add gym_pybullet_drones env

Author: huangshiyu13

examples/gym_pybullet_drones/test_env.py

@@ -17,47 +17,54 @@
 """"""
 import time
 
-import numpy as np
 import gym_pybullet_drones
 import gymnasium as gym
+import numpy as np
 
 from openrl.envs.common import make
+
+
 def test_env():
-    env = gym.make("hover-aviary-v0",gui=False,record=False)
-    print("obs space:",env.observation_space)
-    print("action space:",env.action_space)
+    env = gym.make("hover-aviary-v0", gui=False, record=False)
+    print("obs space:", env.observation_space)
+    print("action space:", env.action_space)
     obs, info = env.reset(seed=42, options={})
-    totoal_step  =0
-    totol_reward = 0.
+    totoal_step = 0
+    totol_reward = 0.0
     while True:
         obs, reward, done, truncated, info = env.step(env.action_space.sample())
-        totoal_step+=1
-        totol_reward+=reward
+        totoal_step += 1
+        totol_reward += reward
         # env.render()
         # time.sleep(1)
         if done:
             break
-    print("total step:",totoal_step)
-    print("total reward:",totol_reward)
+    print("total step:", totoal_step)
+    print("total reward:", totol_reward)
+
 
 def test_vec_env():
-    env = make("pybullet_drones/hover-aviary-v0",env_num=2,gui=False,record=False,asynchronous=True)
-    info,obs = env.reset(seed=0)
+    env = make(
+        "pybullet_drones/hover-aviary-v0",
+        env_num=2,
+        gui=False,
+        record=False,
+        asynchronous=True,
+    )
+    info, obs = env.reset(seed=0)
     totoal_step = 0
-    totol_reward = 0.
+    totol_reward = 0.0
     while True:
-        obs, reward, done, info =  env.step(env.random_action())
-        totoal_step+=1
-        totol_reward+=np.mean(reward)
-        if np.any(done) or totoal_step>100:
+        obs, reward, done, info = env.step(env.random_action())
+        totoal_step += 1
+        totol_reward += np.mean(reward)
+        if np.any(done) or totoal_step > 100:
             break
     env.close()
     print("total step:", totoal_step)
     print("total reward:", totol_reward)
 
 
-
-
-if __name__ == '__main__':
+if __name__ == "__main__":
     test_env()
-    # test_vec_env()
+    # test_vec_env()

examples/gym_pybullet_drones/train_ppo.py

@@ -56,7 +56,6 @@ def evaluation():
         record=False,
     )
 
-
     net = Net(env, cfg=cfg, device="cuda" if torch.cuda.is_available() else "cpu")
     # initialize the trainer
     agent = Agent(
@@ -74,7 +73,7 @@ def evaluation():
     while not np.any(done):
         # Based on environmental observation input, predict next action.
         action, _ = agent.act(obs, deterministic=True)
-        print("action:",action)
+        print("action:", action)
         obs, r, done, info = env.step(action)
         step += 1
         total_reward += np.mean(r)

openrl/envs/common/registration.py

@@ -66,9 +66,11 @@ def make(
         )
     else:
         if id.startswith("pybullet_drones/"):
-
             from openrl.envs.gym_pybullet_drones import make_single_agent_drone_envs
-            env_fns = make_single_agent_drone_envs(id=id,env_num=env_num,render_mode=convert_render_mode,**kwargs)
+
+            env_fns = make_single_agent_drone_envs(
+                id=id, env_num=env_num, render_mode=convert_render_mode, **kwargs
+            )
 
         elif id.startswith("snakes_"):
             from openrl.envs.snake import make_snake_envs

openrl/envs/gym_pybullet_drones/__init__.py

@@ -24,24 +24,22 @@
 from openrl.envs.common import build_envs
 
 
-def make_single_agent_drone_env(id:str,
-                                render_mode,
-                                disable_env_checker,
-                                **kwargs):
+def make_single_agent_drone_env(id: str, render_mode, disable_env_checker, **kwargs):
     import gym_pybullet_drones
+
     prefix = "pybullet_drones/"
     assert id.startswith(prefix), "id must start with pybullet_drones/"
     kwargs.pop("cfg")
 
-    env = gym.envs.registration.make(id[len(prefix):],**kwargs)
+    env = gym.envs.registration.make(id[len(prefix) :], **kwargs)
     return env
 
 
 def make_single_agent_drone_envs(
-        id: str,
-        env_num: int = 1,
-        render_mode: Optional[Union[str, List[str]]] = None,
-        **kwargs,
+    id: str,
+    env_num: int = 1,
+    render_mode: Optional[Union[str, List[str]]] = None,
+    **kwargs,
 ) -> List[Callable[[], Env]]:
     from openrl.envs.wrappers import (  # AutoReset,; DictWrapper,
         RemoveTruncated,