mehdimo / car-price-prediction

A simple end-to-end ML model application using Falsk, Angular, and docker to predict car price

Car Price Prediction App

This is a simple app to show how to create a complete solution that uses a machine learning model for prediction.

Project Structure

root
 +-data: contains the training data.
 |
 +-frontend: frontend application by Angular
 |
 +-model: backend predictor model in python and Flask API.
 

How to Run

Use docker-compose

Note: Make sure you have Docker installed.

• Go in to frontend/src/assets/config.json and change the backendHost to the hosting machine.
• In the root of the project, run docker-compose up
• The two containers will run and your app will be available on the Host IP (port 80).

Use Docker

Both frontend and backend applications come with their own Docker file.

1. Run the model (python):

   • change directory to model/
   • Run docker built -t backend .
   • Run docker run -p 8081:8081 backend
   • The backend app will be running on http://localhost:8081.

2. Run the frontend (angular):

   • change directory to frontend
   • Run docker build -t frontend .
   • Run docker run --rm --name car_price_front_app -p 80:80 frontend
   • The frontend will be running on http://localhost:80.
     • --rm will remove the container after it terminates.
     • --name assigns a name to the container.

Manual

1. Backend

• Navigate to the model/ folder: cd model
• Create virtual env.
  • python -m venv venv
  • source venv/bin/activate
• pip install -r requirements.txt
• cd app
• python app.py

2. Frontend

• Note: Make sure you have Node version 20 (or higher).
• Navigate to frontend/ folder: cd frontend
• Install dependencies: npm install
• ng serve to run your Angular application during development.
  • It's not recommended to use it in production.