🧠 RAG Java Microservice

This project is a Java 24 Spring AI RAG service that integrates with Ollama for LLMs and ChromaDB for vector storage.
It supports:

• Multi-query retrieval
• Chat memory
• Requirements Analyst → Doc Writer agent workflow
• Confluence-ready document generation

---

🚀 Getting Started

1. Clone the repo

git clone https://github.com/akrios-d/rag-java-microservice.git
cd rag-java-microservice

2. Build & Run with Docker Compose

docker compose up --build

This will start:

• rag-service → Spring Boot app (port 8081)
• ollama → LLM backend (port 11434)
• chroma → Vector database (port 8000)

---

🔍 Verify Services

• RAG API → http://localhost:8081/query
• Ollama health → http://localhost:11434/api/tags
• Chroma API → http://localhost:8000/api/v2/heartbeat

---

📡 Example Usage

Ask a question

curl -X POST "http://localhost:8081/query?userId=test&multiQuery=true"      -H "Content-Type: application/json"      -d '{"question":"What is retrieval augmented generation?"}'

Requirements Analyst Mode

curl -X POST "http://localhost:8081/requirements?userId=test"      -H "Content-Type: application/json"      -d '{"input":"I need a system for handling support tickets"}'

Once you type generate, the system switches to Doc Writer and produces a Confluence-ready document.

---

⚡ Models

Pull the embedding model before running queries:

docker exec -it ollama ollama pull all-minilm

And optionally pull LLMs:

docker exec -it ollama ollama pull llama3.2

CREATE EXTENSION IF NOT EXISTS vector;

CREATE TABLE documents (
id UUID PRIMARY KEY,
content TEXT NOT NULL,
metadata JSONB,
embedding VECTOR(384) -- ajuste para o tamanho do embedding do seu modelo
);

---

🔒 Security Notes

• Runs as non-root (distroless base image).
• Minimal runtime environment (no shell, no package manager).
• Data for Ollama and Chroma is persisted in Docker volumes.

---

🛠 Development

Rebuild after changes:

docker compose up --build rag-service