Spring AI Ollama Chat

This project is an example of a Retrieval-Augmented Generation (RAG) application using Ollama and Spring AI. It employs ETL (Extract, Transform, Load) processes to load a PDF into a VectorStore (ChromaDB) with Spring AI, allowing for question-and-answer interactions based on the content of the PDF.

Prerequisites

• JDK 21 or higher
• Maven 3.6.0 or higher
• Docker
• Ollama

Start the required services with Docker Compose (ChromaDB):

cd development
docker-compose up -d

Endpoint

POST /chat/ask

This endpoint receives a question in the request body and returns an answer generated related with the pdf included in the project.

Request

• URL: /chat/ask
• HTTP Method: POST
• Request Body: A plain text containing the question.

Example Request

curl -X POST http://localhost:8080/chat/ask -d 'What Andrew says about MATLAB?'