entity-gym
Prototype for an abstraction over reinforcement learning environments that represent observations as lists of structured objects.
The interface is defined in entity_gym/environment.py.
A number of simple example environments can be found in entity_gym/envs.
After installing entity-gym with pip install -e ., the example environments can be run with python entity_gym/main.py --env=MoveToOrigin|CherryPick|PickMatchingBalls|Minefield|MultiSnake.