remove openai branding, update gym terminology with gymnasium

docs/source/examples/index.rst

@@ -11,4 +11,4 @@ This page lists detailled examples demonstrating how to use this package. They a
     quickstart
     using_a_custom_teambuilder
     connecting_to_showdown_and_challenging_humans
-    rl_with_open_ai_gym_wrapper
+    rl_with_gymnasium_wrapper

docs/source/examples/rl_with_open_ai_gym_wrapper.rst → docs/source/examples/rl_with_gymnasium_wrapper.rst

@@ -1,18 +1,18 @@
-.. _rl_with_open_ai_gym_wrapper:
+.. _rl_with_gymnasium_wrapper:
 
-Reinforcement learning with the OpenAI Gym wrapper
+Reinforcement learning with the Gymnasium wrapper
 ==================================================
 
-The corresponding complete source code can be found `here <https://github.com/hsahovic/poke-env/blob/master/examples/rl_with_new_open_ai_gym_wrapper.py>`__.
+The corresponding complete source code can be found `here <https://github.com/hsahovic/poke-env/blob/master/examples/rl_with_new_gymnasium_wrapper.py>`__.
 
-The goal of this example is to demonstrate how to use the `open ai gym <https://gym.openai.com/>`__ interface proposed by ``EnvPlayer``, and to train a simple deep reinforcement learning agent comparable in performance to the ``MaxDamagePlayer`` we created in :ref:`max_damage_player`.
+The goal of this example is to demonstrate how to use the `farama gymnasium <https://gymnasium.farama.org/>`__ interface proposed by ``EnvPlayer``, and to train a simple deep reinforcement learning agent comparable in performance to the ``MaxDamagePlayer`` we created in :ref:`max_damage_player`.
 
-.. note:: This example necessitates `keras-rl <https://github.com/keras-rl/keras-rl>`__ (compatible with Tensorflow 1.X) or `keras-rl2 <https://github.com/wau/keras-rl2>`__ (Tensorflow 2.X), which implement numerous reinforcement learning algorithms and offer a simple API fully compatible with the Open AI Gym API. You can install them by running ``pip install keras-rl`` or ``pip install keras-rl2``. If you are unsure, ``pip install keras-rl2`` is recommended.
+.. note:: This example necessitates `keras-rl <https://github.com/keras-rl/keras-rl>`__ (compatible with Tensorflow 1.X) or `keras-rl2 <https://github.com/wau/keras-rl2>`__ (Tensorflow 2.X), which implement numerous reinforcement learning algorithms and offer a simple API fully compatible with the Gymnasium API. You can install them by running ``pip install keras-rl`` or ``pip install keras-rl2``. If you are unsure, ``pip install keras-rl2`` is recommended.
 
 Implementing rewards and observations
 *************************************
 
-The open ai gym API provides *rewards* and *observations* for each step of each episode. In our case, each step corresponds to one decision in a battle and battles correspond to episodes.
+The Gymnasium API provides *rewards* and *observations* for each step of each episode. In our case, each step corresponds to one decision in a battle and battles correspond to episodes.
 
 Defining observations
 ^^^^^^^^^^^^^^^^^^^^^
@@ -26,9 +26,9 @@ Observations are embeddings of the current state of the battle. They can be an a
 
 To define our observations, we will create a custom ``embed_battle`` method. It takes one argument, a ``Battle`` object, and returns our embedding.
 
-In addition to this, we also need to describe the embedding to the gym interface.
+In addition to this, we also need to describe the embedding to the gymnasium interface.
 To achieve this, we need to implement the ``describe_embedding`` method where we specify the low bound and the high bound
-for each component of the embedding vector and return them as a ``gym.Space`` object.
+for each component of the embedding vector and return them as a ``gymnasium.Space`` object.
 
 Defining rewards
 ^^^^^^^^^^^^^^^^
@@ -108,7 +108,7 @@ Our player will play the ``gen8randombattle`` format. We can therefore inherit f
 Instantiating and testing a player
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
-Now that our custom class is defined, we can instantiate our RL player and test if it's compliant with the OpenAI gym API.
+Now that our custom class is defined, we can instantiate our RL player and test if it's compliant with the Gymnasium API.
 
