Vector Database

This document describes the vector database functionality in the Nexent SDK.

Overview

The vector database module provides efficient storage and retrieval of high-dimensional vector embeddings with support for:

• Vector similarity search
• Metadata filtering
• Batch operations
• Multiple backend support

Supported Backends

Elasticsearch

High-performance distributed search and analytics engine.

Qdrant

Vector similarity search engine optimized for performance.

FAISS

Library for efficient similarity search and clustering of dense vectors.

Usage

from nexent.vector_database import VectorDB

# Initialize vector database
vector_db = VectorDB(
    backend="elasticsearch",
    host="localhost",
    port=9200
)

# Insert vectors
vector_db.insert(
    vectors=embeddings,
    metadata={"source": "document.pdf"}
)

# Search similar vectors
results = vector_db.search(
    query_vector=query_embedding,
    top_k=10
)

Configuration

The vector database can be configured through:

• Connection parameters
• Index settings
• Search parameters

For detailed configuration options, see the SDK documentation.