File size: 9,921 Bytes
b14c6e3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c18a9d1
 
b14c6e3
c18a9d1
b14c6e3
 
 
 
c18a9d1
 
b14c6e3
c18a9d1
b14c6e3
 
 
 
c18a9d1
 
b14c6e3
c18a9d1
b14c6e3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
"""
Unit Tests for Adaptive Alert Triage Environment

Tests core environment functionality: reset, step, state management.
"""

import pytest
import numpy as np
from adaptive_alert_triage.env import AdaptiveAlertTriageEnv
from adaptive_alert_triage.models import Action, Observation, Reward


class TestEnvironmentBasics:
    """Test basic environment operations."""
    
    def test_initialization(self):
        """Test environment initialization with different tasks."""
        for task_id in ["easy", "medium", "hard"]:
            env = AdaptiveAlertTriageEnv(task_id=task_id, seed=42)
            assert env.task_id == task_id
            assert env.current_step == 0
    
    def test_reset(self):
        """Test environment reset returns valid observation."""
        env = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
        obs = env.reset()
        
        assert isinstance(obs, Observation)
        assert len(obs.alerts) > 0
        assert 0.0 <= obs.system_load <= 1.0
        assert obs.queue_length >= 0
        assert obs.time_remaining == env.max_steps
        assert obs.episode_step == 0
    
    def test_reset_reproducibility(self):
        """Test that same seed produces same initial state."""
        env1 = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
        obs1 = env1.reset()
        
        env2 = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
        obs2 = env2.reset()
        
        assert len(obs1.alerts) == len(obs2.alerts)
        assert obs1.system_load == obs2.system_load
    
    def test_step_basic(self):
        """Test basic step execution."""
        env = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
        obs = env.reset()
        
        # Take action on first alert
        action = Action(alert_id=obs.alerts[0].id, action_type="INVESTIGATE")
        next_obs, reward, done, info = env.step(action)
        
        assert isinstance(next_obs, Observation)
        assert isinstance(reward, Reward)
        assert isinstance(done, bool)
        assert isinstance(info, dict)
        assert env.current_step == 1
    
    def test_step_invalid_alert(self):
        """Test step with invalid alert ID."""
        env = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
        obs = env.reset()
        
        # Invalid alert ID
        action = Action(alert_id="nonexistent_alert", action_type="INVESTIGATE")
        next_obs, reward, done, info = env.step(action)
        
        assert reward.value < 0  # Should be penalized
        assert done  # Should terminate on invalid action
    
    def test_episode_termination(self):
        """Test episode terminates at max_steps."""
        env = AdaptiveAlertTriageEnv(task_id="easy", max_steps=5, seed=42)
        obs = env.reset()
        
        done = False
        steps = 0
        
        while not done and steps < 10:  # Safety limit
            if obs.alerts:
                action = Action(alert_id=obs.alerts[0].id, action_type="INVESTIGATE")
                obs, reward, done, info = env.step(action)
            else:
                break
            steps += 1
        
        assert steps <= 5 or done
    
    def test_state_method(self):
        """Test state() returns complete episode state."""
        env = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
        obs = env.reset()
        
        state = env.state()
        
        assert state.observation.episode_step == obs.episode_step
        assert "true_severities" in state.hidden_state
        assert state.cumulative_reward == 0.0
        assert state.failures_count == 0


class TestTaskConfigurations:
    """Test task-specific configurations."""
    
    def test_easy_task_config(self):
        """Test easy task has correct configuration."""
        env = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
    
        assert env.max_steps == 10
        assert env.max_investigations_per_step is None  # No resource constraint
        assert env.failure_threshold == 2
    
    def test_medium_task_config(self):
        """Test medium task has resource constraints."""
        env = AdaptiveAlertTriageEnv(task_id="medium", seed=42)
    
        assert env.max_steps == 15
        assert env.max_investigations_per_step == 3  # Resource constrained
        assert env.failure_threshold == 3
    
    def test_hard_task_config(self):
        """Test hard task has stricter failure tolerance."""
        env = AdaptiveAlertTriageEnv(task_id="hard", seed=42)
    
        assert env.max_steps == 20
        assert env.max_investigations_per_step == 3
        assert env.failure_threshold == 2  # Stricter
    
    def test_resource_budget_tracking(self):
        """Test resource budget is tracked in medium/hard tasks."""
        env = AdaptiveAlertTriageEnv(task_id="medium", seed=42)
        obs = env.reset()
        
