Struct DistributedIndex 

pub struct DistributedIndex { /* private fields */ }

Distributed vector search index with sharding and replication.

Splits vectors across multiple shards for horizontal scaling. Supports replication for fault tolerance.

Implementations

impl DistributedIndex

pub fn new(shard_config: ShardConfig, search_config: SearchConfig) -> Self

Create a new distributed index.

pub fn build(&mut self, embeddings: &HashMap<String, Vec<f32>>) -> Result<()>

Build the distributed index from embeddings.

Automatically distributes vectors across shards using consistent hashing.

pub fn search(&self, query: &[f32], k: usize) -> Result<Vec<SearchResult>>

Search across all shards and merge results.

Performs a fan-out search to all shards in parallel, then merges and re-ranks results.

pub fn batch_search( &self, queries: &[Vec<f32>], k: usize, ) -> Result<Vec<Vec<SearchResult>>>

Batch search for multiple queries across all shards.

Performs multiple searches in parallel and returns results for each query.

pub fn filtered_search( &self, query: &[f32], k: usize, filter: &Filter, ) -> Result<Vec<SearchResult>>

Search with metadata filtering across all shards.

Applies the filter to each shard before merging results.

pub fn set_metadata(&mut self, entity_id: &str, metadata: Metadata)

Set metadata for an entity across all replica shards.

pub fn get_metadata(&self, entity_id: &str) -> Option<Metadata>

Get metadata for an entity from the primary shard.

pub fn batch_set_metadata(&mut self, metadata_map: &HashMap<String, Metadata>)

Set metadata for multiple entities in batch.

pub fn get_stats(&self) -> Result<DistributedStats>

Get statistics about the distributed index.