_index]['end']:
text = sub_data[sentence_index]['text']
start = sub_data[sentence_index]['start']
end = sub_data[sentence_index]['end']
# split the sentence into two
n_characters = len(text) * ((split_point - start) / (end - start))
text = text.split(' ')
text_before = ''
text_after = ''
for word in text:
if len(text_before) + len(word) <= n_characters:
text_before = text_before + word + ' '
else:
text_after = text_after + word + ' '
# put the split item back to list
sub_data[sentence_index]['end'] = split_point
sub_data[sentence_index]['text'] = text_before[:-1]
sub_data.insert(sentence_index + 1, dict(
text=text_after[:-1],
start=split_point,
end=end
))
split_point_index += 1
sentence_index += 1
elif sub_data[sentence_index]['start'] >= split_point:
split_point_index += 1
else:
sentence_index += 1
# attach to every segment
seg_sub_data = [[] for i in range(len(segment_list))]
for sub in sub_data:
if not sub['text'] == '':
for seg_index in range(len(segment_list)):
seg_start = segment_list[seg_index][0]
seg_end = segment_list[seg_index][1]
sentence_start = sub['start']
sentence_end = sub['end']
if seg_start <= sentence_start and sentence_end <= seg_end:
seg_sub_data[seg_index].append(sub)
# write to json
srt_data_list = []
for seg_index in range(len(segment_list)):
srt_data = dict(
seg_id=video_id + '_' + str(seg_index),
sub=seg_sub_data[seg_index]
)
srt_data_list.append(srt_data)
save_jsonl(srt_data_list, os.path.join(save_path, video_id + '.jsonl'))
def preprocess_subtitles(segment, srt_dir=DOWNLOAD_ROOT, save_path=SUBTITLE_ROOT):
if not os.path.exists(save_path):
os.makedirs(save_path)
print("Start preprocessing srt files from %s ..." % srt_dir)
srt_paths = glob.glob(os.path.join(srt_dir, "*.srt"))
for i in range(len(srt_paths)):
srt = srt_paths[i]
if os.path.exists(save_path + '/' + srt[-19:-7] + '.jsonl'):
srt_paths.remove(srt)
Parallel(n_jobs=32)(delayed(preprocess_subtitles_single_video)
(srt, save_path, segment[os.path.basename(srt).split('.')[0]])
for srt in tqdm(srt_paths, desc="Loop over subtitle files"))
if __name__ == '__main__':
# Get segment info
seg_info = dict()
annotation_list = load_annotation_list()
for anno in annotation_list:
seg_info[anno[0]['videoID']] = anno[1]['segInfo']
preprocess_subtitles(seg_info)