Ensure that consecutive observations are not views onto the same memory addresses

alimuldal · alimuldal · commit 0277e4352c57 · 2018-10-12T20:17:16.000+01:00
This means that observation arrays will not be mutated when the environment steps (google-deepmind#34). PiperOrigin-RevId: 216901616
diff --git a/dm_control/mujoco/engine.py b/dm_control/mujoco/engine.py
@@ -442,31 +442,31 @@ def _physics_state_items(self):
   # Named views of simulation data.
 
   def control(self):
-    """Returns MuJoCo actuation vector as defined in the model."""
-    return self.data.ctrl[:]
+    """Returns a copy of the control signals for the actuators."""
+    return self.data.ctrl[:].copy()
 
   def activation(self):
-    """Returns the internal states of 'filter' or 'integrator' actuators.
+    """Returns a copy of the internal states of actuators.
 
     For details, please refer to
     http://www.mujoco.org/book/computation.html#geActuation
 
     Returns:
       Activations in a numpy array.
     """
-    return self.data.act[:]
+    return self.data.act[:].copy()
 
   def state(self):
     """Returns the full physics state. Alias for `get_physics_state`."""
     return np.concatenate(self._physics_state_items())
 
   def position(self):
-    """Returns generalized positions (system configuration)."""
-    return self.data.qpos[:]
+    """Returns a copy of the generalized positions (system configuration)."""
+    return self.data.qpos[:].copy()
 
   def velocity(self):
-    """Returns generalized velocities."""
-    return self.data.qvel[:]
+    """Returns a copy of the generalized velocities."""
+    return self.data.qvel[:].copy()
 
   def timestep(self):
     """Returns the simulation timestep."""
diff --git a/dm_control/suite/cheetah.py b/dm_control/suite/cheetah.py
@@ -91,7 +91,7 @@ def get_observation(self, physics):
     """Returns an observation of the state, ignoring horizontal position."""
     obs = collections.OrderedDict()
     # Ignores horizontal position to maintain translational invariance.
-    obs['position'] = physics.data.qpos[1:]
+    obs['position'] = physics.data.qpos[1:].copy()
     obs['velocity'] = physics.velocity()
     return obs
 
diff --git a/dm_control/suite/hopper.py b/dm_control/suite/hopper.py
@@ -117,7 +117,7 @@ def get_observation(self, physics):
     """Returns an observation of positions, velocities and touch sensors."""
     obs = collections.OrderedDict()
     # Ignores horizontal position to maintain translational invariance:
-    obs['position'] = physics.data.qpos[1:]
+    obs['position'] = physics.data.qpos[1:].copy()
     obs['velocity'] = physics.velocity()
     obs['touch'] = physics.touch()
     return obs
diff --git a/dm_control/suite/humanoid.py b/dm_control/suite/humanoid.py
@@ -110,19 +110,19 @@ def head_height(self):
 
   def center_of_mass_position(self):
     """Returns position of the center-of-mass."""
-    return self.named.data.subtree_com['torso']
+    return self.named.data.subtree_com['torso'].copy()
 
   def center_of_mass_velocity(self):
     """Returns the velocity of the center-of-mass."""
-    return self.named.data.sensordata['torso_subtreelinvel']
+    return self.named.data.sensordata['torso_subtreelinvel'].copy()
 
   def torso_vertical_orientation(self):
     """Returns the z-projection of the torso orientation matrix."""
     return self.named.data.xmat['torso', ['zx', 'zy', 'zz']]
 
   def joint_angles(self):
     """Returns the state without global orientation or position."""
-    return self.data.qpos[7:]  # Skip the 7 DoFs of the free root joint.
+    return self.data.qpos[7:].copy()  # Skip the 7 DoFs of the free root joint.
 
   def extremities(self):
     """Returns end effector positions in egocentric frame."""
diff --git a/dm_control/suite/humanoid_CMU.py b/dm_control/suite/humanoid_CMU.py
@@ -90,15 +90,15 @@ def center_of_mass_position(self):
 
   def center_of_mass_velocity(self):
     """Returns the velocity of the center-of-mass."""
-    return self.named.data.sensordata['thorax_subtreelinvel']
+    return self.named.data.sensordata['thorax_subtreelinvel'].copy()
 
   def torso_vertical_orientation(self):
     """Returns the z-projection of the thorax orientation matrix."""
     return self.named.data.xmat['thorax', ['zx', 'zy', 'zz']]
 
   def joint_angles(self):
     """Returns the state without global orientation or position."""
-    return self.data.qpos[7:]  # Skip the 7 DoFs of the free root joint.
+    return self.data.qpos[7:].copy()  # Skip the 7 DoFs of the free root joint.
 
   def extremities(self):
     """Returns end effector positions in egocentric frame."""
diff --git a/dm_control/suite/pendulum.py b/dm_control/suite/pendulum.py
@@ -64,7 +64,7 @@ def pole_vertical(self):
 
   def angular_velocity(self):
     """Returns the angular velocity of the pole."""
-    return self.named.data.qvel['hinge']
+    return self.named.data.qvel['hinge'].copy()
 
   def pole_orientation(self):
     """Returns both horizontal and vertical components of pole frame."""
diff --git a/dm_control/suite/swimmer.py b/dm_control/suite/swimmer.py
@@ -163,7 +163,7 @@ def body_velocities(self):
 
   def joints(self):
     """Returns all internal joint angles (excluding root joints)."""
-    return self.data.qpos[3:]
+    return self.data.qpos[3:].copy()
 
 
 class Swimmer(base.Task):
diff --git a/dm_control/suite/tests/domains_test.py b/dm_control/suite/tests/domains_test.py
@@ -204,5 +204,17 @@ def test_task_supports_environment_kwargs(self, domain, task):
     self.assertSetEqual(set(env.observation_spec()),
                         {control.FLAT_OBSERVATION_KEY})
 
+  @parameterized.parameters(*suite.ALL_TASKS)
+  def test_observation_arrays_dont_share_memory(self, domain, task):
+    env = suite.load(domain, task)
+    first_timestep = env.reset()
+    action = np.zeros(env.action_spec().shape)
+    second_timestep = env.step(action)
+    for name, first_array in six.iteritems(first_timestep.observation):
+      second_array = second_timestep.observation[name]
+      self.assertFalse(
+          np.may_share_memory(first_array, second_array),
+          msg='Consecutive observations of {!r} may share memory.'.format(name))
+
 if __name__ == '__main__':
   absltest.main()
