🎓 ScholarChat AI - Smart Textbook Chatbot

A Streamlit-based AI chatbot that allows students to interact with their textbooks using DeepSeek 1.5B LLM and vector-based retrieval for precise answers from uploaded PDFs.

🏗️ Project Architecture

---

🚀 Features

PDF-Based Question Answering – Upload documents and chat with an AI that retrieves relevant content.
DeepSeek-1.5B AI Model – Uses DeepSeek-1.5B for efficient and context-aware responses.
Retrieval-Augmented Generation (RAG) – Uses ChromaDB for vector-based retrieval.
Streamlit Web Interface – Simple, user-friendly chatbot interface.
Supports Multiple PDFs – Process and query multiple textbooks at once.

User Demo

🛠️ Installation & Setup

1️⃣ Clone the Repository

git clone https://github.com/YOUR_GITHUB_USERNAME/ScholarChatAI.git
cd ScholarChatAI

2️⃣ Create a Virtual Environment

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3️⃣ Install Dependencies

pip install -r requirements.txt

4️⃣ Download DeepSeek Model

Ensure you have Ollama installed and DeepSeek-1.5B model downloaded:

ollama pull deepseek-r1:1.5b

5️⃣ Run the Application

streamlit run main.py

---

🖥️ Usage Guide

1️⃣ Upload PDFs

• Click Upload PDF Documents in the sidebar.
• Click Create Knowledge Base to process the documents.

2️⃣ Ask Questions

• Type your query in the chat input.
• The chatbot retrieves relevant content and generates AI responses.

3️⃣ Get Precise Answers

• The AI only answers from the uploaded PDFs.
• If content is missing, it responds:
  "I cannot find relevant information in the provided documents."

---

⚙️ How It Works

🔹 Vector-Based Retrieval (RAG)

• Converts PDF text into vector embeddings using nomic-embed-text.
• Stores them in ChromaDB for fast lookups.
• Uses Maximum Marginal Relevance (MMR) search for precise retrieval.

🔹 AI-Powered Chatbot

• Uses DeepSeek-1.5B via ChatOllama.
• Generates exam-friendly responses with examples.
• Follows structured, educational guidelines.

🏗️ Tech Stack

Technology	Purpose
Python	Main programming language
Streamlit	Web interface for chatbot
LangChain	AI model interaction
ChromaDB	Vector-based document retrieval
Ollama	Local AI model inference
DeepSeek-1.5B	AI model for generating responses

---

📌 TODO & Future Improvements

• ✅ Add support for custom LLM models.
• ✅ Improve retrieval accuracy for multi-document queries.
• 🚀 Implement fine-tuning options for specific subjects.
• 🚀 Add support for non-English documents.

---

🤝 Acknowledgments

• LangChain – For integrating RAG pipelines.
• DeepSeek – For providing high-quality AI models.
• Streamlit – For the intuitive web interface.

---

📜 License

This project is licensed under the MIT License.

---

🌟 Contributing

Pull requests are welcome!
For major changes, open an issue first to discuss your proposal.

1. Fork the repo
2. Create a feature branch (git checkout -b feature-branch)
3. Commit changes (git commit -m "Added new feature")
4. Push to branch (git push origin feature-branch)
5. Open a PR on GitHub

---

📬 Contact

For issues or feedback, open an issue or contact: 📧 Email: [kosamkar.r@northeastern.com]
🌐 GitHub: [github.com/rohit180497]