Setup test model server

Created by: herzogrh

This PR includes the test model server that can be reached both via HTTP and WebSockets. For further documentation, see the Readme.

While setting up the server it became clear that when running large models, the HTTP requests would eventually time-out. That means that when working with HTTP requests, each model server would need to have some sort of queuing service/session management to return the correct results when another request is send from the same client. That to me seems very complicated to implement.

Realizing that, I set up a WebSocket connection to the server which does not time out and could also provide status updates as the model run progresses. There is also a test html file where one can test the WebSocket endpoint.

The test model server was added as a service to the docker-compose file.