🧠 DocuMind — RAG Chatbot

Upload any PDF, DOCX, TXT, MD, XLSX, or CSV file and chat with it using Retrieval-Augmented Generation powered by Groq LLaMA and ChromaDB.

---

📸 Interface

---

✨ Features

Feature	Description
📄 Multi-Format Support	PDF, DOCX, TXT, MD, XLSX, CSV
🔍 Semantic Search	Vector search via ChromaDB + sentence-transformers
🧠 RAG Pipeline	Retrieval-Augmented Generation with source citations
💬 Interactive Chat	Real-time conversation with your documents
🎯 Source Citations	See exactly where each answer comes from
📊 Token Tracking	Daily usage monitoring with budget warnings
🔐 API Security	Key-based authentication on all endpoints
📱 Responsive UI	Clean dark-themed interface with drag-and-drop upload

---

🚀 Quick Start

Windows

setup.bat
venv\Scripts\activate
python app.py

Linux / macOS

chmod +x setup.sh
./setup.sh
source venv/bin/activate
python app.py

Then open http://localhost:5000 🎉

---

📦 Installation

1. Clone the repo

git clone https://github.com/Tayyabah-Rehman/DocuMind.git
cd DocuMind

2. Configure environment

cp .env.example .env
# Edit .env and add your keys

3. Run setup & start

# Windows
setup.bat

# Linux/macOS
chmod +x setup.sh && ./setup.sh

python app.py

---

📁 Project Structure

DocuMind/
├── app.py                     ← Flask app (all routes)
├── config.py                  ← Typed settings loaded from .env
├── requirements.txt
├── .env                       ← Your secrets (never commit this)
├── .env.example               ← Template
├── setup.bat                  ← Windows one-click setup
├── setup.sh                   ← Linux/macOS one-click setup
│
├── utils/
│   ├── __init__.py            ← Exports all public functions
│   ├── security.py            ← File validation, path safety, API key check
│   ├── document_processor.py  ← Text extraction + chunking
│   ├── rag_engine.py          ← ChromaDB embed / store / retrieve
│   └── token_manager.py       ← Daily token usage tracking
│
├── templates/
│   └── index.html
│
└── static/
    ├── css/style.css
    └── js/app.js

---

🌍 Supported File Types

Format	How It's Processed
PDF	Text extraction via PyPDF2
DOCX	Paragraph extraction via python-docx
TXT / MD	Plain text, read directly
XLSX	All sheets via openpyxl
CSV	Rows read as structured text

Any filename is accepted — no naming restrictions.

---

🔐 Environment Variables

Variable	Description	Default
GROQ_API_KEY	Your Groq API key	Required
GROQ_MODEL	LLM model to use	llama-3.3-70b-versatile
SECRET_KEY	Flask session secret	Required
API_KEYS	Comma-separated API keys	Required
DAILY_TOKEN_BUDGET	Max tokens per day	100000
TOKEN_WARN_THRESHOLD	Warning threshold	70000
SKIP_CONFIG_VALIDATION	Set true in development	false
CHROMA_PERSIST_DIR	Vector DB path	./vectorstore
UPLOAD_FOLDER	Upload directory	./uploads

---

🛠️ Tech Stack

Layer	Technology
Backend	Flask (Python)
LLM	Groq LLaMA 3.3 70B
Vector DB	ChromaDB
Embeddings	sentence-transformers (paraphrase-MiniLM-L3-v2)
Document Parsing	LangChain, PyPDF2, python-docx, pandas, openpyxl
Frontend	HTML5, CSS3, Vanilla JS

---

🔑 Key Design Decisions

• Model loaded once at startup — the embedding model is pre-downloaded by the setup script, not on every upload.
• Groq key stays server-side — the frontend never sees your API key.
• ChromaDB collection naming — always doc_ + 12-char MD5 hash, safely within ChromaDB's 3–63 character limit.
• Duplicate detection — re-uploading the same file skips re-embedding automatically.
• Server-side sessions — Flask-Session writes to disk so large document lists never overflow the cookie.

---

🤝 Contributing

1. Fork the repository
2. Create a feature branch: git checkout -b feature/my-feature
3. Commit your changes: git commit -m 'Add my feature'
4. Push: git push origin feature/my-feature
5. Open a Pull Request

---

📄 License

Distributed under the MIT License. See LICENSE for more information.

---

🙏 Acknowledgments

• Groq — lightning-fast LLaMA inference
• LangChain — RAG framework
• ChromaDB — vector storage
• Sentence-Transformers — embeddings

---

⭐ If DocuMind helped you, give it a star on GitHub!