📚 Research Assistant

Research Assistant is a microservice-based web application designed to help students and researchers retrieve relevant information from uploaded documents using Retrieval Augmented Generation (RAG). It seamlessly combines a modern frontend with powerful backend services to manage documents, authenticate users, and generate intelligent, AI-powered responses based on specific document contexts.

---

🌟 Executive Summary (For HR & Non-Technical Readers)

What is this project? This application solves the problem of "information overload" by allowing researchers, students, and professionals to upload lengthy documents and ask direct questions to an AI about those documents. The AI reads the documents and formulates accurate answers based only on the provided text, a process known as Retrieval-Augmented Generation (RAG).

Key Features:

• Secure Authentication: Users can safely register and log in to manage their private research documents.
• Intelligent Document Search (RAG): Extracts and retrieves the most relevant information from uploaded documents rather than relying on generalized AI knowledge.
• Modern Microservices Architecture: The app is broken down into independent, specialized services, making it scalable, resilient, and easier to maintain.

Why this matters: This project demonstrates proficiency in Full-Stack Development, System Design (Microservices), Containerization (Docker), and the implementation of Modern AI/LLM (Large Language Model) techniques. It showcases an end-to-end understanding of how to build, connect, and deploy complex modern web architectures.

---

🛠️ Technical Details (For Engineers)

🏗️ Architecture

The application follows a distributed microservices architecture, fully containerized using Docker and Docker Compose. This ensures environment consistency across development and production deployments.

graph TD
    Client[Frontend Client]
    Auth["Auth Service (Flask)"]
    Doc["Document Service (FastAPI)"]
    RAG["RAG Service (FastAPI)"]
    
    AuthDB[(MySQL Auth DB)]
    DocDB[(Postgres Doc DB)]
    VecDB[(Chroma Vector DB)]
    
    Client -->|/auth| Auth
    Client -->|/api/documents| Doc
    Client -->|/api/rag| RAG
    
    Auth --> AuthDB
    Doc --> DocDB
    Doc --> VecDB
    RAG --> VecDB

Loading

🚀 Tech Stack

• Frontend: Next.js (React), Tailwind CSS, Shadcn UI, Lucide React
• Backend Services:
  • Authentication: Flask, SQLAlchemy (Python)
  • Document Management: FastAPI (Python)
  • RAG Execution: FastAPI (Python), Langchain/LlamaIndex
• Databases:
  • MySQL: Relational storage for user authentication data
  • PostgreSQL: Relational storage for document metadata
  • ChromaDB: Vector database for document embeddings used in semantic search
• Infrastructure: Docker, Docker Compose

---

� Getting Started

Ensure you have Docker and Git installed on your machine.

1. Clone the Repository

git clone <repository-url>
cd Research-Assistant

2. Environment Configuration

Ensure each environment (.env) file is appropriately configured at the root of following directories:

• authentication/.env
• document_service/.env
• rag_service/.env

(Note: API keys for specific AI services like Gemini/OpenAI need to be defined in rag_service/.env according to the required schema.)

3. Run the Services

Spin up all microservices, databases, and networks via Docker Compose:

docker-compose up --build

The application layers will map to the following local endpoints:

• Frontend App: http://localhost:3000
• Auth Service Api: http://localhost:5000
• Document Service Api: http://localhost:6060
• RAG Service Api: http://localhost:7000

---

🔌 API Reference

Authentication Service (:5000/auth)

Method	Endpoint	Description
POST	/login	Authenticate a user and issue a JWT
POST	/register	Register a new user
POST	/verify	Verify the validity of a user's JWT
POST	/logout	Invalidate a user's session

Document Service (:6060/api/documents)

Method	Endpoint	Description
POST	/upload	Persist a document file and its vector embeddings
GET	/fetch/all	Retrieve all metadata for uploaded documents
GET	/fetch/{file_id}	Get metadata for a specific document
DELETE	/delete/{file_id}	Remove a document and its embeddings

RAG Service (:7000/api/rag)

Method	Endpoint	Description
POST	/generate	Generate context-aware AI answers based on query and document constraints

---

💻 Frontend Local Development

If you wish to develop the frontend outside the Docker network:

cd frontend
npm install
npm run dev

The isolated frontend will be available at http://localhost:3000.

---

🔮 Future Improvements & Roadmap

While the MVP is fully functional, there are several areas planned for enhancement:

• AI Model Migration: Switch or modularize the underlying AI language model (e.g., better integration with Gemini, transitioning to a different provider, switching to a more cost-effective open-source model) to reduce latency and improve answer quality.
• Code Refactoring & Cleaning:
  • Standardize error handling and structured logging across all Python microservices.
  • Break down larger monolithic files into smaller modular files.
  • Implement robust unit and integration testing setups.
  • Adopt stricter type linting.
• Advanced features: Introduce background tasks for document processing to enhance responsiveness for large document uploads.