Merge remote-tracking branch 'origin/dev' into 1566-configure-episode-steps-learn-eval

# Conflicts:
#	src/primaite/agents/rllib.py

tests/config/ppo_not_seeded_training_config.yaml

@@ -0,0 +1,155 @@
+# Training Config File
+
+# Sets which agent algorithm framework will be used.
+# Options are:
+# "SB3" (Stable Baselines3)
+# "RLLIB" (Ray RLlib)
+# "CUSTOM" (Custom Agent)
+agent_framework: SB3
+
+# Sets which deep learning framework will be used (by RLlib ONLY).
+# Default is TF (Tensorflow).
+# Options are:
+# "TF" (Tensorflow)
+# TF2 (Tensorflow 2.X)
+# TORCH (PyTorch)
+deep_learning_framework: TF2
+
+# Sets which Agent class will be used.
+# Options are:
+# "A2C" (Advantage Actor Critic coupled with either SB3 or RLLIB agent_framework)
+# "PPO" (Proximal Policy Optimization coupled with either SB3 or RLLIB agent_framework)
+# "HARDCODED" (The HardCoded agents coupled with an ACL or NODE action_type)
+# "DO_NOTHING" (The DoNothing agents coupled with an ACL or NODE action_type)
+# "RANDOM" (primaite.agents.simple.RandomAgent)
+# "DUMMY" (primaite.agents.simple.DummyAgent)
+agent_identifier: PPO
+
+# Sets whether Red Agent POL and IER is randomised.
+# Options are:
+# True
+# False
+random_red_agent: False
+
+# The (integer) seed to be used in random number generation
+# Default is None (null)
+seed: None
+
+# Set whether the agent will be deterministic instead of stochastic
+# Options are:
+# True
+# False
+deterministic: False
+
+# Sets what view of the environment the deterministic hardcoded agent has. The default is BASIC.
+# Options are:
+# "BASIC" (The current observation space only)
+# "FULL" (Full environment view with actions taken and reward feedback)
+hard_coded_agent_view: FULL
+
+# Sets How the Action Space is defined:
+# "NODE"
+# "ACL"
+# "ANY" node and acl actions
+action_type: NODE
+# observation space
+observation_space:
+  # flatten: true
+  components:
+    - name: NODE_LINK_TABLE
+    # - name: NODE_STATUSES
+    # - name: LINK_TRAFFIC_LEVELS
+# Number of episodes to run per session
+num_episodes: 10
+
+# Number of time_steps per episode
+num_steps: 256
+
+# Sets how often the agent will save a checkpoint (every n time episodes).
+# Set to 0 if no checkpoints are required. Default is 10
+checkpoint_every_n_episodes: 0
+
+# Time delay (milliseconds) between steps for CUSTOM agents.
+time_delay: 5
+
+# Type of session to be run. Options are:
+# "TRAIN" (Trains an agent)
+# "EVAL" (Evaluates an agent)
+# "TRAIN_EVAL" (Trains then evaluates an agent)
+session_type: TRAIN_EVAL
+
+# Environment config values
+# The high value for the observation space
+observation_space_high_value: 1000000000
+
+# The Stable Baselines3 learn/eval output verbosity level:
+# Options are:
+# "NONE" (No Output)
+# "INFO" (Info Messages (such as devices and wrappers used))
+# "DEBUG" (All Messages)
+sb3_output_verbose_level: NONE
+
+# Reward values
+# Generic
+all_ok: 0.0000
+# Node Hardware State
+off_should_be_on: -0.001
+off_should_be_resetting: -0.0005
+on_should_be_off: -0.0002
+on_should_be_resetting: -0.0005
+resetting_should_be_on: -0.0005
+resetting_should_be_off: -0.0002
+resetting: -0.0003
+# Node Software or Service State
+good_should_be_patching: 0.0002
+good_should_be_compromised: 0.0005
+good_should_be_overwhelmed: 0.0005
+patching_should_be_good: -0.0005
+patching_should_be_compromised: 0.0002
+patching_should_be_overwhelmed: 0.0002
+patching: -0.0003
+compromised_should_be_good: -0.002
+compromised_should_be_patching: -0.002
+compromised_should_be_overwhelmed: -0.002
+compromised: -0.002
+overwhelmed_should_be_good: -0.002
+overwhelmed_should_be_patching: -0.002
+overwhelmed_should_be_compromised: -0.002
+overwhelmed: -0.002
+# Node File System State
+good_should_be_repairing: 0.0002
+good_should_be_restoring: 0.0002
+good_should_be_corrupt: 0.0005
+good_should_be_destroyed: 0.001
+repairing_should_be_good: -0.0005
+repairing_should_be_restoring: 0.0002
+repairing_should_be_corrupt: 0.0002
+repairing_should_be_destroyed: 0.0000
+repairing: -0.0003
+restoring_should_be_good: -0.001
+restoring_should_be_repairing: -0.0002
+restoring_should_be_corrupt: 0.0001
+restoring_should_be_destroyed: 0.0002
+restoring: -0.0006
+corrupt_should_be_good: -0.001
+corrupt_should_be_repairing: -0.001
+corrupt_should_be_restoring: -0.001
+corrupt_should_be_destroyed: 0.0002
+corrupt: -0.001
+destroyed_should_be_good: -0.002
+destroyed_should_be_repairing: -0.002
+destroyed_should_be_restoring: -0.002
+destroyed_should_be_corrupt: -0.002
+destroyed: -0.002
+scanning: -0.0002
+# IER status
+red_ier_running: -0.0005
+green_ier_blocked: -0.001
+
+# Patching / Reset durations
+os_patching_duration: 5            # The time taken to patch the OS
+node_reset_duration: 5             # The time taken to reset a node (hardware)
+service_patching_duration: 5       # The time taken to patch a service
+file_system_repairing_limit: 5      # The time take to repair the file system
+file_system_restoring_limit: 5      # The time take to restore the file system
+file_system_scanning_limit: 5       # The time taken to scan the file system

