Spaces:
Sleeping
Sleeping
File size: 16,938 Bytes
cd61817 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 | """Tests for Phase B3 — Failure Clustering (classifier + archetype extractor)."""
from __future__ import annotations
import json
from pathlib import Path
from unittest.mock import MagicMock, patch
from ci_triage_env.data.clustering import (
FAMILIES,
Archetype,
ArchetypeExtractor,
LLMClassifier,
RuleBasedClassifier,
classify_all,
)
from ci_triage_env.data.datasets._base import FailureRecord
# ---------------------------------------------------------------------------
# Helpers
# ---------------------------------------------------------------------------
def _record(log_text: str, record_id: str = "test-001") -> FailureRecord:
return FailureRecord(
record_id=record_id,
source_dataset="deflaker",
project="test/project",
log_text=log_text,
)
# ---------------------------------------------------------------------------
# RuleBasedClassifier
# ---------------------------------------------------------------------------
class TestRuleBasedClassifier:
def setup_method(self) -> None:
self.clf = RuleBasedClassifier()
def test_rule_based_oom(self) -> None:
record = _record("kernel: Out of memory: Killed process 123 (pytest)")
family, conf = self.clf.classify(record)
assert family == "infra_resource"
assert conf > 0
def test_rule_based_enospc(self) -> None:
record = _record("write /var/lib/docker: no space left on device")
family, conf = self.clf.classify(record)
assert family == "infra_resource"
assert conf > 0
def test_rule_based_emfile(self) -> None:
record = _record("Error: EMFILE: too many open files, open '/tmp/test'")
family, conf = self.clf.classify(record)
assert family == "infra_resource"
assert conf > 0
def test_rule_based_connection_refused(self) -> None:
record = _record("dial tcp: connection refused localhost:5432")
family, conf = self.clf.classify(record)
assert family == "infra_network"
assert conf > 0
def test_rule_based_dns(self) -> None:
record = _record("getaddrinfo failed: No such host is known")
family, conf = self.clf.classify(record)
assert family == "infra_network"
assert conf > 0
def test_rule_based_race(self) -> None:
record = _record("WARNING: DATA RACE detected in goroutine 42")
family, conf = self.clf.classify(record)
assert family == "race_flake"
assert conf > 0
def test_rule_based_concurrent_map(self) -> None:
record = _record("fatal error: concurrent map writes")
family, conf = self.clf.classify(record)
assert family == "race_flake"
assert conf > 0
def test_rule_based_timeout(self) -> None:
record = _record("context canceled: deadline exceeded after 30s")
family, conf = self.clf.classify(record)
assert family == "timing_flake"
assert conf > 0
def test_rule_based_test_timed_out(self) -> None:
record = _record("Test timed out after 120 seconds")
family, conf = self.clf.classify(record)
assert family == "timing_flake"
assert conf > 0
def test_rule_based_dependency_version_conflict(self) -> None:
record = _record("npm ERR! peer dep missing: react@^17.0.0, version conflict")
family, conf = self.clf.classify(record)
assert family == "dependency_drift"
assert conf > 0
def test_rule_based_cargo_lock(self) -> None:
record = _record("error: failed to select a version for the requirement\nCargo.lock conflict")
family, conf = self.clf.classify(record)
assert family == "dependency_drift"
assert conf > 0
def test_rule_based_unknown_returns_unknown(self) -> None:
record = _record("generic test failure with no specific signal")
family, conf = self.clf.classify(record)
assert family == "unknown"
assert conf == 0.0
def test_rule_based_empty_log(self) -> None:
record = _record("")
family, conf = self.clf.classify(record)
assert family == "unknown"
assert conf == 0.0
def test_ambiguous_when_multiple_families_match(self) -> None:
# Both OOM (infra_resource) and DATA RACE (race_flake) in same log
record = _record(
"Out of memory: Killed process 999\nWARNING: DATA RACE in goroutine 1"
)
family, _conf = self.clf.classify(record)
assert family == "ambiguous"
def test_confidence_in_unit_range(self) -> None:
record = _record("kernel: Out of memory: Killed process 1 (pytest)")
_family, conf = self.clf.classify(record)
assert 0.0 <= conf <= 1.0
def test_real_bug_assertion_error(self) -> None:
record = _record("AssertionError: expected 42, got 0\nassert result == expected")
family, conf = self.clf.classify(record)
assert family == "real_bug"
assert conf > 0
# ---------------------------------------------------------------------------
# Archetype round-trip and extraction
# ---------------------------------------------------------------------------
class TestArchetype:
def test_archetype_round_trip(self) -> None:
arch = Archetype(
archetype_id="infra_resource_001",
family="infra_resource",
pattern_summary="OOM-killer terminated test process",
log_template="[{TIMESTAMP}] Out of memory: Killed process {PID}",
slot_distributions={"TIMESTAMP": ["2024-01-01T00:00:00"], "PID": ["1234"]},
informative_tools_hint=["read_logs:kernel"],
minimal_evidence_hint=["cluster_metrics:5m"],
)
data = arch.model_dump()
restored = Archetype.model_validate(data)
assert restored.archetype_id == arch.archetype_id
