M3u8-To-MP4

Python downloader for saving m3u8 videos to local MP4 files.

QuickStart

Install m3u8_To_MP4 via pip

1. Preparation: configure ffmpeg. (e.g., Win10)

   • Download "release full" build. It will have the largest set of libraries with greater functionality.
   • Extract the contents in the ZIP file to a folder of your choice.
   • To add FFmpeg to Win10 path. (User variables -> Path -> New and add)
   • Verify. Open the Command Prompt or PowerShell window, type ffmpeg, and press Enter.

2. Installation: m3u8_To_MP4

   # via pypi.org
   python -m pip install m3u8_To_MP4
   
   # first clone project, and install.
   git clone https://github.com/songs18/m3u8_To_MP4.git
   python -m pip install ./m3u8_To_MP4

Download a mp4 video

There are two options to download a m3u8 video into a mp4 file: async and multi-threads.

Multi-thread downloader (recommend)

import m3u8_To_MP4

if __name__ == '__main__':
    # 1. Download videos from uri.
    m3u8_To_MP4.multithread_download('http://videoserver.com/playlist.m3u8')

    # 2. Download videos from existing m3u8 files.
    m3u8_To_MP4.multithread_file_download('http://videoserver.com/playlist.m3u8',m3u8_file_path)

    # For compatibility, i reserve this api, but i do not recommend to you again.
    # m3u8_To_MP4.download('http://videoserver.com/playlist.m3u8')

Asynchronous downloader

import m3u8_To_MP4

if __name__ == '__main__':
    # 1. Download mp4 from uri.
    m3u8_To_MP4.async_download('http://videoserver.com/playlist.m3u8')

    # 2. Download mp4 from existing m3u8 files.
    m3u8_To_MP4.async_file_download('http://videoserver.com/playlist.m3u8',m3u8_file_path)

Resuming

If you use default tmp dir, resuming the transfer from the point of interruption will be executed automatically (based on crc32 hashing).

Custom http request header

In some cases, customized http request headers help to match some website requirements. For the available APIs, you can pass in a dictionary type header, which overrides the settings in the program. A simple example is:

import m3u8_To_MP4

if __name__ == '__main__':
    customized_http_header=dict()
    customized_http_header['Referer']='https://videoserver.com/'

    m3u8_To_MP4.multithread_download('http://videoserver.com/playlist.m3u8',customized_http_header=customized_http_header)

Features

1. Treat ffmpeg as a system service to achieve cross-platform.
2. If ffmpeg is not found, archiving is also supported. (v0.1.3 new features)
3. Resume from interruption. (based on crc32 temp directory path)
4. Use system tmp folder.
5. Concurrent requests based on the thread pool.
6. Concurrent requests based on efficient coroutines (v0.1.3 new features).
7. The retry strategy is carried out collectively after the whole cycle is repeated, avoiding the problem of short retry interval.
8. Download videos from existing m3u8 files.
9. Anti-crawler parameters based on customized request headers.
10. Clean codes based on inheritance.

TODO

• Errors: application data after close notify (related to the Python interpreter).
• Extract independent asynchronous http package.
• Support IPv6.
• Compare ffmpeg/avconv/mencoder/moviepy.
• Support bilibili etc.