AI Translation Bot

This project is a demo repository for an AI Bot that helps with translation. The bot is hosted in a FastAPI app and can be accessed from a simple website built with Streamlit or Gradio. The frontend allows users to select two languages (to/from) from the available options: English, Spanish, Chinese, and Japanese. The app targets the OpenAI GPT-4o-mini model hosted in Azure for translation.

Setup and Run the Project

Prerequisites

• Python 3.10 or higher
• Azure account with access to OpenAI GPT-4o-mini
• API keys for Azure OpenAI

Installation

1. Clone the repository:

   git clone https://github.com/jsburckhardt/ai-translation-bot-demo.git
   cd workspace-blank

2. Install the dependencies:

   pip install -r requirements.txt

Running the FastAPI App

1. Navigate to the app directory:

   cd app

2. Start the FastAPI server:

   uvicorn main:app --reload

Running the Frontend

1. Navigate to the frontend directory:

   cd frontend

2. Start the Streamlit or Gradio app:

   streamlit run app.py  # For Streamlit
   # or
   python app.py  # For Gradio

Using the Makefile

You can use the provided Makefile to run the frontend and backend simultaneously.

1. To run the frontend:

   make frontend

2. To run the backend:

   make backend

3. To run both frontend and backend:

   make both

Development Environment

This project includes a .devcontainer setup for development environments. The .devcontainer directory contains the necessary configuration files to set up a development container using Visual Studio Code and Docker.

Feedback Functionality

The project includes feedback functionality implemented in both the backend and frontend. Users can provide feedback on translations using thumbs up or thumbs down buttons.

Testing API Endpoints

You can use the test.rest file to test the API endpoints. The file contains sample requests for the translation and feedback endpoints.

Dependencies

The requirements.txt file lists all the dependencies for the project. Make sure to install them using the following command:

pip install -r requirements.txt

Usage

1. Open your web browser and go to http://localhost:8501 for Streamlit or the appropriate URL for Gradio.
2. Select the languages for translation (to/from).
3. Input the text you want to translate and submit the form.
4. View the translated text on the frontend.
5. Provide feedback with thumbs up or thumbs down.

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

License

This project is licensed under the MIT License.