Module pgvector

Available on crate feature pgvector only.

Integration module for PostgreSQL vector database (pgvector) operations.

This module provides a client interface for vector similarity search operations using pgvector, supporting:

• Vector collection management with configurable schemas
• Efficient vector storage and indexing
• Connection pooling with automatic retries
• Batch operations for optimized performance
• Metadata included in retrieval

The functionality is primarily used through the PgVector client, which implements the [Persist] trait for seamless integration with indexing and query pipelines.

Example

let client = PgVector::builder()
    .db_url("postgresql://localhost:5432/vectors")
    .vector_size(384)
    .build()?;

Structs

MetadataConfig

Configuration for metadata fields in the PostgreSQL table.

PgVector

Represents a Pgvector client with configuration options.

PgVectorBuilder

Builder for PgVector.

VectorConfig

Configuration for vector embedding columns in the PostgreSQL table.

Enums

FieldConfig

Field configuration types supported in the PostgreSQL table schema.