From f00880436934da4d7bab602053e5176711f68a2f Mon Sep 17 00:00:00 2001 From: linyqh Date: Wed, 20 Nov 2024 00:34:11 +0800 Subject: [PATCH] =?UTF-8?q?feat(subtitle):=20=E6=B7=BB=E5=8A=A0=E4=BB=8E?= =?UTF-8?q?=E8=A7=86=E9=A2=91=E6=8F=90=E5=8F=96=E9=9F=B3=E9=A2=91=E5=B9=B6?= =?UTF-8?q?=E7=94=9F=E6=88=90=E5=AD=97=E5=B9=95=E7=9A=84=E5=8A=9F=E8=83=BD?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - 新增 extract_audio_and_create_subtitle函数，用于从视频文件中提取音频并生成字幕文件 - 更新 video_pipeline.py，修改视频下载和处理的相关参数 --- app/services/subtitle.py | 95 ++++++++++++++++++++++++++++++++-------- video_pipeline.py | 14 +++--- 2 files changed, 84 insertions(+), 25 deletions(-) diff --git a/app/services/subtitle.py b/app/services/subtitle.py index f37eb65..7b18e8d 100644 --- a/app/services/subtitle.py +++ b/app/services/subtitle.py @@ -8,6 +8,8 @@ from faster_whisper import WhisperModel from timeit import default_timer as timer from loguru import logger import google.generativeai as genai +from moviepy.editor import VideoFileClip +import os from app.config import config from app.utils import utils @@ -362,29 +364,86 @@ def create_with_gemini(audio_file: str, subtitle_file: str = "", api_key: Option return None +def extract_audio_and_create_subtitle(video_file: str, subtitle_file: str = "") -> Optional[str]: + """ + 从视频文件中提取音频并生成字幕文件。 + + 参数: + - video_file: MP4视频文件的路径 + - subtitle_file: 输出字幕文件的路径（可选）。如果未提供，将根据视频文件名自动生成。 + + 返回: + - str: 生成的字幕文件路径 + - None: 如果处理过程中出现错误 + """ + try: + # 获取视频文件所在目录 + video_dir = os.path.dirname(video_file) + video_name = os.path.splitext(os.path.basename(video_file))[0] + + # 设置音频文件路径 + audio_file = os.path.join(video_dir, f"{video_name}_audio.wav") + + # 如果未指定字幕文件路径，则自动生成 + if not subtitle_file: + subtitle_file = os.path.join(video_dir, f"{video_name}.srt") + + logger.info(f"开始从视频提取音频: {video_file}") + + # 加载视频文件 + video = VideoFileClip(video_file) + + # 提取音频并保存为WAV格式 + logger.info(f"正在提取音频到: {audio_file}") + video.audio.write_audiofile(audio_file, codec='pcm_s16le') + + # 关闭视频文件 + video.close() + + logger.info("音频提取完成，开始生成字幕") + + # 使用create函数生成字幕 + create(audio_file, subtitle_file) + + # 删除临时音频文件 + if os.path.exists(audio_file): + os.remove(audio_file) + logger.info("已清理临时音频文件") + + return subtitle_file + + except Exception as e: + logger.error(f"处理视频文件时出错: {str(e)}") + logger.error(traceback.format_exc()) + return None + + if __name__ == "__main__": - task_id = "test456" + task_id = "12121" task_dir = utils.task_dir(task_id) subtitle_file = f"{task_dir}/subtitle.srt" audio_file = f"{task_dir}/audio.wav" + video_file = f"{task_dir}/duanju_demo.mp4" - subtitles = file_to_subtitles(subtitle_file) - print(subtitles) + extract_audio_and_create_subtitle(video_file, subtitle_file) - # script_file = f"{task_dir}/script.json" - # with open(script_file, "r") as f: - # script_content = f.read() - # s = json.loads(script_content) - # script = s.get("script") - # - # correct(subtitle_file, script) + # subtitles = file_to_subtitles(subtitle_file) + # print(subtitles) - subtitle_file = f"{task_dir}/subtitle111.srt" - create(audio_file, subtitle_file) + # # script_file = f"{task_dir}/script.json" + # # with open(script_file, "r") as f: + # # script_content = f.read() + # # s = json.loads(script_content) + # # script = s.get("script") + # # + # # correct(subtitle_file, script) - # # 使用Gemini模型处理音频 - # gemini_api_key = config.app.get("gemini_api_key") # 请替换为实际的API密钥 - # gemini_subtitle_file = create_with_gemini(audio_file, api_key=gemini_api_key) - # - # if gemini_subtitle_file: - # print(f"Gemini生成的字幕文件: {gemini_subtitle_file}") + # subtitle_file = f"{task_dir}/subtitle111.srt" + # create(audio_file, subtitle_file) + + # # # 使用Gemini模型处理音频 + # # gemini_api_key = config.app.get("gemini_api_key") # 请替换为实际的API密钥 + # # gemini_subtitle_file = create_with_gemini(audio_file, api_key=gemini_api_key) + # # + # # if gemini_subtitle_file: + # # print(f"Gemini生成的字幕文件: {gemini_subtitle_file}") diff --git a/video_pipeline.py b/video_pipeline.py index 3e35544..5dca576 100644 --- a/video_pipeline.py +++ b/video_pipeline.py @@ -110,12 +110,12 @@ class VideoPipeline: """运行完整的pipeline""" try: current_path = os.path.dirname(os.path.abspath(__file__)) - video_path = os.path.join(current_path, "resource", "videos", video_name) + video_path = os.path.join(current_path, "resource", "videos", f"{video_name}.mp4") # 判断视频是否存在 if not os.path.exists(video_path): # 1. 下载视频 print(f"视频不存在, 开始下载视频: {video_path}") - download_result = self.download_video(youtube_url=youtube_url, resolution="1080p", output_format="mp4", rename=video_name) + download_result = self.download_video(url=youtube_url, resolution="1080p", output_format="mp4", rename=video_name) video_path = download_result["output_path"] else: print(f"视频已存在: {video_path}") @@ -168,12 +168,12 @@ class VideoPipeline: if __name__ == "__main__": pipeline = VideoPipeline() result = pipeline.run_pipeline( - task_id="test_123", + task_id="test_111901", script_name="test.json", - youtube_url="https://www.youtube.com/watch?v=Kenm35gdqtk", - video_name="test.mp4", - skip_seconds=0, - threshold=30, + youtube_url="https://www.youtube.com/watch?v=vLJ7Yed6FQ4", + video_name="2024-11-19-01", + skip_seconds=50, + threshold=35, vision_batch_size=10, vision_llm_provider="gemini", voice_name="zh-CN-YunjianNeural",