🧠 Semantic Search with Spring AI, Ollama, and Milvus

This project implements a semantic search API for product-related data using:

• Spring AI
• Ollama as the local LLM/embedding provider
• Milvus as the vector store backend
• Structured product data (e.g. title, description, features, price, category)

✨ Features

• 🔍 Semantic product search over embedded documents
• 💬 Chat-style querying with structured JSON responses
• ⚙️ Embedding generation endpoint
• 📦 Integration with Milvus for vector similarity search
• 🧠 Uses Ollama models locally (e.g., mxbai-embed-large, llama3.2)

System Design

---

🧰 Tech Stack

Component	Tool / Library
LLM & Embeddings	Ollama (llama3.2, mxbai-embed-large)
Embedding store	Milvus vector DB
Framework	Spring Boot + Spring AI
REST API	Spring MVC (@RestController)
Data Format	JSON

---

🚀 Getting Started

Prerequisites

• Java 17+
• Docker (for running Milvus)
• Ollama installed and running locally

Run Ollama Models

brew install ollama

# Pull models
ollama pull mxbai-embed-large
ollama pull llama3.2

# Check available models
ollama ls

Run Milvus

docker-compose up -d

Run Source

mvn spring-boot:run

🧪 API Endpoints

Chat with model

Request:

curl --location 'http://localhost:8080/api/ollama/chat?query=Captial%20of%20Vietnam'

Response:

{
    "answer": "Hanoi"
}

Embedding token

Request:

curl --location 'http://localhost:8080/api/ollama/embedding?query=smartphone'

Response:

{
    "embedding": [
        -0.018574331,
        -0.05374644,
        -0.05071582,
        ...]
}

Semantic search from dataset

Request:

curl --location 'http://localhost:8080/api/ollama/search?query=headphone'

Response:

{
    "title": "Bluetooth Wireless Earbuds",
    "description": "Noise-canceling earbuds with up to 30 hours battery life and fast charging support. Sweat and water-resistant.",
    "price": "$49.99",
    "category": "Electronics",
    "features": "Touch controls, Charging case, Bluetooth 5.3, Built-in microphone"
}

Contact

• If you have any questions, please open an issue on GitHub or contacting me via email: ducan1406@gmail.com.