Python implements m3u8 to download mp4 video principle and source code

python download m3u8 video

Use python to download m3u8 video files (MP4). The following is the principle of the m3u8 file organized by me based on my personal understanding, as well as the download principle and the source code of the download. If there are any mistakes, please comment and point out

m3u8 video file principle

Now the size of a video ranges from a few hundred M to several G. It is impossible for us to wait for a video to be downloaded before watching it when playing a video on the Internet. Instead, it uses the method of "buffering" while watching, and the m3u8 video format is commonly used to achieve this method

1. Slice mp4 video file into multiple ts video files
2. The web page starts from video loading , downloads ts files in sequence, and realizes playing downloaded ts segment videos while downloading ts segment video files
3. Each ts video file is equivalent to a part of the original mp4 video file
4. The m3u8 file is actually a list file, which stores all the ts file paths sliced ​​by an mp4 file
5. Loading videos on the web page is to download the ts segment video files sequentially according to the corresponding ts list address in m3u8 to realize "buffering" while playing the "buffered" video (that is, the downloaded ts segment video)
   

Python download m3u8 video principle

After the above study, we already know that the mp4 file is actually sliced ​​into multiple ts files, and then the download path of the ts file is sorted into a list and saved in our m3u8 file

Now we should understand that the m3u8 file actually just stores the download path of the ts file we need , not a video file.

Therefore, we only need to download the ts files sequentially according to the ts link list in the m3u8 file . Then, we stitch all the downloaded ts video clip files together, which is our complete mp4 video file

You may have noticed the encryption method in the picture above. In fact, some m3u8 ts video clips are encrypted, but some are not. For encrypted videos, AES encryption is generally used . We only need to download the AES encrypted key according to the key file link given in the m3u8 list , and then decrypt it according to the key .

Aes encryption/decryption will not be repeated here, you can understand it by yourself

Note: Some ts link addresses in m3u8 files are incomplete and need to be spliced ​​into complete links

Python implementation source code

1. Install the module
   requests request module by yourself, pip install requests
   AES decryption module, pip install pycryptodome (do not install the wrong module~)
2. Change as needed
   The maximum number of threads in the thread pool can be adjusted as needed max_workers = 20
   The number of reissue requests can be adjusted as needed max_request = 5
3. If the ts link is incomplete, you need to change the ts link request header ts_url_title = ''
   by yourself (since this situation does not exist in general m3u8 files, so laziness has not been processed, and the missing link header is generally the header of the m3u8 link)

# m3u8视频下载

import os
import re
import time
import shutil
import requests
from concurrent.futures import ThreadPoolExecutor, wait
from Crypto.Cipher import AES


# UA伪装
headers = {
    
    
    'User-Agent': 'Mozilla/5.0 (SymbianOS/9.4; Series60/5.0 NokiaN97-1/20.0.019; Profile/MIDP-2.1 Configuration/CLDC-1.1) AppleWebKit/525 (KHTML like Gecko) BrowserNG/7.1.18124'
}


def download_mp4(mp4_file_path, ts_url_list, ts_url_title):
    '''下载ts文件并写入mp4文件

    :param mp4_file_path: mp4文件名
    :param ts_url_list: ts请求链接列表
    :return:
    '''

    # 判断文件是否存在，存在则先清空
    if os.path.exists(mp4_file_path):
        with open(mp4_file_path, 'w') as fp:
            fp.write('')

    # 创建存放ts的文件夹
    if not os.path.exists('ts'):
        os.mkdir('ts')

    print('开始下载{}...'.format(mp4_file_path))

    excutor = ThreadPoolExecutor(max_workers=20)  # 线程池

    len_list = len(ts_url_list)  # ts链接总数
    all_tasks = [excutor.submit(lambda args: download_ts(*args), (ts_url_id, len_list, ts_url_list, ts_url_title))
                 for ts_url_id in range(len_list)]  # 创建任务

    wait(all_tasks)  # 等待所有任务执行完成

    # 检测ts数目是否正确
    if len(os.listdir('ts')) == len_list:
        pass
    else:
        print('ts文件部分缺失...')

        # 删除存放ts的临时文件
        shutil.rmtree('ts')

        return ''

