From 70bea0c9e179f5997ba4db5aab4903184f28458d Mon Sep 17 00:00:00 2001 From: eust-w Date: Sat, 11 Jul 2026 06:13:43 +0800 Subject: [PATCH] test: re-audit reconstruction evidence across builds --- scripts/acceptance_poc_cluster.py | 19 +++++++-- tests/test_production_platform.py | 67 +++++++++++++++++++++++++++++++ 2 files changed, 83 insertions(+), 3 deletions(-) diff --git a/scripts/acceptance_poc_cluster.py b/scripts/acceptance_poc_cluster.py index 0852a16..dc0b594 100755 --- a/scripts/acceptance_poc_cluster.py +++ b/scripts/acceptance_poc_cluster.py @@ -524,8 +524,9 @@ def resume_reconstruction_evidence(run: AcceptanceRun, report_path: Path) -> tup previous = json.loads(report_path.read_text(encoding="utf-8")) if previous.get("schema") != "real2sim.poc_acceptance_report.v1": raise AcceptanceError(f"unsupported resume report schema: {previous.get('schema')}") - if previous.get("build_id") != run.build_id: - raise AcceptanceError(f"resume report build_id {previous.get('build_id')} does not match {run.build_id}") + source_build_id = str(previous.get("build_id") or "") + if not source_build_id: + raise AcceptanceError("resume report does not identify its source build") prepared = next((item for item in previous.get("datasets", []) if item.get("role") == "playroom_prepared"), None) scenes = [item for item in previous.get("scenes", []) if str(item.get("role") or "").startswith("playroom_reconstruction_")] @@ -568,7 +569,19 @@ def resume_reconstruction_evidence(run: AcceptanceRun, report_path: Path) -> tup run.report["datasets"] = [item for item in previous.get("datasets", []) if item.get("role") in {"playroom_upload", "playroom_prepared"}] run.report["scenes"] = scenes run.report["storage"] = dict(previous.get("storage") or {}) - run.report["resumed_from"] = str(report_path.resolve()) + run.report["resumed_evidence"] = { + "report": str(report_path.resolve()), + "source_build_id": source_build_id, + "target_build_id": run.build_id, + "cross_build": source_build_id != run.build_id, + "reaudited_task_ids": [ + str(task["id"]) + for task_type, count in required_counts.items() + for task in by_type[task_type][:count] + ], + "prepared_dataset_id": prepared_id, + "scene_ids": scene_ids, + } run.save() return prepared_id, scene_ids diff --git a/tests/test_production_platform.py b/tests/test_production_platform.py index 7d16f32..68c5e0c 100644 --- a/tests/test_production_platform.py +++ b/tests/test_production_platform.py @@ -387,6 +387,73 @@ def test_source_revision_rejects_unverified_archive_marker(tmp_path: Path) -> No assert read_source_revision(root) == "" +def test_acceptance_reaudits_reconstruction_evidence_across_builds(tmp_path: Path) -> None: + script_path = ROOT / "scripts/acceptance_poc_cluster.py" + spec = importlib.util.spec_from_file_location("real2sim_acceptance_poc_cluster", script_path) + assert spec and spec.loader + module = importlib.util.module_from_spec(spec) + spec.loader.exec_module(module) + tasks = [ + {"id": "media", "type": "media_prepare"}, + {"id": "preflight", "type": "reconstruction_preflight"}, + {"id": "recon-a", "type": "scene_reconstruction"}, + {"id": "recon-b", "type": "scene_reconstruction"}, + {"id": "sim-a", "type": "simulator_validation"}, + {"id": "sim-b", "type": "simulator_validation"}, + ] + report_path = tmp_path / "prior-report.json" + report_path.write_text( + json.dumps( + { + "schema": "real2sim.poc_acceptance_report.v1", + "build_id": "prior-build", + "tasks": tasks, + "datasets": [{"id": "prepared", "role": "playroom_prepared"}], + "scenes": [ + {"id": "scene-a", "role": "playroom_reconstruction_1"}, + {"id": "scene-b", "role": "playroom_reconstruction_2"}, + ], + "storage": {"source_image_count": 225}, + } + ), + encoding="utf-8", + ) + + class FakeClient: + def get(self, path: str) -> dict: + if path == "/api/datasets/prepared/manifest": + return {"frame_count": 225} + return {"assets": [{"name": "scene.ply"}, {"name": "scene.xml"}]} + + class FakeRun: + build_id = "current-build" + client = FakeClient() + + def __init__(self) -> None: + self.report = {"datasets": [], "scenes": [], "storage": {}, "checks": []} + self.audited: list[str] = [] + + def wait_task(self, task_id: str, required_artifacts: tuple[str, ...]) -> dict: + assert required_artifacts + self.audited.append(task_id) + return {"id": task_id, "status": "succeeded"} + + def check(self, name: str, ok: bool, detail=None) -> None: + assert ok, (name, detail) + self.report["checks"].append({"name": name, "ok": ok}) + + def save(self) -> None: + return None + + run = FakeRun() + prepared_id, scene_ids = module.resume_reconstruction_evidence(run, report_path) + assert prepared_id == "prepared" + assert scene_ids == ["scene-a", "scene-b"] + assert run.audited == ["media", "preflight", "recon-a", "recon-b", "sim-a", "sim-b"] + assert run.report["resumed_evidence"]["cross_build"] is True + assert run.report["resumed_evidence"]["source_build_id"] == "prior-build" + + def test_motrixsim_runtime_scene_uses_current_python_when_module_is_installed(monkeypatch: pytest.MonkeyPatch) -> None: script_path = ROOT / "scripts/generate_motrixsim_runtime_scene.py" spec = importlib.util.spec_from_file_location("real2sim_generate_motrixsim_runtime_scene", script_path)