Pyright versioning update. Fix #2700 (#2739)

* Moved pygame imports into render

* Formatting

* Make pygame optional for box2d, try to make formatting work

* fix tests, fix pre-commit.

* Update ci linter config.

* fix type hints for latest pyright version and backward compatibility with numpy <= 1.21.5

* pre-commit.

Co-authored-by: Ariel Kwiatkowski <ariel.j.kwiatkowski@gmail.com>
Co-authored-by: Gianluca De Cola <gianluca.decola@ags-it.com>

.github/workflows/lint_python.yml

@@ -8,13 +8,15 @@ jobs:
     strategy:
       matrix:
         python-platform: ["Linux"]
-        python-version: ["3.7"]
+        python-version: ["3.7", "3.8", "3.9", "3.10"]
       fail-fast: false
     env:
-      PYRIGHT_VERSION: 1.1.204
+      PYRIGHT_VERSION: 1.1.235
     steps:
       - uses: actions/checkout@v2
       - uses: actions/setup-python@v2
+        with:
+          python-version: ${{ matrix.python-version }}
       - run: pip install -e .[nomujoco]
       - uses: jakebailey/pyright-action@v1
         with:

gym/core.py

@@ -78,7 +78,7 @@ class Env(Generic[ObsType, ActType]):
             reward (float) : amount of reward returned after previous action
             done (bool): whether the episode has ended, in which case further :meth:`step` calls will return undefined results. A done signal may be emitted for different reasons: Maybe the task underlying the environment was solved successfully, a certain timelimit was exceeded, or the physics simulation has entered an invalid state. ``info`` may contain additional information regarding the reason for a ``done`` signal.
             info (dict): contains auxiliary diagnostic information (helpful for debugging, learning, and logging). This might, for instance, contain:
-                
+
                 - metrics that describe the agent's performance or
                 - state variables that are hidden from observations or
                 - information that distinguishes truncation and termination or
@@ -106,7 +106,8 @@ class Env(Generic[ObsType, ActType]):
         integer seed right after initialization and then never again.
 
         Args:
-            seed (int or None): The seed that is used to initialize the environment's PRNG. If the environment does not already have a PRNG and ``seed=None`` (the default option) is passed, a seed will be chosen from some source of entropy (e.g. timestamp or /dev/urandom). However, if the environment already has a PRNG and ``seed=None`` is pased, the PRNG will *not* be reset. If you pass an integer, the PRNG will be reset even if it already exists. Usually, you want to pass an integer *right after the environment has been initialized and then never again*. Please refer to the minimal example above to see this paradigm in action.
+            seed (int or None): The seed that is used to initialize the environment's PRNG. If the environment does not already have a PRNG and ``seed=None`` (the default option) is passed, a seed will be chosen from some source of entropy (e.g. timestamp or /dev/urandom). However, if the environment already has a PRNG and ``seed=None`` is passed, the PRNG will *not* be reset.
+                                If you pass an integer, the PRNG will be reset even if it already exists. Usually, you want to pass an integer *right after the environment has been initialized and then never again*. Please refer to the minimal example above to see this paradigm in action.
             return_info (bool): If true, return additional information along with initial observation. This info should be analogous to the info returned in :meth:`step`
             options (dict or None): Additional information to specify how the environment is reset (optional, depending on the specific environment)
 
@@ -135,7 +136,7 @@ class Env(Generic[ObsType, ActType]):
         - ansi: Return a string (str) or StringIO.StringIO containing a
           terminal-style text representation. The text can include newlines
           and ANSI escape sequences (e.g. for colors).
-                
+
         Note:
             Make sure that your class's metadata 'render_modes' key includes
               the list of supported modes. It's recommended to call super()

gym/envs/classic_control/continuous_mountain_car.py

@@ -125,7 +125,7 @@ class Continuous_MountainCarEnv(gym.Env):
             low=self.low_state, high=self.high_state, dtype=np.float32
         )
 
-    def step(self, action):
+    def step(self, action: np.ndarray):
 
         position = self.state[0]
         velocity = self.state[1]

gym/envs/classic_control/mountain_car.py

@@ -112,7 +112,7 @@ class MountainCarEnv(gym.Env):
         self.action_space = spaces.Discrete(3)
         self.observation_space = spaces.Box(self.low, self.high, dtype=np.float32)
 
-    def step(self, action):
+    def step(self, action: int):
         assert self.action_space.contains(
             action
         ), f"{action!r} ({type(action)}) invalid"

gym/spaces/box.py

@@ -29,12 +29,12 @@ class Box(Space[np.ndarray]):
     There are two common use cases:
 
     * Identical bound for each dimension::
-        
+
         >>> Box(low=-1.0, high=2.0, shape=(3, 4), dtype=np.float32)
         Box(3, 4)
 
