diff --git a/app/services/material.py b/app/services/material.py index ab0dab0..bab1aba 100644 --- a/app/services/material.py +++ b/app/services/material.py @@ -251,16 +251,41 @@ def download_videos( return video_paths +def time_to_seconds(time_str: str) -> float: + """ + 将时间字符串转换为秒数 + 支持格式： + 1. "MM:SS" (分:秒) + 2. "SS" (纯秒数) + """ + parts = time_str.split(':') + if len(parts) == 2: + minutes, seconds = map(float, parts) + return minutes * 60 + seconds + return float(time_str) + + +def format_timestamp(seconds: float) -> str: + """ + 将秒数转换为 "MM:SS" 格式的时间字符串 + """ + minutes = int(seconds) // 60 + secs = int(seconds) % 60 + return f"{minutes:02d}:{secs:02d}" + + def save_clip_video(timestamp: str, origin_video: str, save_dir: str = "") -> dict: """ 保存剪辑后的视频 Args: - timestamp: 需要裁剪的单个时间戳，如：'00:36-00:40' + timestamp: 需要裁剪的单个时间戳，支持两种格式： + 1. '00:36-00:40' (分:秒-分:秒) + 2. 'SS-SS' (秒-秒) origin_video: 原视频路径 save_dir: 存储目录 Returns: - 裁剪后的视频路径 + 裁剪后的视频路径，格式为 {timestamp: video_path} """ if not save_dir: save_dir = utils.storage_dir("cache_videos") @@ -276,35 +301,64 @@ def save_clip_video(timestamp: str, origin_video: str, save_dir: str = "") -> di return {timestamp: video_path} try: + # 先加载视频获取总时长 + video = VideoFileClip(origin_video) + total_duration = video.duration + + # 获取目标时间段 + start_str, end_str = timestamp.split('-') + start = time_to_seconds(start_str) + end = time_to_seconds(end_str) + + # 验证时间段是否有效 + if start >= total_duration: + logger.warning(f"起始时间 {format_timestamp(start)} ({start:.2f}秒) 超出视频总时长 {format_timestamp(total_duration)} ({total_duration:.2f}秒)") + video.close() + return {} + + if end > total_duration: + logger.warning(f"结束时间 {format_timestamp(end)} ({end:.2f}秒) 超出视频总时长 {format_timestamp(total_duration)} ({total_duration:.2f}秒)，将自动调整为视频结尾") + end = total_duration + + if end <= start: + logger.warning(f"结束时间 {format_timestamp(end)} 必须大于起始时间 {format_timestamp(start)}") + video.close() + return {} + # 剪辑视频 - start, end = utils.split_timestamp(timestamp) - video = VideoFileClip(origin_video).subclip(start, end) + duration = end - start + logger.info(f"开始剪辑视频: {format_timestamp(start)} - {format_timestamp(end)}，时长 {format_timestamp(duration)}") + subclip = video.subclip(start, end) - # 检查视频是否有音频轨道 - if video.audio is not None: - video.write_videofile(video_path, logger=None) # 有音频时正常处理 - else: - # 没有音频时使用不同的写入方式 - video.write_videofile(video_path, audio=False, logger=None) - - video.close() # 确保关闭视频文件 - - if os.path.getsize(video_path) > 0 and os.path.exists(video_path): - try: - clip = VideoFileClip(video_path) - duration = clip.duration - fps = clip.fps - clip.close() - if duration > 0 and fps > 0: - return {timestamp: video_path} - except Exception as e: - logger.warning(f"视频文件验证失败: {video_path} => {str(e)}") - if os.path.exists(video_path): - os.remove(video_path) + try: + # 检查视频是否有音频轨道并写入文件 + subclip.write_videofile(video_path, audio=(subclip.audio is not None), logger=None) + + # 验证生成的视频文件 + if os.path.exists(video_path) and os.path.getsize(video_path) > 0: + with VideoFileClip(video_path) as clip: + if clip.duration > 0 and clip.fps > 0: + return {timestamp: video_path} + + raise ValueError("视频文件验证失败") + + except Exception as e: + logger.warning(f"视频文件处理失败: {video_path} => {str(e)}") + if os.path.exists(video_path): + os.remove(video_path) + except Exception as e: - logger.warning(f"视频剪辑失败: {str(e)}") + logger.warning(f"视频剪辑失败: \n{str(traceback.format_exc())}") if os.path.exists(video_path): os.remove(video_path) + finally: + # 确保视频对象被正确关闭 + try: + video.close() + if 'subclip' in locals(): + subclip.close() + except: + pass return {} @@ -426,6 +480,4 @@ def merge_videos(video_paths, ost_list): if __name__ == "__main__": - download_videos( - "test123", ["Money Exchange Medium"], audio_duration=100, source="pixabay" - ) + save_clip_video('00:50-01:41', 'E:\\projects\\NarratoAI\\resource\\videos\\WeChat_20241110144511.mp4') diff --git a/app/utils/script_generator.py b/app/utils/script_generator.py index 72e7408..2bec834 100644 --- a/app/utils/script_generator.py +++ b/app/utils/script_generator.py @@ -326,7 +326,7 @@ class ScriptProcessor: start_seconds = time_to_seconds(start_str) end_seconds = time_to_seconds(end_str) duration = end_seconds - start_seconds - word_count = int(duration / 0.2) + word_count = int(duration / 0.35) return word_count except Exception as e: diff --git a/app/utils/utils.py b/app/utils/utils.py index 4880e6c..bcebe4c 100644 --- a/app/utils/utils.py +++ b/app/utils/utils.py @@ -395,6 +395,8 @@ def cut_video(params, progress_callback=None): time_list = [i['timestamp'] for i in video_script_list] total_clips = len(time_list) + + print("time_list", time_list) def clip_progress(current, total): progress = int((current / total) * 100) @@ -413,12 +415,16 @@ def cut_video(params, progress_callback=None): st.session_state['subclip_videos'] = subclip_videos + print("list:", subclip_videos) + for i, video_script in enumerate(video_script_list): + print(i) + print(video_script) try: video_script['path'] = subclip_videos[video_script['timestamp']] except KeyError as err: logger.error(f"裁剪视频失败: {err}") - raise ValueError(f"裁剪视频失败: {err}") + # raise ValueError(f"裁剪视频失败: {err}") return task_id, subclip_videos diff --git a/app/utils/video_processor.py b/app/utils/video_processor.py index 6822b46..eb0da75 100644 --- a/app/utils/video_processor.py +++ b/app/utils/video_processor.py @@ -65,16 +65,44 @@ class VideoProcessor: shot_boundaries.append(i) return shot_boundaries + def filter_keyframes_by_time(self, keyframes: List[np.ndarray], + keyframe_indices: List[int]) -> Tuple[List[np.ndarray], List[int]]: + """ + 过滤关键帧，确保每秒最多只有一个关键帧 + + Args: + keyframes: 关键帧列表 + keyframe_indices: 关键帧索引列表 + + Returns: + Tuple[List[np.ndarray], List[int]]: 过滤后的关键帧列表和对应的帧索引 + """ + if not keyframes or not keyframe_indices: + return keyframes, keyframe_indices + + filtered_frames = [] + filtered_indices = [] + last_second = -1 + + for frame, idx in zip(keyframes, keyframe_indices): + current_second = idx // self.fps + if current_second != last_second: + filtered_frames.append(frame) + filtered_indices.append(idx) + last_second = current_second + + return filtered_frames, filtered_indices + def extract_keyframes(self, frames: List[np.ndarray], shot_boundaries: List[int]) -> Tuple[List[np.ndarray], List[int]]: """ - 从每个镜头中提取关键帧 + 从每个镜头中提取关键帧，并确保每秒最多一个关键帧 Args: frames: 视频帧列表 shot_boundaries: 镜头边界列表 Returns: - Tuple[List[np.ndarray], List[int]]: 关���帧列表和对应的帧索引 + Tuple[List[np.ndarray], List[int]]: 关键帧列表和对应的帧索引 """ keyframes = [] keyframe_indices = [] @@ -92,7 +120,10 @@ class VideoProcessor: keyframes.append(shot_frames[center_idx]) keyframe_indices.append(start + center_idx) - return keyframes, keyframe_indices + # 过滤每秒多余的关键帧 + filtered_keyframes, filtered_indices = self.filter_keyframes_by_time(keyframes, keyframe_indices) + + return filtered_keyframes, filtered_indices def save_keyframes(self, keyframes: List[np.ndarray], keyframe_indices: List[int], output_dir: str) -> None: @@ -206,4 +237,4 @@ class VideoProcessor: # 调整关键帧索引，加上跳过的帧数 adjusted_indices = [idx + skip_frames for idx in keyframe_indices] - self.save_keyframes(keyframes, adjusted_indices, output_dir) \ No newline at end of file + self.save_keyframes(keyframes, adjusted_indices, output_dir) diff --git a/app/utils/vision_analyzer.py b/app/utils/vision_analyzer.py index 060b29a..eb143ec 100644 --- a/app/utils/vision_analyzer.py +++ b/app/utils/vision_analyzer.py @@ -100,7 +100,7 @@ class VisionAnalyzer: except Exception as e: retry_count += 1 - error_msg = f"批次 {i // batch_size} 处理出错: {str(e)}\n{traceback.format_exc()}" + error_msg = f"批次 {i // batch_size} 处理出错: {str(e)}" logger.error(error_msg) if retry_count >= 3: @@ -111,7 +111,7 @@ class VisionAnalyzer: 'model_used': self.model_name }) else: - logger.info(f"批次 {i // batch_size} 处理失败，等待60秒后重试...") + logger.info(f"批次 {i // batch_size} 处理失败，等待60秒后重试当前批次...") await asyncio.sleep(60) pbar.update(1) diff --git a/config.example.toml b/config.example.toml index 3b5af4f..d7e5e45 100644 --- a/config.example.toml +++ b/config.example.toml @@ -1,11 +1,11 @@ [app] - project_version="0.3.2" + project_version="0.3.4" # 支持视频理解的大模型提供商 # gemini # NarratoAPI # qwen2-vl (待增加) vision_llm_provider="gemini" - vision_batch_size = 5 + vision_batch_size = 7 vision_analysis_prompt = "你是资深视频内容分析专家，擅长分析视频画面信息，分析下面视频画面内容，只输出客观的画面描述不要给任何总结或评价" ########## Vision Gemini API Key diff --git a/webui/components/script_settings.py b/webui/components/script_settings.py index ff8cb1a..0b3d5d1 100644 --- a/webui/components/script_settings.py +++ b/webui/components/script_settings.py @@ -18,6 +18,69 @@ from app.utils import utils, check_script, vision_analyzer, video_processor from webui.utils import file_utils +def get_batch_timestamps(batch_files, prev_batch_files=None): + """ + 获取一批文件的时间戳范围 + 返回: (first_timestamp, last_timestamp, timestamp_range) + + 文件名格式: keyframe_001253_000050.jpg + 其中 000050 表示 00:00:50 (50秒) + 000101 表示 00:01:01 (1分1秒) + + Args: + batch_files: 当前批次的文件列表 + prev_batch_files: 上一个批次的文件列表，用于处理单张图片的情况 + """ + if not batch_files: + logger.warning("Empty batch files") + return "00:00", "00:00", "00:00-00:00" + + # 如果当前批次只有一张图片，且有上一个批次的文件，则使用上一批次的最后一张作为首帧 + if len(batch_files) == 1 and prev_batch_files and len(prev_batch_files) > 0: + first_frame = os.path.basename(prev_batch_files[-1]) + last_frame = os.path.basename(batch_files[0]) + logger.debug(f"单张图片批次，使用上一批次最后一帧作为首帧: {first_frame}") + else: + # 提取首帧和尾帧的时间戳 + first_frame = os.path.basename(batch_files[0]) + last_frame = os.path.basename(batch_files[-1]) + + # 从文件名中提取时间信息 + first_time = first_frame.split('_')[2].replace('.jpg', '') # 000050 + last_time = last_frame.split('_')[2].replace('.jpg', '') # 000101 + + # 转换为分:秒格式 + def format_timestamp(time_str): + # 时间格式为 MMSS，如 0050 表示 00:50, 0101 表示 01:01 + if len(time_str) < 4: + logger.warning(f"Invalid timestamp format: {time_str}") + return "00:00" + + minutes = int(time_str[-4:-2]) # 取后4位的前2位作为分钟 + seconds = int(time_str[-2:]) # 取后2位作为秒数 + + # 处理进位 + if seconds >= 60: + minutes += seconds // 60 + seconds = seconds % 60 + + return f"{minutes:02d}:{seconds:02d}" + + first_timestamp = format_timestamp(first_time) + last_timestamp = format_timestamp(last_time) + timestamp_range = f"{first_timestamp}-{last_timestamp}" + + logger.debug(f"解析时间戳: {first_frame} -> {first_timestamp}, {last_frame} -> {last_timestamp}") + return first_timestamp, last_timestamp, timestamp_range + +def get_batch_files(keyframe_files, result, batch_size=5): + """ + 获取当前批次的图片文件 + """ + batch_start = result['batch_index'] * batch_size + batch_end = min(batch_start + batch_size, len(keyframe_files)) + return keyframe_files[batch_start:batch_end] + def render_script_panel(tr): """渲染脚本配置面板""" with st.container(border=True): @@ -238,7 +301,7 @@ def generate_script(tr, params): # 检查是否已经提取过关键帧 keyframe_files = [] if os.path.exists(video_keyframes_dir): - # 获取已有的关键帧文件 + # 取已有的关键帧文件 for filename in sorted(os.listdir(video_keyframes_dir)): if filename.endswith('.jpg'): keyframe_files.append(os.path.join(video_keyframes_dir, filename)) @@ -282,7 +345,7 @@ def generate_script(tr, params): except Exception as cleanup_err: logger.error(f"清理失败的关键帧目录时出错: {cleanup_err}") - raise Exception(f"关键帧提取失败: {str(e)}") + raise Exception(f"关键帧提取��败: {str(e)}") # 根据不同的 LLM 提供商处理 vision_llm_provider = st.session_state.get('vision_llm_providers').lower() @@ -327,38 +390,27 @@ def generate_script(tr, params): # 合并所有批次的析结果 frame_analysis = "" + prev_batch_files = None + for result in results: if 'error' in result: logger.warning(f"批次 {result['batch_index']} 处理出现警告: {result['error']}") continue - # 获取当前批次的图片文件 - batch_start = result['batch_index'] * config.app.get("vision_batch_size", 5) - batch_end = min(batch_start + config.app.get("vision_batch_size", 5), len(keyframe_files)) - batch_files = keyframe_files[batch_start:batch_end] + batch_files = get_batch_files(keyframe_files, result, config.app.get("vision_batch_size", 5)) + logger.debug(f"批次 {result['batch_index']} 处理完成，共 {len(batch_files)} 张图片") + logger.debug(batch_files) - # 提取首帧和尾帧的时间戳 - first_frame = os.path.basename(batch_files[0]) - last_frame = os.path.basename(batch_files[-1]) - - # 从文件名中提取时间信息 - first_time = first_frame.split('_')[2].replace('.jpg', '') # 000002 - last_time = last_frame.split('_')[2].replace('.jpg', '') # 000005 - - # 转换为分:秒格式 - def format_timestamp(time_str): - seconds = int(time_str) - minutes = seconds // 60 - seconds = seconds % 60 - return f"{minutes:02d}:{seconds:02d}" - - first_timestamp = format_timestamp(first_time) - last_timestamp = format_timestamp(last_time) + first_timestamp, last_timestamp, _ = get_batch_timestamps(batch_files, prev_batch_files) + logger.debug(f"处理时间戳: {first_timestamp}-{last_timestamp}") # 添加带时间戳的分析结果 frame_analysis += f"\n=== {first_timestamp}-{last_timestamp} ===\n" frame_analysis += result['response'] frame_analysis += "\n" + + # 更新上一个批次的文件 + prev_batch_files = batch_files if not frame_analysis.strip(): raise Exception("未能生成有效的帧分析结果") @@ -378,45 +430,27 @@ def generate_script(tr, params): # 构建帧内容列表 frame_content_list = [] + prev_batch_files = None + for i, result in enumerate(results): if 'error' in result: continue - # 获取当前批次的图片文件 - batch_start = result['batch_index'] * config.app.get("vision_batch_size", 5) - batch_end = min(batch_start + config.app.get("vision_batch_size", 5), len(keyframe_files)) - batch_files = keyframe_files[batch_start:batch_end] - - # 提取首帧和尾帧的时间戳 - first_frame = os.path.basename(batch_files[0]) - last_frame = os.path.basename(batch_files[-1]) - - # 从文件名中提取时间信息 - first_time = first_frame.split('_')[2].replace('.jpg', '') # 000002 - last_time = last_frame.split('_')[2].replace('.jpg', '') # 000005 - - # 转换为分:秒格式 - def format_timestamp(time_str): - seconds = int(time_str) - minutes = seconds // 60 - seconds = seconds % 60 - return f"{minutes:02d}:{seconds:02d}" - - first_timestamp = format_timestamp(first_time) - last_timestamp = format_timestamp(last_time) - - # 构建时间戳范围字符串 (MM:SS-MM:SS 格式) - timestamp_range = f"{first_timestamp}-{last_timestamp}" + batch_files = get_batch_files(keyframe_files, result, config.app.get("vision_batch_size", 5)) + _, _, timestamp_range = get_batch_timestamps(batch_files, prev_batch_files) frame_content = { - "timestamp": timestamp_range, # 使用时间范围格式 "MM:SS-MM:SS" - "picture": result['response'], # 图片分析结果 - "narration": "", # 将由 ScriptProcessor 生成 - "OST": 2 # 默认值 + "timestamp": timestamp_range, + "picture": result['response'], + "narration": "", + "OST": 2 } frame_content_list.append(frame_content) logger.debug(f"添加帧内容: 时间范围={timestamp_range}, 分析结果长度={len(result['response'])}") + + # 更新上一个批次的文件 + prev_batch_files = batch_files if not frame_content_list: raise Exception("没有有效的帧内容可以处理") @@ -442,13 +476,15 @@ def generate_script(tr, params): ) adapter = HTTPAdapter(max_retries=retry_strategy) session.mount("https://", adapter) - - response = session.post( - f"{config.app.get('narrato_api_url')}/video/config", - params=api_params, - timeout=30, - verify=True # 启用证书验证 - ) + try: + response = session.post( + f"{config.app.get('narrato_api_url')}/video/config", + params=api_params, + timeout=30, + verify=True # 启用证书验证 + ) + except: + pass custom_prompt = st.session_state.get('custom_prompt', '') processor = ScriptProcessor( @@ -621,7 +657,7 @@ def save_script(tr, video_clip_json_details): # 显示成功消息 st.success(tr("Script saved successfully")) - # 强制重新加载页面��更新选择框 + # 强制重新加载页面更新选择框 time.sleep(0.5) # 给一点时间让用户看到成功消息 st.rerun()