Flask Model Server

This project hosts a Flask server that serves predictions from multiple models.

---

How to Run the Server

1. Set up a Virtual Environment

   python -m venv venv

2. Activate the Virtual Environment

   • On macOS/Linux:

     source venv/bin/activate

   • On Windows:

     venv\Scripts\activate

3. Install Dependencies

   pip install -r requirements.txt

4. Run the Server

   flask run

---

How to Add a New Model

1. Create a New File inside the models/ Directory

2. Define a Function with the @register_model Decorator

   Example:

   from .registry import register_model
   from PIL import Image
   import numpy as np
   
   
   @register_model("xo")  # Replace "xo" with your desired action name
   def xo_model(img: Image.Image) -> str:
       result = np.array(detect_tic_tac_toe(img)).flatten().tolist()
   
       m = {"X": 1, "O": 2, "-": 0}
   
       logger.info(f"Result: {result}")
   
       return ",".join([str(m[item]) for item in result])

   • The function must:
     • Take a PIL.Image.Image as input.
     • Return a comma-separated string as output.

3. Model Files

   • If your model requires any additional files, add them to the model_files/ directory.
   • Use clear and descriptive file names based on your model's purpose.

   Example directory structure:

   model_files/
     ├── xo.pt
     ├── some_other_model_file.pt
   

---

How to Call the API

• Method: POST
• Endpoint: /predict
• Query Parameter: action (the action name you registered)
• Request Body: raw image binary buffer (NOT multipart/form-data)