assert restored.family == arch.family
assert restored.log_template == arch.log_template
def test_archetype_json_round_trip(self) -> None:
arch = Archetype(
archetype_id="race_flake_001",
family="race_flake",
pattern_summary="Concurrent map write panic",
log_template="fatal error: concurrent map writes at {TIMESTAMP}",
slot_distributions={"TIMESTAMP": ["2024-06-01T12:00:00"]},
informative_tools_hint=["read_logs:full"],
minimal_evidence_hint=["read_logs:full"],
)
restored = Archetype.model_validate_json(arch.model_dump_json())
assert restored.family == "race_flake"
class TestArchetypeExtractor:
def setup_method(self) -> None:
self.extractor = ArchetypeExtractor()
def _make_records(self, log_texts: list[str]) -> list[FailureRecord]:
return [
_record(text, record_id=f"rec-{i:03d}")
for i, text in enumerate(log_texts)
]
def test_extract_returns_archetypes(self) -> None:
records = self._make_records([
"kernel: Out of memory: Killed process 123 (pytest) 2024-01-01T10:00:00",
"kernel: Out of memory: Killed process 456 (cargo) 2024-01-02T11:00:00",
"Out of memory: cannot allocate 4096kB",
"OOMKilled: container exceeded memory limit 2048MB",
"killed by OS: out of memory at 2024-01-03T09:00:00",
])
archetypes = self.extractor.extract(records, "infra_resource", n_archetypes=4)
assert len(archetypes) >= 1
for arch in archetypes:
assert arch.family == "infra_resource"
assert arch.archetype_id.startswith("infra_resource_")
assert arch.log_template
assert arch.informative_tools_hint
assert arch.minimal_evidence_hint
def test_extract_with_single_record(self) -> None:
records = self._make_records(["context canceled: deadline exceeded"])
archetypes = self.extractor.extract(records, "timing_flake", n_archetypes=4)
assert len(archetypes) == 1
assert archetypes[0].family == "timing_flake"
def test_extract_empty_records_returns_empty(self) -> None:
archetypes = self.extractor.extract([], "real_bug", n_archetypes=4)
assert archetypes == []
def test_extract_slots_present_in_template(self) -> None:
records = self._make_records([
"2024-03-15T08:30:00 OOMKilled process 99999 after 120.5s",
])
archetypes = self.extractor.extract(records, "infra_resource", n_archetypes=1)
assert len(archetypes) == 1
template = archetypes[0].log_template
# At least one slot should have been extracted
assert "{" in template
def test_slot_distributions_are_populated(self) -> None:
records = self._make_records([
"2024-01-01T10:00:00 Out of memory: Killed process 1234",
"2024-02-01T11:00:00 Out of memory: Killed process 5678",
])
archetypes = self.extractor.extract(records, "infra_resource", n_archetypes=1)
assert len(archetypes) >= 1
# At least one slot family should have distribution values
all_vals = []
for arch in archetypes:
for vals in arch.slot_distributions.values():
all_vals.extend(vals)
assert len(all_vals) > 0
def test_archetype_extraction_from_fixture(self) -> None:
"""Given 5 fixture FailureRecords, extract at least 1 archetype with slots."""
records = self._make_records([
"fatal error: concurrent map writes at goroutine 12",
"WARNING: DATA RACE in goroutine 42 on address 0xc000deadbeef",
"data race detected: read at 0x00c0001234ab by goroutine 7",
"deadlock detected: all goroutines are asleep",
"fatal error: concurrent map iteration and map write",
])
archetypes = self.extractor.extract(records, "race_flake", n_archetypes=4)
assert len(archetypes) >= 1
# Check that at least one archetype has slot placeholders
has_slots = any("{" in a.log_template for a in archetypes)
assert has_slots
def test_n_archetypes_cap(self) -> None:
"""Never return more archetypes than records."""
records = self._make_records(["timeout exceeded", "test timed out after 60s"])
archetypes = self.extractor.extract(records, "timing_flake", n_archetypes=10)
assert len(archetypes) <= len(records)
# ---------------------------------------------------------------------------
# classify_all
# ---------------------------------------------------------------------------
class TestClassifyAll:
def test_classify_all_no_records(self) -> None:
by_family = classify_all([])
assert set(by_family.keys()) == set(FAMILIES)
assert all(len(v) == 0 for v in by_family.values())
def test_classify_all_routes_correctly(self) -> None:
records = [
_record("Out of memory: Killed process 1", "oom-1"),
_record("WARNING: DATA RACE in goroutine 99", "race-1"),
_record("context canceled: deadline exceeded", "timeout-1"),
]
by_family = classify_all(records)
assert len(by_family["infra_resource"]) >= 1
assert len(by_family["race_flake"]) >= 1
assert len(by_family["timing_flake"]) >= 1
def test_classify_all_unknown_falls_into_real_bug(self) -> None:
records = [_record("some completely generic text with no signal", "unknown-1")]
by_family = classify_all(records, openai_api_key=None)
# unknowns with no LLM go to real_bug
assert len(by_family["real_bug"]) == 1
def test_classify_all_returns_all_families(self) -> None:
by_family = classify_all([_record("OOM error")])
assert set(by_family.keys()) == set(FAMILIES)
def test_classify_all_writes_per_family_files(self, tmp_path: Path) -> None:
"""classify_all + extract puts archetype files under each family dir."""
