Add open ai gym environments

alr_envs/open_ai/continuous_mountain_car/__init__.py

@@ -0,0 +1,2 @@
+from alr_envs.open_ai.continuous_mountain_car.positional_wrapper import PositionalWrapper
+from alr_envs.open_ai.continuous_mountain_car.mp_wrapper import MPWrapper

alr_envs/open_ai/continuous_mountain_car/mp_wrapper.py

@@ -0,0 +1,17 @@
+from typing import Union
+
+from mp_env_api.env_wrappers.mp_env_wrapper import MPEnvWrapper
+
+
+class MPWrapper(MPEnvWrapper):
+    @property
+    def start_pos(self):
+        raise ValueError("Start position is not available")
+
+    @property
+    def goal_pos(self):
+        raise ValueError("Goal position is not available and has to be learnt based on the environment.")
+
+    @property
+    def dt(self) -> Union[float, int]:
+        return 1.

alr_envs/open_ai/continuous_mountain_car/positional_wrapper.py

@@ -0,0 +1,13 @@
+from typing import Union
+import numpy as np
+from mp_env_api.env_wrappers.positional_env_wrapper import PositionalEnvWrapper
+
+
+class PositionalWrapper(PositionalEnvWrapper):
+    @property
+    def current_vel(self) -> Union[float, int, np.ndarray]:
+        return np.array([self.state[1]])
+
+    @property
+    def current_pos(self) -> Union[float, int, np.ndarray]:
+        return np.array([self.state[0]])

alr_envs/open_ai/fetch/__init__.py

@@ -0,0 +1,2 @@
+from alr_envs.open_ai.fetch.positional_wrapper import PositionalWrapper
+from alr_envs.open_ai.fetch.mp_wrapper import MPWrapper

alr_envs/open_ai/fetch/mp_wrapper.py

@@ -0,0 +1,18 @@
+from typing import Union
+
+from gym import spaces
+from mp_env_api.env_wrappers.mp_env_wrapper import MPEnvWrapper
+
+
+class MPWrapper(MPEnvWrapper):
+    @property
+    def start_pos(self):
+        return self.initial_gripper_xpos
+
+    @property
+    def goal_pos(self):
+        raise ValueError("Goal position is not available and has to be learnt based on the environment.")
+
+    @property
+    def dt(self) -> Union[float, int]:
+        return self.env.dt

alr_envs/open_ai/fetch/positional_wrapper.py

@@ -0,0 +1,13 @@
+from typing import Union
+import numpy as np
+from mp_env_api.env_wrappers.positional_env_wrapper import PositionalEnvWrapper
+
+
+class PositionalWrapper(PositionalEnvWrapper):
+    @property
+    def current_vel(self) -> Union[float, int, np.ndarray]:
+        return self._get_obs()["observation"][-5:-1]
+
+    @property
+    def current_pos(self) -> Union[float, int, np.ndarray]:
+        return self._get_obs()["observation"][:4]

alr_envs/open_ai/reacher_v2/__init__.py

@@ -0,0 +1,2 @@
+from alr_envs.open_ai.reacher_v2.positional_wrapper import PositionalWrapper
+from alr_envs.open_ai.reacher_v2.mp_wrapper import MPWrapper

alr_envs/open_ai/reacher_v2/mp_wrapper.py

@@ -0,0 +1,18 @@
+from typing import Union
+
+from mp_env_api.env_wrappers.mp_env_wrapper import MPEnvWrapper
+
+
+class MPWrapper(MPEnvWrapper):
+
+    @property
+    def start_pos(self):
+        raise ValueError("Start position is not available")
+
+    @property
+    def goal_pos(self):
+        return self.goal
+
+    @property
+    def dt(self) -> Union[float, int]:
+        return self.env.dt

alr_envs/open_ai/reacher_v2/positional_wrapper.py

@@ -0,0 +1,13 @@
+from typing import Union
+import numpy as np
+from mp_env_api.env_wrappers.positional_env_wrapper import PositionalEnvWrapper
+
+
+class PositionalWrapper(PositionalEnvWrapper):
+    @property
+    def current_vel(self) -> Union[float, int, np.ndarray]:
+        return self.sim.data.qvel[:2]
+
+    @property
+    def current_pos(self) -> Union[float, int, np.ndarray]:
+        return self.sim.data.qpos[:2]