sinergym.utils.wrappers.LoggerWrapper

class sinergym.utils.wrappers.LoggerWrapper(*args: Any, **kwargs: Any)

__init__(env: ~typing.Any, logger_class: ~typing.Callable = <class 'sinergym.utils.logger.CSVLogger'>, monitor_header: ~typing.List[str] | None = None, progress_header: ~typing.List[str] | None = None, flag: bool = True)

CSVLogger to log interactions with environment.

Parameters:

• env (Any) – Original Gym environment.

• logger_class (CSVLogger) – CSV Logger class to use to log all information.

• monitor_header – Header for monitor.csv in each episode. Default is None (default format).

• progress_header – Header for progress.csv in whole simulation. Default is None (default format).

• flag (bool, optional) – State of logger (activate or deactivate). Defaults to True.

Methods

__init__(env[, logger_class, ...])	CSVLogger to log interactions with environment.
activate_logger()	Activate logger if its flag False.
close()	Recording last episode summary and close env.
deactivate_logger()	Deactivate logger if its flag True.
reset()	Resets the environment.
step(action)	Step the environment.

activate_logger() → None

Activate logger if its flag False.

close() → None

Recording last episode summary and close env.

deactivate_logger() → None

Deactivate logger if its flag True.

reset() → ndarray

Resets the environment. Recording episode summary in logger

Returns:

First observation given

Return type:

np.ndarray

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

Step the environment. Logging new information

Parameters:

action (Union[int, np.ndarray]) – Action executed in step

Returns:

Tuple with next observation, reward, bool for terminated episode and dict with extra information.

Return type:

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