! Index Params */

Represents a retrieved document with optional metadata, fields, and vectors. This immutable data class encapsulates the result of a search or ret

Represents a scalar (non-vector) field in a collection schema. A `FieldSchema` defines the name, data type, nullability, and optional inverte

! JSON Value */

Represents a search query for a specific field in a collection. A `Query` can be constructed for either vector search or full-text search, bu

! Index Query Meta */

Represents a vector field in a collection schema. A `VectorSchema` defines the name, data type, dimensionality, and index configuration for a

* Field schema */

! JSON String */

Defines the structure of a collection in Zvec. A collection schema specifies the name of the collection and its fields, including both scalar

! Half-Precision Floating Point */

HNSW builder (adds one specific field: ef_search)

Full-text search query parameters. Attributes: query_string (Optional[str]): FTS query expression (e.g. '+vector -slow "exact

Sparse text embedding function using Qwen (DashScope) API. This class provides text-to-sparse-vector embedding capabilities using Alibaba Clo

Dense text embedding function using OpenAI API. This class provides text-to-vector embedding capabilities using OpenAI's embedding models. It

TODO: maybe a base class for quantizer?

BM25-based sparse embedding function using DashText SDK. This class provides text-to-sparse-vector embedding capabilities using the DashText

Re-ranker using Sentence Transformer cross-encoder models for semantic re-ranking. This re-ranker leverages pre-trained cross-encoder models to p

FLAT builder (no extra fields, just inherits base functionality)

Default local dense embedding using all-MiniLM-L6-v2 model. This is the default implementation for dense text embedding that uses the ``all-M

Re-ranker using Qwen (DashScope) cross-encoder API for semantic re-ranking. This re-ranker leverages DashScope's TextReRank service to perform

Re-ranker using Reciprocal Rank Fusion (RRF) for multi-vector search. RRF combines results from multiple vector queries without requiring rel

Default local sparse embedding using SPLADE model. This class provides sparse vector embedding using the SPLADE (SParse Lexical AnD Expansion

Re-ranker that combines scores using per-sub-query weights. Each sub-query's score is normalized by metric type (automatic when used via coll

Dense text embedding function using Qwen (DashScope) API. This class provides text-to-vector embedding capabilities using Alibaba Cloud's Das

! AiLego Blob Wrap */

Unified query executor that routes based on query count and reranker type.

Vamana builder (adds ef_search field)

! Cube class */

! Math Helper */

Neighbor

Re-ranker that delegates to a user-provided callback. The callback receives sub-query results, field schemas, and topn. Args: callba

Deprecated alias for Query. Use Query instead.

! Distance module */

HNSW-Rabitq builder (adds ef_search field)

MultiChunkCluster

Lightweight MemoryBlock for buffer pool mode: release on destruction

IVF builder (adds specific fields: nprobe, l1QueryParam, l2QueryParam)

! Location of Vector in Storage Segment */

* Collection stats */

! File Stream Writer */

! Index Meta */

! Index Plugin */