 .. code-block:: python
 
@@ -340,7 +340,7 @@ To use the ``cross_evaluate`` method, the strategy is the same to the one used f
 Final result
 ************
 
-Running the `whole file <https://github.com/hsahovic/poke-env/blob/master/examples/rl_with_new_open_ai_gym_wrapper.py>`__ should take a couple of minutes and print something similar to this:
+Running the `whole file <https://github.com/hsahovic/poke-env/blob/master/examples/rl_with_gymnasium_wrapper.py>`__ should take a couple of minutes and print something similar to this:
 
 .. code-block:: console
 

docs/source/getting_started.rst

@@ -41,7 +41,7 @@ Agents in ``poke-env`` are instances of the ``Player`` class. Explore the follow
 
 - Basic agent: :ref:`/examples/cross_evaluate_random_players.ipynb`
 - Advanced agent: :ref:`max_damage_player`
-- RL agent: :ref:`rl_with_open_ai_gym_wrapper`
+- RL agent: :ref:`rl_with_gymnasium_wrapper`
 - Using teams: :ref:`ou_max_player`
 - Custom team builder: :ref:`using_a_custom_teambuilder`
 

docs/source/index.rst

@@ -6,7 +6,7 @@ Poke-env: A Python Interface for Training Reinforcement Learning Pokémon Bots
 
 Poke-env provides an environment for engaging in `Pokémon Showdown <https://pokemonshowdown.com/>`__ battles with a focus on reinforcement learning. 
 
-It boasts a straightforward API for handling Pokémon, Battles, Moves, and other battle-centric objects, alongside an `OpenAI Gym <https://gym.openai.com/>`__ interface for training agents.
+It boasts a straightforward API for handling Pokémon, Battles, Moves, and other battle-centric objects, alongside a `Farama Gymnasium <https://gymnasium.farama.org/>`__ interface for training agents.
 
 .. attention:: While poke-env aims to support all Pokémon generations, it was primarily developed with the latest generations in mind. If you discover any missing or incorrect functionalities for earlier generations, please `open an issue <https://github.com/hsahovic/poke-env/issues>`__ to help improve the library.
 

docs/source/modules/player.rst

@@ -21,10 +21,10 @@ Player
    :undoc-members:
    :show-inheritance:
 
-OpenAIGymEnv
+GymnasiumEnv
 ************
 
-.. automodule:: poke_env.player.openai_api
+.. automodule:: poke_env.player.gymnasium_api
    :members:
    :undoc-members:
    :show-inheritance:

examples/openai_example.py → examples/gymnasium_example.py

@@ -7,13 +7,13 @@
 from poke_env.environment.abstract_battle import AbstractBattle
 from poke_env.player import (
     Gen8EnvSinglePlayer,
+    GymnasiumEnv,
     ObservationType,
-    OpenAIGymEnv,
     RandomPlayer,
 )
 
 
-class TestEnv(OpenAIGymEnv):
+class TestEnv(GymnasiumEnv):
     def __init__(self, **kwargs):
         self.opponent = RandomPlayer(
             battle_format="gen8randombattle",
@@ -66,31 +66,31 @@ def describe_embedding(self) -> Space:
         return Box(np.array([0, 0]), np.array([6, 6]), dtype=int)
 
 
-def openai_api():
-    gym_env = TestEnv(
+def gymnasium_api():
+    gymnasium_env = TestEnv(
         battle_format="gen8randombattle",
         server_configuration=LocalhostServerConfiguration,
         start_challenging=True,
     )
-    check_env(gym_env)
-    gym_env.close()
+    check_env(gymnasium_env)
+    gymnasium_env.close()
 
 
 def env_player():
     opponent = RandomPlayer(
         battle_format="gen8randombattle",
         server_configuration=LocalhostServerConfiguration,
     )
-    gym_env = Gen8(
+    gymnasium_env = Gen8(
         battle_format="gen8randombattle",
         server_configuration=LocalhostServerConfiguration,
         start_challenging=True,
         opponent=opponent,
     )
-    check_env(gym_env)
-    gym_env.close()
+    check_env(gymnasium_env)
+    gymnasium_env.close()
 
 
 if __name__ == "__main__":
-    openai_api()
+    gymnasium_api()
     env_player()

examples/rl_with_new_open_ai_gym_wrapper.py → examples/rl_with_gymnasium_wrapper.py

@@ -72,7 +72,7 @@ def describe_embedding(self) -> Space:
 
 async def main():
     # First test the environment to ensure the class is consistent
-    # with the OpenAI API
+    # with the Gymnasium API
     opponent = RandomPlayer(battle_format="gen8randombattle")
     test_env = SimpleRLPlayer(
         battle_format="gen8randombattle", start_challenging=True, opponent=opponent

integration_tests/test_env_player.py

@@ -90,7 +90,7 @@ def play_function(player, n_battles):
 
 
 @pytest.mark.timeout(30)
-def test_random_gym_players_gen4():
+def test_random_gymnasium_players_gen4():
     random_player = RandomPlayer(battle_format="gen4randombattle", log_level=25)
     env_player = RandomGen4EnvPlayer(
         log_level=25, opponent=random_player, start_challenging=False
@@ -100,7 +100,7 @@ def test_random_gym_players_gen4():
 