    # ts合并为mp4文件
    print('ts文件下载完成，正在合并ts文件...')
    for ts_url_id in range(len_list):
        ts_file_name = 'ts/{}.ts'.format(ts_url_id)
        with open(ts_file_name, 'rb') as fp:
            ts_content = fp.read()  # 读取ts数据

        with open(mp4_file_path, 'ab') as fp:
            fp.write(ts_content)  # 将ts数据追加写入文件

    print('ts文件合并成功！')

    # 删除存放ts的临时文件
    shutil.rmtree('ts')

    return 1


def download_ts(ts_url_id, len_list, ts_url_list, ts_url_title):
    ''' 请求下载ts文件

    :param ts_url_id: 分区ts的id
    :param len_list: ts个数
    :param ts_url_list: 存放ts的列表
    :param ts_url_title: ts链接拼接的头部
    :return:
    '''

    print('{}/{}开始下载'.format(ts_url_id, len_list - 1))

    # 请求不成功补发请求,最大补发次数为
    max_request = 5  # 最大补发请求次数
    for i in range(max_request):
        try:
            response = requests.get(url=ts_url_title + ts_url_list[ts_url_id],
                                    headers=headers, timeout=(5, 20))  # 请求获取ts数据
            if response.status_code == 200:
                ts_content = response.content
                break
        except:
            if i == max_request - 1:
                print('{}/{}下载失败'.format(ts_url_id, len_list - 1))
                return ''
            else:
                print('{}/{}下载失败，正在补发请求...'.format(ts_url_id, len_list - 1))

    ts_file_name = 'ts/{}.ts'.format(ts_url_id)

    with open(ts_file_name, 'wb') as fp:
        fp.write(ts_content)  # 将ts数据写入文件

    print('{}/{}下载完成'.format(ts_url_id, len_list - 1))


def deciphering(key, fileName):
    '''对aes加密视频进行解密

    :param key: aes解密密钥
    :param fileName: 需要解密的文件
    :return:
    '''

    # 读取原文件
    with open(fileName, 'rb') as fp:
        part = fp.read()

    # aes解密需要的偏移量
    iv = b'0000000000000000'

    # 解密数据
    plain_data = AES.new(key, AES.MODE_CBC, iv).decrypt(part)

    # 将解密数据写入文件
    with open(fileName, 'wb') as fp:
        fp.write(plain_data)

    print('视频解密完成！')


def timer(start_time, end_time, mp4_file_name):
    '''计时器

    :param start_time: 开始时间
    :param end_time: 结束时间
    :return:
    '''

    spend_second = end_time - start_time
    hour = str(int(spend_second / (60 * 60)))
    minute = str(int(spend_second / 60))
    second = str(int(spend_second % 60))
    spend_time = '{}h{}m{}s'.format(hour, minute, second)

    print('{}下载完成！用时：{}'.format(mp4_file_name, spend_time))


def start(m3u8_url, mp4_file_name, ts_url_title):
    '''开始

    :param m3u8_url: m3u8链接
    :param mp4_file_path: 下载后的视频名称
    :return:
    '''

    # 开始计时
    start_time = time.time()

    # 创建目录文件
    if not os.path.exists('mv'):
        os.mkdir('mv')

    # 视频保存路径
    mp4_file_path = 'mv/' + mp4_file_name + '.mp4'

    # 获取m3u8内容
    m3u8_file = requests.get(url=m3u8_url, headers=headers).text

    # 整理ts列表
    ts_url_list = re.findall(',\n(.*?)\n#', m3u8_file)

    # 下载ts，并拼接为mp4文件
    mp4 = download_mp4(mp4_file_path, ts_url_list, ts_url_title)

    # 判断是否存在加密
    if mp4 and re.search('#EXT-X-KEY', m3u8_file):
        print('{}视频存在加密，正在对其进行解密，请稍后...'.format(mp4_file_path))

        # 获取key
        key_url = re.search('#EXT-X-KEY:(.*URI="(.*)")\n', m3u8_file)[2]  # 获取key的url
        key = requests.get(url=key_url, headers=headers).content  # 请求获取key

        # 解密视频
        deciphering(key, mp4_file_path)

    # 计时结束
    end_time = time.time()

    # 耗时统计
    timer(start_time, end_time, mp4_file_name)


if __name__ == '__main__':
    # m3u8 链接
    m3u8_url = input('请输入m3u8链接：')

    # ts链接头
    ts_url_title = ''

    # mp4 名称
    mp4_file_name = input('请输入视频名称：')

    # 执行下载
    start(m3u8_url, mp4_file_name, ts_url_title)

Hope to help you~