Gymnasium/gym/wrappers/frame_stack.py

from collections import deque
import numpy as np

from gym.spaces import Box
from gym import Wrapper


class LazyFrames(object):
    r"""Ensures common frames are only stored once to optimize memory use. 

    To further reduce the memory use, it is optionally to turn on lz4 to 
    compress the observations.

    .. note::

        This object should only be converted to numpy array just before forward pass. 

    """
    def __init__(self, frames, lz4_compress=False):
        if lz4_compress:
            from lz4.block import compress
            self.frame_shape = frames[0].shape
            self.dtype = frames[0].dtype
            frames = [compress(frame) for frame in frames]
        self._frames = frames
        self.lz4_compress = lz4_compress

    def __array__(self, dtype=None):
        if self.lz4_compress:
            from lz4.block import decompress
            frames = [np.frombuffer(decompress(frame), dtype=self.dtype).reshape(self.frame_shape) for frame in self._frames]
        else:
            frames = self._frames
        out = np.stack(frames, axis=0)
        if dtype is not None:
            out = out.astype(dtype)
        return out

    def __len__(self):
        return len(self.__array__())

    def __getitem__(self, i):
        return self.__array__()[i]

    def __eq__(self, other):
        return self.__array__() == other

    @property
    def shape(self):
        return self.__array__().shape


class FrameStack(Wrapper):
    r"""Observation wrapper that stacks the observations in a rolling manner. 

    For example, if the number of stacks is 4, then the returned observation contains
    the most recent 4 observations. For environment 'Pendulum-v0', the original observation
    is an array with shape [3], so if we stack 4 observations, the processed observation
    has shape [3, 4]. 

    .. note::

        To be memory efficient, the stacked observations are wrapped by :class:`LazyFrame`.

    .. note::

        The observation space must be `Box` type. If one uses `Dict`
        as observation space, it should apply `FlattenDictWrapper` at first. 

    Example::

        >>> import gym
        >>> env = gym.make('PongNoFrameskip-v0')
        >>> env = FrameStack(env, 4)
        >>> env.observation_space
        Box(4, 210, 160, 3)

    Args:
        env (Env): environment object
        num_stack (int): number of stacks

    """
    def __init__(self, env, num_stack, lz4_compress=False):
        super(FrameStack, self).__init__(env)
        self.num_stack = num_stack
        self.lz4_compress = lz4_compress

        self.frames = deque(maxlen=num_stack)

        low = np.repeat(self.observation_space.low[np.newaxis, ...], num_stack, axis=0)
        high = np.repeat(self.observation_space.high[np.newaxis, ...], num_stack, axis=0)
        self.observation_space = Box(low=low, high=high, dtype=self.observation_space.dtype)

    def _get_observation(self):
        assert len(self.frames) == self.num_stack, (len(self.frames), self.num_stack)
        return LazyFrames(list(self.frames), self.lz4_compress)

    def step(self, action):
        observation, reward, done, info = self.env.step(action)
        self.frames.append(observation)
        return self._get_observation(), reward, done, info

    def reset(self, **kwargs):
        observation = self.env.reset(**kwargs)
        [self.frames.append(observation) for _ in range(self.num_stack)]
        return self._get_observation()