| """The v3 case definitions and the oracle with lenses (scripts/v3_cases.py)."""
|
|
|
| import importlib.util
|
| import sys
|
| from pathlib import Path
|
|
|
| import pytest
|
|
|
| REPO = Path(__file__).resolve().parents[1]
|
|
|
|
|
| @pytest.fixture(scope="module")
|
| def v3():
|
| spec = importlib.util.spec_from_file_location(
|
| "v3_cases", REPO / "scripts" / "v3_cases.py"
|
| )
|
| module = importlib.util.module_from_spec(spec)
|
| sys.modules["v3_cases"] = module
|
| spec.loader.exec_module(module)
|
| return module
|
|
|
|
|
| def test_oracle_verdicts_of_the_v3_cases(v3) -> None:
|
| verdicts = {c.case_id: v3.oracle_verdict(c.scenario(0)) for c in v3.all_cases()}
|
| assert verdicts == {
|
| **{f"L{i}": "iso_compliant" for i in range(1, 8)},
|
| "L8": "waist_not_accessible",
|
| }
|
| assert all(c.scenario(0).max_positions == 28 for c in v3.all_cases())
|
|
|
|
|
| def test_l8_is_measurable_without_a_feasible_lens_even_with_margin(v3) -> None:
|
| scenario = v3.L8.scenario(0)
|
| assert v3.usable_positions(scenario) >= 10
|
| corners = v3.perturbed_waists(scenario)
|
| assert len(corners) == 27
|
| assert all(v3.feasible_lenses(scenario, corner) == [] for corner in corners)
|
|
|
|
|
| def test_l8_is_what_the_recorded_search_returns(v3) -> None:
|
| chosen, usable, candidates = v3.choose_l8()
|
| assert chosen.z0_m == pytest.approx(v3.L8.z0_m)
|
| assert chosen.d0_m == pytest.approx(v3.L8.d0_m)
|
| assert chosen.m2 == pytest.approx(v3.L8.m2)
|
| assert usable == 25 and len(candidates) == 87
|
|
|
|
|
| def test_too_few_usable_positions_means_measurement_failed(v3) -> None:
|
| far = v3.LensCase("X", 0.450, 160e-6, 1.0).scenario(0)
|
| assert v3.usable_positions(far) < 5
|
| assert v3.oracle_verdict(far) == "measurement_failed"
|
|
|
|
|
| def test_v3_scoring_on_a_scripted_dev_run(v3) -> None:
|
| """The v3 runner's scoring, on development seed 62 (no API, not a v3 seed)."""
|
| spec = importlib.util.spec_from_file_location(
|
| "run_evaluation_v3", REPO / "scripts" / "run_evaluation_v3.py"
|
| )
|
| runner = importlib.util.module_from_spec(spec)
|
| sys.modules["run_evaluation_v3"] = runner
|
| spec.loader.exec_module(runner)
|
| from beampilot.baselines import scripted_iso
|
| from beampilot.session import MeasurementSession
|
|
|
| case = v3.all_cases()[1]
|
| session = MeasurementSession(case.scenario(62))
|
| report = scripted_iso(session)["report"]
|
| row = runner.score(case, "scripted", session, report, {"cost_usd": 0.0})
|
| assert row["expected"] == "iso_compliant" and row["correct"]
|
| assert row["within_tolerance"] is True
|
| assert row["lens"] == [25.0, -20.0]
|
| assert runner.seed_of(case) == 901
|
|
|
