remove 'mode' from render() for all dummy envs (test suite)

fancy_gym/envs/classic_control/hole_reacher/hole_reacher.py

@@ -178,9 +178,7 @@ class HoleReacherEnv(BaseReacherDirectEnv):
 
         return False
 
-    def render(self, mode=None):
-        if mode==None:
-            mode = self.render_mode
+    def render(self):
         if self.fig is None:
             # Create base figure once on the beginning. Afterwards only update
             plt.ion()

fancy_gym/envs/classic_control/simple_reacher/simple_reacher.py

@@ -98,9 +98,7 @@ class SimpleReacherEnv(BaseReacherTorqueEnv):
     def _check_collisions(self) -> bool:
         return self._check_self_collision()
 
-    def render(self, mode=None):  # pragma: no cover
-        if mode==None:
-            mode = self.render_mode
+    def render(self):  # pragma: no cover
         if self.fig is None:
             # Create base figure once on the beginning. Afterwards only update
             plt.ion()

fancy_gym/envs/classic_control/viapoint_reacher/viapoint_reacher.py

@@ -123,9 +123,7 @@ class ViaPointReacherEnv(BaseReacherDirectEnv):
     def _check_collisions(self) -> bool:
         return self._check_self_collision()
 
-    def render(self, mode=None):
-        if mode==None:
-            mode = self.render_mode
+    def render(self):
         goal_pos = self._goal.T
         via_pos = self._via_point.T
 

fancy_gym/examples/examples_movement_primitives.py

@@ -16,7 +16,7 @@ def example_mp(env_name="fancy_ProMP/HoleReacher-v0", seed=1, iterations=1, rend
     """
     # Equivalent to gym, we have a make function which can be used to create environments.
     # It takes care of seeding and enables the use of a variety of external environments using the gym interface.
-    env = gym.make(env_name, render_mode='human' if render else None)
+    env = gym.make(env_name)
 
     returns = 0
     # env.render(mode=None)
@@ -26,6 +26,14 @@ def example_mp(env_name="fancy_ProMP/HoleReacher-v0", seed=1, iterations=1, rend
     for i in range(iterations):
 
         if render and i % 1 == 0:
+            # This renders the full MP trajectory
+            # It is only required to call render() once in the beginning, which renders every consecutive trajectory.
+            # Resetting to no rendering, can be achieved by render(mode=None).
+            # It is also possible to change the mode multiple times when
+            # e.g. only every second trajectory should be displayed, such as here
+            # Just make sure the correct mode is set before executing the step.
+            env.render(mode="human")
+        else:
             env.render()
 
         # Now the action space is not the raw action but the parametrization of the trajectory generator,
@@ -41,19 +49,14 @@ def example_mp(env_name="fancy_ProMP/HoleReacher-v0", seed=1, iterations=1, rend
         if terminated or truncated:
             print(reward)
             obs = env.reset()
-    env.close()
 
 
 def example_custom_mp(env_name="fancy_ProMP/Reacher5d-v0", seed=1, iterations=1, render=True):
     """
-    Example for running a custom movement primitive based environments.
-    Our already registered environments follow the same structure.
-    Hence, this also allows to adjust hyperparameters of the movement primitives.
-    Yet, we recommend the method above if you are just interested in changing those parameters for existing tasks.
-    We appreciate PRs for custom environments (especially MP wrappers of existing tasks) 
-    for our repo: https://github.com/ALRhub/fancy_gym/
+    Example for running a movement primitive based environment, which is already registered
     Args:
-        seed: seed
+        env_name: DMP env_id
+        seed: seed for deterministic behaviour
         iterations: Number of rollout steps to run
         render: Render the episode
 
@@ -62,14 +65,14 @@ def example_custom_mp(env_name="fancy_ProMP/Reacher5d-v0", seed=1, iterations=1,
     """
     # Changing the arguments of the black box env is possible by providing them to gym through mp_config_override.
     # E.g. here for way to many basis functions
-    env = gym.make(env_name, seed, mp_config_override={'basis_generator_kwargs': {'num_basis': 1000}}, render_mode='human' if render else None)
+    env = gym.make(env_name, seed, mp_config_override={'basis_generator_kwargs': {'num_basis': 1000}})
 
     returns = 0
     obs = env.reset()
 
     # This time rendering every trajectory
     if render:
-        env.render()
+        env.render(mode="human")
 
     # number of samples/full trajectories (multiple environment steps)
     for i in range(iterations):
@@ -81,47 +84,8 @@ def example_custom_mp(env_name="fancy_ProMP/Reacher5d-v0", seed=1, iterations=1,
             print(i, reward)
             obs = env.reset()
 
