关于机器学习:下载网课的ts视频

本来筹备用爬虫，开多线程，去下载一个个 ts 片段，最初用 ffmpeg 合成残缺的 ts 视频的。

from concurrent.futures import ThreadPoolExecutor
import requests
import logging
import re
import os

url = 'http://v3.julyedu.com/video/259/6390/01a311da6a2cd91-'


def download(name):
    str_name = "%05d" % name
    print(str_name)
    file_name = str_name + '.ts'
    print(url + file_name)
    try:
        res = requests.get(url=url + file_name, timeout=15)
        content = res.content

        with open(r'%s' % file_name, 'wb')as f:
            f.write(content)
            print(file_name + '\x1b[1;30;42m download success \033[0m')
            num = name // 20
            print(file_name + 'download complete,' + 'download' +
                  '%s %% %s' % (name / 11, '>' * num))

    except Exception as e:
        print(file_name + '\x1b[1;30;41m download fail \033[0m')
        print(e)
        name = re.findall('(\d+).ts', file_name)[0]
        print(name + 'download fail')

        my_log = logging.getLogger('lo')
        my_log.setLevel(logging.DEBUG)
        file = logging.FileHandler('error.log', encoding='utf-8')
        file.setLevel(logging.ERROR)
        my_log_fmt = logging.Formatter('%(asctime)s-%(levelname)s:%(message)s')
        file.setFormatter(my_log_fmt)
        my_log.addHandler(file)
        my_log.error(file_name + 'download fail')
        my_log.error(e)

        download(int(name))


p = ThreadPoolExecutor(2)
for name in range(1, 556 + 1):
    p.submit(download, name)

# win: copy /b *.ts video.ts

# ffmpeg -allowed_extensions ALL -i HdNz1kaz.m3u8 -c copy new.mp4
# https://blog.csdn.net/weixin_34190136/article/details/85989221

但最初发现，间接用 vlc 的串流，和网上的 m3u8 文件，就能够把残缺的 ts 视频下载下来了。成果还比片段拼成的视频​​连贯。。