Key.step

Key.step(action: Action, front_object: Self | None = None, walkable: bool = False) → float

Compute grid object changes after single timestep.

The default step behaviour is idle and gives 0 reward.

Parameters:

• action (Action) – Action taken in a grid environment by the RL agent.

• front_object (GridObject | None, optional) – Grid object currently in front of the RL agent, by default None.

• walkable (bool, optional) – Wheter the agent can move over the cell in front, by default False.

Returns:

Optional reward, which is independent of the environment task and episode termination.

Return type:

float