-    env.close()
     return obs
 
-class Custom_MPWrapper(fancy_gym.envs.mujoco.reacher.MPWrapper):
-    mp_config = {
-        'ProMP': {
-                'trajectory_generator_kwargs':  {
-                    'trajectory_generator_type': 'promp',
-                    'weights_scale': 2
-                },
-                'phase_generator_kwargs': {
-                    'phase_generator_type': 'linear'
-                },
-                'controller_kwargs': {
-                    'controller_type': 'velocity'
-                },
-                'basis_generator_kwargs': {
-                    'basis_generator_type': 'zero_rbf',
-                    'num_basis': 5,
-                    'num_basis_zero_start': 1
-                }
-        },
-        'DMP': {
-            'trajectory_generator_kwargs': {
-                'trajectory_generator_type': 'dmp',
-                'weights_scale': 500
-            },
-            'phase_generator_kwargs': {
-                'phase_generator_type': 'exp',
-                'alpha_phase': 2.5
-            },
-            'controller_kwargs': {
-                'controller_type': 'velocity'
-            },
-            'basis_generator_kwargs': {
-                'basis_generator_type': 'rbf',
-                'num_basis': 5
-            }
-        }
-    }
-
 
 def example_fully_custom_mp(seed=1, iterations=1, render=True):
     """
@@ -141,92 +105,37 @@ def example_fully_custom_mp(seed=1, iterations=1, render=True):
     """
 
     base_env_id = "fancy/Reacher5d-v0"
-    custom_env_id = "fancy/Reacher5d-Custom-v0"
-    custom_env_id_DMP = "fancy_DMP/Reacher5d-Custom-v0"
-    custom_env_id_ProMP = "fancy_ProMP/Reacher5d-Custom-v0"
 
-    fancy_gym.upgrade(custom_env_id, mp_wrapper=Custom_MPWrapper, add_mp_types=['ProMP', 'DMP'], base_id=base_env_id)
+    # Replace this wrapper with the custom wrapper for your environment by inheriting from the RawInterfaceWrapper.
+    # You can also add other gym.Wrappers in case they are needed.
+    wrappers = [fancy_gym.envs.mujoco.reacher.MPWrapper]
 
-    env = gym.make(custom_env_id_ProMP, render_mode='human' if render else None)
+    # For a ProMP
+    trajectory_generator_kwargs = {'trajectory_generator_type': 'promp',
+                                   'weights_scale': 2}
+    phase_generator_kwargs = {'phase_generator_type': 'linear'}
+    controller_kwargs = {'controller_type': 'velocity'}
+    basis_generator_kwargs = {'basis_generator_type': 'zero_rbf',
+                              'num_basis': 5,
+                              'num_basis_zero_start': 1
+                              }
 
-    rewards = 0
-    obs = env.reset()
+    # # For a DMP
+    # trajectory_generator_kwargs = {'trajectory_generator_type': 'dmp',
+    #                                'weights_scale': 500}
+    # phase_generator_kwargs = {'phase_generator_type': 'exp',
+    #                           'alpha_phase': 2.5}
+    # controller_kwargs = {'controller_type': 'velocity'}
+    # basis_generator_kwargs = {'basis_generator_type': 'rbf',
+    #                           'num_basis': 5
+    #                           }
+    env = fancy_gym.make_bb(env_id=base_env_id, wrappers=wrappers, black_box_kwargs={},
+                            traj_gen_kwargs=trajectory_generator_kwargs, controller_kwargs=controller_kwargs,
+                            phase_kwargs=phase_generator_kwargs, basis_kwargs=basis_generator_kwargs,
+                            seed=seed)
 
