Environment

class env_wrapper.chickai_env_wrapper.ChickAIEnvWrapper(run_id: str, env_path=None, base_port=5004, **kwargs)[source]

Wrapper class for the ChickAI environment.

Parameters:

run_id (str) – The ID of the current run.
env_path (str) – The path to the Unity environment executable.
base_port (int) – The base port number for the Unity environment.
**kwargs – Additional keyword arguments.

env

The wrapped Unity environment.

Type:: UnityToGymWrapper

mode

The mode of the environment.

Type:: str

step()[source]: Step the environment for one timestep.

log()[source]: Write to the log file.

close()[source]: Close the environment.

reset()[source]: Reset the environment.

steps_from_eps()[source]: Get the number of steps from the given episode.

close()[source]

Close the environment.

Returns:: None

log(msg: str) → None[source]

Write a message to the log file.

Parameters:: msg (str) – The message to write.
Returns:: None

reset(seed: int | None = None, **kwargs)[source]

Reset the environment.

Parameters:

seed (int) – The random seed for the environment.
**kwargs – Additional keyword arguments.

Returns:

The initial state of the environment.

step(action)[source]

Step the environment for one timestep.

Parameters:: action – The action to take in the environment.
Returns:: A tuple containing the next state, reward, done flag, and additional information.
Return type:: tuple

abstract steps_from_eps(eps)[source]

Get the number of steps from the given episode.

Parameters:: eps – The episode.
Returns:: The number of steps.
Return type:: int

class env_wrapper.parsing_env_wrapper.ParsingEnv(run_id: str, env_path=None, base_port=5004, **kwargs)[source]

A class representing the Parsing Environment.

numb_conditions

The number of test conditions in the environment.

Type:: int

__init__(self, run_id: str, env_path=None, base_port=5004, **kwargs): Initializes the ParsingEnv object.

steps_from_eps(self, eps)[source]: Calculates the number of steps based on the number of episodes.

steps_from_eps(eps)[source]

Calculates the number of steps based on the number of episodes.

Parameters:: eps (int) – The number of episodes.
Returns:: The total number of steps.
Return type:: int

class env_wrapper.dvs_wrapper.DVSWrapper(env, change_threshold=60, kernel_size=(3, 3), sigma=1)[source]

A gym observation wrapper that performs Dynamic Vision Sensor (DVS) transformation on the environment observations.

Parameters:

env (gym.Env) – The environment to wrap.
change_threshold (int) – The threshold value for detecting changes in pixel intensity.
kernel_size (tuple) – The size of the Gaussian kernel used for blurring.
sigma (float) – The standard deviation of the Gaussian kernel.

change_threshold

The threshold value for detecting changes in pixel intensity.

Type:: int

kernel_size

The size of the Gaussian kernel used for blurring.

Type:: tuple

sigma

The standard deviation of the Gaussian kernel.

Type:: float

num_stack

The number of frames to stack.

Type:: int

env

The wrapped environment.

Type:: gym.Env

stack

A deque to store the stacked frames.

Type:: collections.deque

shape

The shape of the observation space.

Type:: tuple

observation_space

The modified observation space.

Type:: gym.spaces.Box

create_grayscale(image)[source]: Converts an image to grayscale.

gaussianDiff(previous, current)[source]: Computes the difference between two images using Gaussian blur.

observation(obs)[source]: Performs the DVS transformation on the observation.

threshold(change)[source]: Applies a threshold to the change map.

reset(**kwargs)[source]: Resets the environment and returns the initial observation.

create_grayscale(image)[source]

Converts an image to grayscale.

Parameters:: image (numpy.ndarray) – The input image.
Returns:: The grayscale image.
Return type:: numpy.ndarray

gaussianDiff(previous, current)[source]

Computes the difference between two images using Gaussian blur.

Parameters:

previous (numpy.ndarray) – The previous image.
current (numpy.ndarray) – The current image.

Returns:

The difference map.

Return type:

numpy.ndarray

observation(obs)[source]

Performs the DVS transformation on the observation.

Parameters:: obs (list) – The list of stacked frames.
Returns:: The transformed observation.
Return type:: numpy.ndarray

reset(**kwargs)[source]

Resets the environment and returns the initial observation.

Parameters:: **kwargs – Additional keyword arguments for resetting the environment.
Returns:: The initial observation.
Return type:: numpy.ndarray

threshold(change)[source]

Applies a threshold to the change map.

Parameters:: change (numpy.ndarray) – The change map.
Returns:: The thresholded change map.
Return type:: numpy.ndarray

class env_wrapper.observation_wrapper.ObservationWrapper(env)[source]: Gym env wrapper that transpose visual observations to (C,H,W).