rename create random to random action

ml-agents-envs/mlagents_envs/base_env.py

             return np.zeros((n_agents, self.continuous_size), dtype=np.float32)
         return np.zeros((n_agents, self.discrete_size), dtype=np.int32)
 
-    def create_random(self, n_agents: int) -> np.ndarray:
+    def random_action(self, n_agents: int) -> np.ndarray:
         """
         Generates a numpy array corresponding to a random action (either discrete
         or continuous) for a number of agents.

ml-agents-envs/mlagents_envs/tests/test_steps.py

     specs = ActionSpec.create_continuous(action_len)
     zero_action = specs.create_empty(4)
     assert np.array_equal(zero_action, np.zeros((4, action_len), dtype=np.float32))
-    random_action = specs.create_random(4)
+    random_action = specs.random_action(4)
     assert random_action.dtype == np.float32
     assert random_action.shape == (4, action_len)
     assert np.min(random_action) >= -1
     zero_action = specs.create_empty(4)
     assert np.array_equal(zero_action, np.zeros((4, len(action_shape)), dtype=np.int32))
 
-    random_action = specs.create_random(4)
+    random_action = specs.random_action(4)
     assert random_action.dtype == np.int32
     assert random_action.shape == (4, len(action_shape))
     assert np.min(random_action) >= 0

ml-agents/mlagents/trainers/tests/torch/test_reward_providers/utils.py

     next_observations = [
         np.random.normal(size=shape) for shape in behavior_spec.observation_shapes
     ]
-    action = behavior_spec.action_spec.create_random(1)[0, :]
+    action = behavior_spec.action_spec.random_action(1)[0, :]
     for _ in range(number):
         curr_split_obs = SplitObservations.from_observations(curr_observations)
         next_split_obs = SplitObservations.from_observations(next_observations)