rose-game-ai-reference

ROSE game example of a machine learining self driving car module.

This is a reference implementation of a self driving modules for the ROSE game.

This example uses machine learning algorithms, and requires pytorch when running locally.

ROSE project: https://github.com/RedHat-Israel/ROSE

ROSE self driving car module template: https://github.com/RedHat-Israel/rose-game-ai

Requirements

Requires	Version
Podman (or Docker)	>= 4.8	For running containerized
Python	>= 3.9	For running the code loally

ROSE game components

Component	Reference
Game engine	https://github.com/RedHat-Israel/rose-game-engine
Game web based user interface	https://github.com/RedHat-Israel/rose-game-web-ui
Self driving car module	https://github.com/RedHat-Israel/rose-game-ai
Self driving car module example	https://github.com/RedHat-Israel/rose-game-ai-reference

Running a self driving module reference example

podman run --rm --network host -it quay.io/rose/rose-game-ai-reference:latest

Running a self driving module locally

Clone this repository, and make sure you have a game engine running.

Install requirements:

# Install requirements
pip install -r requirements.txt
pip install -r requirements-dev.txt

Write your own driving module, you can use the file mydriver.py:

vi mydriver.py

Run using mydriver.py as the driving module:

python main.py --driver mydriver.py

Running ROSE game components containerized

Running the game engine ( on http://127.0.0.1:8880 )

podman run --rm --network host -it quay.io/rose/rose-game-engine:latest

Running the game web based user interface ( on http://127.0.0.1:8080 )

podman run --rm --network host -it quay.io/rose/rose-game-web-ui:latest

Running community contributed driver ( on http://127.0.0.1:8082 )

You can use community drivers to compare and evaluate your driver during the development process.

podman run --rm --network host -it quay.io/yaacov/rose-go-driver:latest --port 8082

Running your self driving module, requires a local mydriver.py file with your driving module. ( on http://127.0.0.1:8081 )

# NOTE: will mount mydriver.py from local directory into the container file system
podman run --rm --network host -it \
  -v $(pwd)/:/driver:z \
  -e DRIVER=/driver/mydriver.py \
  -e PORT=8081 \
  quay.io/rose/rose-game-ai-reference:latest

Testing your driver

Send car and track information by uring POST request to your driver ( running on http://127.0.0.1:8081 ):

curl -X POST -H "Content-Type: application/json" -d '{
            "info": {
                "car": {
                    "x": 3,
                    "y": 8
                }
            },
            "track": [
                ["", "", "bike", "", "", ""],
                ["", "crack", "", "", "trash", ""],
                ["", "", "penguin", "", "", "water"],
                ["", "water", "", "trash", "", ""],
                ["barrier", "", "", "", "bike", ""],
                ["", "", "trash", "", "", ""],
                ["", "crack", "", "", "", "bike"],
                ["", "", "", "penguin", "water", ""],
                ["", "", "bike", "", "", ""]
            ]
        }' http://localhost:8081/

The response in JSON format should include the car name and the recommended action:

{
  "info": {
    "name": "Go Cart",
    "action": "pickup"
  }
}