🚀 ChatRAG ( Chatbot AI Agent & Knowledge Base Injector)

LLM-powered tool for working with your datafiles and other data files locally on your own system or server.

Also make your LLM and other models more knowledgeable using this data injector.

📌 Description

This project uses Python for both front-end and back-end, connecting to your LLM for easy access. For ease of access, this project requires an API or another endpoint like Ollama, LM Studio, etc.

• Users can access their personal private files on their own local systems using LLMs for fast understanding and advanced reasoning.
• Can run on any LLM that your system supports (results may vary from model to model).
• Read the requirements file to see technologies used.

Demo

https://drive.google.com/file/d/1DkBTtHKwj6ADI5ju6SYtKZYwiiBjwpsX/view?usp=sharing

🌟 Features

• ✨ Real-time data processing, understanding, and advanced reasoning.
• 🚀 Multi-platform support.
• 🔐 Full security as everything happens locally.
• 📊 Customizable LLMs data and can be used for other file types and files.

📦 Installation

• Recommended to run on WSL (Windows Subsystem for Linux) or a Linux distro for the best CORS (Cross-Origin Resource Sharing) experience, as Windows CORS may be limited depending on user customization and settings.

# Clone the repository
git clone https://github.com/Angad2005/ChatPDF.git

# Navigate to the directory
cd ChatPDF

# Optional: Create and activate a virtual environment
python3 -m venv env          # Create virtual environment
source env/bin/activate      # On Windows use: .\env\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Run tool
streamlit run main.py

# Optional:
# Change API endpoints in models.py file
# OpenAI for direct API integration
def get_llm():
    return OpenAI(
        openai_api_key="not-needed",
        openai_api_base="http://192.168.29.31:1234/v1",
        model_name="lmstudio-ai/your-model-name"
    )

Example of LLM Endpoints

LM Studio

LLM_BACKEND=lmstudio  
LMSTUDIO_API_KEY=your_lmstudio_api_key  # Usually "EMPTY" if authentication is disabled  
LMSTUDIO_ENDPOINT=http://localhost:1234/v1  

Ollama

LLM_BACKEND=ollama  
OLLAMA_HOST=http://localhost:11434  
OLLAMA_MODEL=llama3  # Replace with your desired model name (e.g., "mistral", "phi3")  

🧪 Usage

# Start the app
streamlit run main.py

⚙️ Configuration

File Upload Settings

Edit main.py to customize file types:

# File upload (supports multiple formats)
uploaded_files = st.file_uploader(
    "Upload Documents",
    type=["pdf", "docx", "txt", "csv", "xlsx"],  # Add/remove formats here
    accept_multiple_files=True
)

LLM Backend Configuration

Update models.py to set your LLM backend.

LM Studio Example

def get_llm():
    return OpenAI(
        openai_api_key="EMPTY",  # Use "EMPTY" if authentication is disabled
        openai_api_base="http://localhost:1234/v1",
        model_name="lmstudio-ai/your-model-name"
    )

Ollama Example

def get_llm():
    return OpenAI(
        openai_api_key="EMPTY",
        openai_api_base="http://localhost:11434/v1",
        model_name="llama3"  # Replace with your desired model
    )

🧩 Supported LLM Backends

LM Studio

LLM_BACKEND=lmstudio  
LMSTUDIO_API_KEY=your_lmstudio_api_key  # Usually "EMPTY"
LMSTUDIO_ENDPOINT=http://localhost:1234/v1  

Ollama

LLM_BACKEND=ollama  
OLLAMA_HOST=http://localhost:11434  
OLLAMA_MODEL=llama3  # Replace with your desired model (e.g., "mistral", "phi3")

🤝 Contributing

Contributions are welcome! Whether you're fixing bugs, improving documentation, or adding new features, follow these guidelines:

1. Fork the Repository
   Click the "Fork" button at the top-right of the GitHub page.

2. Create a Feature Branch

   git checkout -b feature/your-feature-name

3. Make Changes

   • Follow PEP8 style guidelines.
   • Write clear, descriptive commit messages.
   • Test your changes thoroughly.

4. Submit a Pull Request Push your branch to your fork:

   git push origin feature/your-feature-name

   Open a pull request on the original repository with:

   • A clear title and description
   • Screenshots (if applicable)
   • Relevant issue number (if fixing a bug)

5. Code of Conduct
   We follow the Contributor Covenant Code of Conduct. Please report any unacceptable behavior.

6. Reporting Issues

   • Check the GitHub Issues tab first to avoid duplicates.
   • Include:
     • Detailed steps to reproduce
     • Expected vs. actual behavior
     • Screenshots/logs (if relevant)

📄 License

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