Chatbot UI

A web-based chatbot interface that uses Groq's LLM models to provide intelligent responses with PDF document processing support and multi-language capabilities.

Deployed live on Render: https://ngo-chatbot.onrender.com (Due to limited Render resources, the website might take some time to run optimally)

Features

• 🤖 Integrated with multiple LLM models through Groq API
  • LLaMA 3 8B & 70B
  • Mixtral 8x7B
  • Gemma 7B
• 📁 PDF document analysis and context extraction
• 🌐 Automatic language detection and translation
• 💬 Persistent conversation history
• 🎨 Clean, modern dark-themed UI
• 🔄 Model switching without losing context
• To switch languages use terms like {"speak in", "talk in", "reply in", "respond in", "use", "switch to", "change to", "change language to", "habla en", "parle en", "sprich in", "parla in"}

Getting Started

Prerequisites

• Python 3.8+
• Groq API key - Get one here

Installation

1. Clone the repository:

   git clone https://github.com/SpontaneousSecret/ngo_chatbot

2. Install dependencies:

   pip install -r requirements.txt

3. Create a .env file in the project root with your Groq API key:

   GROQ_API_KEY="your_groq_api_key_here"
   

Running the Application

Start the FastAPI server:

uvicorn main:app --reload

The application will be available at http://localhost:8000.

Usage

Web Interface

1. Open your browser and navigate to http://localhost:8000
2. Type your message in the input box and press Enter or click the send button
3. Upload PDFs for document analysis by using the attach button
4. Change models via the dropdown in the top right corner

API Endpoints

• GET / - Serves the web interface
• GET /models - Lists all available models
• POST /chat - Sends a message to the chatbot
• GET /conversations - Lists all conversations
• GET /conversations/{conversation_id} - Gets a specific conversation
• DELETE /conversations/{conversation_id} - Deletes a conversation
• PUT /conversations/{conversation_id}/model - Changes the model for a conversation

API Reference

Chat Endpoint

POST /chat

Parameters:

• message (form) - The user's message
• pdf (file, optional) - PDF file to provide context
• model_id (form, default: "llama3-8b") - Model ID to use
• conversation_id (form, optional) - ID of existing conversation

Response:

{
  "response": "The bot's response",
  "conversation_id": "unique-conversation-id",
  "model_id": "llama3-70b"
}

Architecture

• Frontend: HTML, CSS, JavaScript
• Backend: FastAPI (Python)
• LLM Provider: Groq API
• Document Processing: PDFPlumber
• Language Processing: LangDetect, Deep-Translator

Project Structure

├── main.py                # FastAPI application
├── requirements.txt       # Python dependencies
├── .env                   # Environment variables (create this)
├── static/                # Static web files
│   ├── index.html         # Web interface
│   ├── style.css          # CSS styles
│   └── script.js          # Frontend JavaScript
└── tools/                 # Utility modules
    ├── pdf_tool.py        # PDF processing utilities
    └── language_tool.py   # Language detection and translation

Development

Adding New Models

To add a new model, update the AVAILABLE_MODELS dictionary in main.py:

AVAILABLE_MODELS = {
    "new-model": {
        "id": "model-id-from-groq",
        "provider": "groq",
        "max_tokens": 8192,
        "description": "Description of the model"
    },
    # ... existing models
}

Extending the UI

The frontend UI is built with vanilla JavaScript. To extend it:

1. Modify the HTML structure in static/index.html
2. Update the styles in static/style.css
3. Add functionality in static/script.js

Future Improvements

• Add authentication for user accounts
• Implement file attachments besides PDF
• Add search functionality for conversation history
• Support for streaming responses
• Database integration for persistent storage

License

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

Acknowledgments

• FastAPI
• Groq
• PDFPlumber