https://github.com/gms-001/rag-based-chatbot-tds

RAG-powered Virtual Teaching Assistant for IIT Madras’ Tools in Data Science (TDS) course. Built with LangChain, LangSmith, and Streamlit, this chatbot provides context-aware answers from course materials, manages user sessions with SQLite, and supports real-time chat streaming.
https://github.com/gms-001/rag-based-chatbot-tds

gemini-api langchain langgraph langsmith python rag-chatbot selenium sqlite

Last synced: about 1 month ago
JSON representation

RAG-powered Virtual Teaching Assistant for IIT Madras’ Tools in Data Science (TDS) course. Built with LangChain, LangSmith, and Streamlit, this chatbot provides context-aware answers from course materials, manages user sessions with SQLite, and supports real-time chat streaming.

• Host: GitHub
• URL: https://github.com/gms-001/rag-based-chatbot-tds
• Owner: GmS-001
• License: mit
• Created: 2025-08-21T14:59:57.000Z (about 2 months ago)
• Default Branch: main
• Last Pushed: 2025-09-01T19:30:23.000Z (about 1 month ago)
• Last Synced: 2025-09-01T20:44:19.475Z (about 1 month ago)
• Topics: gemini-api, langchain, langgraph, langsmith, python, rag-chatbot, selenium, sqlite
• Language: Jupyter Notebook
• Homepage:
• Size: 5.54 MB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# ✨ RAG-Based Virtual Teaching Assistant

A **Retrieval-Augmented Generation (RAG)** powered chatbot built for the **Tools in Data Science (TDS)** course at **IIT Madras**.  

This virtual assistant provides **accurate, context-aware answers** to students' queries using course materials, while also being capable of answering general knowledge questions gracefully.  

Built with **LangGraph**, **LangChain**, **LangSmith**, **Streamlit**, **Selenium** and a **SQLite backend**, this project demonstrates a **production-grade RAG application** with authentication, conversation threading, and persistent chat history.

### ✨Live Website : https://rag-based-chatbot-tds-gms.streamlit.app/

---

## 🚀 Features

- **Context-Aware Responses (RAG)**  

  Answers are based on retrieved chunks from course content first, ensuring accuracy and reliability.

- **Fallback General Knowledge Answers**  

  When no course context is available, the bot clearly states this and provides helpful responses using general reasoning.

- **LangSmith-Backed LLM Tracing & Analysis**  

  Used **LangSmith** to monitor, debug, and optimize LLM calls for better transparency and reliability in production.

- **Interactive Streamlit Frontend**  

  - Chat-like UI for students.  

  - Live message streaming.  

  - Sidebar for switching between chat threads.  

- **User Authentication & Conversation Management**  

  - Login/registration system with secure password storage.  

  - Each user has multiple conversation threads, all persisted in a SQLite database.

- **Well-Structured Backend**  

  - Modular design (`frontend`, `backend`, `database_utils`, `prompts`) for scalability.  

  - Easy to extend with new models, retrievers, or UI features.

---

## 🗂️ Project Structure

```plaintext

├── backend_basic.py      # Core RAG pipeline: retrieval & LLM integration

├── database_utils.py     # SQLite utilities: user auth, chat history, threads

├── frontend.py           # Streamlit app for chat interface

├── prompts.py            # System prompts and configurations

├── requirements.txt      # Python dependencies

├── README.md             # Project documentation

└── .gitignore            # Ignored files (db, cache, venv, etc.)