joschkabirk / lunchbot

Bot that extracts the lunch menu from the DESY canteen website and posts a message in a dedicated Mattermost channel.

lunchbot

Bot that extracts the lunch menu from the DESY canteen website and posts a message in a dedicated Mattermost channel.

Example:

At the moment, this implementation is designed to represent only one execution of the bot. To run it periodically, use e.g. cron.

Overall, the whole setup is kinda DESY-specific due to the fact that the DESY Sync&Share is used to store the menu preview images.

Usage

Requirements

The easiest way to run this bot is to use the dedicated Docker image. The image is available on Docker Hub at jobirk/lunchbot.

Furthermore, you have to set the following environment variables in a .env file:

API_TO_USE="openai" # options are "openai" and "huggingface"
HUGGINGFACE_API_URL="<your Hugging Face API URL>"
HUGGINGFACE_API_TOKEN="<your Hugging Face API token>"
ALSTERFOOD_WEBSITE_URL="https://desy.myalsterfood.de/"
OPENAI_API_KEY="<your OpenAI API key>"
MATTERMOST_WEBHOOK_URL="<your Mattermost webhook URL>"
MATTERMOST_WEBHOOK_URL_ALERT="<your Mattermost webhook URL for alert messages>"
USE_OPENAI_IMAGE_URL="<if set to 'true', the bot will use the OpenAI image API to generate images (which expire after 1 hour)>"
IMAGE_CLOUD_UPLOAD_URL="<your image cloud upload URL>"
IMAGE_CLOUD_UPLOAD_TOKEN="<your image cloud upload token>"
IMAGE_CLOUD_DOWNLOAD_URL="<your image cloud download URL>"
MESSAGE_PREFIX="<your message prefix>"
MESSAGE_SUFFIX_MON="<your message suffix for mondays>"
MESSAGE_SUFFIX_TUE="<your message suffix for tuesdays>"
...
MESSAGE_SUFFIX_SUN="<your message suffix for sundays>"
MATTERMOST_USERNAME="<username displayed for the bot>"
SYSTEM_CONTENT="<description of the system that describes the food>"
DESCRIPTION_SUFFIX="<suffix to put after "Description" in the table header>"
ALERT_PREFIX="<prefix for the alert message>"

Run the bot

Inside the container (and repo), just run:

python scripts/run_lunchbot.py

This will run the bot once. To run it periodically, just use e.g. cron to run the script every day at a specific time.