added new mp wrappers to all environments

alr_envs/alr/mujoco/half_cheetah_jump/new_mp_wrapper.py

@@ -0,0 +1,24 @@
+from typing import Tuple, Union
+
+import numpy as np
+
+from mp_env_api import MPEnvWrapper
+
+from alr_envs.mp.raw_interface_wrapper import RawInterfaceWrapper
+
+
+class MPWrapper(RawInterfaceWrapper):
+    def context_mask(self):
+        return np.hstack([
+            [False] * 17,
+            [True] # goal height
+        ])
+
+    @property
+    def current_pos(self) -> Union[float, int, np.ndarray]:
+        return self.env.sim.data.qpos[3:9].copy()
+
+    @property
+    def current_vel(self) -> Union[float, int, np.ndarray, Tuple]:
+        return self.env.sim.data.qvel[3:9].copy()
+

alr_envs/alr/mujoco/hopper_throw/mp_wrapper.py

@@ -10,7 +10,7 @@ class MPWrapper(MPEnvWrapper):
     def active_obs(self):
         return np.hstack([
             [False] * 17,
-            [True] # goal pos
+            [True]  # goal pos
         ])
 
     @property

alr_envs/alr/mujoco/hopper_throw/new_mp_wrapper.py

@@ -0,0 +1,27 @@
+from typing import Tuple, Union
+
+import numpy as np
+
+from mp_env_api import MPEnvWrapper
+
+from alr_envs.mp.raw_interface_wrapper import RawInterfaceWrapper
+
+
+class MPWrapper(RawInterfaceWrapper):
+    def context_mask(self):
+        return np.hstack([
+            [False] * 17,
+            [True] # goal pos
+        ])
+
+    @property
+    def current_pos(self) -> Union[float, int, np.ndarray]:
+        return self.env.sim.data.qpos[3:6].copy()
+
+    @property
+    def current_vel(self) -> Union[float, int, np.ndarray, Tuple]:
+        return self.env.sim.data.qvel[3:6].copy()
+
+    @property
+    def dt(self) -> Union[float, int]:
+        return self.env.dt

alr_envs/alr/mujoco/reacher/new_mp_wrapper.py

@@ -2,8 +2,10 @@ from alr_envs.mp.black_box_wrapper import BlackBoxWrapper
 from typing import Union, Tuple
 import numpy as np
 
+from alr_envs.mp.raw_interface_wrapper import RawInterfaceWrapper
 
-class MPWrapper(BlackBoxWrapper):
+
+class MPWrapper(RawInterfaceWrapper):
 
     @property
     def current_pos(self) -> Union[float, int, np.ndarray, Tuple]:
@@ -12,7 +14,7 @@ class MPWrapper(BlackBoxWrapper):
     def current_vel(self) -> Union[float, int, np.ndarray, Tuple]:
         return self.env.sim.data.qvel.flat[:self.env.n_links]
 
-    def get_context_mask(self):
+    def context_mask(self):
         return np.concatenate([
             [False] * self.env.n_links,  # cos
             [False] * self.env.n_links,  # sin
@@ -21,4 +23,4 @@ class MPWrapper(BlackBoxWrapper):
             [False] * 3,  # goal distance
             # self.get_body_com("target"),  # only return target to make problem harder
             [False],  # step
-        ])
+        ])

alr_envs/alr/mujoco/walker_2d_jump/new_mp_wrapper.py

@@ -0,0 +1,28 @@
+from typing import Tuple, Union
+
+import numpy as np
+
+from mp_env_api import MPEnvWrapper
+
+from alr_envs.mp.raw_interface_wrapper import RawInterfaceWrapper
+
+
+class MPWrapper(RawInterfaceWrapper):
+    def context_mask(self):
+        return np.hstack([
+            [False] * 17,
+            [True] # goal pos
+        ])
+
+    @property
+    def current_pos(self) -> Union[float, int, np.ndarray]:
+        return self.env.sim.data.qpos[3:9].copy()
+
+    @property
+    def current_vel(self) -> Union[float, int, np.ndarray, Tuple]:
+        return self.env.sim.data.qvel[3:9].copy()
+
+    @property
+    def goal_pos(self) -> Union[float, int, np.ndarray, Tuple]:
+        raise ValueError("Goal position is not available and has to be learnt based on the environment.")
+