records = [
_record("Out of memory: Killed process 1", "oom-1"),
_record("connection refused to db:5432", "net-1"),
_record("deadline exceeded in test", "timeout-1"),
]
by_family = classify_all(records)
extractor = ArchetypeExtractor()
for family, recs in by_family.items():
if not recs:
continue
archetypes = extractor.extract(recs, family, n_archetypes=2)
family_dir = tmp_path / family
family_dir.mkdir(parents=True, exist_ok=True)
(family_dir / "archetypes.json").write_text(
json.dumps([a.model_dump() for a in archetypes], indent=2)
)
written_families = [d.name for d in tmp_path.iterdir() if d.is_dir()]
assert len(written_families) >= 1
for d in tmp_path.iterdir():
arch_file = d / "archetypes.json"
assert arch_file.exists()
data = json.loads(arch_file.read_text())
assert isinstance(data, list)
assert len(data) >= 1
# ---------------------------------------------------------------------------
# LLMClassifier (mocked)
# ---------------------------------------------------------------------------
class TestLLMClassifier:
def _make_mock_client(self, label: str = "infra_resource") -> MagicMock:
mock_response = MagicMock()
mock_response.choices[0].message.content = label
mock_response.usage.prompt_tokens = 100
mock_response.usage.completion_tokens = 5
return mock_response
def test_llm_classifier_respects_budget(self) -> None:
"""LLMClassifier stops calling after budget exhausted."""
records = [
_record(f"generic unknown failure {i}", f"rec-{i}")
for i in range(5)
]
mock_response = self._make_mock_client("real_bug")
with patch("ci_triage_env.data.clustering.classifier.LLMClassifier.__init__", return_value=None):
clf = LLMClassifier.__new__(LLMClassifier)
clf.budget = 0.0 # already over budget
clf.spent = 0.0
clf.model = "gpt-4o-mini"
mock_client = MagicMock()
mock_client.chat.completions.create.return_value = mock_response
clf.client = mock_client
results = clf.classify_batch(records)
# All results should be ("unknown", 0.0) since budget is exhausted
assert all(family == "unknown" for family, _ in results)
# Client should NOT have been called
mock_client.chat.completions.create.assert_not_called()
def test_llm_classifier_classifies_records(self) -> None:
"""LLMClassifier calls the API and returns valid labels."""
records = [_record("some unknown log text", "unk-1")]
mock_response = self._make_mock_client("real_bug")
with patch("ci_triage_env.data.clustering.classifier.LLMClassifier.__init__", return_value=None):
clf = LLMClassifier.__new__(LLMClassifier)
clf.budget = 5.0
clf.spent = 0.0
clf.model = "gpt-4o-mini"
mock_client = MagicMock()
mock_client.chat.completions.create.return_value = mock_response
clf.client = mock_client
results = clf.classify_batch(records)
assert len(results) == 1
family, conf = results[0]
assert family == "real_bug"
assert conf == 0.7
def test_llm_classifier_invalid_label_falls_back_to_unknown(self) -> None:
records = [_record("something", "x-1")]
mock_response = self._make_mock_client("not_a_valid_label")
with patch("ci_triage_env.data.clustering.classifier.LLMClassifier.__init__", return_value=None):
clf = LLMClassifier.__new__(LLMClassifier)
clf.budget = 5.0
clf.spent = 0.0
clf.model = "gpt-4o-mini"
mock_client = MagicMock()
mock_client.chat.completions.create.return_value = mock_response
clf.client = mock_client
results = clf.classify_batch(records)
family, _ = results[0]
assert family == "unknown"
def test_classify_all_uses_llm_for_unknowns(self) -> None:
"""classify_all routes unknown records through LLM when api_key provided."""
records = [_record("no signal here at all", "unk-1")]
with (
patch("ci_triage_env.data.clustering.classifier.LLMClassifier") as MockLLMClass,
):
mock_llm = MagicMock()
mock_llm.classify_batch.return_value = [("timing_flake", 0.7)]
mock_llm.spent = 0.0001
MockLLMClass.return_value = mock_llm
by_family = classify_all(records, openai_api_key="sk-fake-key")
assert len(by_family["timing_flake"]) == 1
MockLLMClass.assert_called_once_with("sk-fake-key")
|