Struct BrainIndex 

pub struct BrainIndex { /* private fields */ }

Implementations

impl BrainIndex

pub fn open(brain_path: impl AsRef<Path>) -> Result<Self>

pub fn rebuild(&self, embedder: Option<&dyn Embedder>) -> Result<RebuildStats>

Walk the brain directory, parse markdown files, and repopulate the index, then embed any new or changed entries (skipped entirely if embedder is None, or if embedding a given entry fails — indexing is never blocked by the embedding step).

pub fn query( &self, text: &str, embedder: Option<&dyn Embedder>, filters: QueryFilters, ) -> Result<Vec<BrainEntry>>

Hybrid full-text + semantic search. When text is non-empty and embedder is Some, blends the existing FTS5 keyword ranking with vector similarity via Reciprocal Rank Fusion. Falls back to exactly today’s keyword-only (or filter-only) behavior when text is empty or embedder is None.

pub fn get(&self, id: &str) -> Result<Option<BrainEntry>>

Fetch a single entry by its relative path id.

pub fn backlinks(&self, id: &str) -> Result<Vec<BrainEntry>>

Entries whose links resolve to id (i.e. entries that link to id).

Resolution rule (matches the app’s render-time wikilink resolution): a raw link target t resolves to entry e when t == e.name OR t == e.id OR stem(t) == stem(e.id) OR stem(t) == e.name.

pub fn outlinks(&self, id: &str) -> Result<Vec<BrainEntry>>

Resolved targets of id’s own links (i.e. entries that id links to).

pub fn links_all(&self) -> Result<Vec<(String, String)>>

The full resolved (from_id, to_id) edge list, for pinboard rendering.

pub fn related(&self, id: &str, limit: usize) -> Result<Vec<BrainEntry>>

Entries related to id, ranked: direct links (either direction) first, then co-citation (entries reachable via a shared linked target or a shared linking source), then entries sharing at least one tag. Deduplicated, excludes id itself, capped at limit.