Vector Shop - AI-Powered E-commerce with Weaviate

A demonstration e-commerce application showcasing advanced vector search capabilities using Weaviate Java Client v6, Spring Boot 3.x, and vanilla JavaScript.

Features

• 3 Search Types:

  • Keyword Search (BM25): Traditional keyword-based search
  • Similarity Search (Vector): Semantic search using text embeddings
  • Hybrid Search: Combines BM25 and vector search for best results

• 20 Product Catalog: Pre-loaded with diverse items across clothes, shoes, and accessories

• Modern UI: Clean, responsive interface with tabbed search modes

• Real-time Results: Instant search results with product cards

• Weaviate Client v6: Built with the latest Weaviate Java Client v6 API

Prerequisites

1. Java 21 and Maven+
2. Weaviate Cloud free Sandbox cluster
3. Optional for running Weaviate locally: Docker Compose

Setting Up Weaviate (optional)

Docker Compose

The project includes a docker-compose.yml file configured with Weaviate 1.34.0 and text2vec-transformers.

Start Weaviate:

docker-compose up -d

This will start:

• Weaviate on port 8080 (HTTP) and 50051 (gRPC)
• text2vec-transformers inference service for text vectorization

Wait about 30-60 seconds for all services to initialize.

Verify Weaviate is running:

curl http://localhost:8080/v1/.well-known/ready

You should see: {"status":"healthy"}

Running the application

1. Clone or navigate to the project directory:

cd vector-shop

2. Build the project:

mvn clean install

3. Run the application:

mvn spring-boot:run

The application will:

• Start on port 8081
• Connect to Weaviate on localhost:8080
• Automatically create the schema
• Import 20 products from products.json

4. Access the application:

Open your browser and navigate to:

http://localhost:8081

Project structure

vector-shop/
├── src/
│   └── main/
│       ├── java/com/example/vectorshop/
│       │   ├── VectorShopApplication.java
│       │   ├── config/
│       │   │   └── WeaviateConfig.java
│       │   ├── controller/
│       │   │   ├── SearchController.java
│       │   │   └── SetupController.java
│       │   ├── model/
│       │   │   ├── ConnectionRequest.java
│       │   │   ├── GenerativeSearchResponse.java
│       │   │   ├── Product.java
│       │   │   └── SetupResponse.java
│       │   └── service/
│       │       ├── SearchService.java
│       │       └── SetupService.java
│       └── resources/
│           ├── application.properties
│           ├── products.json
│           └── static/
│               ├── index.html
│               ├── script.js
│               ├── styles.css
│               └── code/ (code snippet examples)
├── docker-compose.yml
├── pom.xml
├── LICENSE.md
└── README.md

API Endpoints

• GET /api/search/keyword?query={query} - Keyword search
• GET /api/search/similarity?query={query} - Similarity search
• GET /api/search/hybrid?query={query} - Hybrid search
• GET /api/search/image?imageUrl={url} - Image search
• GET /api/search/generative?query={query} - Generative search

Configuration

Edit src/main/resources/application.properties:

# Server Configuration
server.port=8081

# Weaviate Configuration
weaviate.host=localhost
weaviate.port=8080
weaviate.scheme=http

Technologies Used

• Backend: Java 21, Spring Boot 3.2.1, Maven
• Vector Database: Weaviate 1.34.0
• Vectorizers: text2vec-weaviate (for Weaviate Cloud) & text2vec-transformers (for local instance)
• Frontend: HTML5, CSS3, JavaScript
• Libraries:
  • Weaviate Java Client v6 (6.0.0-M2)
  • Jackson (JSON processing)
  • Lombok (code generation)

Example Searches

• Keyword: "blue jeans", "running shoes", "winter"
• Similarity: "comfortable casual wear", "athletic footwear"
• Hybrid: "stylish accessories for summer"

Troubleshooting

Port 8080 already in use

```bash # Find and stop the process lsof -ti:8080 | xargs kill -9 ```

Port 8081 already in use

Edit src/main/resources/application.properties and change:

server.port=8082

Data not loading

Check the application logs for errors. The seeder runs automatically on startup.

Weaviate not responding

```bash # Restart Weaviate docker-compose restart weaviate ```

Connection Refused Error

• Ensure Weaviate is running on port 8080
• Check Docker containers: docker ps

No Search Results

• Check Weaviate logs: docker logs weaviate
• Verify data was seeded in application logs

Generative Search Not Working

• This feature requires OpenAI API key
• Error messages will indicate if the API key is missing or invalid

Image Search Not Working

• Image search in v6 requires base64-encoded images
• URL-based image search is not directly supported in the current implementation

License

MIT License - Free to use and modify.