← Về thư mục
scripts/claim_audit_calibration.py
- GoldSetValidationError · class · L74-L81 — class GoldSetValidationError(ValueError)
- validate_gold_set · function · L84-L147 — def validate_gold_set(tuples: list[dict[str, Any]]) -> None
- _resolve_constraint_rule_text · function · L150-L184 — def _resolve_constraint_rule_text(tup: dict[str, Any]) -> str
- _derive_constraint_scope · function · L187-L212 — def _derive_constraint_scope(constraint_id: str) -> str
- _breakdown_covers_expected · function · L215-L285 — def _breakdown_covers_expected( breakdown: Any, expected_sub_claims: Any ) -> bool
- _item_text · function · L246-L248 — def _item_text(item: dict[str, Any]) -> str
- _has_tokens · function · L250-L253 — def _has_tokens(text_l: str, tokens: Any) -> bool
- _matches · function · L255-L270 — def _matches(exp_idx: int, item: dict[str, Any]) -> bool
- _zero_division_safe · function · L288-L301 — def _zero_division_safe(numerator: int, denominator: int) -> float
- run_calibration · function · L304-L476 — def run_calibration( gold_set: list[dict[str, Any]], *, judge_fn: Callable[..., dict[str, Any]], thresholds: dict[str, float] | None = None, ) -> dict[str, Any]
- _accumulate_one_vs_rest · function · L479-L508 — def _accumulate_one_vs_rest( stats: dict[str, int], *, positive_class: str, expected_label: str, actual_label: str, ) -> None