mirror of
https://github.com/hanzoai/insights.git
synced 2026-10-30 16:15:31 +00:00
The tree was a debranded snapshot of upstream taken around 2026-02-20, and upstream has moved about five months since. This replaces it with upstream master, mechanically re-derived: bin/debrand does the rename, and the pipeline and squashed migrations are restored from the shipped tree. Not a merge. The previous import replaced the tree wholesale in a single commit, so there is no shared content history to merge against. Reproducing the rename on the new tree is what actually converges. What survives: .hanzo/workflows, the Dockerfiles and bin/docker* entrypoints, compose files, LLM.md, and insights/migrations. The squashed initial migration is load-bearing — the live database records it by name, so upstream's ~918 individual migrations cannot come back. Nine products are dropped as unreferenced. Twenty-one more that we do not run are kept anyway because the core app still imports them; unpicking those is a refactor and does not belong in a rebase. This base does not build yet, and the reasons are not in the rename: upstream now wants Python 3.13.13 against our 3.12.12 and Django 5.2 against our 4.2, and the dependency manifests still name upstream packages that we republish under our own names. Those are decisions, not renames.
217 lines
8.3 KiB
Python
217 lines
8.3 KiB
Python
from __future__ import annotations
|
|
|
|
import sys
|
|
import json
|
|
import tempfile
|
|
import importlib.util
|
|
from pathlib import Path
|
|
from types import ModuleType
|
|
|
|
import unittest
|
|
|
|
from parameterized import parameterized
|
|
|
|
SCRIPT_PATH = Path(__file__).with_name("test_selection_verdict.py")
|
|
|
|
|
|
def _load_verdict_module() -> ModuleType:
|
|
spec = importlib.util.spec_from_file_location("test_selection_verdict", SCRIPT_PATH)
|
|
assert spec is not None
|
|
assert spec.loader is not None
|
|
module = importlib.util.module_from_spec(spec)
|
|
sys.modules[spec.name] = module
|
|
spec.loader.exec_module(module)
|
|
return module
|
|
|
|
|
|
def _write_junit(junit_dir: Path, name: str, content: str) -> None:
|
|
junit_dir.mkdir(parents=True, exist_ok=True)
|
|
(junit_dir / name).write_text(content)
|
|
|
|
|
|
def _write_selection(path: Path, *, combined_tests: list[str], full_run_reasons: list[str] | None = None) -> None:
|
|
selection = {
|
|
"combined": {"tests": combined_tests, "count": len(combined_tests)},
|
|
"ast": {"tests": combined_tests, "full_run_reasons": full_run_reasons or []},
|
|
"snob": {"tests": [], "count": 0, "status": "ok"},
|
|
}
|
|
path.write_text(json.dumps(selection))
|
|
|
|
|
|
JUNIT_FAILURE = """<?xml version="1.0" encoding="utf-8"?>
|
|
<testsuites>
|
|
<testsuite name="pytest" tests="2" failures="1" errors="0">
|
|
<testcase classname="insights.api.test.test_foo.TestFoo" name="test_pass" time="0.1"/>
|
|
<testcase classname="insights.api.test.test_foo.TestFoo" name="test_fail" time="0.2">
|
|
<failure message="assert False">Traceback...</failure>
|
|
</testcase>
|
|
</testsuite>
|
|
</testsuites>
|
|
"""
|
|
|
|
JUNIT_NESTED_FAILURE = """<?xml version="1.0" encoding="utf-8"?>
|
|
<testsuites>
|
|
<testsuite name="outer" tests="1" failures="1">
|
|
<testsuite name="inner" tests="1" failures="1">
|
|
<testcase classname="insights.api.test.test_nested.TestNested" name="test_x">
|
|
<error message="boom">Traceback...</error>
|
|
</testcase>
|
|
</testsuite>
|
|
</testsuite>
|
|
</testsuites>
|
|
"""
|
|
|
|
JUNIT_ALL_PASS = """<?xml version="1.0" encoding="utf-8"?>
|
|
<testsuites>
|
|
<testsuite name="pytest" tests="1" failures="0">
|
|
<testcase classname="insights.api.test.test_bar.TestBar" name="test_ok"/>
|
|
</testsuite>
|
|
</testsuites>
|
|
"""
|
|
|
|
|
|
class TestClassnameToFilepath(unittest.TestCase):
|
|
@parameterized.expand(
|
|
[
|
|
(
|
|
"standard_django_class",
|
|
"insights.api.test.test_web_experiment.TestWebExperiment",
|
|
"insights/api/test/test_web_experiment.py",
|
|
),
|
|
(
|
|
"function_test_no_class",
|
|
"insights.api.test.test_module",
|
|
"insights/api/test/test_module.py",
|
|
),
|
|
(
|
|
"ee_path",
|
|
"ee.datastore.queries.test.test_cohort.TestCohort",
|
|
"ee/datastore/queries/test/test_cohort.py",
|
|
),
|
|
(
|
|
"products_path",
|
|
"products.feature_flags.backend.test.test_api.TestFlags",
|
|
"products/feature_flags/backend/test/test_api.py",
|
|
),
|
|
]
|
|
)
|
|
def test_maps_classname(self, _name: str, classname: str, expected: str) -> None:
|
|
verdict = _load_verdict_module()
|
|
self.assertEqual(verdict.classname_to_filepath(classname), expected)
|
|
|
|
|
|
class TestParseJunitFailures(unittest.TestCase):
|
|
