Gymnasium/gym/wrappers/frame_stack.py

from collections import deque
from typing import Optional

import numpy as np

from gym import ObservationWrapper
from gym.spaces import Box


class LazyFrames:
    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.

    Args:
        lz4_compress (bool): use lz4 to compress the frames internally

    """
    __slots__ = ("frame_shape", "dtype", "shape", "lz4_compress", "_frames")

    def __init__(self, frames, lz4_compress=False):
        self.frame_shape = tuple(frames[0].shape)
        self.shape = (len(frames),) + self.frame_shape
        self.dtype = frames[0].dtype
        if lz4_compress:
            from lz4.block import compress

            frames = [compress(frame) for frame in frames]
        self._frames = frames
        self.lz4_compress = lz4_compress

    def __array__(self, dtype=None):
        arr = self[:]
        if dtype is not None:
            return arr.astype(dtype)
        return arr

    def __len__(self):
        return self.shape[0]

    def __getitem__(self, int_or_slice):
        if isinstance(int_or_slice, int):
            return self._check_decompress(self._frames[int_or_slice])  # single frame
        return np.stack(
            [self._check_decompress(f) for f in self._frames[int_or_slice]], axis=0
        )

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

    def _check_decompress(self, frame):
        if self.lz4_compress:
            from lz4.block import decompress

            return np.frombuffer(decompress(frame), dtype=self.dtype).reshape(
                self.frame_shape
            )
        return frame


class FrameStack(ObservationWrapper):
    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-v1', the original observation
    is an array with shape [3], so if we stack 4 observations, the processed observation
    has shape [4, 3].

    .. 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
        lz4_compress (bool): use lz4 to compress the frames internally

    """

    def __init__(self, env, num_stack, lz4_compress=False):
        super().__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 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.observation(), reward, done, info

    def reset(self, **kwargs):
        if kwargs.get("return_info", False):
            obs, info = self.env.reset(**kwargs)
        else:
            obs = self.env.reset(**kwargs)
            info = None  # Unused
        [self.frames.append(obs) for _ in range(self.num_stack)]

        if kwargs.get("return_info", False):
            return self.observation(), info
        else:
            return self.observation()