[skip ci] adding should_still_train method to ppo

ml-agents/mlagents/trainers/ppo/trainer.py

         self.seed = seed
         self.policy: NNPolicy = None  # type: ignore
 
+    def should_still_train(self) -> bool:
+        """
+        Returns whether or not the trainer should train. A Trainer could
+        stop training if it wasn't training to begin with, or if max_steps
+        is reached.
+        """
+        return self.is_training and self.get_step <= self.trainer_settings.max_steps
+
     def _process_trajectory(self, trajectory: Trajectory) -> None:
         """
         Takes a trajectory and processes it, putting it into the update buffer.