mirror of
https://github.com/linyqh/NarratoAI.git
synced 2026-01-06 22:38:11 +00:00
feat(subtitle): 实现字幕文件合并功能
- 重构 subtitle_merger.py，增加对 editedTimeRange 的解析和排序 - 更新 merge_subtitle_files 函数，支持处理字幕项列表 - 在 task.py 中集成字幕合并逻辑
This commit is contained in:
linyq
parent
fcf568963a
commit
6001e8038f
@ -38,15 +38,20 @@ def format_time(td):
|
||||
return f"{hours:02d}:{minutes:02d}:{seconds:02d},{milliseconds:03d}"
|
||||
|
||||
|
||||
def extract_time_range_from_filename(filename):
|
||||
"""从文件名中提取时间范围"""
|
||||
pattern = r'subtitle_(\d{2})_(\d{2})_(\d{2})-(\d{2})_(\d{2})_(\d{2})'
|
||||
match = re.search(pattern, filename)
|
||||
|
||||
if not match:
|
||||
def parse_edited_time_range(time_range_str):
|
||||
"""从editedTimeRange字符串中提取时间范围"""
|
||||
if not time_range_str:
|
||||
return None, None
|
||||
|
||||
start_h, start_m, start_s, end_h, end_m, end_s = map(int, match.groups())
|
||||
parts = time_range_str.split('-')
|
||||
if len(parts) != 2:
|
||||
return None, None
|
||||
|
||||
start_time_str, end_time_str = parts
|
||||
|
||||
# 将HH:MM:SS格式转换为timedelta
|
||||
start_h, start_m, start_s = map(int, start_time_str.split(':'))
|
||||
end_h, end_m, end_s = map(int, end_time_str.split(':'))
|
||||
|
||||
start_time = timedelta(hours=start_h, minutes=start_m, seconds=start_s)
|
||||
end_time = timedelta(hours=end_h, minutes=end_m, seconds=end_s)
|
||||
@ -54,33 +59,36 @@ def extract_time_range_from_filename(filename):
|
||||
return start_time, end_time
|
||||
|
||||
|
||||
def merge_subtitle_files(subtitle_files, output_file=None):
|
||||
def merge_subtitle_files(subtitle_items, output_file=None):
|
||||
"""
|
||||
合并多个SRT字幕文件
|
||||
|
||||
参数:
|
||||
subtitle_files: 包含SRT文件路径的列表
|
||||
subtitle_items: 字典列表，每个字典包含subtitle文件路径和editedTimeRange
|
||||
output_file: 输出文件的路径，如果为None则自动生成
|
||||
|
||||
返回:
|
||||
合并后的字幕文件路径
|
||||
"""
|
||||
# 按文件名中的开始时间排序
|
||||
sorted_files = sorted(subtitle_files,
|
||||
key=lambda x: extract_time_range_from_filename(x)[0])
|
||||
# 按照editedTimeRange的开始时间排序
|
||||
sorted_items = sorted(subtitle_items,
|
||||
key=lambda x: parse_edited_time_range(x.get('editedTimeRange', ''))[0] or timedelta())
|
||||
|
||||
merged_subtitles = []
|
||||
subtitle_index = 1
|
||||
|
||||
for file_path in sorted_files:
|
||||
# 从文件名获取起始时间偏移
|
||||
offset_time, _ = extract_time_range_from_filename(file_path)
|
||||
for item in sorted_items:
|
||||
if not item.get('subtitle') or not os.path.exists(item.get('subtitle')):
|
||||
continue
|
||||
|
||||
# 从editedTimeRange获取起始时间偏移
|
||||
offset_time, _ = parse_edited_time_range(item.get('editedTimeRange', ''))
|
||||
|
||||
if offset_time is None:
|
||||
print(f"警告: 无法从文件名 {os.path.basename(file_path)} 中提取时间范围，跳过该文件")
|
||||
print(f"警告: 无法从项目 {item.get('_id')} 的editedTimeRange中提取时间范围，跳过该项")
|
||||
continue
|
||||
|
||||
with open(file_path, 'r', encoding='utf-8') as file:
|
||||
with open(item['subtitle'], 'r', encoding='utf-8') as file:
|
||||
content = file.read()
|
||||
|
||||
# 解析字幕文件
|
||||
@ -117,24 +125,23 @@ def merge_subtitle_files(subtitle_files, output_file=None):
|
||||
merged_subtitles.append('\n'.join(new_block))
|
||||
subtitle_index += 1
|
||||
|
||||
# 合并所有字幕块
|
||||
merged_content = '\n\n'.join(merged_subtitles)
|
||||
|
||||
# 确定输出文件路径
|
||||
if output_file is None:
|
||||
# 自动生成输出文件名
|
||||
first_file_path = sorted_files[0]
|
||||
last_file_path = sorted_files[-1]
|
||||
_, first_end = extract_time_range_from_filename(first_file_path)
