hk_bot

A simple Telegram bot framework for wind/kite sports updates. Commands are auto-discovered plugins — add a file to src/commands/ and it just works.

Commands

Command	Description
/woo	WOO Leaderboard
/windguru	Wind Forecast
/hkr	HKR Reviews
/iksurfmag	IKSurf News
/youtube	Latest video from configured YouTube channels

YouTube Russian Voiceover

When a YouTube video is fetched (via /youtube or embedded in /iksurfmag), the bot automatically:

1. Extracts auto-generated English subtitles via yt-dlp
2. Falls back to faster-whisper (tiny model, ~39 MB) if no subtitles exist
3. Translates the transcript to Russian via MyMemory (free)
4. Generates Russian TTS audio via gTTS (free, requires internet)
5. Mixes the TTS audio into the video using ffmpeg

Two messages are sent: the voiceover version (with caption) followed by the original. If any step fails, the original video is sent silently.

Project Structure

hk_bot/
├── src/
│   ├── bot.py                  # Entry point
│   ├── api/
│   │   ├── abstract_request_command.py   # Base class for all commands
│   │   ├── abstract_cron_command.py      # Mixin for scheduled commands
│   │   └── abstract_news_command.py      # Mixin for news-style commands
│   └── commands/               # Auto-discovered command plugins
│       ├── woo_command.py
│       ├── windguru_command.py
│       ├── hkr_command.py
│       └── iksurfmag_command.py
├── requirements.txt
└── .env.example

Adding a Command

1. Create src/commands/my_command.py
2. Define a class extending AbstractRequestCommand
3. Set NAME (slash command) and LABEL, implement async def run() -> str

from api.abstract_request_command import AbstractRequestCommand

class MyCommand(AbstractRequestCommand):
    NAME = "mycommand"
    LABEL = "My Command"

    async def run(self) -> str:
        return "Hello!"

The command is picked up automatically — no registration needed.

Setup

# System dependency (required for voiceover)
apt-get install -y ffmpeg   # Debian/Ubuntu/Raspberry Pi OS

pip install -r requirements.txt
cp .env.example .env
# Edit .env with your TELEGRAM_BOT_TOKEN
python src/bot.py

Note: On first run, faster-whisper will download the Whisper tiny model (~39 MB) to ~/.cache/ automatically. This only happens once.

Deploy

Push to main — GitHub Actions deploys automatically to the LXC container.

License

MIT