-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathpipline.py
205 lines (179 loc) · 7.18 KB
/
pipline.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
from base.media_transcribe import transcribe_batch
from base.sub_translate import sub_translate
from base.sub_segment import sub_segment
from base.sub_optimize import sub_optimize
from base.path_request import Directory
from base.files_find import Files
from base.srt_generate import generate_bilingual_srt
from base.srt2ass import convert_srt_to_ass
import whisperX.whisperx as wsx
from config import *
import os
import json
import threading
import logging
import colorlog # 引入 colorlog 模块
# 配置日志
def setup_logger():
# 创建一个日志记录器
logger = logging.getLogger("pipline")
logger.setLevel(logging.DEBUG)
# 创建日志目录（如果不存在）
log_dir = "log"
if not os.path.exists(log_dir):
os.makedirs(log_dir)
# 创建一个控制台处理器，并设置颜色格式
console_handler = logging.StreamHandler()
console_formatter = colorlog.ColoredFormatter(
"%(log_color)s%(asctime)s - %(levelname)s - %(message)s",
datefmt="%Y-%m-%d %H:%M:%S",
log_colors={
"DEBUG": "cyan",
"INFO": "green",
"WARNING": "yellow",
"ERROR": "red",
"CRITICAL": "red,bg_white",
},
)
console_handler.setFormatter(console_formatter)
# 创建一个文件处理器，并设置格式
file_handler = logging.FileHandler(os.path.join(log_dir, "pipline.log"))
file_formatter = logging.Formatter("%(asctime)s - %(levelname)s - %(message)s")
file_handler.setFormatter(file_formatter)
# 将处理器添加到日志记录器
logger.addHandler(console_handler)
logger.addHandler(file_handler)
return logger
logger = setup_logger()
# 使用锁来确保只有一个 process_transcript 在运行
process_lock = threading.Lock()
def process_transcript(aligned_transcript, input_path, output_path):
with process_lock:
logging.info(f"文件 {input_path} 字幕翻译开始")
sub_translate(
aligned_transcript,
api_key=api_key,
base_url=base_url,
model=llm_model,
src_lang=src_lang,
dst_lang=dst_lang,
media_title=os.path.basename(input_path),
context_window=10,
batch_size=10,
thread_count=10,
)
logging.info(f"文件 {input_path} 字幕翻译结束")
with open(
os.path.join("output/translated_transcripts", output_path) + ".json",
"w",
encoding="utf-8",
) as f:
json.dump(aligned_transcript, f, ensure_ascii=False)
logging.info(f"文件 {input_path} 字幕分割开始")
sub_segment(
aligned_transcript,
api_key=api_key,
base_url=base_url,
model=llm_model,
word_limit=12,
thread_count=20,
)
logging.info(f"文件 {input_path} 字幕分割结束")
sub_optimize(aligned_transcript, src_lang=src_lang, dst_lang=dst_lang)
logging.info(f"文件 {input_path} 字幕翻译优化")
with open(
os.path.join("output/segmented_transcripts", output_path) + ".json",
"w",
encoding="utf-8",
) as f:
json.dump(aligned_transcript, f, ensure_ascii=False)
srt_path = os.path.join("output/srt", output_path) + ".srt"
generate_bilingual_srt(aligned_transcript, srt_path)
logging.info(f"文件 {input_path} 生成 srt 字幕")
convert_srt_to_ass(
srt_path,
os.path.join("output/ass", output_path) + ".ass",
original_style=original_style,
translated_style=translated_style,
)
logging.info(f"文件 {input_path} 生成 ass 字幕")
def run():
src_dir = Directory("选择视频文件所在文件夹")
input_paths, output_paths = Files(
src_dir,
"",
extensions=[".webm", ".mkv", ".flv", ".mp4", ".mp3", ".flac", ".ogg", ".wav"],
mkdir=False,
)
file_count = len(input_paths)
# 创建输出需要的文件夹
for output_path in output_paths:
output_dir = os.path.dirname(os.path.join("output/transcripts", output_path))
if not os.path.exists(output_dir):
os.makedirs(output_dir)
output_dir = os.path.dirname(
os.path.join("output/aligned_transcripts", output_path)
)
if not os.path.exists(output_dir):
os.makedirs(output_dir)
output_dir = os.path.dirname(
os.path.join("output/translated_transcripts", output_path)
)
if not os.path.exists(output_dir):
os.makedirs(output_dir)
output_dir = os.path.dirname(
os.path.join("output/segmented_transcripts", output_path)
)
if not os.path.exists(output_dir):
os.makedirs(output_dir)
output_dir = os.path.dirname(os.path.join("output/srt", output_path))
if not os.path.exists(output_dir):
os.makedirs(output_dir)
output_dir = os.path.dirname(os.path.join("output/ass", output_path))
if not os.path.exists(output_dir):
os.makedirs(output_dir)
# 去除路径的扩展名
for i in range(file_count):
output_paths[i], _ = os.path.splitext(output_paths[i])
whisper_model = wsx.load_model(
whisper_model_type,
device,
compute_type=compute_type,
download_root=whisper_model_dir,
)
align_model, metadata = wsx.load_align_model(
language_code=transcibe_lang, device=device, model_dir=align_model_dir
)
for i in range(file_count):
transcript_path = os.path.join("output/transcripts", output_paths[i]) + ".json"
aligned_transcript_path = (
os.path.join("output/aligned_transcripts", output_paths[i]) + ".json"
)
# 检查转录文件是否已经存在
if os.path.exists(transcript_path) and os.path.exists(aligned_transcript_path):
logging.info(f"文件 {input_paths[i]} 的转录结果已存在，直接读取")
with open(transcript_path, "r", encoding="utf-8") as f:
transcript = json.load(f)
with open(aligned_transcript_path, "r", encoding="utf-8") as f:
aligned_transcript = json.load(f)
else:
logging.info(f"文件 {input_paths[i]} Whisper 转录开始")
transcript, aligned_transcript = transcribe_batch(
input_paths[i],
device=device,
whisper_model=whisper_model,
align_model=align_model,
metadata=metadata,
batch_size=10,
language=transcibe_lang,
)
logging.info(f"文件 {input_paths[i]} Whisper 转录结束")
with open(transcript_path, "w", encoding="utf-8") as f:
json.dump(transcript, f, ensure_ascii=False)
with open(aligned_transcript_path, "w", encoding="utf-8") as f:
json.dump(aligned_transcript, f, ensure_ascii=False)
# 启动一个新线程来处理转录后的任务
threading.Thread(
target=process_transcript,
args=(aligned_transcript, input_paths[i], output_paths[i]),
).start()