Fix bug with construct_curr_info (#2490)

* Fix bug with construct_curr_info
* Add more tests

ml-agents/mlagents/trainers/rl_trainer.py

         :return: curr_info: Reconstructed BrainInfo to match agents of next_info.
         """
         visual_observations: List[List[Any]] = [
-            []
+            [] for _ in next_info.visual_observations
         ]  # TODO add types to brain.py methods
         vector_observations = []
         text_observations = []

ml-agents/mlagents/trainers/tests/test_rl_trainer.py

 
 @mock.patch("mlagents.trainers.rl_trainer.RLTrainer.add_policy_outputs")
 @mock.patch("mlagents.trainers.rl_trainer.RLTrainer.add_rewards_outputs")
-def test_rl_trainer(add_policy_outputs, add_rewards_outputs):
+@pytest.mark.parametrize("num_vis_obs", [0, 1, 2], ids=["vec", "1 viz", "2 viz"])
+def test_rl_trainer(add_policy_outputs, add_rewards_outputs, num_vis_obs):
     trainer = create_rl_trainer()
     trainer.policy = create_mock_policy()
     fake_action_outputs = {
         num_agents=2,
         num_vector_observations=8,
         num_vector_acts=2,
-        num_vis_observations=1,
+        num_vis_observations=num_vis_obs,
     )
     trainer.add_experiences(
         create_mock_all_brain_info(mock_braininfo),
         num_agents=1,
         num_vector_observations=8,
         num_vector_acts=2,
-        num_vis_observations=1,
+        num_vis_observations=num_vis_obs,
+    assert len(brain_info.visual_observations) == num_vis_obs
+    assert len(brain_info.vector_observations) == 1
+    assert len(brain_info.previous_vector_actions) == 1
 
     # Test end episode
     trainer.end_episode()