demo-rag-chat4docs-python

This app is a chatbot for uploaded documents with Retrieval-Augmented Generation (RAG). It works with different LLMs, e.g., OpenAI, Google Generative AI, and models hosted by HuggingFace and Ollama.

Prerequisites

• Python 3.9 or higher+
• OpenAI API Key
• HuggingFace API token
• Ollama installed and the models in the code pulled
• Setting up the Environment Variables
  • OPENAI_API_KEY: OpenAI API Key
  • OPENAI_BASE_URL: OpenAI Base URL. Optional
  • HUGGINGFACEHUB_API_TOKEN: HuggingFace API token
  • GOOGLE_API_KEY: Google Generative AI API Key`

Get Started

1. Clone this repo
2. Install Python
3. Run the following command in the terminal

   pip install -r requirements.txt
   

4. Start the app by running the following command

   streamlit run app.py
   

5. Upload your documents, which can be in PDF, DOCX, TXT, or format.
6. Ask questions about your document.

Built With

• Streamlit
• LangChain
• OpenAI
• Faiss
• Google Generative AI
• Ollama
• HuggingFace

Next Steps

• Supporting more LLMs
• Supporting more embedding models
• Supporting more vector stores
• Supporting more document formats
• Fixing bugs
• Refactoring the code