Module query_shape

Structs§

classify
contains_all_content_tokens: Return whether every normalized query content token occurs in the supplied text slice. Exact-tier fusion supplies at most three consecutive source lines and exits as soon as every requested token has been observed.
extract_content_tokens: Unique lowercase natural-language tokens with stopwords and words shorter than the three-character trigram minimum removed.
extract_lexical_tokens: Tokens supplied to the trigram search during hybrid retrieval and a lexical retry after zero results. Natural-language queries contribute normalized content words with stopwords removed. Regex retries prefer explicit code-shaped tokens before falling back to identifier-shaped text.
extract_short_nl_lexical_tokens: Legacy wrapper that delegates natural-language token extraction.
extract_tokens
looks_like_regex
pre_tier_exempt