/GeminiChat

GeminiChat is a Chainlit app that uses Google's Gemini model for real-time text generation, featuring an interactive chat interface and robust error handling.

Primary LanguagePython

GeminiChat

GeminiChat is a Chainlit application that integrates with Google's Gemini model to provide real-time text generation capabilities. This app uses Chainlit for the chat interface and Google's Generative AI (Gemini) for generating responses.

Features

  • Real-time text generation using the Gemini model.
  • Intermediate response to indicate that the request is being processed.
  • Error handling for model response issues.

Prerequisites

  1. Python 3.8+: Ensure you have Python 3.8 or higher installed.
  2. Chainlit: Install Chainlit using pip.
  3. Google Generative AI: Obtain a Google API key for accessing Gemini.

Installation

  1. Clone the Repository:

    git clone https://github.com/msnabiel/GeminiChat.git
cd GeminiChat

  2. Install Dependencies:

    pip install chainlit google-generativeai

  3. Set Up Environment Variables:

    • Obtain your Google API key from the Google Cloud Console.
    • Set the environment variable GOOGLE_API_KEY with your API key: 
      export GOOGLE_API_KEY="your-google-api-key"

Configuration

  • Google Generative AI: The app uses the gemini-1.5-flash model from Google's Generative AI. Ensure you have the correct API key and model name.

Usage

  1. Run the Application:

    chainlit run app.py

  2. Interact with the App:

    • Open the Chainlit interface in your browser at http://localhost:8000.
    • Send messages to receive generated responses from the Gemini model.

Code Explanation

  • Imports: Import necessary libraries for Chainlit, Google Generative AI, and asynchronous operations.
  • Model Initialization: Configure and initialize the Gemini model using the API key.
  • Response Generator: Define an asynchronous function to generate responses from the model.
  • Main Function: Handle incoming messages, send an intermediate response, generate the final answer, and update the intermediate response with the final result.

Example

Here is a brief example of how the app functions:

  1. User sends a message.
  2. The app responds with "Processing your request...".
  3. The app generates a response using the Gemini model.
  4. The intermediate response is updated with the final answer.

Error Handling

If the model fails to generate a response, the app will return an error message indicating the issue.

License

This project is licensed under the MIT License.