rldurham.Recorder

class rldurham.Recorder(env, info=True, video=False, logs=False, ignore_existing=False, key='recorder', video_folder='videos', video_prefix='xxxx00-agent-video', full_stats=False, smoothing=None)

Bases: Wrapper, RecordConstructorArgs

Public Data Attributes:

Inherited from Wrapper

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

Inherited from Env

metadata
render_mode
spec
unwrapped	Returns the base non-wrapped environment.
np_random_seed	Returns the environment's internal _np_random_seed that if not set will first initialise with a random int as seed.
np_random	Returns the environment's internal _np_random that if not set will initialise with a random seed.
action_space
observation_space

Public Methods:

__init__(env[, info, video, logs, ...])	Wrapper to record statistics, videos, and logs.
add_stats(info[, ignore_existing])
step(action)	Uses the step() of the env that can be overwritten to change the returned data.
reset(*[, seed, options])	Uses the reset() of the env that can be overwritten to change the returned data.
close()	Closes the wrapper and env.
write_log([folder, file])
__getattr__(item)

Inherited from Wrapper

__init__(env)	Wraps an environment to allow a modular transformation of the step() and reset() methods.
step(action)	Uses the step() of the env that can be overwritten to change the returned data.
reset(*[, seed, options])	Uses the reset() of the env that can be overwritten to change the returned data.
render()	Uses the render() of the env that can be overwritten to change the returned data.
close()	Closes the wrapper and env.
wrapper_spec(**kwargs)	Generates a WrapperSpec for the wrappers.
has_wrapper_attr(name)	Checks if the given attribute is within the wrapper or its environment.
get_wrapper_attr(name)	Gets an attribute from the wrapper and lower environments if name doesn't exist in this object.
set_wrapper_attr(name, value, *[, force])	Sets an attribute on this wrapper or lower environment if name is already defined.
__str__()	Returns the wrapper name and the env representation string.
__repr__()	Returns the string representation of the wrapper.
class_name()	Returns the class name of the wrapper.

Inherited from Env

step(action)	Run one timestep of the environment's dynamics using the agent actions.
reset(*[, seed, options])	Resets the environment to an initial internal state, returning an initial observation and info.
render()	Compute the render frames as specified by render_mode during the initialization of the environment.
close()	After the user has finished using the environment, close contains the code necessary to "clean up" the environment.
__str__()	Returns a string of the environment with spec id's if :attr:`spec.
__enter__()	Support with-statement for the environment.
__exit__(*args)	Support with-statement for the environment and closes the environment.
has_wrapper_attr(name)	Checks if the attribute name exists in the environment.
get_wrapper_attr(name)	Gets the attribute name from the environment.
set_wrapper_attr(name, value, *[, force])	Sets the attribute name on the environment with value, see Wrapper.set_wrapper_attr for more info.

Inherited from Generic

__class_getitem__	Parameterizes a generic class.
__init_subclass__	Function to initialize subclasses.

Inherited from RecordConstructorArgs

__init__(*[, _disable_deepcopy])

Records all arguments passed to constructor to _saved_kwargs.

Private Data Attributes:

Inherited from Wrapper

_np_random	This code will never be run due to __getattr__ being called prior this.
_action_space
_observation_space
_metadata
_cached_spec

Inherited from Env

_np_random
_np_random_seed

Inherited from RecordConstructorArgs

_saved_kwargs

Private Methods:

_video_episode_trigger(episode_id)
_video_name_func()
_finish_episode()

---

__annotations__ = {}

classmethod __class_getitem__()

Parameterizes a generic class.

At least, parameterizing a generic class is the main thing this method does. For example, for some generic class Foo, this is called when we do Foo[int] - there, with cls=Foo and params=int.

However, note that this method is also called when defining generic classes in the first place with class Foo[T]: ….

__dict__ = mappingproxy({'__module__': 'rldurham', '__init__': <function Recorder.__init__>, '_video_episode_trigger': <function Recorder._video_episode_trigger>, '_video_name_func': <function Recorder._video_name_func>, 'add_stats': <function Recorder.add_stats>, 'step': <function Recorder.step>, '_finish_episode': <function Recorder._finish_episode>, 'reset': <function Recorder.reset>, 'close': <function Recorder.close>, 'write_log': <function Recorder.write_log>, '__doc__': None, '__parameters__': (), '__getattr__': <function transparent_wrapper.<locals>.__getattr__>, '__annotations__': {}, '__slotnames__': []})

__enter__()

Support with-statement for the environment.

__exit__(*args)

Support with-statement for the environment and closes the environment.