|
||||
_, last_end = extract_time_range_from_filename(last_file_path)
|
||||
dir_path = os.path.dirname(sorted_items[0]['subtitle'])
|
||||
first_start = parse_edited_time_range(sorted_items[0]['editedTimeRange'])[0]
|
||||
last_end = parse_edited_time_range(sorted_items[-1]['editedTimeRange'])[1]
|
||||
|
||||
dir_path = os.path.dirname(first_file_path)
|
||||
first_start_str = os.path.basename(first_file_path).split('-')[0].replace('subtitle_', '')
|
||||
first_start_h, first_start_m, first_start_s = int(first_start.seconds // 3600), int((first_start.seconds % 3600) // 60), int(first_start.seconds % 60)
|
||||
last_end_h, last_end_m, last_end_s = int(last_end.seconds // 3600), int((last_end.seconds % 3600) // 60), int(last_end.seconds % 60)
|
||||
|
||||
first_start_str = f"{first_start_h:02d}_{first_start_m:02d}_{first_start_s:02d}"
|
||||
last_end_str = f"{last_end_h:02d}_{last_end_m:02d}_{last_end_s:02d}"
|
||||
|
||||
output_file = os.path.join(dir_path, f"merged_subtitle_{first_start_str}-{last_end_str}.srt")
|
||||
|
||||
# 合并所有字幕块
|
||||
merged_content = '\n\n'.join(merged_subtitles)
|
||||
|
||||
# 写入合并后的内容
|
||||
with open(output_file, 'w', encoding='utf-8') as file:
|
||||
file.write(merged_content)
|
||||
@ -143,12 +150,53 @@ def merge_subtitle_files(subtitle_files, output_file=None):
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
subtitle_files = [
|
||||
"/Users/apple/Desktop/home/NarratoAI/storage/tasks/qyn2-2-demo/subtitle_00_00_00-00_01_15.srt",
|
||||
"/Users/apple/Desktop/home/NarratoAI/storage/tasks/qyn2-2-demo/subtitle_00_01_15-00_04_40.srt",
|
||||
"/Users/apple/Desktop/home/NarratoAI/storage/tasks/qyn2-2-demo/subtitle_00_04_58-00_05_45.srt",
|
||||
"/Users/apple/Desktop/home/NarratoAI/storage/tasks/qyn2-2-demo/subtitle_00_05_45-00_06_00.srt",
|
||||
# 测试数据
|
||||
test_data = [
|
||||
{'picture': '【解说】好的，各位，欢迎回到我的频道！《庆余年 2》刚开播就给了我们一个王炸！范闲在北齐"死"了？这怎么可能！',
|
||||
'timestamp': '00:00:00-00:01:15',
|
||||
'narration': '好的各位，欢迎回到我的频道！《庆余年 2》刚开播就给了我们一个王炸！范闲在北齐"死"了？这怎么可能！上集片尾那个巨大的悬念，这一集就立刻揭晓了！范闲假死归来，他面临的第一个，也是最大的难关，就是如何面对他最敬爱的，同时也是最可怕的那个人——庆帝！',
|
||||
'OST': 0,
|
||||
'_id': 1,
|
||||
'audio': '/Users/apple/Desktop/home/NarratoAI/storage/tasks/qyn2-2-demo/audio_00_00_00-00_01_15.mp3',
|
||||
'subtitle': '/Users/apple/Desktop/home/NarratoAI/storage/tasks/qyn2-2-demo/subtitle_00_00_00-00_01_15.srt',
|
||||
'sourceTimeRange': '00:00:00-00:00:26',
|
||||
'duration': 26,
|
||||
'editedTimeRange': '00:00:00-00:00:26'
|
||||
},
|
||||
{'picture': '【解说】上一集我们看到，范闲在北齐遭遇了惊天变故，生死不明！',
|
||||
'timestamp': '00:01:15-00:04:40',
|
||||
'narration': '但我们都知道，他绝不可能就这么轻易退场！第二集一开场，范闲就已经秘密回到了京都。他的生死传闻，可不像我们想象中那样只是小范围流传，而是…',
|
||||
'OST': 0,
|
||||
'_id': 2,
|
||||
'audio': '/Users/apple/Desktop/home/NarratoAI/storage/tasks/qyn2-2-demo/audio_00_01_15-00_04_40.mp3',
|
||||
'subtitle': '/Users/apple/Desktop/home/NarratoAI/storage/tasks/qyn2-2-demo/subtitle_00_01_15-00_04_40.srt',
|
||||
'sourceTimeRange': '00:01:15-00:01:29',
|
||||
'duration': 14,
|
||||
'editedTimeRange': '00:00:26-00:00:40'
|
||||
},
|
||||
{'picture': '【解说】"欺君之罪"！在封建王朝，这可是抄家灭族的大罪！搁一般人，肯定脚底抹油溜之大吉了。',
|
||||
'timestamp': '00:04:58-00:05:45',
|
||||
'narration': '"欺君之罪"！在封建王朝，这可是抄家灭族的大罪！搁一般人，肯定脚底抹油溜之大吉了。但范闲是谁啊？他偏要反其道而行之！他竟然决定，直接去见庆帝！冒着天大的风险，用"假死"这个事实去赌庆帝的态度！',
|
||||
'OST': 0,
|
||||
'_id': 4,
|
||||
'audio': '/Users/apple/Desktop/home/NarratoAI/storage/tasks/qyn2-2-demo/audio_00_04_58-00_05_45.mp3',
