Open Image API

This API provides several endpoints to extract text from images using OpenAI's GPT-4 with Vision model and generate an image of a store menu using OpenAI's DALL-E model. Additionally, it offers an endpoint to create variations of input images.

Table of Contents

• Setup
  • Prerequisites
  • Installation
• Usage
  • Running the Server
  • Endpoints
• Contributing
• License

Setup

Prerequisites

Before running the Flask API, make sure you have the following installed:

• Python 3.x
• pip (Python package installer)
• Docker (optional)

Installation

1. Clone this repository to your local machine:

   git clone https://github.com/YounoussaBen/OpenAI_OCR.git
   cd OpenAI_OCR

2. Create a virtual environment named env:

   python3 -m venv env

3. Activate the virtual environment:

   source env/bin/activate

4. Install dependencies using pip within the virtual environment:

   pip install -r requirements.txt

5. Set up your environment variables:

   Create a .env file in the root directory of your project and add the following:

   OPENAI_API_KEY=your_openai_api_key_here
   

Usage

Running the Server

You can run the Flask server using either Python directly or Gunicorn. Additionally, there's a script available to simplify the process.

Running with Python

To run the Flask server with Python, execute the following command:

python app.py

The server will start running on http://localhost:5000 by default.

Running with Gunicorn

You can also run the Flask server with Gunicorn. Use the provided run.sh script:

./run.sh

The server will start running with Gunicorn on http://localhost:5000 by default.

Running with Docker

Alternatively, you can run the Flask server using Docker:

docker build -t my-flask-app .
docker run -p 5000:5000 my-flask-app

Endpoints

Extract Text from Image

• URL: /extract_text
• Method: POST
• Request Body:
  • image (file): An image file to extract text from.
• Response: Returns a JSON object with the extracted text or an error.

Example Request:

curl -X POST http://localhost:5000/extract_text -F "image=@path_to_your_image.jpg"

Example Response:

{
  "extracted_text": "The extracted text will be here."
}

Generate Store Menu Image

• URL: /generate_menu_image
• Method: POST
• Request Body:
  • products (list): A list of products, where each product is a dictionary with a name and price key.
  • description (string): A description of the shop.
• Response: Returns a JSON object with the URL of the generated menu image or an error.

Example Request:

{
  "products": [
    {"name": "Coffee", "price": 2.5},
    {"name": "Sandwich", "price": 5.0},
    {"name": "Salad", "price": 7.0},
    {"name": "Juice", "price": 3.5}
  ],
  "description": "A charming little bakery known for its fresh bread and pastries."
}

Example Response:

{
  "image_url": "URL_of_the_generated_image"
}

Create Image Variation

• URL: /create_image_variation
• Method: POST
• Request Body:
  • image (file): An image file for which a variation is to be created.
• Response: Returns a JSON object with the URL of the generated variation image or an error.

Example Request:

curl -X POST http://localhost:5000/create_image_variation -F "image=@path_to_your_image.jpg"

Example Response:

{
  "image_url": "URL_of_the_generated_variation_image"
}

Contributing

Contributions are welcome! If you find any issues or want to add new features, feel free to open a pull request.

License

This project is licensed under the MIT License.