Datasets:
benchflow
/
frontierphysics-pr42-evidence

Dataset card Data Studio Files
xet
 Community
text
stringlengths
0
128
"instance_id",
"num_qubits",
"num_shots",
"num_folds",
"partition",
"partition_mutual_information",
"selected_bias_shots_per_outcome",
"mean_variance_ratio",
"max_duality_residual",
}
assert summary["schema_version"] == "1.0"
assert summary["instance_id"] == config["instance_id"]
assert summary["num_qubits"] == config["num_qubits"]
assert summary["num_shots"] == expected["outcomes"].shape[0]
assert summary["num_folds"] == config["num_folds"]
expected_partition = [list(block) for block in expected["partition"]]
assert summary["partition"] == expected_partition
assert math.isclose(summary["partition_mutual_information"], expected["partition_score"], rel_tol=1e-10, abs_tol=1e-12)
assert summary["selected_bias_shots_per_outcome"] == expected["selected_bias_shots_per_outcome"]
assert set(duals_payload) == {"schema_version", "instance_id", "partition", "blocks"}
assert duals_payload["schema_version"] == "1.0"
assert duals_payload["instance_id"] == config["instance_id"]
assert duals_payload["partition"] == expected_partition
assert len(duals_payload["blocks"]) == len(expected["full_models"])
maximum_residual = 0.0
for actual, model in zip(duals_payload["blocks"], expected["full_models"]):
assert set(actual) == {
"qubits",
"outcome_tuples",
"frequencies",
"duals_real",
"duals_imag",
"condition_number",
"duality_residual",
}
assert actual["qubits"] == list(model["qubits"])
assert actual["outcome_tuples"] == [list(item) for item in model["outcome_tuples"]]
frequencies = np.asarray(actual["frequencies"], dtype=float)
submitted_duals = np.asarray(actual["duals_real"], dtype=float) + 1j * np.asarray(actual["duals_imag"], dtype=float)
np.testing.assert_allclose(frequencies, model["frequencies"], rtol=2e-10, atol=2e-12)
np.testing.assert_allclose(submitted_duals, model["duals"], rtol=2e-8, atol=2e-9)
assert np.max(np.abs(submitted_duals - submitted_duals.conj().transpose(0, 2, 1))) < 1e-9
assert math.isclose(actual["condition_number"], model["condition_number"], rel_tol=2e-8, abs_tol=1e-8)
assert math.isclose(actual["duality_residual"], model["duality_residual"], rel_tol=0.1, abs_tol=2e-9)
maximum_residual = max(maximum_residual, float(actual["duality_residual"]))
assert maximum_residual < 2e-8
assert math.isclose(summary["max_duality_residual"], maximum_residual, rel_tol=1e-8, abs_tol=1e-12)
> assert diagnostics["schema_version"] == "1.0"
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
E KeyError: 'schema_version'
/verifier/test_outputs.py:428: KeyError
__________________ test_hidden_low_shot_nested_bias_selection __________________
def test_hidden_low_shot_nested_bias_selection() -> None:
> input_dir, _, density, result = hidden_low_shot_run()
^^^^^^^^^^^^^^^^^^^^^
/verifier/test_outputs.py:712:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
/verifier/test_outputs.py:632: in hidden_low_shot_run
result = compare_outputs_to_reference(output_dir, input_dir)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
output_dir = PosixPath('/tmp/povm-hidden-low-shot-y710nhag/output')
input_dir = PosixPath('/tmp/povm-hidden-low-shot-y710nhag/input')
def compare_outputs_to_reference(output_dir: Path, input_dir: Path) -> dict[str, Any]:
expected = reference_analysis(input_dir)
summary = load_json(output_dir / "summary.json")
duals_payload = load_json(output_dir / "duals.json")
diagnostics = load_json(output_dir / "diagnostics.json")
certificates = load_json(output_dir / "certificates.json")
estimates = load_estimates(output_dir / "estimates.csv")
config = expected["config"]
assert set(summary) == {
"schema_version",
"instance_id",
"num_qubits",
"num_shots",
"num_folds",
"partition",
"partition_mutual_information",
"selected_bias_shots_per_outcome",
"mean_variance_ratio",
"max_duality_residual",
}
assert summary["schema_version"] == "1.0"
assert summary["instance_id"] == config["instance_id"]
assert summary["num_qubits"] == config["num_qubits"]
assert summary["num_shots"] == expected["outcomes"].shape[0]
assert summary["num_folds"] == config["num_folds"]
expected_partition = [list(block) for block in expected["partition"]]
assert summary["partition"] == expected_partition
assert math.isclose(summary["partition_mutual_information"], expected["partition_score"], rel_tol=1e-10, abs_tol=1e-12)
assert summary["selected_bias_shots_per_outcome"] == expected["selected_bias_shots_per_outcome"]