__getattr__(item)

__init__(env, info=True, video=False, logs=False, ignore_existing=False, key='recorder', video_folder='videos', video_prefix='xxxx00-agent-video', full_stats=False, smoothing=None)

Wrapper to record statistics, videos, and logs.

The statistics added to info are

“idx”: index of this episode “length”: length of this episode “r_sum”: sum of rewards of this episode “r_mean”: mean reward of this episode “r_std”: standard deviation of rewards of this episode

If smoothings is not None:

“length_”: smoothed length of episodes (i.e. smoothed version of “length”) “r_sum_”: sum (over smoothing window) of sum (over single episodes) of rewards “r_mean_”: average (over smoothing window) of sum (over single episodes) of rewards (i.e. smoothed

version of “r_sum”)

“r_std_”: standard deviation of sums over rewards (i.e. the std associated with “r_mean_”)

if full_stats:

“full”: list of [(obs, reward, terminated, truncated, info), …] for episode

Parameters:

• env – environment

• info – populate info dict with statistics at each step

• video – wrap env with VideoRecorder and activate recording (can be deactivated by setting video = False on the Recorder object)

• logs – keep logs of episode statistics (these are different stats from the ones in info and useful for evaluating training performance over episodes)

• ignore_existing – silently skip adding stats in info if an entry with corresponding key already exists (otherwise raises a RuntimeError)

• key – key used to store stats in info dict

• video_folder – folder for storing videos

• video_prefix – prefix for video files

• full_stats – add full stats of (obs, reward, terminated, truncated, info) for each step of the episode

• smoothing – add stats smoothed over this many time steps

classmethod __init_subclass__()

Function to initialize subclasses.

__module__ = 'rldurham'

__orig_bases__ = (gymnasium.core.Env[~WrapperObsType, ~WrapperActType], typing.Generic[~WrapperObsType, ~WrapperActType, ~ObsType, ~ActType])

__parameters__ = ()

__repr__()

Returns the string representation of the wrapper.

__slotnames__ = []

__str__()

Returns the wrapper name and the env representation string.

__weakref__

list of weak references to the object

_action_space: Space[WrapperActType] | None

_cached_spec: EnvSpec | None

_finish_episode()

_metadata: dict[str, Any] | None

property _np_random

This code will never be run due to __getattr__ being called prior this.

It seems that @property overwrites the variable (_np_random) meaning that __getattr__ gets called with the missing variable.

_np_random_seed: int | None = None

_observation_space: Space[WrapperObsType] | None

_video_episode_trigger(episode_id)

_video_name_func()

property action_space: Space[ActType] | Space[WrapperActType]

Return the Env action_space unless overwritten then the wrapper action_space is used.

add_stats(info, ignore_existing=None)

classmethod class_name()

Returns the class name of the wrapper.

Return type:

str

close()

Closes the wrapper and env.

get_wrapper_attr(name)

Gets an attribute from the wrapper and lower environments if name doesn’t exist in this object.

Return type:

Any

Args:

name: The variable name to get

Returns:

The variable with name in wrapper or lower environments

has_wrapper_attr(name)

Checks if the given attribute is within the wrapper or its environment.

Return type:

bool

property metadata: dict[str, Any]

Returns the Env metadata.

property np_random: Generator

Returns the Env np_random attribute.

property np_random_seed: int | None

Returns the base environment’s np_random_seed.

property observation_space: Space[ObsType] | Space[WrapperObsType]

Return the Env observation_space unless overwritten then the wrapper observation_space is used.

render()

Uses the render() of the env that can be overwritten to change the returned data.

Return type:

Union[TypeVar(RenderFrame), list[TypeVar(RenderFrame)], None]

property render_mode: str | None

Returns the Env render_mode.

reset(*, seed=None, options=None)

Uses the reset() of the env that can be overwritten to change the returned data.

set_wrapper_attr(name, value, *, force=True)

Sets an attribute on this wrapper or lower environment if name is already defined.

Return type:

bool

Args:

name: The variable name value: The new variable value force: Whether to create the attribute on this wrapper if it does not exists on the

lower environment instead of raising an exception

Returns:

If the variable has been set in this or a lower wrapper.

property spec: EnvSpec | None

Returns the Env spec attribute with the WrapperSpec if the wrapper inherits from EzPickle.

step(action)

Uses the step() of the env that can be overwritten to change the returned data.

property unwrapped: Env[ObsType, ActType]

Returns the base environment of the wrapper.

This will be the bare gymnasium.Env environment, underneath all layers of wrappers.

classmethod wrapper_spec(**kwargs)

Generates a WrapperSpec for the wrappers.

Return type:

WrapperSpec

write_log(folder='logs', file='xxxx00-agent-log.txt')