DocuMind: Your Intelligent Document Assistant 🧠

DocuMind is an advanced document processing and question-answering system that leverages the power of Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) to provide intelligent responses based on your documents and web content. (Note: This has only been tested on MacOS, might not work on other operating systems - it's Open Source so feel free to download and modify it to get it to work on your end)

Features 🌟

Document Processing: Upload and process various document formats (TXT, DOCX, PDF).
Web Crawling: Crawl websites and add their content to the knowledge base.
Intelligent Chatbot: Ask questions about your documents and get informed responses.
RAG System: Utilizes a Retrieval-Augmented Generation system for accurate and context-aware answers.
LaTeX Rendering: Supports rendering of LaTeX equations in chat responses.
Code Highlighting: Displays code snippets with syntax highlighting.

Screenshots 📸

Requirements 📋

Python 3.8+
Streamlit
OpenAI API (compatible with LM Studio)
FAISS
PyPDF2
docx2txt
tiktoken
latex2mathml
And other dependencies listed in requirements.txt

Setup 🚀

Clone the repository:

git clone https://github.com/yorkiedev/documind.git
cd documind

Install the required dependencies:
```
pip install -r requirements.txt
```
Set up LM Studio:
- Ensure LM Studio is running and accessible at http://localhost:1234/v1
- Configure your LM Studio with appropriate models and presets
Run the Streamlit app:
```
streamlit run ragcrawler.py
```

Usage 📚

Document Upload:
- Use the sidebar to upload your documents (TXT, DOCX, PDF formats supported).
- The system will process and index the documents automatically.
Web Crawling:
- Enter a URL in the sidebar and specify the maximum number of pages to crawl.
- Click "Crawl and Add to RAG" to add the web content to your knowledge base.
Chatting with DocuMind:
- Type your questions in the chat input at the bottom of the page.
- DocuMind will provide answers based on the content of your documents and crawled web pages.
Viewing Responses:
- Responses may include formatted text, LaTeX equations, and code snippets.
- LaTeX equations are rendered inline for easy reading.
- Code snippets are displayed with syntax highlighting.
Managing Documents:
- Use the "Clear All Documents" button in the sidebar to reset your knowledge base.

Customization 🛠️

Modify the EMBEDDING_MODEL and CHAT_MODEL constants in the script to use different language models.
Adjust the chunk size and overlap in the DocumentProcessor.split_text method to fine-tune document processing.
Customize the system prompt in the main chat loop to change the AI's behavior and personality.

Contributing 🤝

Contributions to DocuMind are welcome! Please feel free to submit pull requests, create issues or spread the word.

License 📄

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments 👏

Thanks to the Streamlit team for their amazing framework.
Kudos to the developers of FAISS, PyPDF2, and other libraries used in this project.
Special thanks to the LM Studio community for their support and resources.

Happy document processing and chatting with DocuMind! If you have any questions or run into issues, please open an issue on GitHub. Enjoy! 🎉

YorkieDev / DocuMind