Spaces:
Configuration error
Configuration error
| """Numeral-script normalization without value repair.""" | |
| from .contracts import SourceScript | |
| WESTERN = "0123456789" | |
| ARABIC_INDIC = "٠١٢٣٤٥٦٧٨٩" | |
| EXTENDED_ARABIC_INDIC = "۰۱۲۳۴۵۶۷۸۹" | |
| TRANSLATION = str.maketrans( | |
| ARABIC_INDIC + EXTENDED_ARABIC_INDIC + "٬٫،", | |
| WESTERN + WESTERN + ",.,", | |
| ) | |
| ALLOWED = frozenset(WESTERN + "/-.,:% ") | |
| def _source_script(text: str) -> SourceScript: | |
| scripts = set() | |
| if any(char in WESTERN for char in text): | |
| scripts.add("western") | |
| if any(char in ARABIC_INDIC for char in text): | |
| scripts.add("arabic_indic") | |
| if any(char in EXTENDED_ARABIC_INDIC for char in text): | |
| scripts.add("extended_arabic_indic") | |
| if len(scripts) > 1: | |
| return "mixed" | |
| return next(iter(scripts), "western") # type: ignore[return-value] | |
| def normalize_numeric_text(text: str) -> tuple[str, SourceScript]: | |
| stripped = text.strip() | |
| script = _source_script(stripped) | |
| normalized = stripped.translate(TRANSLATION) | |
| unsupported = next((char for char in normalized if char not in ALLOWED), None) | |
| if unsupported is not None: | |
| raise ValueError("unsupported character in numeric evidence") | |
| if not any(char.isdigit() for char in normalized): | |
| raise ValueError("numeric evidence contains no digits") | |
| return normalized, script | |