Viewers

causal_world.viewers.task_viewer.get_world(task_generator_id, task_params, world_params, enable_visualization=False, env_wrappers=array([], dtype=float64), env_wrappers_args=array([], dtype=float64))

Returns a particular CausalWorld instance with optional wrappers

Parameters

• task_generator_id – (str) id of the task of the environment

• task_params – (dict) task params of the environment

• world_params – (dict) world_params of the environment

• enable_visualization – (bool) if GUI visualization is enabled

• env_wrappers – (list) a list of gym wrappers

• env_wrappers_args – (list) a list of kwargs for the gym wrappers

Returns

(CausalWorld) a CausalWorld environment instance

causal_world.viewers.task_viewer.record_video(env, policy, file_name, number_of_resets=1, max_time_steps=None)

Records a video of a policy for a specified environment :param env: (causal_world.CausalWorld) the environment to use for

recording.

Parameters

• policy – the policy to be evaluated

• file_name – (str) full path where the video is being stored.

• number_of_resets – (int) the number of resets/episodes to be viewed

• max_time_steps – (int) the maximum number of time steps per episode

Returns

causal_world.viewers.task_viewer.record_video_of_episode(episode, file_name, env_wrappers=array([], dtype=float64), env_wrappers_args=array([], dtype=float64))

Records a video of a logged episode for a specified environment

Parameters

• episode – (Episode) the logged episode

• file_name – (str) full path where the video is being stored.

• env_wrappers – (list) a list of gym wrappers

• env_wrappers_args – (list) a list of kwargs for the gym wrappers

Returns

causal_world.viewers.task_viewer.record_video_of_policy(task, world_params, policy_fn, file_name, number_of_resets, max_time_steps=100, env_wrappers=array([], dtype=float64), env_wrappers_args=array([], dtype=float64))

Records a video of a policy for a specified environment

Parameters

• task – (Task) the task of the environment

• world_params – (dict) the world_params of the environment

• policy_fn – the policy to be evaluated

• file_name – (str) full path where the video is being stored.

• number_of_resets – (int) the number of resets/episodes to be viewed

• max_time_steps – (int) the maximum number of time steps per episode

• env_wrappers – (list) a list of gym wrappers

• env_wrappers_args – (list) a list of kwargs for the gym wrappers

Returns

causal_world.viewers.task_viewer.record_video_of_random_policy(task, world_params, file_name, number_of_resets, max_time_steps=100, env_wrappers=array([], dtype=float64), env_wrappers_args=array([], dtype=float64))

Records a video of a random policy for a specified environment

Parameters

• task – (Task) the task of the environment

• world_params – (dict) the world_params of the environment

• file_name – (str) full path where the video is being stored.

• number_of_resets – (int) the number of resets/episodes to be viewed

• max_time_steps – (int) the maximum number of time steps per episode

• env_wrappers – (list) a list of gym wrappers

• env_wrappers_args – (list) a list of kwargs for the gym wrappers

Returns

causal_world.viewers.task_viewer.view_episode(episode, env_wrappers=array([], dtype=float64), env_wrappers_args=array([], dtype=float64))

Visualizes a logged episode in the GUI

Parameters

• episode – (Episode) the logged episode

• env_wrappers – (list) a list of gym wrappers

• env_wrappers_args – (list) a list of kwargs for the gym wrappers

Returns

causal_world.viewers.task_viewer.view_policy(task, world_params, policy_fn, max_time_steps, number_of_resets, env_wrappers=array([], dtype=float64), env_wrappers_args=array([], dtype=float64))

Visualizes a policy for a specified environment in the GUI

Parameters

• task – (Task) the task of the environment

• world_params – (dict) the world_params of the environment

• policy_fn – the policy to be evaluated

• max_time_steps – (int) the maximum number of time steps per episode

• number_of_resets – (int) the number of resets/episodes to be viewed

• env_wrappers – (list) a list of gym wrappers

• env_wrappers_args – (list) a list of kwargs for the gym wrappers

Returns