def test_returns_zero_when_dir_missing(self) -> None:
|
|
verdict = _load_verdict_module()
|
|
with tempfile.TemporaryDirectory() as root:
|
|
failures, total, seen = verdict.parse_junit_failures(Path(root) / "missing")
|
|
self.assertEqual(failures, [])
|
|
self.assertEqual(total, 0)
|
|
self.assertEqual(seen, 0)
|
|
|
|
def test_extracts_failures_with_classname_mapping(self) -> None:
|
|
verdict = _load_verdict_module()
|
|
with tempfile.TemporaryDirectory() as root:
|
|
junit_dir = Path(root)
|
|
_write_junit(junit_dir, "junit-core.xml", JUNIT_FAILURE)
|
|
failures, total, seen = verdict.parse_junit_failures(junit_dir)
|
|
self.assertEqual(failures, ["insights/api/test/test_foo.py"])
|
|
self.assertEqual(total, 2)
|
|
self.assertEqual(seen, 1)
|
|
|
|
def test_finds_testcases_in_nested_testsuites(self) -> None:
|
|
verdict = _load_verdict_module()
|
|
with tempfile.TemporaryDirectory() as root:
|
|
junit_dir = Path(root)
|
|
_write_junit(junit_dir, "junit.xml", JUNIT_NESTED_FAILURE)
|
|
failures, total, _seen = verdict.parse_junit_failures(junit_dir)
|
|
self.assertEqual(failures, ["insights/api/test/test_nested.py"])
|
|
self.assertEqual(total, 1)
|
|
|
|
def test_skips_malformed_xml(self) -> None:
|
|
verdict = _load_verdict_module()
|
|
with tempfile.TemporaryDirectory() as root:
|
|
junit_dir = Path(root)
|
|
_write_junit(junit_dir, "good.xml", JUNIT_ALL_PASS)
|
|
_write_junit(junit_dir, "bad.xml", "<not valid xml")
|
|
failures, total, seen = verdict.parse_junit_failures(junit_dir)
|
|
self.assertEqual(failures, [])
|
|
self.assertEqual(total, 1)
|
|
self.assertEqual(seen, 2)
|
|
|
|
|
|
class TestComputeVerdict(unittest.TestCase):
|
|
def _run(
|
|
self,
|
|
*,
|
|
combined_tests: list[str],
|
|
full_run_reasons: list[str] | None = None,
|
|
junit_files: list[tuple[str, str]] | None = None,
|
|
) -> dict[str, object]:
|
|
verdict = _load_verdict_module()
|
|
with tempfile.TemporaryDirectory() as root:
|
|
tmp_path = Path(root)
|
|
selection_path = tmp_path / "selection.json"
|
|
_write_selection(selection_path, combined_tests=combined_tests, full_run_reasons=full_run_reasons)
|
|
|
|
junit_dir = tmp_path / "junit"
|
|
for name, content in junit_files or []:
|
|
_write_junit(junit_dir, name, content)
|
|
|
|
return verdict.compute_verdict(selection_path, junit_dir)
|
|
|
|
def test_unknown_when_no_junit_xmls_found(self) -> None:
|
|
result = self._run(combined_tests=["insights/api/test/test_foo.py"])
|
|
self.assertEqual(result["backend_conclusion"], "unknown")
|
|
self.assertIsNone(result["recall"])
|
|
self.assertEqual(result["caught"], [])
|
|
self.assertEqual(result["missed"], [])
|
|
self.assertEqual(result["junit_xml_files_seen"], 0)
|
|
|
|
def test_success_when_all_pass(self) -> None:
|
|
result = self._run(
|
|
combined_tests=["insights/api/test/test_bar.py"],
|
|
junit_files=[("junit.xml", JUNIT_ALL_PASS)],
|
|
)
|
|
self.assertEqual(result["backend_conclusion"], "success")
|
|
self.assertIsNone(result["recall"])
|
|
self.assertEqual(result["failure_count"], 0)
|
|
|
|
def test_failure_caught_when_in_selection(self) -> None:
|
|
result = self._run(
|
|
combined_tests=["insights/api/test/test_foo.py", "insights/api/test/test_other.py"],
|
|
junit_files=[("junit.xml", JUNIT_FAILURE)],
|
|
)
|
|
self.assertEqual(result["backend_conclusion"], "failure")
|
|
self.assertEqual(result["recall"], 1.0)
|
|
self.assertEqual(result["caught"], ["insights/api/test/test_foo.py"])
|
|
self.assertEqual(result["missed"], [])
|
|
|
|
def test_failure_missed_when_not_in_selection(self) -> None:
|
|
result = self._run(
|
|
combined_tests=["insights/api/test/test_unrelated.py"],
|
|
junit_files=[("junit.xml", JUNIT_FAILURE)],
|
|
)
|
|
self.assertEqual(result["backend_conclusion"], "failure")
|
|
self.assertEqual(result["recall"], 0.0)
|
|
self.assertEqual(result["caught"], [])
|
|
self.assertEqual(result["missed"], ["insights/api/test/test_foo.py"])
|
|
|
|
def test_full_run_treats_failures_as_caught(self) -> None:
|
|
# Selection's combined.tests does NOT include the failing file, but
|
|
# full_run_reasons indicates the selector opted into running everything.
|
|
result = self._run(
|
|
combined_tests=["insights/api/test/test_unrelated.py"],
|
|
full_run_reasons=["conftest.py matches full-run pattern"],
|
|
junit_files=[("junit.xml", JUNIT_FAILURE)],
|
|
)
|
|
self.assertEqual(result["backend_conclusion"], "failure")
|
|
self.assertTrue(result["full_run_triggered"])
|
|
self.assertEqual(result["recall"], 1.0)
|
|
self.assertEqual(result["caught"], ["insights/api/test/test_foo.py"])
|
|
self.assertEqual(result["missed"], [])
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|