#901 - Dropped temp_primaite_sessiion_2 from conftest.py.

- Re-added the hard-coded mean rewards per episode values from a rpe-trained agent to the deterministic test in test_seeding_and_deterministic_session.py
- Partially tidies up some tests in test_observation_space.py; Still some work to be done on this at a later date.

tests/test_seeding_and_deterministic_session.py

@@ -1,5 +1,3 @@
-import time
-
 import pytest as pytest
 
 from primaite.config.lay_down_config import dos_very_basic_config_path
@@ -11,45 +9,45 @@ from tests import TEST_CONFIG_ROOT
     [[TEST_CONFIG_ROOT / "ppo_seeded_training_config.yaml", dos_very_basic_config_path()]],
     indirect=True,
 )
-@pytest.mark.parametrize(
-    "temp_primaite_session_2",
-    [[TEST_CONFIG_ROOT / "ppo_seeded_training_config.yaml", dos_very_basic_config_path()]],
-    indirect=True,
-)
-def test_seeded_learning(temp_primaite_session, temp_primaite_session_2):
-    """Test running seeded learning produces the same output when ran twice."""
+def test_seeded_learning(temp_primaite_session):
+    """
+    Test running seeded learning produces the same output when ran twice.
+
+    .. note::
+
+        If this is failing, the hard-coded expected_mean_reward_per_episode
+        from a pre-trained agent will probably need to be updated. If the
+        env changes and those changed how this agent is trained, chances are
+        the mean rewards are going to be different.
+
+        Run the test, but print out the session.learn_av_reward_per_episode()
+        before comparing it. Then copy the printed dict and replace the
+        expected_mean_reward_per_episode with those values. The test should
+        now work. If not, then you've got a bug :).
     """
     expected_mean_reward_per_episode = {
-        1: -90.703125,
-        2: -91.15234375,
-        3: -87.5,
-        4: -92.2265625,
-        5: -94.6875,
-        6: -91.19140625,
-        7: -88.984375,
-        8: -88.3203125,
-        9: -112.79296875,
-        10: -100.01953125,
+        1: -33.90625,
+        2: -32.32421875,
+        3: -25.234375,
+        4: -30.15625,
+        5: -27.1484375,
+        6: -29.609375,
+        7: -29.921875,
+        8: -29.3359375,
+        9: -28.046875,
+        10: -27.24609375,
     }
-    """
+
     with temp_primaite_session as session:
-        assert session._training_config.seed == 67890, (
-            "Expected output is based upon a agent that was trained with " "seed 67890"
-        )
+        assert (
+            session._training_config.seed == 67890
+        ), "Expected output is based upon a agent that was trained with seed 67890"
         session.learn()
-        actual_mean_reward_per_episode_run_1 = session.learn_av_reward_per_episode()
-    time.sleep(2)
-    with temp_primaite_session_2 as session:
-        assert session._training_config.seed == 67890, (
-            "Expected output is based upon a agent that was trained with " "seed 67890"
-        )
-        session.learn()
-        actual_mean_reward_per_episode_run_2 = session.learn_av_reward_per_episode()
 
-    assert actual_mean_reward_per_episode_run_1 == actual_mean_reward_per_episode_run_2
+        assert expected_mean_reward_per_episode == session.learn_av_reward_per_episode()
 
 
-@pytest.mark.skip(reason="Inconsistent results. Needs someone with RL " "knowledge to investigate further.")
+@pytest.mark.skip(reason="Inconsistent results. Needs someone with RL knowledge to investigate further.")
 @pytest.mark.parametrize(
     "temp_primaite_session",
     [[TEST_CONFIG_ROOT / "ppo_seeded_training_config.yaml", dos_very_basic_config_path()]],