pub fn normalize_sentence(s: &str) -> StringExpand description
Normalize a sentence for map_store lookup. Rules:
- ASCII alphanumerics are kept verbatim (case preserved).
'and,are dropped without inserting a space, so contractions (we've→weve) and thousands separators (1,000→1000) stay one token.- Any other character (whitespace,
_,-,.,!, non-ASCII, …) is treated as a word boundary; runs collapse to a single space and leading/trailing spaces are trimmed.