 
 @pytest.mark.timeout(30)
-def test_random_gym_players_gen5():
+def test_random_gymnasium_players_gen5():
     random_player = RandomPlayer(battle_format="gen5randombattle", log_level=25)
     env_player = RandomGen5EnvPlayer(
         log_level=25, opponent=random_player, start_challenging=False
@@ -110,7 +110,7 @@ def test_random_gym_players_gen5():
 
 
 @pytest.mark.timeout(30)
-def test_random_gym_players_gen6():
+def test_random_gymnasium_players_gen6():
     random_player = RandomPlayer(battle_format="gen6randombattle", log_level=25)
     env_player = RandomGen6EnvPlayer(
         log_level=25, opponent=random_player, start_challenging=False
@@ -120,7 +120,7 @@ def test_random_gym_players_gen6():
 
 
 @pytest.mark.timeout(30)
-def test_random_gym_players_gen7():
+def test_random_gymnasium_players_gen7():
     random_player = RandomPlayer(battle_format="gen7randombattle", log_level=25)
     env_player = RandomGen7EnvPlayer(
         log_level=25, opponent=random_player, start_challenging=False
@@ -130,7 +130,7 @@ def test_random_gym_players_gen7():
 
 
 @pytest.mark.timeout(30)
-def test_random_gym_players_gen8():
+def test_random_gymnasium_players_gen8():
     random_player = RandomPlayer(battle_format="gen8randombattle", log_level=25)
     env_player = RandomGen8EnvPlayer(
         log_level=25, opponent=random_player, start_challenging=False
@@ -140,7 +140,7 @@ def test_random_gym_players_gen8():
 
 
 @pytest.mark.timeout(30)
-def test_random_gym_players_gen9():
+def test_random_gymnasium_players_gen9():
     random_player = RandomPlayer(battle_format="gen9randombattle", log_level=25)
     env_player = RandomGen9EnvPlayer(
         log_level=25, opponent=random_player, start_challenging=False

src/poke_env/player/__init__.py

@@ -2,7 +2,7 @@
 """
 
 from poke_env.concurrency import POKE_LOOP
-from poke_env.player import env_player, openai_api, player, random_player, utils
+from poke_env.player import env_player, gymnasium_api, player, random_player, utils
 from poke_env.player.baselines import MaxBasePowerPlayer, SimpleHeuristicsPlayer
 from poke_env.player.battle_order import (
     BattleOrder,
@@ -19,7 +19,7 @@
     Gen8EnvSinglePlayer,
     Gen9EnvSinglePlayer,
 )
-from poke_env.player.openai_api import ActType, ObsType, OpenAIGymEnv
+from poke_env.player.gymnasium_api import ActType, GymnasiumEnv, ObsType
 from poke_env.player.player import Player
 from poke_env.player.random_player import RandomPlayer
 from poke_env.player.utils import (
@@ -32,7 +32,7 @@
 
 __all__ = [
     "env_player",
-    "openai_api",
+    "gymnasium_api",
     "player",
     "random_player",
     "utils",
@@ -47,7 +47,7 @@
     "Gen8EnvSinglePlayer",
     "Gen9EnvSinglePlayer",
     "POKE_LOOP",
-    "OpenAIGymEnv",
+    "GymnasiumEnv",
     "PSClient",
     "Player",
     "RandomPlayer",

src/poke_env/player/env_player.py

@@ -1,4 +1,4 @@
-"""This module defines a player class exposing the Open AI Gym API with utility functions.
+"""This module defines a player class exposing the Gymnasium API with utility functions.
 """
 
 from abc import ABC
@@ -8,15 +8,15 @@
 
 from poke_env.environment.abstract_battle import AbstractBattle
 from poke_env.player.battle_order import BattleOrder, ForfeitBattleOrder
-from poke_env.player.openai_api import ActType, ObsType, OpenAIGymEnv
+from poke_env.player.gymnasium_api import ActType, GymnasiumEnv, ObsType
 from poke_env.player.player import Player
 from poke_env.ps_client.account_configuration import AccountConfiguration
 from poke_env.ps_client.server_configuration import ServerConfiguration
 from poke_env.teambuilder.teambuilder import Teambuilder
 
 
-class EnvPlayer(OpenAIGymEnv[ObsType, ActType], ABC):
-    """Player exposing the Open AI Gym Env API."""
+class EnvPlayer(GymnasiumEnv[ObsType, ActType], ABC):
+    """Player exposing the Gymnasium Env API."""
 
