benchmark/objective_integrity._kind_recall_problems has two guard branches that are never exercised by the test suite:
- line 188 —
if not isinstance(objective, dict): return ["objective is not a dict"]
- line 195 —
if not _is_ratio(value): return [f"kind_recall={value!r} is not a ratio in [0, 1]"]
test_kind_recall_problems_helper covers the empty and bool-kind_recall cases, but not these two — even though the sibling helper's test, test_recall_field_problems_helper, already covers the same non-dict and out-of-range branches for _recall_field_problems.
Ask
Add the two missing assertions to test_kind_recall_problems_helper (a non-dict objective and an out-of-range kind_recall), mirroring the sibling helper's coverage. Tests only.
benchmark/objective_integrity._kind_recall_problemshas two guard branches that are never exercised by the test suite:if not isinstance(objective, dict): return ["objective is not a dict"]if not _is_ratio(value): return [f"kind_recall={value!r} is not a ratio in [0, 1]"]test_kind_recall_problems_helpercovers the empty andbool-kind_recallcases, but not these two — even though the sibling helper's test,test_recall_field_problems_helper, already covers the same non-dict and out-of-range branches for_recall_field_problems.Ask
Add the two missing assertions to
test_kind_recall_problems_helper(a non-dict objective and an out-of-rangekind_recall), mirroring the sibling helper's coverage. Tests only.