pub fn normalize_query(text: &str) -> StringExpand description
Normalize text for fuzzy matching: NFC + lowercase + strip Polish diacritics.
NFC canonical composition is applied first so NFD-stored variants (e.g.
o + combining acute) collapse to the same code points as NFC-stored
composed variants (ó) before diacritic mapping. Without this step,
źródło typed via NFD on one platform would refuse to match the same
word typed via NFC on another.
Maps: ą→a, ć→c, ę→e, ł→l, ń→n, ó→o, ś→s, ź→z, ż→z Enables “wdrozenie” to match “wdrożenie”, “zrodlo” to match “źródło”, etc.