In preparation for reinforcement learning and other AI magic, we are currently overhauling the entire backend and middleware. This includes:
- Message queue based on RabbitMQ [done]
- Transition to a PostgreSQL database that stores all scenarios, actions, games, etc. instead of using a file-based approach [done]
- Rewriting the entire backend to C#, replacing the Python code [done]
Alpha players: Please be aware that this might temporarily (re)introduce errors that were not there is earlier versions. Hang in there, we are working on it!
Update (May 2024): Overhaul done & most bugs squished!