Spaces:
Runtime error
Runtime error
File size: 8,616 Bytes
f36b499 | 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 | """Tests for trajectory analytics (issue #22)."""
from __future__ import annotations
import json
from pathlib import Path
import pytest
from open_range.training.analytics import TrajectoryAnalyzer
# ---------------------------------------------------------------------------
# Fixtures
# ---------------------------------------------------------------------------
def _make_record(
episode_id: str = "ep-001",
role: str = "red",
reward: float = 0.5,
outcome: str = "red_win",
tier: int = 1,
vuln_class: str | None = None,
n_steps: int = 3,
) -> dict:
"""Build a single JSONL record matching TrajectoryLogger output format."""
messages = [
{"role": "system", "content": "You are a pentester."},
]
for i in range(n_steps):
messages.append({"role": "user", "content": f"observation {i}"})
messages.append({"role": "assistant", "content": f"command {i}"})
record = {
"episode_id": episode_id,
"snapshot_id": "snap-001",
"tier": tier,
"role": role,
"messages": messages,
"reward": reward,
"outcome": outcome,
}
if vuln_class is not None:
record["vuln_class"] = vuln_class
return record
@pytest.fixture
def sample_jsonl(tmp_path) -> Path:
"""Create a sample JSONL file with multiple records."""
records = [
_make_record("ep-001", "red", 1.0, "red_win", vuln_class="sqli", n_steps=4),
_make_record("ep-001", "blue", 0.2, "red_win", n_steps=4),
_make_record("ep-002", "red", 0.0, "timeout", vuln_class="sqli", n_steps=6),
_make_record("ep-002", "blue", 0.8, "timeout", n_steps=6),
_make_record("ep-003", "red", 0.5, "red_win", vuln_class="xss", n_steps=3),
_make_record("ep-003", "blue", 0.3, "red_win", n_steps=3),
]
path = tmp_path / "trajectories.jsonl"
with open(path, "w") as f:
for r in records:
f.write(json.dumps(r) + "\n")
return path
@pytest.fixture
def compare_jsonl(tmp_path) -> Path:
"""Create a second JSONL file for comparison."""
records = [
_make_record("ep-101", "red", 1.0, "red_win", vuln_class="sqli", n_steps=2),
_make_record("ep-101", "blue", 0.5, "red_win", n_steps=2),
_make_record("ep-102", "red", 1.0, "red_win", vuln_class="xss", n_steps=3),
_make_record("ep-102", "blue", 0.7, "red_win", n_steps=3),
]
path = tmp_path / "compare.jsonl"
with open(path, "w") as f:
for r in records:
f.write(json.dumps(r) + "\n")
return path
# ---------------------------------------------------------------------------
# Tests: loading
# ---------------------------------------------------------------------------
class TestLoading:
def test_load_returns_count(self, sample_jsonl):
analyzer = TrajectoryAnalyzer()
count = analyzer.load(sample_jsonl)
assert count == 6
def test_records_accessible(self, sample_jsonl):
analyzer = TrajectoryAnalyzer()
analyzer.load(sample_jsonl)
assert len(analyzer.records) == 6
def test_load_accumulates(self, sample_jsonl, compare_jsonl):
analyzer = TrajectoryAnalyzer()
analyzer.load(sample_jsonl)
analyzer.load(compare_jsonl)
assert len(analyzer.records) == 10
def test_empty_analyzer(self):
analyzer = TrajectoryAnalyzer()
s = analyzer.summary()
assert s["total_episodes"] == 0
assert s["total_records"] == 0
# ---------------------------------------------------------------------------
# Tests: summary
# ---------------------------------------------------------------------------
class TestSummary:
def test_total_episodes(self, sample_jsonl):
analyzer = TrajectoryAnalyzer()
analyzer.load(sample_jsonl)
s = analyzer.summary()
assert s["total_episodes"] == 3 # ep-001, ep-002, ep-003
def test_total_records(self, sample_jsonl):
analyzer = TrajectoryAnalyzer()
analyzer.load(sample_jsonl)
s = analyzer.summary()
assert s["total_records"] == 6
def test_outcomes(self, sample_jsonl):
