LukeBot 2.0

Czech Discord bot built with py-cord and Django ORM. Provides entertainment and utility commands for a Czech Discord community.

Requirements

• Python 3.14+
• uv
• Chrome + ChromeDriver (for lunch parsers using Selenium)

Setup

# Clone & install dependencies
uv sync

# Copy env template and fill in values
cp env.example .env

# Apply database migrations
uv run python manage.py migrate

# Run the bot
uv run python main.py

Environment Variables

Variable	Description
BOT_SECRET	Discord bot token
SETTINGS_SECRET_KEY	Django secret key
DISCORD_SERVER_IDS	Comma-separated server IDs for slash command sync
REDDIT_CLIENT	Reddit API client ID
REDDIT_SECRET	Reddit API client secret
GIPHY_CLIENT	Giphy API token
LUNCH_ROOMS	Comma-separated Discord channel IDs for lunch menus

Docker

cp env.example .env
# fill in .env
docker compose up -d

Running Tests

uv run pytest

Tests use an in-memory SQLite database and mock all external APIs (Reddit, Giphy, HTTP requests).

Project Structure

bot/                  # Core layer (permissions, plugin loader, DB helpers)
plugins/              # Bot plugins (each is a py-cord Cog)
  core/               # Runtime plugin management (reload/activate/deactivate)
  lunch/              # Daily lunch menus scraped from restaurant websites
    parsers/          # One file per restaurant (knoflik, radegast, budha, ...)
  misc/               # Fun commands (ping, sexymetr, gay_calculator, ...)
  reddit/             # Auto-sends images from configured subreddits
  responses/          # Passive listener — reacts to Czech text patterns
data/                 # Django app with SQLite models
tests/                # pytest test suite

Architecture Notes

• Plugin loading happens in LukeBot.setup_hook() via bot.plugin_loader.all_plugins() — loads each plugins/<name>/<name>.py that exposes setup(bot).
• Permissions are hardcoded in bot/constants.py (Discord user/role IDs). Moving these to the database is a known TODO.
• Reddit uses a RedditManager class: PRAW calls run in a thread pool (asyncio.to_thread), URL validation uses aiohttp, images are bulk-inserted via bulk_create(ignore_conflicts=True).
• RunningCommand model prevents duplicate concurrent execution of long-running commands.
• Lunch parsers run at 10:10 AM Europe/Prague time and scrape restaurant websites. vlk and orel require Chrome (Selenium).