🧠 Exam Helper

A powerful exam preparation assistant that allows students to upload their notes in the form of PDF, DOCX, or PPTX documents, submit multiple previous year questions in a single batch, and receive context-aware, accurate answers and a list of important topics using Retrieval-Augmented Generation (RAG) with LLMs.

🚀 Key Features

• Multi-Document Upload: Supports PDF, DOCX, and PPTX formats.
• Multi-Question Input: Add multiple previous year questions at once — answers are batch-processed for speed and clarity.
• RAG Pipeline: Embeds documents using open-source models (e.g. bge-base-en), stores them in a vector store using langgraph's InMemoryStore, and retrieves context based on the questions.
• LLM Integration: Uses Groq + llama-3.1-8b-instant for grounded answer generation while using the context provided by RAG.
• Important Topics: Extracts most frequently occuring topics from the answers and displays a concise summary.
• Automatic Cleanup: All uploaded files are removed after processing to keep your workspace clean.

💻 Technologies Used

• Frontend - HTML, CSS, JS
• Backend - Flask(Python), RAG Pipeline
• LLM Stack - LangChain, Groq, HuggingFace Embeddings

📦 File Structure

rag-project/
├── app.py                      # Flask app entry point
├── .env                        # API key (not tracked by Git)
├── uploads/                    # Temporary upload folder (auto-cleaned)
├── templates/
│   └── index.html              # Frontend HTML
├── static/
│   ├── css/style.css           # Custom dark theme styles
│   └── js/script.js            # AJAX + DOM updates
├── llm_utils/
│   ├── __init__.py
│   └── pipeline.py             # RAG + embedding logic
└── requirements.txt            # Python dependencies

⚙️ Local Setup Instructions

1. Clone the repository:

git clone https://github.com/your-username/rag-doc-qa.git
cd rag-doc-qa

2. Create a virtual environment and install dependencies:

python -m venv venv
source venv/bin/activate  # For Windows: venv\Scripts\activate
pip install -r requirements.txt

3. Set up your .env file:

GROQ_API_KEY=your_groq_key_here

4. Run the app:

flask run

📬 Future Enhancements

• Generate important questions based on previous year papers
• Add handwritten notes support with OCR
• Add multi-user session support
• Enable document history and downloads
• Summarization refinement via separate LLM
• Deploy to cloud

📸 Working Images

An example use case (I have used notes for my Mechanical Engineering Course):