Simplify examples for building multi-agent apps with Docker/Kubernetes

[nerdai]

The pattern for building multi-agent apps to be orchestrated by Docker Compose or Kubernetes involves the following steps:

1. Build a multi-agent app Python package containing three modules: core_services, agent_services, additional_services
2. Any service built in those modules also contain its launch() logic and gets executed if __name__ == "__main__"
3. This allows for a single Docker image for all of the services, and we just need to use a command that points to the correct service in order to build that service in Docker Compose or Kubernetes.

The code in the examples: kafka/pig-latin-translation or rabbitmq/multi-agent-app-rabbitmq utilize this pattern and feel a tad heavy. The point of this issue is to slim down the process/code for these examples and generally our method for building multi-agent apps for deployment.