data_prepare/json_down.py

import os
import json  # 新增：用于读取JSON文件
import yt_dlp
from concurrent.futures import ThreadPoolExecutor, as_completed
from tqdm import tqdm

# ================= 配置区域 =================
# 修改：直接读取你提供的JSON文件路径
JSON_DATA_PATH = "/data/shuimu.chen/videomarathon/panda70m_mc_subset_100k.json"
SAVE_ROOT = "/data/shuimu.chen/videomarathon/100k_videos"
MAX_WORKERS = 1  # 保持单线程比较稳

# 【请确认】cookies.txt 文件的绝对路径
COOKIE_PATH = "/data/shuimu.chen/TimeSearch-R/data_prepare/cookies.txt" 
# ===========================================

def process_video(item):
    """
    单个视频下载任务函数 (完全不变)
    """
    video_rel_path = item['rel_path']
    url = item['url']
    
    # 构建保存路径
    # base_name 类似于 "category/video_id"
    base_name = os.path.splitext(video_rel_path)[0]
    full_save_path_no_ext = os.path.join(SAVE_ROOT, base_name)
    
    # 你的脚本用来检查的最终路径（带 .mp4）
    final_file_path = full_save_path_no_ext + ".mp4"

    # 检查文件是否已存在
    if os.path.exists(final_file_path):
        return "skipped"

    # 确保目录存在
    os.makedirs(os.path.dirname(final_file_path), exist_ok=True)

    # yt-dlp 配置
    ydl_opts = {
        'format': 'bestvideo[height<=480]+bestaudio/best[height<=480]',
        'sleep_interval': 5, # # 每次下载完至少睡 5 秒
        'max_sleep_interval': 15, # # 最多随机睡到 30 秒
        'merge_output_format': 'mp4',
        'sleep_interval_requests': 1, # 甚至连请求视频信息也要睡 5 秒
        # 'js_runtimes': ['node'],
        # 【核心修改】这里加上了 .%(ext)s
        # 意思就是：文件名 = 你的路径 + . + 扩展名(mp4)
        'outtmpl': f"{full_save_path_no_ext}.%(ext)s",
        'js_runtimes': ['node:/data/shuimu.chen/TimeSearch-R/data_prepare/node-v24.12.0-linux-x64/bin/node'],
        
        'quiet': False, # 建议开启 False 以便调试，稳定后可改为 True
        'no_warnings': False,
        'ignoreerrors': True,
        'cookiefile': COOKIE_PATH,  
    }

    try:
        with yt_dlp.YoutubeDL(ydl_opts) as ydl:
            ydl.download([url])
        
        # 再次检查文件是否生成
        if os.path.exists(final_file_path):
            return "success"
        else:
            return "failed"
    except Exception as e:
        return "error"

def main():
    # 0. 检查 cookie
    if not os.path.exists(COOKIE_PATH):
        print(f"❌ 错误: 找不到 Cookie 文件: {COOKIE_PATH}")
        print("请上传 cookies.txt 后再重试！")
        return

    # 1. 加载和筛选数据
    print(f"正在加载JSON数据文件: {JSON_DATA_PATH} ...")
    
    # 修改：直接加载你提供的JSON文件
    with open(JSON_DATA_PATH, 'r', encoding='utf-8') as f:
        dataset = json.load(f)
    
    print(f"已加载 {len(dataset)} 条数据")
    
    unique_videos = {}
    print("正在从JSON数据中筛选唯一视频任务 (Panda-70M + Multiple Choice) ...")
    
    # 修改：简化筛选逻辑，因为你的JSON已经是MC样本了
    for item in tqdm(dataset, desc="筛选唯一视频"):
        # 你提供的JSON里都是Panda-70M的MC样本，这里可以简化判断
        video_rel_path = item.get('video')
        url = item.get('URL')
        
        if video_rel_path and url:
            if video_rel_path not in unique_videos:
                unique_videos[video_rel_path] = url

    tasks = [{'rel_path': k, 'url': v} for k, v in unique_videos.items()]
    print(f"筛选完成！共 {len(tasks)} 个唯一视频任务。")
    print(f"注: 原JSON中有 {len(dataset)} 条问答对，涉及 {len(tasks)} 个唯一视频。")
    
    # 2. 多线程下载
    print(f"\n开始并发下载 (Workers={MAX_WORKERS}, Cookie已启用)...")
    
    results = {
        "success": 0,
        "skipped": 0,
        "failed": 0,
        "error": 0
    }

    with ThreadPoolExecutor(max_workers=MAX_WORKERS) as executor:
        future_to_video = {executor.submit(process_video, task): task for task in tasks}
        
        for future in tqdm(as_completed(future_to_video), total=len(tasks), desc="下载进度"):
            status = future.result()
            results[status] += 1

    # 3. 总结
    print("\n" + "=" * 40)
    print("处理完毕")
    print("=" * 40)
    print(f"JSON中问答对总数: {len(dataset)}")
    print(f"唯一视频任务数: {len(tasks)}")
    print(f"跳过(已存在): {results['skipped']}")
    print(f"下载成功: {results['success']}")
    print(f"下载失败: {results['failed'] + results['error']}")
    print(f"数据保存在: {os.path.abspath(SAVE_ROOT)}")
    
    # 额外统计信息
    if results['success'] > 0:
        coverage_rate = results['success'] / len(tasks) * 100
        print(f"视频覆盖率: {coverage_rate:.1f}%")

if __name__ == "__main__":
    main()