Wrappers

navground.learning.wrappers

class MaskWrapper(env: ParallelEnv[int, Observation, Action], observation_indices: Iterable[int] = (), observation_keys: Iterable[str] = (), action_indices: Iterable[int] = ())

Bases: BaseParallelWrapper[int, dict[str, ndarray[tuple[Any, …], dtype[Any]]] | ndarray[tuple[Any, …], dtype[Any]], ndarray[tuple[Any, …], dtype[Any]]]

This wrapper masks some of the observations and/or actions

Parameters:

• env (ParallelEnv[int, Observation, Action]) – The wrapped environment

• observation_indices (Iterable[int]) – Which observation indices to keep (if array)

• observation_keys (Iterable[str]) – Which observation keys to keep (if dict)

• action_indices (Iterable[int]) – Which action indices to keep (if array)

class NameWrapper(env: BaseParallelEnv)

Bases: BaseParallelWrapper[str, dict[str, ndarray[tuple[Any, …], dtype[Any]]] | ndarray[tuple[Any, …], dtype[Any]], ndarray[tuple[Any, …], dtype[Any]]]

This wrapper renames the agent using their navground tags as prefix “tag1-tag2-…-tagn” or “agent” (if no tag is set).

Indices are appended to the predix, e.g. agent_0, agent_1, … .

Parameters:

env (BaseParallelEnv) – The wrapped environment