use-vector-store

Vector-store and embedding modeling primitives for RustUse.

Experimental

use-vector-store is experimental while use-nosql remains below 0.3.0.

Example

use use_vector_store::{Embedding, SimilarityMetric, VectorDimension, VectorId, VectorRecord};

let record = VectorRecord::new(VectorId::new("review_1"), Embedding::new(vec![0.1, 0.2]))
    .with_dimension(VectorDimension::new(2))?
    .with_similarity_metric(SimilarityMetric::Cosine);

assert_eq!(record.dimension(), Some(VectorDimension::new(2)));
# Ok::<(), use_vector_store::InvalidDimensionError>(())

Scope

• Vector identifiers, embeddings, dimensions, similarity metrics, metadata, and records.
• Dimension consistency validation.

Non-goals

• Qdrant, Pinecone, Weaviate, or other vector database clients.
• Approximate nearest-neighbor indexes.
• Embedding generation.

License

Licensed under either Apache-2.0 or MIT.