ezdb-utpal-raina 1.0.0

EzDB B-Class - Free & open source vector database for semantic search

EzDB B-Class (Basic) - Free & Open Source Vector Database

EzDB is a lightweight, easy-to-use vector database built in Python for semantic search and AI applications.

📦 This is EzDB B-Class (Basic) - The free, open-source version. For production features, see EzDB Professional and EzDB Enterprise.

⚡ Quick Links

• 📖 Getting Started Guide - Learn the basics
• 🚀 API Documentation - Complete API reference
• 🐳 Docker Quick Start - Deploy in 60 seconds
• 📦 Docker Guide - Complete Docker documentation
• 🏗️ Product Tiers - B-Class, Professional, Enterprise
• 🚀 Deployment Guide - Production deployment

Features

• Vector Storage: Store and search high-dimensional vectors efficiently
• Multiple Metrics: Cosine similarity, Euclidean distance, Dot product
• Fast Indexing: HNSW (Hierarchical Navigable Small World) for ANN search
• Metadata Filtering: Search with metadata constraints
• Persistence: Save/load databases to disk
• Dual Mode: Use as embedded library OR REST API server
• REST API: Full HTTP API with Python/JavaScript/Go clients
• Collections: Manage multiple isolated vector collections
• Docker Ready: Easy deployment with Docker and docker-compose

Installation

pip install -r requirements.txt

Quick Start

Embedded Mode (Python Library)

from ezdb import EzDB

# Create a new database
db = EzDB(dimension=384)

# Insert vectors with metadata
db.insert(
    vector=[0.1, 0.2, 0.3, ...],
    metadata={"text": "Hello world", "category": "greeting"}
)

# Search for similar vectors
results = db.search(query_vector=[0.1, 0.2, 0.3, ...], top_k=5)

# Save and load
db.save("my_database.ezdb")
db = EzDB.load("my_database.ezdb")

REST API Server (Network Access)

Start Server:

pip install -r requirements-server.txt
python -m uvicorn ezdb.server.app:app --host 0.0.0.0 --port 8000

Use Client:

from ezdb.client import EzDBClient

client = EzDBClient("http://localhost:8000")
client.insert(vector=[0.1, 0.2, 0.3, ...], metadata={"text": "Hello"})
results = client.search(vector=[0.1, 0.2, 0.3, ...], top_k=5)

Interactive API Docs: http://localhost:8000/docs

See API.md for complete API documentation.

Architecture

• Storage Engine: Efficient in-memory vector storage with metadata
• Indexing: HNSW (Hierarchical Navigable Small World) for fast ANN search
• Similarity: Cosine, Euclidean, and Dot Product metrics
• Persistence: JSON-based serialization for easy inspection

Use Cases

• Semantic search
• Recommendation systems
• RAG (Retrieval Augmented Generation)
• Document similarity
• Image search
• Anomaly detection