sinergym.utils.wrappers.WandBLogger

class sinergym.utils.wrappers.WandBLogger(env: Env, entity: str | None = None, project_name: str | None = None, run_name: str | None = None, group: str | None = None, job_type: str | None = None, tags: List[str] | None = None, episode_percentage: float = 0.9, save_code: bool = False, dump_frequency: int = 1000, artifact_save: bool = True, artifact_type: str = 'output', excluded_info_keys: List[str] = ['reward', 'action', 'timestep', 'month', 'day', 'hour', 'time_elapsed(hours)', 'reward_weight', 'is_raining'], excluded_episode_summary_keys: List[str] = ['terminated', 'truncated'])

__init__(env: Env, entity: str | None = None, project_name: str | None = None, run_name: str | None = None, group: str | None = None, job_type: str | None = None, tags: List[str] | None = None, episode_percentage: float = 0.9, save_code: bool = False, dump_frequency: int = 1000, artifact_save: bool = True, artifact_type: str = 'output', excluded_info_keys: List[str] = ['reward', 'action', 'timestep', 'month', 'day', 'hour', 'time_elapsed(hours)', 'reward_weight', 'is_raining'], excluded_episode_summary_keys: List[str] = ['terminated', 'truncated'])

Wrapper to log data in WandB platform. It is required to be wrapped by a BaseLoggerWrapper child class previously.

Parameters:

• env (Env) – Original Sinergym environment.

• entity (Optional[str]) – The entity to which the project belongs. If you are using a previous wandb run, it is not necessary to specify it. Defaults to None.

• project_name (Optional[str]) – The project name. If you are using a previous wandb run, it is not necessary to specify it. Defaults to None.

• run_name (Optional[str]) – The name of the run. Defaults to None (Sinergym env name + wandb unique identifier).

• group (Optional[str]) – The name of the group to which the run belongs. Defaults to None.

• job_type (Optional[str]) – The type of job. Defaults to None.

• tags (Optional[List[str]]) – List of tags for the run. Defaults to None.

• episode_percentage (float) – Percentage of episode which must be completed to log episode summary. Defaults to 0.9.

• save_code (bool) – Whether to save the code in the run. Defaults to False.

• dump_frequency (int) – Frequency to dump log in platform. Defaults to 1000.

• artifact_save (bool) – Whether to save artifacts in WandB. Defaults to True.

• artifact_type (str) – Type of artifact to save. Defaults to ‘output’.

• excluded_info_keys (List[str]) – List of keys to exclude from info dictionary. Defaults to [‘reward’, ‘action’, ‘timestep’, ‘month’, ‘day’, ‘hour’, ‘time_elapsed(hours)’, ‘reward_weight’, ‘is_raining’].

• excluded_episode_summary_keys (List[str]) – List of keys to exclude from episode summary. Defaults to [‘terminated’, ‘truncated’].

Methods

__init__(env[, entity, project_name, ...])	Wrapper to log data in WandB platform.
class_name()	Returns the class name of the wrapper.
close()	Recording last episode summary and close env.
get_wrapper_attr(name)	Gets an attribute from the wrapper and lower environments if name doesn't exist in this object.
has_wrapper_attr(name)	Checks if the given attribute is within the wrapper or its environment.
render()	Uses the render() of the env that can be overwritten to change the returned data.
reset([seed, options])	Reset the environment.
save_artifact()	Save sinergym output as artifact in WandB platform.
set_wandb_finish(wandb_finish)	Set if WandB run must be finished when environment is closed.
set_wrapper_attr(name, value)	Sets an attribute on this wrapper or lower environment if name is already defined.
step(action)	Sends action to the environment.
wandb_log()	Log last step information in WandB platform.
wandb_log_summary()	Log episode summary in WandB platform.
wrapper_spec(**kwargs)	Generates a WrapperSpec for the wrappers.

Attributes

action_space	Return the Env action_space unless overwritten then the wrapper action_space is used.
logger
metadata	Returns the Env metadata.
np_random	Returns the Env np_random attribute.
np_random_seed	Returns the base environment's np_random_seed.
observation_space	Return the Env observation_space unless overwritten then the wrapper observation_space is used.
render_mode	Returns the Env render_mode.
spec	Returns the Env spec attribute with the WrapperSpec if the wrapper inherits from EzPickle.
unwrapped	Returns the base environment of the wrapper.

close() → None

Recording last episode summary and close env.

Parameters:

wandb_finish (bool) – Whether to finish WandB run. Defaults to True.

logger = <Logger WRAPPER WandBLogger (INFO)>

reset(seed: int | None = None, options: Dict[str, Any] | None = None) → Tuple[ndarray, Dict[str, Any]]

Reset the environment. Recording episode summary in WandB platform if it is not the first episode.

Parameters:

• seed (Optional[int]) – The seed that is used to initialize the environment’s episode (np_random). if value is None, a seed will be chosen from some source of entropy. Defaults to None.

• options (Optional[Dict[str, Any]]) – Additional information to specify how the environment is reset. Defaults to None.

Returns:

Current observation and info context with additional information.

Return type:

Tuple[np.ndarray,Dict[str,Any]]

save_artifact() → None

Save sinergym output as artifact in WandB platform.

set_wandb_finish(wandb_finish: bool) → None

Set if WandB run must be finished when environment is closed.

Parameters:

wandb_finish (bool) – Whether to finish WandB run.

step(action: int | ndarray) → Tuple[ndarray, float, bool, bool, Dict[str, Any]]

Sends action to the environment. Logging new interaction information in WandB platform.

Parameters:

action (Union[int, float, np.integer, np.ndarray, List[Any], Tuple[Any]]) – Action selected by the agent.

Returns:

Observation for next timestep, reward obtained, Whether the episode has ended or not, Whether episode has been truncated or not, and a dictionary with extra information

Return type:

Tuple[np.ndarray, float, bool, Dict[str, Any]]

wandb_log() → None

Log last step information in WandB platform.

wandb_log_summary() → None

Log episode summary in WandB platform.