CSE 587 - Rainfall Prediction Project

This repository contains a Dockerized application for predicting whether it will rain in Australia. The application consists of a frontend and a backend, both running in separate Docker containers.

Installation and Setup

Prerequisites

• Docker installed on your machine

Steps

1. Clone this repository:

   git clone https://github.com/imounish/cse587-rainfall-prediction.git

2. Navigate to the cloned repository:

   cd cse587-rainfall-prediction

3. Create a .streamlit/secrets.toml file in the current working directory cse587-rainfall-prediction and add the following secrets:

   predict_path = "predict"
   
   [backend.local]
   url = "http://backend:8000"

4. Build and run the Docker containers using Docker Compose:

   docker-compose up --build

5. Once the containers are up and running, you can access the application in your web browser at http://localhost:8501.

Directory Structure

• frontend: Contains the frontend code written in Streamlit.
• backend/: Contains the backend code written in Python FastAPI.
• docker-compose.yml: Docker Compose file to orchestrate the containers.

Usage

• Visit http://localhost:8501 in your web browser to access the Weather Prediction App.
• Enter the required input features in the frontend UI and submit the form.
• The frontend will make a POST request to the backend API endpoint, which will predict whether it will rain based on the input features.

API Endpoint

• The backend server exposes an API endpoint at /predict which accepts POST requests containing input features and returns the prediction.

Notes

• Make sure to create the .streamlit/secrets.toml file as mentioned in the installation steps to securely store the API endpoint URL.

License

This project is licensed under the MIT License - see the LICENSE file for details.