Expand description
Title + full-text search over the doc tree.
Two code paths:
- Title search —
fuzzy-matcher::SkimMatcherV2against the index’s entry titles. Fast, no preprocessing, suitable for the “I know roughly what the page is called” query. - Full-text search —
tantivyin-memory (RamDirectory) index built once atSearchIndex::build. Matches tokens in the plain-text projection of every page’s body and returns top-N matches with a 140-character snippet.
The DocsBrowser TUI and the DocsServer::search_handler both
call into SearchIndex.
Structs§
- Full
Text Hit - Result of a full-text search.
- Search
Index - In-memory tantivy-backed full-text index over the doc tree.
- Title
Hit - Result of a title search — one row per hit, ordered by descending fuzzy score.