|
||||
'subtitle': '/Users/apple/Desktop/home/NarratoAI/storage/tasks/qyn2-2-demo/subtitle_00_04_58-00_05_45.srt',
|
||||
'sourceTimeRange': '00:04:58-00:05:20',
|
||||
'duration': 22,
|
||||
'editedTimeRange': '00:00:57-00:01:19'
|
||||
},
|
||||
{'picture': '【解说】但想见庆帝，哪有那么容易？范闲艺高人胆大，竟然选择了最激进的方式——闯宫！',
|
||||
'timestamp': '00:05:45-00:06:00',
|
||||
'narration': '但想见庆帝，哪有那么容易？范闲艺高人胆大，竟然选择了最激进的方式——闯宫！',
|
||||
'OST': 0,
|
||||
'_id': 5,
|
||||
'audio': '/Users/apple/Desktop/home/NarratoAI/storage/tasks/qyn2-2-demo/audio_00_05_45-00_06_00.mp3',
|
||||
'subtitle': '/Users/apple/Desktop/home/NarratoAI/storage/tasks/qyn2-2-demo/subtitle_00_05_45-00_06_00.srt',
|
||||
'sourceTimeRange': '00:05:45-00:05:53',
|
||||
'duration': 8,
|
||||
'editedTimeRange': '00:01:19-00:01:27'
|
||||
}
|
||||
]
|
||||
|
||||
output_file = merge_subtitle_files(subtitle_files)
|
||||
output_file = merge_subtitle_files(test_data)
|
||||
print(f"字幕文件已合并至: {output_file}")
|
||||
|
||||
@ -250,7 +250,7 @@ def start_subclip(task_id: str, params: VideoClipParams, subclip_path_videos: di
|
||||
subclip_clip_result = {
|
||||
tts_result['_id']: tts_result['subtitle_file'] for tts_result in tts_results
|
||||
}
|
||||
list_script = update_script.update_script_timestamps(list_script, video_clip_result, tts_clip_result, subclip_clip_result)
|
||||
new_script_list = update_script.update_script_timestamps(list_script, video_clip_result, tts_clip_result, subclip_clip_result)
|
||||
|
||||
sm.state.update_task(task_id, state=const.TASK_STATE_PROCESSING, progress=60)
|
||||
|
||||
@ -258,24 +258,19 @@ def start_subclip(task_id: str, params: VideoClipParams, subclip_path_videos: di
|
||||
4. 合并音频和字幕
|
||||
"""
|
||||
logger.info("\n\n## 4. 合并音频和字幕")
|
||||
subtitle_files = [
|
||||
tts_result["subtitle_file"] for tts_result in tts_results
|
||||
]
|
||||
total_duration = sum([script["duration"] for script in list_script])
|
||||
total_duration = sum([script["duration"] for script in new_script_list])
|
||||
if tts_results:
|
||||
try:
|
||||
# 合并音频文件
|
||||
merged_audio_path = audio_merger.merge_audio_files(
|
||||
task_id=task_id,
|
||||
total_duration=total_duration,
|
||||
list_script=list_script
|
||||
list_script=new_script_list
|
||||
)
|
||||
logger.info(f"音频文件合并成功->{merged_audio_path}")
|
||||
# # 合并字幕文件
|
||||
# merged_subtitle_path = subtitle_merger.merge_subtitle_files(
|
||||
# subtitle_files=subtitle_files,
|
||||
# )
|
||||
# logger.info(f"字幕文件合并成功->{merged_subtitle_path}")
|
||||
# 合并字幕文件
|
||||
merged_subtitle_path = subtitle_merger.merge_subtitle_files(new_script_list)
|
||||
logger.info(f"字幕文件合并成功->{merged_subtitle_path}")
|
||||
except Exception as e:
|
||||
logger.error(f"合并音频文件失败: {str(e)}")
|
||||
merged_audio_path = ""
|
||||
@ -292,10 +287,10 @@ def start_subclip(task_id: str, params: VideoClipParams, subclip_path_videos: di
|
||||
|
||||
combined_video_path = path.join(utils.task_dir(task_id), f"merger.mp4")
|
||||
logger.info(f"\n\n## 5. 合并视频: => {combined_video_path}")
|
||||
|
||||
videos_clips = [new_script['video_path'] for new_script in new_script_list]
|
||||
merger_video.combine_clip_videos(
|
||||
output_video_path=combined_video_path,
|
||||
video_paths=subclip_videos,
|
||||
video_paths=videos_clips,
|
||||
video_ost_list=video_ost,
|
||||
video_aspect=params.video_aspect,
|
||||
threads=params.n_threads # 多线程
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user