hermes-core 1.8.78

//! Vector indexing data structures
//!
//! This module provides a modular architecture for vector search:
//!
//! ## Module Structure
//!
//! - `ivf` - Core IVF (Inverted File Index) infrastructure
//!   - `CoarseCentroids` - k-means clustering for coarse quantization
//!   - `SoarConfig` / `MultiAssignment` - SOAR geometry-aware assignment
//!
//! - `quantization` - Residual product quantization
//!   - `PQCodebook` - the index-global OPQ codebook
//!
//! - `index` - Segment payloads for the production ANN implementations
//!   - `IVFPQIndex` - float vectors with residual PQ codes
//!   - `BinaryIvfIndex` - exact packed binary vectors
//!
//! ## SOAR (Spilling with Orthogonality-Amplified Residuals)
//!
//! The IVF module includes Google's SOAR algorithm for improved recall:
//! - Assigns vectors to multiple clusters (primary + secondary)
//! - Secondary clusters chosen to have orthogonal residuals
//! - Improves recall by 5-15% with ~1.3-2x storage overhead

pub mod index;
pub mod ivf;
mod kmeans;
pub mod quantization;

// IVF core
pub use ivf::{CoarseCentroids, CoarseConfig, IvfProbePlan, MultiAssignment, SoarConfig};

// Quantization
pub use quantization::{DistanceTable, PQCodebook, PQConfig};

// Indexes
pub use index::{
    BinaryCoarseQuantizer, BinaryIvfConfig, BinaryIvfIndex, IVFPQConfig, IVFPQIndex, IvfPqQueryPlan,
};