Spaces:
Running on Zero
Running on Zero
| """Typo-tolerant matching for course search (pure, tested). 'phyy' still finds 'Physics'.""" | |
| import difflib | |
| import re | |
| from typing import Optional, Sequence | |
| _WORD = re.compile(r"[a-z0-9\u1200-\u137f]+") | |
| def words(text: str) -> list[str]: | |
| return _WORD.findall(text.lower()) | |
| def _common_prefix(a: str, b: str) -> int: | |
| n = 0 | |
| for x, y in zip(a, b): | |
| if x != y: | |
| break | |
| n += 1 | |
| return n | |
| def _token_score(token: str, candidate_words: Sequence[str]) -> int: | |
| best = 0 | |
| for w in candidate_words: | |
| if token == w: | |
| return 4 | |
| if w.startswith(token): | |
| best = max(best, 3) | |
| elif token in w: | |
| best = max(best, 2) | |
| else: | |
| p = _common_prefix(token, w) | |
| if (p >= 3 and p / len(token) >= 0.6) or (len(token) >= 4 and difflib.SequenceMatcher(None, token, w).ratio() >= 0.75): | |
| best = max(best, 1) | |
| return best | |
| def match_score(tokens: Sequence[str], texts: Sequence[str]) -> Optional[int]: | |
| """None if ANY token matches nothing; otherwise a higher score = better match.""" | |
| candidate_words = [w for t in texts for w in words(t)] | |
| total = 0 | |
| for token in tokens: | |
| score = _token_score(token, candidate_words) | |
| if score == 0: | |
| return None | |
| total += score | |
| return total | |