     _ACTION_SPACE: List[int] = []
     _DEFAULT_BATTLE_FORMAT = "gen8randombattle"

src/poke_env/player/openai_api.py → src/poke_env/player/gymnasium_api.py

@@ -1,4 +1,4 @@
-"""This module defines a player class with the OpenAI API on the main thread.
+"""This module defines a player class with the Gymnasium API on the main thread.
 For a black-box implementation consider using the module env_player.
 """
 
@@ -62,7 +62,7 @@ class _AsyncPlayer(Generic[ObsType, ActType], Player):
 
     def __init__(
         self,
-        user_funcs: OpenAIGymEnv[ObsType, ActType],
+        user_funcs: GymnasiumEnv[ObsType, ActType],
         username: str,
         **kwargs: Any,
     ):
@@ -94,12 +94,12 @@ def _battle_finished_callback(self, battle: AbstractBattle):
         asyncio.run_coroutine_threadsafe(self.observations.async_put(to_put), POKE_LOOP)
 
 
-class OpenAIGymEnv(
+class GymnasiumEnv(
     Env[ObsType, ActType],
     ABC,
 ):
     """
-    Base class implementing the OpenAI Gym API on the main thread.
+    Base class implementing the Gymnasium API on the main thread.
     """
 
     _INIT_RETRIES = 100
@@ -239,7 +239,7 @@ def action_to_move(self, action: int, battle: AbstractBattle) -> BattleOrder:
     def embed_battle(self, battle: AbstractBattle) -> ObsType:
         """
         Returns the embedding of the current battle state in a format compatible with
-        the OpenAI gym API.
+        the Gymnasium API.
 
         :param battle: The current battle state.
         :type battle: AbstractBattle
@@ -416,7 +416,7 @@ def close(self, purge: bool = True):
     def background_send_challenge(self, username: str):
         """
         Sends a single challenge specified player. The function immediately returns
-        to allow use of the OpenAI gym API.
+        to allow use of the Gymnasium API.
 
         :param username: The username of the player to challenge.
         :type username: str
@@ -434,7 +434,7 @@ def background_send_challenge(self, username: str):
     def background_accept_challenge(self, username: str):
         """
         Accepts a single challenge specified player. The function immediately returns
-        to allow use of the OpenAI gym API.
+        to allow use of the Gymnasium API.
 
         :param username: The username of the player to challenge.
         :type username: str

unit_tests/player/test_env_player.py

@@ -19,7 +19,7 @@
     Gen9EnvSinglePlayer,
     RandomPlayer,
 )
-from poke_env.player.openai_api import _AsyncPlayer
+from poke_env.player.gymnasium_api import _AsyncPlayer
 
 account_configuration = AccountConfiguration("username", "password")
 server_configuration = ServerConfiguration("server.url", "auth.url")
@@ -42,15 +42,15 @@ def embed_battle(self, battle):
 
 
 def test_init():
-    gym_env = CustomEnvPlayer(
+    gymnasium_env = CustomEnvPlayer(
         None,
         account_configuration=account_configuration,
         server_configuration=server_configuration,
         start_listening=False,
         battle_format="gen7randombattles",
     )
-    player = gym_env.agent
-    assert isinstance(gym_env, CustomEnvPlayer)
+    player = gymnasium_env.agent
+    assert isinstance(gymnasium_env, CustomEnvPlayer)
     assert isinstance(player, _AsyncPlayer)
 
 
@@ -60,11 +60,11 @@ async def __call__(self, *args, **kwargs):
 
 
 @patch(
-    "poke_env.player.openai_api._AsyncQueue.async_get",
+    "poke_env.player.gymnasium_api._AsyncQueue.async_get",
     return_value=2,
     new_callable=AsyncMock,
 )
-@patch("poke_env.player.openai_api._AsyncQueue.async_put", new_callable=AsyncMock)
+@patch("poke_env.player.gymnasium_api._AsyncQueue.async_put", new_callable=AsyncMock)
 def test_choose_move(queue_put_mock, queue_get_mock):
     player = CustomEnvPlayer(
         None,

unit_tests/player/test_openai.py → unit_tests/player/test_gymnasium.py

@@ -10,14 +10,14 @@
     ActType,
     BattleOrder,
     ForfeitBattleOrder,
+    GymnasiumEnv,
     ObsType,
-    OpenAIGymEnv,
     Player,
 )
-from poke_env.player.openai_api import _AsyncPlayer, _AsyncQueue
+from poke_env.player.gymnasium_api import _AsyncPlayer, _AsyncQueue
 
 
-class DummyEnv(OpenAIGymEnv[ObsType, ActType]):
+class DummyEnv(GymnasiumEnv[ObsType, ActType]):
     def __init__(self, *args, **kwargs):
         self.opponent = None
         super().__init__(*args, **kwargs)