        # Resource budget should be visible in observation
        assert obs.resource_budget is not None
        assert obs.resource_budget == 3


class TestObservability:
    """Test partial observability - hidden vs visible attributes."""
    
    def test_hidden_attributes_not_exposed(self):
        """Test that true_severity and correlations are hidden from agent."""
        env = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
        obs = env.reset()
        
        # Get full state (includes hidden info)
        state = env.state()
        
        # Check that observation doesn't expose hidden attributes
        for alert in obs.alerts:
            # These should be zeroed out or default in observation
            assert alert.true_severity == 0.0  # Hidden
            assert alert.is_correlated == False  # Hidden
        
        # But they should exist in hidden state
        assert "true_severities" in state.hidden_state
    
    def test_visible_attributes_noisy(self):
        """Test that visible_severity differs from true_severity (noise)."""
        env = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
        env.reset()
        
        # Access internal alerts (with true values)
        internal_alerts = env.alerts
        
        if internal_alerts:
            # Visible and true severity should differ due to noise
            # (This is probabilistic, so we check that at least some differ)
            has_noise = any(
                abs(a.visible_severity - a.true_severity) > 0.01
                for a in internal_alerts
            )
            assert has_noise, "Expected observation noise in severity"


class TestAlertDynamics:
    """Test alert generation and aging dynamics."""
    
    def test_alerts_age_over_time(self):
        """Test that unhandled alerts age."""
        env = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
        obs = env.reset()
        
        if len(obs.alerts) < 2:
            pytest.skip("Need multiple alerts for this test")
        
        # Handle only first alert
        first_alert_id = obs.alerts[0].id
        second_alert_initial_age = obs.alerts[1].age
        
        action = Action(alert_id=first_alert_id, action_type="INVESTIGATE")
        next_obs, _, _, _ = env.step(action)
        
        # Second alert should still exist and be older
        second_alert_new = next((a for a in next_obs.alerts if a.age > second_alert_initial_age), None)
        assert second_alert_new is not None, "Alert should have aged"
    
    def test_new_alerts_generated(self):
        """Test that new alerts are generated over time."""
        env = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
        obs = env.reset()
        
        initial_alert_ids = {a.id for a in obs.alerts}
        
        # Take several steps
        for _ in range(5):
            if obs.alerts:
                action = Action(alert_id=obs.alerts[0].id, action_type="INVESTIGATE")
                obs, _, done, _ = env.step(action)
                if done:
                    break
        
        # Should have some new alert IDs
        current_alert_ids = {a.id for a in obs.alerts}
        new_alerts = current_alert_ids - initial_alert_ids
        
        # Probabilistic, but likely to have new alerts after 5 steps
        assert len(new_alerts) > 0, "Expected new alerts to be generated"


class TestRewardSignals:
    """Test reward calculation."""
    
    def test_positive_reward_for_critical(self):
        """Test that investigating critical alerts gives positive reward."""
        env = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
        obs = env.reset()
        
        # Find a high-severity alert (visible)
        high_sev_alert = max(obs.alerts, key=lambda a: a.visible_severity)
        
        action = Action(alert_id=high_sev_alert.id, action_type="INVESTIGATE")
        _, reward, _, _ = env.step(action)
        
        # Likely to be positive for high visible severity
        # (Not guaranteed due to noise, but statistically likely)
        assert isinstance(reward.value, float)
    
    def test_reward_components(self):
        """Test that reward has component breakdown."""
        env = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
        obs = env.reset()
        
        action = Action(alert_id=obs.alerts[0].id, action_type="INVESTIGATE")
        _, reward, _, _ = env.step(action)
        
        assert isinstance(reward.components, dict)
        assert len(reward.components) > 0


def test_render(capsys):
    """Test render method produces output."""
    env = AdaptiveAlertTriageEnv(task_id="easy", seed=42)
    obs = env.reset()
    
    # Test human mode (prints)
    env.render(mode="human")
    captured = capsys.readouterr()
    assert "Step" in captured.out
    assert "Failures" in captured.out
    
    # Test ansi mode (returns string)
    output = env.render(mode="ansi")
    assert isinstance(output, str)
    assert "Step" in output


if __name__ == "__main__":
    # Run tests with pytest
    pytest.main([__file__, "-v"])