title	Rag Chatbot
emoji	🤖
colorFrom	blue
colorTo	purple
sdk	docker
app_port	8501
tags	streamlit
pinned	false
short_description	RAG Chatbot that answers questions about any PDF or TXT
license	mit

🤖 RAG Chatbot

Ask questions about any document — get instant AI-powered answers.

🔗 Try it live — no install needed

Note: if the demo shows a temporary scheduling error, it's a known Hugging Face infrastructure issue, not a bug in the app. Refresh in a bit, or run locally using the steps below.

---

---

📌 Overview

A RAG (Retrieval-Augmented Generation) chatbot that answers questions based on the content of any document. Supports both PDF and TXT files, uses semantic search (via sentence-transformers) to find the most relevant context, and generates answers using Flan-T5, an instruction-tuned model that paraphrases rather than copying text verbatim. Deployed live on Hugging Face Spaces with Docker — no API key needed.

Core idea: Upload your document. Ask it anything. Get instant answers.

---

✨ Features

• 📄 PDF & TXT Support — Upload any document format
• 🧠 Semantic Search — Finds relevant context using sentence embeddings, not just keyword matching
• 🌐 Web UI — Clean Streamlit interface, live on Hugging Face Spaces
• 💻 CLI Mode — Terminal-based chat, for local/offline use
• 🐳 Dockerized Deployment — Fully containerized and reproducible
• 🔒 Self-contained — No external API key required to run

---

🚀 Live Demo

huggingface.co/spaces/zain-cs/rag-chatbot

Upload a PDF or TXT file, then ask questions about its content directly in your browser.

---

🖥️ Running Locally

1. Clone the repository

git clone https://github.com/zain-cs/rag-chatbots.git
cd rag-chatbots

2. Create and activate a virtual environment

python -m venv venv

venv\Scripts\activate      # Windows
source venv/bin/activate   # Mac/Linux

3. Install dependencies

pip install -r requirements.txt

4a. Run the Web UI

streamlit run app.py

4b. Run the CLI chatbot

python src/chatbot.py

---

🔍 How RAG Works

Your Document (PDF or TXT)
          │
          ▼
Chunk the document into passages
          │
          ▼
Embed chunks + question with sentence-transformers
          │
          ▼
Retrieve the most semantically relevant chunk
          │
          ▼
Feed chunk + question to Flan-T5
          │
          ▼
Generated Answer

---

🗂️ Project Structure

📦 rag-chatbots
┣ 📂 data
┃ ┗ 📄 sample.txt              ← Sample document
┣ 📂 src
┃ ┣ 🐍 chatbot.py              ← CLI chatbot
┃ ┣ 🐍 pdf_loader.py           ← PDF & TXT loader
┃ ┗ 🐍 semantic_search.py      ← Embedding-based chunk retrieval
┣ 🐍 app.py                    ← Streamlit web UI (deployed entry point)
┣ 🐳 Dockerfile                ← Container config for Hugging Face Spaces
┣ 📄 .gitignore
┣ 📄 requirements.txt
┗ 📄 README.md

---

🛠️ Tech Stack

Tool	Purpose
Python 3.10+	Core language
Flan-T5 (base)	Local instruction-tuned text generation model
Sentence-Transformers (MiniLM)	Semantic embeddings for retrieval
HuggingFace Transformers	Model loading and inference
PyTorch	Deep learning backend
Streamlit	Web UI framework
PyMuPDF	PDF text extraction
Docker	Containerized deployment

---

🗺️ Roadmap

• Document loading and chunking
• Interactive CLI chat loop
• PDF & TXT support
• Streamlit web interface
• Semantic search with embeddings
• Deploy on Hugging Face Spaces
• Upgrade generation model to Flan-T5 (instruction-tuned) for higher answer quality
• Experiment with larger models (Flan-T5-large or an API-based LLM) for even better answer quality

---

👤 Author

Zain — @zain-cs

Open to freelance ML and AI projects.

---

📄 License

This project is licensed under the MIT License — free to use, modify, and build on it.