     * Independent bound for each dimension::
-        
+
         >>> Box(low=np.array([-1.0, -2.0]), high=np.array([2.0, 4.0]), dtype=np.float32)
         Box(2,)
 
@@ -66,9 +66,9 @@ class Box(Space[np.ndarray]):
 
         # Capture the boundedness information before replacing np.inf with get_inf
         _low = np.full(shape, low, dtype=float) if np.isscalar(low) else low
-        self.bounded_below = -np.inf < _low
+        self.bounded_below = -np.inf < _low  # type: ignore
         _high = np.full(shape, high, dtype=float) if np.isscalar(high) else high
-        self.bounded_above = np.inf > _high
+        self.bounded_above = np.inf > _high  # type: ignore
 
         low = _broadcast(low, dtype, shape, inf_sign="-")  # type: ignore
         high = _broadcast(high, dtype, shape, inf_sign="+")  # type: ignore

gym/spaces/dict.py

@@ -18,7 +18,7 @@ class Dict(Space[TypingDict[str, Space]], Mapping):
         self.observation_space = spaces.Dict({"position": spaces.Discrete(2), "velocity": spaces.Discrete(3)})
 
     Example usage [nested]::
-    
+
         self.nested_observation_space = spaces.Dict({
             'sensors':  spaces.Dict({
                 'position': spaces.Box(low=-100, high=100, shape=(3,)),

gym/spaces/multi_binary.py

@@ -35,7 +35,8 @@ class MultiBinary(Space[np.ndarray]):
             self.n = n = int(n)
             input_n = (n,)
 
-        assert (np.asarray(input_n) > 0).all(), "n (counts) have to be positive"
+        # n (counts) have to be positive
+        assert (np.asarray(input_n) > 0).all()  # type: ignore
 
         super().__init__(input_n, np.int8, seed)
 

gym/spaces/multi_discrete.py

@@ -1,6 +1,7 @@
 from __future__ import annotations
 
 from collections.abc import Sequence
+from typing import Iterable
 
 import numpy as np
 
@@ -13,9 +14,9 @@ from .space import Space
 class MultiDiscrete(Space[np.ndarray]):
     """
     The multi-discrete action space consists of a series of discrete action spaces with different number of actions in each. It is useful to represent game controllers or keyboards where each key can be represented as a discrete action space. It is parametrized by passing an array of positive integers specifying number of actions for each discrete action space.
-    
+
-    Note: 
+    Note:
-    
+
         Some environment wrappers assume a value of 0 always represents the NOOP action.
 
     e.g. Nintendo Game Controller - Can be conceptualized as 3 discrete action spaces:
@@ -52,7 +53,7 @@ class MultiDiscrete(Space[np.ndarray]):
         # is within correct bounds for space dtype (even though x does not have to be unsigned)
         return bool(x.shape == self.shape and (0 <= x).all() and (x < self.nvec).all())
 
-    def to_jsonable(self, sample_n):
+    def to_jsonable(self, sample_n: Iterable[np.ndarray]):
         return [sample.tolist() for sample in sample_n]
 
     def from_jsonable(self, sample_n):
@@ -66,7 +67,7 @@ class MultiDiscrete(Space[np.ndarray]):
         if nvec.ndim == 0:
             subspace = Discrete(nvec)
         else:
-            subspace = MultiDiscrete(nvec, self.dtype)
+            subspace = MultiDiscrete(nvec, self.dtype)  # type: ignore
         subspace.np_random.bit_generator.state = self.np_random.bit_generator.state
         return subspace
 

gym/spaces/tuple.py

@@ -72,7 +72,7 @@ class Tuple(Space[tuple], Sequence):
     def __repr__(self) -> str:
         return "Tuple(" + ", ".join([str(s) for s in self.spaces]) + ")"
 
-    def to_jsonable(self, sample_n) -> list:
+    def to_jsonable(self, sample_n: Sequence) -> list:
         # serialize as list-repr of tuple of vectors
         return [
             space.to_jsonable([sample[i] for sample in sample_n])

gym/spaces/utils.py

@@ -19,7 +19,7 @@ def flatdim(space: Space) -> int:
     the space is not defined in ``gym.spaces``.
 
     Example usage::
-   
+
         >>> s = spaces.Dict({"position": spaces.Discrete(2), "velocity": spaces.Discrete(3)})
         >>> spaces.flatdim(s)
         5
@@ -134,7 +134,7 @@ def _unflatten_multidiscrete(space: MultiDiscrete, x: np.ndarray) -> np.ndarray:
     offsets[1:] = np.cumsum(space.nvec.flatten())
 
     (indices,) = np.nonzero(x)
-    return np.asarray(indices - offsets[:-1], dtype=space.dtype).reshape(space.shape)
+    return np.asarray(indices - offsets[:-1], dtype=space.dtype).reshape(space.shape)  # type: ignore
 
 
 @unflatten.register(Tuple)