hora-search/hora
Hora
Hora is a vector similarity search library written in Rust designed for efficient approximate nearest neighbor retrieval in high-dimensional data spaces. It functions as a compiled search engine that organizes vector spaces using hierarchical graphs, trees, and product quantization, while computing proximity through various mathematical distance formulas including Euclidean, Manhattan, cosine, and dot product calculations.
The library supports high-performance execution through hardware-accelerated instruction sets and concurrent thread dispatch across multiple hardware processors. To accommodate diverse environments, it includes cross-platform support, single-threading configuration modes, and foreign function interface bindings that expose search routines to Python, Java, JavaScript, and other external programming environments.
Additionally, the software compiles directly into WebAssembly bytecode for execution within web browsers and edge environments, and can operate as a standalone network service to process remote queries from external applications.
Features
- Approximate Nearest Neighbor Search - Provides a high-performance vector search library written in Rust for indexing and querying high-dimensional data efficiently.
- Vector Search Indexes - Constructs high-performance search indexes using hierarchical graphs, trees, and product quantization to organize multi-dimensional vectors.
- Vector Search - Exposes high-performance vector indexing and similarity search capabilities to various programming environments.
- Vector Search Engines - Functions as a compiled search engine providing fast approximate nearest neighbor retrieval across multiple programming languages.
- Vector Similarity Search - Calculates vector distances and executes similarity queries using graph and tree index structures.
- Embedding Pairwise Distance Calculators - Computes similarity scores between numerical embeddings using metrics like Euclidean, cosine, and dot product calculations.
- Vector Distance Metrics - Computes similarity scores between high-dimensional vectors using Euclidean, Manhattan, cosine, and dot product distance calculations.
- Multi-Dimensional Data Indexing - Organizes high-dimensional vector spaces using interchangeable data structures like graphs, trees, and inverted files.
- SIMD-Accelerated Arithmetic - Executes parallel vector distance computations using single-instruction-multiple-data hardware instructions to maximize search throughput.
- Cross-Platform Deployment Targets - Executes code across major desktop and mobile operating systems and web browsers without heavy external dependencies.
- Standalone Service Deployments - Exposes vector search capabilities over a standalone network service so external applications can query indexes remotely.
- Foreign Function Bindings - Exposes high-performance native search routines to external runtime environments through foreign function interface wrappers.
- Multi-Language Bindings - Integrates high-speed vector search capabilities into Python, JavaScript, Java, and other tech stacks using native programming language bindings.
- Hardware-Accelerated Numerical Libraries - Executes vector distance computations using hardware-accelerated single instruction multiple data processing to maximize throughput.
- Customizable Distance Metrics - Computes vector proximity dynamically using pluggable mathematical formulas including Euclidean, Manhattan, cosine, and dot product calculations.
- WebAssembly Compilation - Compiles core search routines directly into portable WebAssembly bytecode for execution within web browsers and edge environments.
Star history
How this analysis was created: This summary and feature list are AI-generated from collected project material and can contain mistakes. Stars, license and language are imported from GitHub. Inclusion does not mean that we have tested or audited this project. Check the source documentation for any feature you depend on. Learn more on our About page.
