New version of playlist4whisper: Playlist GUI that plays a livestream with AI transcription, multi-instance/multi-user support, and customizable channel settings

[antor44]

"Playlist4Whisper" is an application that displays a playlist for "livestream_video.sh". It plays an online video and uses AI technology to transcribe the audio into text. It supports multi-instance and multi-user execution, and allows for changing options per channel and global options.

https://github.com/antor44/livestream_video

Author: Antonio R. Version: 1.60 License: GPL 3.0

Usage:

python playlist4whisper.py

-Support for IPTV, YouTube and Twitch

The program will load the default playlists playlist_iptv.m3u, playlist_youtube.m3u, and playlist_twitch.m3u, and will store options in config_xxx.json.

When updating, you may need to delete your previous configuration files in the installation directory: config_iptv.json, config_youtube.json, and config_twitch.json.

For multi-instances with SMPlayer: Go to Preferences - Interface - Instances, and turn off the option to use only one instance.

The application has a simple GUI using Python and the Tkinter library. It transcribes livestreams by feeding the output of ffmpeg to whisper.cpp, based on "livestream.sh" from whisper.cpp.

This program depends on other Linux programs and their libraries, such as Python, whisper.cpp and mpv. For example, Ubuntu Linux users can install the following packages:

sudo apt-get install mpv smplayer ffmpeg python3-tk

For YouTube yt-dlp is required (https://github.com/yt-dlp/yt-dlp) For Twitch streamlink is required (https://streamlink.github.io/)

playlist4whisper.py, livestream_video.sh, and the default playlist_xxx.m3u files must be located in the same directory as whisper-cpp. The main executable of whisper.cpp, which is the primary example, should be in the same directory with the default executable name 'main'. Additionally, the Whisper model file from OpenAI should be placed in the "models" subdirectory with the correct format and name, as specified in the Whisper.cpp repository. This can be done using terminal commands such as the following examples:

make tiny.en

make small

---

You can run play4whisper.py on macOS by following these steps:

Install Homebrew by visiting https://brew.sh/ and following the installation instructions.

Once Homebrew is installed, open a terminal and install the required dependencies. Run the following commands:

brew install python3
brew install [email protected]
brew install make
brew install xquartz
brew install xterm
brew install mpv

Next, install the necessary Python packages using pip3. Run the following commands:
pip3 install yt-dlp
pip3 install streamlink

Download some models and compile whisper-cpp following the instructions provided in the documentation at https://github.com/ggerganov/whisper.cpp . If you encounter an "Illegal instruction: 4" error during compilation, you can resolve it by deleting line 67 "CFLAGS += -mf16c" in the Makefile.

playlist4whisper.py, livestream_video.sh, and the default playlist_xxx.m3u files must be located in the same directory as whisper-cpp.

Finally, you can run the GUI by executing the following command in the terminal:

python3 playlist4whisper.py

Please note that on macOS, only xterm terminal and mpv videoplayer are supported.