analyzer = TrajectoryAnalyzer()
analyzer.load(sample_jsonl)
s = analyzer.summary()
assert s["outcomes"]["red_win"] == 2
assert s["outcomes"]["timeout"] == 1
def test_avg_reward(self, sample_jsonl):
analyzer = TrajectoryAnalyzer()
analyzer.load(sample_jsonl)
s = analyzer.summary()
# (1.0 + 0.2 + 0.0 + 0.8 + 0.5 + 0.3) / 6 = 2.8 / 6 = 0.4667
assert abs(s["avg_reward"] - 0.4667) < 0.01
def test_avg_steps(self, sample_jsonl):
analyzer = TrajectoryAnalyzer()
analyzer.load(sample_jsonl)
s = analyzer.summary()
# Steps are counted as assistant messages: 4, 4, 6, 6, 3, 3
expected = (4 + 4 + 6 + 6 + 3 + 3) / 6
assert abs(s["avg_steps"] - expected) < 0.1
def test_per_role_stats(self, sample_jsonl):
analyzer = TrajectoryAnalyzer()
analyzer.load(sample_jsonl)
s = analyzer.summary()
assert "red" in s["per_role"]
assert "blue" in s["per_role"]
assert s["per_role"]["red"]["count"] == 3
assert s["per_role"]["blue"]["count"] == 3
# ---------------------------------------------------------------------------
# Tests: by_vuln_class
# ---------------------------------------------------------------------------
class TestByVulnClass:
def test_vuln_class_breakdown(self, sample_jsonl):
analyzer = TrajectoryAnalyzer()
analyzer.load(sample_jsonl)
vc = analyzer.by_vuln_class()
assert "sqli" in vc
assert "xss" in vc
# sqli: ep-001 (red_win), ep-002 (timeout) -> 2 attempts, 1 solve
assert vc["sqli"]["attempts"] == 2
assert vc["sqli"]["solves"] == 1
assert abs(vc["sqli"]["solve_rate"] - 0.5) < 0.01
# xss: ep-003 (red_win) -> 1 attempt, 1 solve
assert vc["xss"]["attempts"] == 1
assert vc["xss"]["solves"] == 1
assert vc["xss"]["solve_rate"] == 1.0
def test_no_vuln_class_returns_empty(self, tmp_path):
"""Records without vuln_class produce empty breakdown."""
path = tmp_path / "no_vuln.jsonl"
record = _make_record(vuln_class=None)
with open(path, "w") as f:
f.write(json.dumps(record) + "\n")
analyzer = TrajectoryAnalyzer()
analyzer.load(path)
assert analyzer.by_vuln_class() == {}
# ---------------------------------------------------------------------------
# Tests: compare
# ---------------------------------------------------------------------------
class TestCompare:
def test_compare_produces_diffs(self, sample_jsonl, compare_jsonl):
a1 = TrajectoryAnalyzer()
a1.load(sample_jsonl)
a2 = TrajectoryAnalyzer()
a2.load(compare_jsonl)
diff = a1.compare(a2)
assert "avg_reward_diff" in diff
assert "total_episodes_diff" in diff
assert diff["total_episodes_diff"] == 2 - 3 # 2 episodes vs 3
def test_compare_reward_direction(self, sample_jsonl, compare_jsonl):
a1 = TrajectoryAnalyzer()
a1.load(sample_jsonl)
a2 = TrajectoryAnalyzer()
a2.load(compare_jsonl)
diff = a1.compare(a2)
# compare has higher rewards on average
assert diff["avg_reward_compare"] > diff["avg_reward_baseline"]
# ---------------------------------------------------------------------------
# Tests: report
# ---------------------------------------------------------------------------
class TestReport:
def test_report_is_string(self, sample_jsonl):
analyzer = TrajectoryAnalyzer()
analyzer.load(sample_jsonl)
report = analyzer.report()
assert isinstance(report, str)
assert "OpenRange Trajectory Analysis Report" in report
def test_report_contains_stats(self, sample_jsonl):
analyzer = TrajectoryAnalyzer()
analyzer.load(sample_jsonl)
report = analyzer.report()
assert "Total episodes:" in report
assert "Average reward:" in report
assert "Outcomes:" in report
def test_report_contains_vuln_breakdown(self, sample_jsonl):
analyzer = TrajectoryAnalyzer()
analyzer.load(sample_jsonl)
report = analyzer.report()
assert "sqli" in report
assert "xss" in report
def test_empty_report(self):
analyzer = TrajectoryAnalyzer()
report = analyzer.report()
assert "Total episodes: 0" in report
|