feat: depricate step done return value (#98)

This commit depricates the step `done` return value. This was done for consistency with gym v0.25.0
(see openai/gym@907b1b2).

BREAKING CHANGE: step now returns 5 values instead of 4.

examples/use_simzoo.py

@@ -9,8 +9,8 @@
 # TODO: [ ] Check ex3_ekf.
 # TODO: [ ] Check cartpole.
 
-# ENV_NAME = "Oscillator-v1"
-ENV_NAME = "Ex3EKF-v1"
+ENV_NAME = "Oscillator-v1"
+# ENV_NAME = "Ex3EKF-v1"
 
 if __name__ == "__main__":
     env = gym.make(ENV_NAME, render_mode="human")

simzoo/envs/biological/oscillator/oscillator.py

@@ -199,7 +199,10 @@ def step(self, action):
 
                 - obs (:obj:`numpy.ndarray`): The current state
                 - cost (:obj:`numpy.float64`): The current cost.
-                - done (:obj:`bool`): Whether the episode was done.
+                - terminated (:obj:`bool`): Whether the episode was done.
+                - truncated (:obj:`bool`): Whether the episode was truncated. This value
+                    is set by wrappers when for example a time limit is reached or the
+                    agent goes out of bounds.
                 - info_dict (:obj:`dict`): Dictionary with additional information.
         """
         # Clip action if needed
@@ -299,16 +302,14 @@ def step(self, action):
         # cost = (abs(p1 - r1)) ** 0.2
 
         # Define stopping criteria
-        if cost > self.reward_range.high or cost < self.reward_range.low:
-            done = True
-        else:
-            done = False
+        terminated = bool(cost > self.reward_range.high or cost < self.reward_range.low)
 
         # Return state, cost, done and reference
         return (
             np.array([m1, m2, m3, p1, p2, p3, r1, p1 - r1]),
             cost,
-            done,
+            terminated,
+            False,
             dict(reference=r1, state_of_interest=p1 - r1),
         )
 

simzoo/envs/classic_control/cart_pole_cost/cart_pole_cost.py

@@ -310,7 +310,10 @@ def step(self, action):
 
                 - obs (:obj:`numpy.ndarray`): The current state
                 - cost (:obj:`numpy.float64`): The current cost.
-                - done (:obj:`bool`): Whether the episode was done.
+                - terminated (:obj:`bool`): Whether the episode was done.
+                - truncated (:obj:`bool`): Whether the episode was truncated. This value
+                    is set by wrappers when for example a time limit is reached or the
+                    agent goes out of bounds.
                 - info_dict (:obj:`dict`): Dictionary with additional information.
         """
         # Clip action if needed
@@ -378,13 +381,15 @@ def step(self, action):
         cost, ref = self.cost(x, theta)
 
         # Define stopping criteria
-        if (
+        terminated = bool(
             abs(x) > self.x_threshold
             or abs(theta) > self.theta_threshold_radians
             or cost > self.reward_range.high
             or cost < self.reward_range.low
-        ):
-            done = True
+        )
+
+        # Calculate stopping cost
+        if terminated:
             cost = 100.0
 
             # Throw warning if already done
@@ -399,16 +404,15 @@ def step(self, action):
                         "True' -- any further steps are undefined behavior."
                     )
                 self.steps_beyond_done += 1
-        else:
-            done = False
 
-        # Return state, cost, done and info_dict
+        # Return state, cost, terminated and info_dict
         violation_of_constraint = bool(abs(x) > self.const_pos)
         violation_of_x_threshold = bool(x < -self.x_threshold or x > self.x_threshold)
         return (
             self.state,
             cost,
-            done,
+            terminated,
+            False,
             dict(
                 cons_pos=self.const_pos,
                 cons_theta=self.theta_threshold_radians,

simzoo/envs/classic_control/ex3_ekf/ex3_ekf.py

@@ -185,7 +185,10 @@ def step(self, action):
 
                 - obs (:obj:`numpy.ndarray`): The current state
                 - cost (:obj:`numpy.float64`): The current cost.
-                - done (:obj:`bool`): Whether the episode was done.
+                - terminated (:obj:`bool`): Whether the episode was done.
+                - truncated (:obj:`bool`): Whether the episode was truncated. This value
+                    is set by wrappers when for example a time limit is reached or the
+                    agent goes out of bounds.
                 - info_dict (:obj:`dict`): Dictionary with additional information.
         """
         # Clip action if needed
@@ -250,10 +253,7 @@ def step(self, action):
         # cost = np.abs(hat_x_1 - x_1)**1 + np.abs(hat_x_2 - x_2)**1
 
         # Define stopping criteria
-        if cost > self.reward_range.high or cost < self.reward_range.low:
-            done = True
-        else:
-            done = False
+        terminated = bool(cost > self.reward_range.high or cost < self.reward_range.low)
 
         # Update state
         self.state = np.array([hat_x_1, hat_x_2, x_1, x_2])
@@ -264,7 +264,8 @@ def step(self, action):
         return (
             np.array([hat_x_1, hat_x_2, x_1, x_2]),
             cost,
-            done,
+            terminated,
+            False,
             dict(
                 reference=y_1,
                 state_of_interest=np.array([hat_x_1 - x_1, hat_x_2 - x_2]),