Tap30 Demand Prediction API

This project is an API that uses a demand prediction model to predict the demand for Tap30 service. The API is implemented using FastAPI framework.

Prerequisites

• Python 3.8+
• Required libraries:
  • FastAPI
  • Uvicorn
  • XGBoost (for the model)
  • joblib
  • numpy
  • mysql-connector-python (for MariaDB connection)

Installation

To set up the project and install required libraries, follow these steps:

1. Create a virtual environment (optional):

   python -m venv venv

2. Activate the virtual environment:

   • For Windows:

     .\venv\Scripts\activate

   • For macOS/Linux:

     source venv/bin/activate

3. Install required libraries:

   pip install fastapi uvicorn joblib numpy scikit-learn mysql-connector-python

4. Place your trained model file (model.joblib) in the project directory.

Using the API

1. Running the Server

To start the server, use the following command:

uvicorn main:app --reload

The server will now be running at http://127.0.0.1:8000.

2. Sending a Request to the API

To send data to the API and get predictions, you can use tools like Postman or curl.

POST Request:

• URL: http://127.0.0.1:8000/predict
• Method: POST
• Request body (JSON):

  {
      "day": 100,
      "hour": 8,
      "row": 2,
      "col": 6,
  }

Response:

On success, the response will look like this:

{
  "input_features": {
    "day": 100,
    "hour": 8,
    "row": 2,
    "col": 6,
  },
  "predicted_demand": 50
}

3. Error Messages

If the input data is missing required fields, the API will return an error message. For example, if a required field is missing:

{
  "detail": [
    {
      "loc": ["body", "day"],
      "msg": "Field required",
      "type": "missing"
    },
    {
      "loc": ["body", "hour"],
      "msg": "Field required",
      "type": "missing"
    }
  ]
}

Important Notes

1. If you are using MariaDB to store data, you need to configure the database connection details in the main.py file.

2. This project is set up to run locally, but you can deploy it on a server or platforms like Heroku or AWS for production use.

Files

• main.py: FastAPI code for processing the prediction and returning results.
• model.joblib: The trained model file saved using XGBoost or any other model you have used.
• requirements.txt: A list of Python dependencies for quick installation.

Development and Contributing

• If you have any suggestions or issues, feel free to open an issue on the GitHub repository.
• Contributions to the development of this project are welcome.

Resources

• FastAPI Documentation
• Uvicorn Documentation
• XGBoost Documentation