     if render:
-        env.render()
-
-    # number of samples/full trajectories (multiple environment steps)
-    for i in range(iterations):
-        ac = env.action_space.sample()
-        obs, reward, terminated, truncated, info = env.step(ac)
-        rewards += reward
-
-        if terminated or truncated:
-            print(rewards)
-            rewards = 0
-            obs = env.reset()
-
-    try: # Some mujoco-based envs don't correlcty implement .close
-        env.close()
-    except:
-        pass
-
-
-def example_fully_custom_mp_alternative(seed=1, iterations=1, render=True):
-    """
-    Instead of defining the mp_args in a new custom MP_Wrapper, they can also be provided during registration.
-    Args:
-        seed: seed
-        iterations: Number of rollout steps to run
-        render: Render the episode
-
-    Returns:
-
-    """
-
-    base_env_id = "fancy/Reacher5d-v0"
-    custom_env_id = "fancy/Reacher5d-Custom-v0"
-    custom_env_id_ProMP = "fancy_ProMP/Reacher5d-Custom-v0"
-
-    fancy_gym.upgrade(custom_env_id, mp_wrapper=fancy_gym.envs.mujoco.reacher.MPWrapper, add_mp_types=['ProMP'], base_id=base_env_id, mp_config_override=     {'ProMP': {
-                'trajectory_generator_kwargs':  {
-                    'trajectory_generator_type': 'promp',
-                    'weights_scale': 2
-                },
-                'phase_generator_kwargs': {
-                    'phase_generator_type': 'linear'
-                },
-                'controller_kwargs': {
-                    'controller_type': 'velocity'
-                },
-                'basis_generator_kwargs': {
-                    'basis_generator_type': 'zero_rbf',
-                    'num_basis': 5,
-                    'num_basis_zero_start': 1
-                }
-        }})
-
-    env = gym.make(custom_env_id_ProMP, render_mode='human' if render else None)
-
-    rewards = 0
-    obs = env.reset()
-
-    if render:
-        env.render()
-
-    # number of samples/full trajectories (multiple environment steps)
-    for i in range(iterations):
-        ac = env.action_space.sample()
-        obs, reward, terminated, truncated, info = env.step(ac)
-        rewards += reward
-
-        if terminated or truncated:
-            print(rewards)
-            rewards = 0
-            obs = env.reset()
-
-    if render:
-        env.render()
+        env.render(mode="human")
 
     rewards = 0
     obs = env.reset()
@@ -242,13 +151,8 @@ def example_fully_custom_mp_alternative(seed=1, iterations=1, render=True):
             rewards = 0
             obs = env.reset()
 
-    try: # Some mujoco-based envs don't correlcty implement .close
-        env.close()
-    except:
-        pass
 
-
-def main():
+if __name__ == '__main__':
     render = False
     # DMP
     example_mp("fancy_DMP/HoleReacher-v0", seed=10, iterations=5, render=render)
@@ -268,7 +172,3 @@ def main():
 
     # Custom MP
     example_fully_custom_mp(seed=10, iterations=1, render=render)
-    example_fully_custom_mp_alternative(seed=10, iterations=1, render=render)
-
-if __name__=='__main__':
-    main()

test/test_all_gym_builtin_envs.py

@@ -12,14 +12,13 @@ GYM_IDS = [spec.id for spec in gym.envs.registry.values() if
            not isinstance(spec.entry_point, Callable) and
            "fancy_gym" not in spec.entry_point and 'make_bb_env_helper' not in spec.entry_point
            and 'jax' not in spec.id.lower()
-           and 'shimmy' not in spec.id.lower()
-           and 'ale_py' not in spec.id.lower()
-           and 'tabular' not in spec.id.lower()
+           and 'jax' not in spec.id.lower()
            and not re.match(r'GymV2.Environment', spec.id)
            ]
 GYM_MP_IDS = fancy_gym.ALL_DMC_MOVEMENT_PRIMITIVE_ENVIRONMENTS['all']
 SEED = 1
 
+
 @pytest.mark.parametrize('env_id', GYM_IDS)
 def test_step_gym_functionality(env_id: str):
     """Tests that step environments run without errors using random actions."""

test/test_black_box.py

@@ -41,7 +41,7 @@ class ToyEnv(gym.Env):
         obs, reward, terminated, truncated, info = np.array([-1]), 1, False, False, {}
         return obs, reward, terminated, truncated, info
 
-    def render(self, mode="human"):
+    def render(self):
         pass
 
 
@@ -103,7 +103,7 @@ def test_verbosity(mp_type: str, env_wrap: Tuple[str, Type[RawInterfaceWrapper]]
 
     env_step = make(env_id)
     env_step.reset()
-    _obs, _reward, _terminated, _truncated, info = env_step.step(env_step.action_space.sample())
+    _obs, _reward, _terminated, _truncated, info = env.step(env.action_space.sample())
     info_keys_step = info.keys()
 
     assert all(e in info_keys for e in info_keys_step)

test/test_fancy_registry.py

@@ -33,7 +33,7 @@ class ToyEnv(gym.Env):
         obs, reward, terminated, truncated, info = np.array([-1]), 1, False, False, {}
         return obs, reward, terminated, truncated, info
 
-    def render(self, mode="human"):
+    def render(self):
         pass
 
 

test/test_replanning_sequencing.py

@@ -37,7 +37,7 @@ class ToyEnv(gym.Env):
         obs, reward, terminated, truncated, info = np.array([-1]), 1, False, False, {}
         return obs, reward, terminated, truncated, info
 
-    def render(self, mode="human"):
+    def render(self):
         pass