Struct SqliteStore 

pub struct SqliteStore { /* private fields */ }

SQLite-backed store with sqlite-vec KNN search.

Implementations

impl SqliteStore

pub async fn connect(url: &str, dim: usize) -> Result<Self>

Connect to (creating if missing) the SQLite database at url (e.g. sqlite://data/rag.db or sqlite::memory:), expecting dim-dimensional embeddings.

Trait Implementations

impl VectorStore for SqliteStore

fn migrate<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Create tables/indexes if they do not exist. Safe to call repeatedly.

fn upsert_document<'life0, 'life1, 'async_trait>( &'life0 self, doc: &'life1 Document, ) -> Pin<Box<dyn Future<Output = Result<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Insert or replace a document row (keyed by id).

fn find_document_by_hash<'life0, 'life1, 'async_trait>( &'life0 self, hash: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<Option<String>>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Return the id of an existing document with this content hash, if any. Used to skip re-ingesting unchanged documents.

fn insert_chunks<'life0, 'life1, 'async_trait>( &'life0 self, chunks: &'life1 [Chunk], ) -> Pin<Box<dyn Future<Output = Result<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Bulk-insert chunks (each must carry a populated embedding).

fn vector_search<'life0, 'life1, 'async_trait>( &'life0 self, query: &'life1 [f32], k: usize, ) -> Pin<Box<dyn Future<Output = Result<Vec<Scored>>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Dense search: the k chunks whose embeddings are most cosine-similar to query. Returned chunks omit their embedding vector to keep payloads small.

fn all_chunks<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<Vec<Chunk>>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Every chunk (id, doc_id, ordinal, text, metadata) with embedding == None. Feeds the BM25 keyword index; keep it modest for large corpora.

fn count_chunks<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<usize>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Total number of stored chunks.

fn count_documents<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<usize>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Total number of stored documents.

fn list_documents<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<Vec<Document>>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Every stored document with its metadata (including processing metrics).

fn delete_document<'life0, 'life1, 'async_trait>( &'life0 self, doc_id: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Delete one document and all of its chunks (used to roll back a failed streaming ingest). Deleting an unknown id is not an error.

fn delete_documents_by_source<'life0, 'life1, 'async_trait>( &'life0 self, source_uri: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Delete every document (and chunks) originating from source_uri — stale rows from interrupted runs or previous versions of a changed file.

fn clear<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Remove all documents and chunks.