Expand description
Qdrant vector database integration.
QdrantClientis the direct client.bqadds binary-quantization utilities and a BQ-optimized client.
Re-exports§
pub use client::QdrantClient;pub use client::VectorDbClient;pub use error::VectorDbError;pub use mock::MockVectorDbClient;mockpub use mock::cosine_similarity;mockpub use model::SearchResult;pub use model::VectorPoint;pub use model::embedding_bytes_to_f32;pub use model::f32_to_embedding_bytes;pub use model::generate_point_id;
Modules§
- bq
- Binary quantization helpers and BQ client. Binary quantization helpers and Qdrant BQ client.
- client
- Direct Qdrant client and trait.
- error
- Vector DB error types.
- mock
mock - In-memory vector DB mock (enabled with
mockfeature). - model
- Shared vector DB model types.
- rescoring
- Full-precision rescoring utilities. Full-precision rescoring of candidates.
Enums§
- Write
Consistency - Write-consistency mode for Qdrant upserts/deletes.
Constants§
- DEFAULT_
COLLECTION_ NAME - Default non-BQ collection name.
- DEFAULT_
VECTOR_ SIZE - Default vector size for non-BQ collections.