vicinity

Approximate nearest-neighbor search. Each algorithm is a separate feature flag -- depend on only what you use.

Default distance is cosine. L2, angular, and inner product via DistanceMetric. For cosine, enable auto_normalize to L2-normalize on insert.

Which index?

Which index?
├── General purpose: HNSW (default)
├── Memory constrained: HNSW + RaBitQ (SymphonyQG) or IVF-PQ
├── Disk-backed / large scale: DiskANN
├── Filtered search: ACORN (metadata filters) or Curator (label filters)
├── Batch/static: IVF-PQ or IVF-AVQ
└── Sparse vectors: SparseMIPS

For high-dimensional data (d ≥ 256), prefer SQ4U or SymphonyQG over plain HNSW — quantized graph traversal reduces distance computation cost. At low dimensions (d ≤ 25), plain HNSW wins; quantization overhead outweighs savings.

Install

Each algorithm is a separate feature. Enable what you need:

[dependencies]
vicinity = { version = "0.6", features = ["hnsw"] }          # graph index
# vicinity = { version = "0.6", features = ["ivf_pq"] }      # compressed index
# vicinity = { version = "0.6", features = ["nsw"] }         # flat graph

Usage

HNSW

High recall, in-memory. Best default choice.

use vicinity::hnsw::HNSWIndex;

let mut index = HNSWIndex::builder(128).m(16).ef_search(50).build()?;
index.add_slice(0, &[0.1; 128])?;
index.add_slice(1, &[0.2; 128])?;
index.build()?;

let results = index.search(&[0.1; 128], 5, 50)?;
// results: Vec<(doc_id, distance)>

IVF-PQ

Compressed index. 32–64× less memory than HNSW, lower recall. Use for datasets that don't fit in RAM.

use vicinity::ivf_pq::{IVFPQIndex, IVFPQParams};

let params = IVFPQParams { num_clusters: 256, num_codebooks: 8, nprobe: 16, ..Default::default() };
let mut index = IVFPQIndex::new(128, params)?;
index.add_slice(0, &[0.1; 128])?;
index.add_slice(1, &[0.2; 128])?;
index.build()?;

let results = index.search(&[0.1; 128], 5)?;

Persistence

Save and load indexes with the serde feature:

[dependencies]
vicinity = { version = "0.6", features = ["hnsw", "serde"] }

// Save
index.save_to_file("my_index.json")?;

// Load
let index = HNSWIndex::load_from_file("my_index.json")?;

See examples/06_save_and_load.rs for a full example.

Benchmark

GloVe-25 (1.18M vectors, 25-d, cosine), Apple Silicon, single-threaded:

Summary at best recall per algorithm:

Algorithm	Recall@10	QPS
HNSW (M=16)	100.0%	2,857
Vamana	100.0%	1,177
DiskANN	100.0%	1,029
NSW (M=16)	99.2%	1,288
IVF-PQ	98.7%	69
IVF-AVQ	90.9%	194
RP-Forest	58.5%	4,221

Full numbers and SIFT-128 results in docs/benchmark-results.md.

Supported Algorithms

Each algorithm has a named feature flag:

Algorithm	Feature	Notes
HNSW	hnsw (default)	Best recall/QPS balance for in-memory search
SQ4U	hnsw + sq4	HNSW with 4-bit quantized graph traversal + exact rerank; benefits high-d data
SymphonyQG	hnsw + ivf_rabitq	HNSW with RaBitQ quantized graph traversal; cheap approximate beam search + exact rerank
NSW	nsw	Flat small-world graph; competitive with HNSW on high-d data
Vamana	vamana	DiskANN-style robust pruning; fast search, higher build time
NSG	nsg	Monotonic RNG pruning; 50K cap due to O(n) ensure_connectivity
EMG	emg	Multi-scale graph with alpha-pruning
FINGER	finger	Projection-based distance lower bounds for search pruning
PiPNN	pipnn	Partition-then-refine with HashPrune; reduces I/O during build
FreshGraph	fresh_graph	Streaming insert/delete with tombstones
IVF-PQ	ivf_pq	Compressed index; 32-64x less memory, lower recall
IVF-AVQ	ivf_avq	Anisotropic VQ + reranking; inner product search
IVF-RaBitQ	ivf_rabitq	RaBitQ binary quantization; provable error bounds
RpQuant	rp_quant	Random projection + scalar quantization
BinaryFlat	binary_index	1-bit quantization + full-precision rerank
Curator	curator	K-means tree with per-label Bloom filters; low-selectivity filtered search
FilteredGraph	filtered_graph	Predicate-filtered graph search (AND/OR metadata filters)
ACORN	hnsw	Filtered HNSW search with subgraph sampling (SIGMOD 2024)
ESG	esg	Range-filtered search over numeric attributes
SparseMIPS	sparse_mips	Graph index for sparse vectors (SPLADE/BM25)
LEMUR	lemur	Late-interaction retrieval (multi-vector MIPS); inference-only
DiskANN	diskann	Vamana + SSD I/O layout; experimental
SNG	sng	Small navigable graph; O(n^2) construction
DEG	hnsw	Density-adaptive edge budgets (submodule of hnsw); O(n^2)
KD-Tree	kdtree	Exact NN; fast for d <= 20
Ball Tree	balltree	Exact NN; slightly better than KD-Tree for d=20-50
RP-Forest	rptree	Approximate; fast build, moderate recall

Quantization: RaBitQ, SAQ (quantization feature, via qntz crate). PQ is part of ivf_pq.

See docs.rs for the full API.

Documentation

• User guide -- quick start, distance metrics, LID, common pitfalls
• Benchmarks -- recall/QPS tables across datasets
• ANN landscape -- algorithmic principles, math foundations, research context
• References -- bibliography for every algorithm in the crate

License

MIT OR Apache-2.0