WhatSongIsThatBot

Identify songs in seconds right from Telegram. Not only through audio files but also via voice messages and videos!

Setup

Install python dependencies:

git clone https://github.com/Nachtalb/WhatSongIsThatBot
cd WhatSongIsThatBot
pip install -r requirements.txt

Install additional dependency songrec according to its instruction.

Copy over the sample config and adjust it's contents.

cp config.sample.json config.json

Name	Description
token	Telegram bot token acquired via BotFather
songrec	Path to songrec's binary

With webhook

To run in webhook mode, add these additional settings to your config.json. {token} will be automatically filled in.

{
    // ...

    "webhook": {
        "port": 9001,
        "host": "0.0.0.0",
        "path": "{token}",
        "url": "https://example.com"
    }
}

Like this the hook will be started on 0.0.0.0:9001/your_token and Telegram connects via https://example.com/your_token.

As a service

You can copy the wsit.service over to /etc/systemd/system/ or ~/.config/systemd/user/ to run the bot as a service. Once you have copied the file over, you have to adjust the paths inside it. Then:

# If it's inside ~/.config/systemd/user/ folder
systemctl --user daemon-reload               # load the service
systemctl --user enable --now wsit.service   # start the service on boot up and now

# Otherwise on system level
sudo systemctl daemon-reload
sudo systemctl enable --now wsit.service

Usage

There are no fancy option or whatever needed, just run the main.py file.

python main.py

old.py?

I first tried to implement the whole scheme with ShazamAPI. An unofficial packaged that aims to do the same as songrec programmatically in python. But after implementing and testing it, I dropped it. Compared with songrec it's very very slow and much more inaccurate. But I left it in here as reference material.