Struct ContextualEmbeddingSearch 

pub struct ContextualEmbeddingSearch { /* private fields */ }

Context-aware embedding search engine with query expansion, negative suppression, and diversity-aware re-ranking.

Implementations

impl ContextualEmbeddingSearch

pub fn new() -> Self

Create an empty search engine.

pub fn add_document(&mut self, doc: SearchDoc) -> Result<(), SearchError>

Add a document to the index.

The first document sets the expected embedding dimension for all subsequent insertions and queries.

pub fn remove_document(&mut self, id: &str) -> Result<(), SearchError>

Remove a document from the index by ID.

pub fn search( &mut self, query: &[f64], context: &SearchContext, config: &SearchConfig, ) -> Result<Vec<ContextualResult>, SearchError>

Search the index.

Errors

• SearchError::IndexEmpty — no documents indexed.
• SearchError::DimensionMismatch — query dimension doesn’t match index.
• SearchError::ConfigurationError — top_k == 0 or rerank_top_n == 0.

pub fn update_context( &self, context: &mut SearchContext, query: &[f64], query_text: String, )

Record a new query into context.

Adds query_text to query_history and appends the corresponding embedding.

pub fn batch_search( &mut self, queries: &[Vec<f64>], context: &SearchContext, config: &SearchConfig, ) -> Result<Vec<Vec<ContextualResult>>, SearchError>

Execute multiple independent searches sharing the same context and config.

pub fn stats(&self) -> SearchStats

Return accumulated search statistics.

pub fn len(&self) -> usize

Number of indexed documents.

pub fn is_empty(&self) -> bool

true if no documents are indexed.

pub fn dimension(&self) -> Option<usize>

Current embedding dimensionality, or None if the index is empty.

Trait Implementations

impl Default for ContextualEmbeddingSearch

fn default() -> Self

Returns the “default value” for a type.