Techietalks AI - Multi-Model Chat Assistant with RAG (Retrieval-Augmented Generation)

Welcome to the Techietalks AI repository! This project is a multi-model AI-powered chat assistant built using Streamlit, Pydantic_AI, and multiple AI models like OpenAI, DeepSeek, and Gemini. It allows users to interact with different AI models, switch between them, and enjoy a conversational experience. Additionally, this version includes Retrieval-Augmented Generation (RAG) functionality, enabling the assistant to answer questions based on uploaded PDF documents.

This repository contains all the necessary files to set up and run the chat assistant locally or in a Docker container. Below, you'll find a detailed explanation of the project and instructions to get started.

📁 Repository Structure

Here’s a breakdown of the files in this repository:

1. .DS_Store: A macOS-specific file that stores folder attributes (e.g., icon positions). You can ignore this file.
2. .env: A file to store environment variables, such as API keys. Important: Add your OPENAI_API_KEY, DEEPSEEK_API_KEY, and GEMINI_API_KEY here to authenticate with the respective APIs.
3. .gitignore: Specifies files and folders that Git should ignore (e.g., .env to avoid exposing sensitive information).
4. Dockerfile: Contains instructions to build a Docker image for the Streamlit application. It sets up a Python environment, installs dependencies, and runs the Streamlit app.
5. app.py: The main application file. It contains the code for the Streamlit-based chat interface, integrates with multiple AI models (OpenAI, DeepSeek, and Gemini), and includes RAG functionality for processing PDF documents.
6. docker-compose.yml: A configuration file to run the application using Docker Compose. It sets up two services: the chatbot (Streamlit app) and the webserver (NGINX for serving static content).
7. requirements.txt: Lists all Python dependencies required to run the application.
8. sree.txt: A placeholder text file (likely for personal notes or testing).
9. data/: A directory containing:
   • pdfs/: Stores uploaded PDF documents.
   • vector_store/: Stores vector embeddings of the PDF documents for RAG functionality.
10. web/static/index.html: A static HTML file served by the NGINX web server. You can customize this file to add additional content or documentation.

🚀 How It Works

This project uses the following technologies:

• Streamlit: A framework for building interactive web applications with Python. It powers the chat interface.
• Pydantic_AI: A library that helps structure and validate AI responses.
• OpenAI API: A powerful AI model for generating conversational responses.
• DeepSeek API: An AI model that provides conversational responses.
• Gemini API: Another AI model for generating responses.
• Retrieval-Augmented Generation (RAG): A technique that combines retrieval of relevant information from documents (e.g., PDFs) with generative AI to provide context-aware answers.
• NGINX: A web server used to serve static content (e.g., HTML files).
• Rich: A library used for pretty-printing debug information (e.g., conversation history).

When you type a question into the chat interface, the app sends it to the selected AI model (OpenAI, DeepSeek, or Gemini), processes the response, and displays it in a conversational format. It also maintains a history of the conversation, allowing the AI to provide more context-aware answers. The "New Chat" button resets the conversation, clearing the history and starting fresh.

The RAG functionality allows users to upload PDF documents, which are processed and used to provide context-aware answers. The app uses ChromaDB to store vector embeddings of the documents and retrieve relevant information when answering questions.

🛠️ Setup Instructions

Follow these steps to set up and run the project:

Prerequisites

1. Python 3.11: Ensure Python is installed on your system.
2. Docker: Required to run the application in containers.
3. API Keys: Obtain API keys for OpenAI, DeepSeek, and Gemini from their respective websites.

Steps

1. Clone the Repository: bash git clone https://github.com/schogini/techietalksai.git cd techietalksai/0007

2. Set Up Environment Variables:

   • Create a .env file in the project directory.
   • Add your API keys: plaintext OPENAI_API_KEY=your_openai_api_key_here DEEPSEEK_API_KEY=your_deepseek_api_key_here GEMINI_API_KEY=your_gemini_api_key_here

3. Run the Application:

   • Using Docker Compose: bash docker-compose up --build
   • Once the containers are running:
     • Access the Chat Assistant at http://localhost:8502.
     • Access the Web Server (static content) at http://localhost:8503.

4. Upload PDFs:

   • Use the sidebar in the chat interface to upload PDF documents. The app will process them and use their content to provide context-aware answers.

📄 License

This project is open-source and available under the MIT License. Feel free to use, modify, and distribute it as needed.

📧 Contact

For AI consultancy, training, and development, contact Schogini Systems Private Limited at https://www.schogini.com.

👨‍💻 Author

Sreeprakash Neelakantan
- Website: https://www.schogini.com
- GitHub: https://github.com/schogini/techietalksai.git

Enjoy using the Techietalks AI Multi-Model Chat Assistant with RAG! If you have any questions or feedback, feel free to reach out. 😊