LabMate-AI - A AI Assistant for Tinkerers' Lab, PREC Loni

Welcome to the AI Lab Assistant repository of Tinkerers' Lab, PREC Loni! 🚀 This project is a Spring Boot-based RAG (Retrieval-Augmented Generation) LLM model, designed to assist students and researchers by providing contextual responses based on academic materials, research papers, and lab resources.

✨ Features

• RAG-Powered AI: Combines retrieval-based search with a powerful LLM for accurate and contextual responses.
• Spring Boot Backend: A robust and scalable backend built using Java Spring Boot.
• Vector Database Integration: Uses pgvector with PostgreSQL to store and retrieve embeddings efficiently.
• Local Embedding Generation: Utilizes nomic-embed-text to generate embeddings without relying on external API keys.
• PDF Processing: Scans and processes PDFs to extract relevant knowledge for AI-assisted answers.

📁 Repository Structure

├── src/main/java/com/TinkerersLab/LabAssistant  # Backend Spring Boot code
│   ├── config/                                 # Configuration files
│   ├── controller/                             # REST API controllers
│   ├── service/                                # Business logic
│   ├── model/                                  # Data models
│   ├── util/                                   # Utility functions
│   ├── LabAssistantApplication.java            # Application Entrypoint
│
├── data/                                       # PDF and document storage
├── Dockerfile                                  # Docker setup for containerization
├── README.md                                   # Project documentation
├── comose.yaml                                 # Docker configuration of services
├── pom.xml                                     # Project dependency configuration
└── .env                                        # Environment variables

🚀 Getting Started

Prerequisites

• Java 17+
• Spring Boot 3+
• PostgreSQL with pgvector
• Docker (optional, for containerized deployment)
• Ollama

Installation & Setup

1. Clone the Repository

   git clone https://github.com/tinkererslab/LabMate-AI.git
   cd LabMate-AI

2. Set Up the Database

   • Install PostgreSQL and enable pgvector extension.
   • Create a new database and update .env with database credentials.

3. Run the Application

   ./LabMate-AI/mvnw spring-boot:run

   or

   docker compose up

4. Add PDF file for context

   cp ~/input-context.pdf /opt/docs/.

   or

   cp data/input.zip /opt/docs;
   cd /opt/docs;
   unzip input.zip

   (you have have multiple pdf files at /opt/docs)

5. Access the API

   • API runs on http://localhost:8080/api/v1/llm/ask and http://localhost:8080/api/v1/llm/askAndStream?query=this+is+my+query
   • add request body as

   {
      "query": "you query"
   }

6. Access the UI

   • Visit http://localhost:8080 on you web browser

📌 Future Enhancements

• Fine-tuning the LLM for improved contextual accuracy.
• Integration with Chat UI for a seamless user experience.
• Multi-Document Support to process various academic file formats.

🔄 Customizing for Personal Use

This AI Lab Assistant can be used as a personal AI assistant for any purpose! Simply replace the contents of the input PDF files in the data/ directory with your own documents. The system will process and retrieve information based on your specific files, making it a versatile tool for various applications.

🤝 Contributing

We welcome contributions from the community! Feel free to fork this repository, submit pull requests, or open issues.

📜 License

This project is open-source and available under the MIT License.

---

🚀 Built with passion at Tinkerers' Lab, PREC Loni. Let's innovate together! 🔥

Contact

For any inquiries or feedback, please reach out to us at:

• Email: tl@pravaraengg.org.in or
• Maintainer : email , github
• GitHub: Tinkerer's Lab PREC

---