tests/config/ppo_seeded_training_config.yaml

@@ -0,0 +1,155 @@
+# Training Config File
+
+# Sets which agent algorithm framework will be used.
+# Options are:
+# "SB3" (Stable Baselines3)
+# "RLLIB" (Ray RLlib)
+# "CUSTOM" (Custom Agent)
+agent_framework: SB3
+
+# Sets which deep learning framework will be used (by RLlib ONLY).
+# Default is TF (Tensorflow).
+# Options are:
+# "TF" (Tensorflow)
+# TF2 (Tensorflow 2.X)
+# TORCH (PyTorch)
+deep_learning_framework: TF2
+
+# Sets which Agent class will be used.
+# Options are:
+# "A2C" (Advantage Actor Critic coupled with either SB3 or RLLIB agent_framework)
+# "PPO" (Proximal Policy Optimization coupled with either SB3 or RLLIB agent_framework)
+# "HARDCODED" (The HardCoded agents coupled with an ACL or NODE action_type)
+# "DO_NOTHING" (The DoNothing agents coupled with an ACL or NODE action_type)
+# "RANDOM" (primaite.agents.simple.RandomAgent)
+# "DUMMY" (primaite.agents.simple.DummyAgent)
+agent_identifier: PPO
+
+# Sets whether Red Agent POL and IER is randomised.
+# Options are:
+# True
+# False
+random_red_agent: False
+
+# The (integer) seed to be used in random number generation
+# Default is None (null)
+seed: 67890
+
+# Set whether the agent will be deterministic instead of stochastic
+# Options are:
+# True
+# False
+deterministic: True
+
+# Sets what view of the environment the deterministic hardcoded agent has. The default is BASIC.
+# Options are:
+# "BASIC" (The current observation space only)
+# "FULL" (Full environment view with actions taken and reward feedback)
+hard_coded_agent_view: FULL
+
+# Sets How the Action Space is defined:
+# "NODE"
+# "ACL"
+# "ANY" node and acl actions
+action_type: NODE
+# observation space
+observation_space:
+  # flatten: true
+  components:
+    - name: NODE_LINK_TABLE
+    # - name: NODE_STATUSES
+    # - name: LINK_TRAFFIC_LEVELS
+# Number of episodes to run per session
+num_episodes: 10
+
+# Number of time_steps per episode
+num_steps: 256
+
+# Sets how often the agent will save a checkpoint (every n time episodes).
+# Set to 0 if no checkpoints are required. Default is 10
+checkpoint_every_n_episodes: 0
+
+# Time delay (milliseconds) between steps for CUSTOM agents.
+time_delay: 5
+
+# Type of session to be run. Options are:
+# "TRAIN" (Trains an agent)
+# "EVAL" (Evaluates an agent)
+# "TRAIN_EVAL" (Trains then evaluates an agent)
+session_type: TRAIN_EVAL
+
+# Environment config values
+# The high value for the observation space
+observation_space_high_value: 1000000000
+
+# The Stable Baselines3 learn/eval output verbosity level:
+# Options are:
+# "NONE" (No Output)
+# "INFO" (Info Messages (such as devices and wrappers used))
+# "DEBUG" (All Messages)
+sb3_output_verbose_level: NONE
+
+# Reward values
+# Generic
+all_ok: 0
+# Node Hardware State
+off_should_be_on: -10
+off_should_be_resetting: -5
+on_should_be_off: -2
+on_should_be_resetting: -5
+resetting_should_be_on: -5
+resetting_should_be_off: -2
+resetting: -3
+# Node Software or Service State
+good_should_be_patching: 2
+good_should_be_compromised: 5
+good_should_be_overwhelmed: 5
+patching_should_be_good: -5
+patching_should_be_compromised: 2
+patching_should_be_overwhelmed: 2
+patching: -3
+compromised_should_be_good: -20
+compromised_should_be_patching: -20
+compromised_should_be_overwhelmed: -20
+compromised: -20
+overwhelmed_should_be_good: -20
+overwhelmed_should_be_patching: -20
+overwhelmed_should_be_compromised: -20
+overwhelmed: -20
+# Node File System State
+good_should_be_repairing: 2
+good_should_be_restoring: 2
+good_should_be_corrupt: 5
+good_should_be_destroyed: 10
+repairing_should_be_good: -5
+repairing_should_be_restoring: 2
+repairing_should_be_corrupt: 2
+repairing_should_be_destroyed: 0
+repairing: -3
+restoring_should_be_good: -10
+restoring_should_be_repairing: -2
+restoring_should_be_corrupt: 1
+restoring_should_be_destroyed: 2
+restoring: -6
+corrupt_should_be_good: -10
+corrupt_should_be_repairing: -10
+corrupt_should_be_restoring: -10
+corrupt_should_be_destroyed: 2
+corrupt: -10
+destroyed_should_be_good: -20
+destroyed_should_be_repairing: -20
+destroyed_should_be_restoring: -20
+destroyed_should_be_corrupt: -20
+destroyed: -20
+scanning: -2
+# IER status
+red_ier_running: -5
+green_ier_blocked: -10
+
+# Patching / Reset durations
+os_patching_duration: 5            # The time taken to patch the OS
+node_reset_duration: 5             # The time taken to reset a node (hardware)
+service_patching_duration: 5       # The time taken to patch a service
+file_system_repairing_limit: 5      # The time take to repair the file system
+file_system_restoring_limit: 5      # The time take to restore the file system
+file_system_scanning_limit: 5       # The time taken to scan the file system

tests/conftest.py

@@ -58,7 +58,6 @@ class TempPrimaiteSession(PrimaiteSession):
 
     def __exit__(self, type, value, tb):
         shutil.rmtree(self.session_path)
-        shutil.rmtree(self.session_path.parent)
         _LOGGER.debug(f"Deleted temp session directory: {self.session_path}")
 
 

tests/mock_and_patch/get_session_path_mock.py

@@ -1,6 +1,7 @@
 import tempfile
 from datetime import datetime
 from pathlib import Path
+from uuid import uuid4
 
 from primaite import getLogger
 
@@ -14,9 +15,7 @@ def get_temp_session_path(session_timestamp: datetime) -> Path:
     :param session_timestamp: This is the datetime that the session started.
     :return: The session directory path.
     """
-    date_dir = session_timestamp.strftime("%Y-%m-%d")
-    session_path = session_timestamp.strftime("%Y-%m-%d_%H-%M-%S")
-    session_path = Path(tempfile.gettempdir()) / "primaite" / date_dir / session_path
+    session_path = Path(tempfile.gettempdir()) / "primaite" / str(uuid4())
     session_path.mkdir(exist_ok=True, parents=True)
     _LOGGER.debug(f"Created temp session directory: {session_path}")
     return session_path

tests/test_primaite_session.py

@@ -33,6 +33,9 @@ def test_primaite_session(temp_primaite_session):
         # Check that the network png file exists
         assert (session_path / f"network_{session.timestamp_str}.png").exists()
 
+        # Check that the saved agent exists
+        assert session._agent_session._saved_agent_path.exists()
+
         # Check that both the transactions and av reward csv files exist
         for file in session.learning_path.iterdir():
             if file.suffix == ".csv":

tests/test_seeding_and_deterministic_session.py

@@ -0,0 +1,49 @@
+import pytest as pytest
+
+from primaite.config.lay_down_config import dos_very_basic_config_path
+from tests import TEST_CONFIG_ROOT
+
+
+@pytest.mark.parametrize(
+    "temp_primaite_session",
+    [[TEST_CONFIG_ROOT / "ppo_seeded_training_config.yaml", dos_very_basic_config_path()]],
+    indirect=True,
+)
+def test_seeded_learning(temp_primaite_session):
+    """Test running seeded learning produces the same output when ran twice."""
+    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,
+    }
+    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"
+        )
+        session.learn()
+        actual_mean_reward_per_episode = session.learn_av_reward_per_episode()
+
+    assert actual_mean_reward_per_episode == expected_mean_reward_per_episode
+
+
+@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()]],
+    indirect=True,
+)
+def test_deterministic_evaluation(temp_primaite_session):
+    """Test running deterministic evaluation gives same av eward per episode."""
+    with temp_primaite_session as session:
+        # do stuff
+        session.learn()
+        session.evaluate()
+        eval_mean_reward = session.eval_av_reward_per_episode_csv()
+        assert len(set(eval_mean_reward.values())) == 1