Skip to content

Latest commit

 

History

History
50 lines (39 loc) · 1.78 KB

README.md

File metadata and controls

50 lines (39 loc) · 1.78 KB

Feature Engineering Pipeline

This project streamlines the process of feature engineering using an automated tool named Featuretools. It's designed to efficiently generate meaningful features from relational datasets, facilitating more effective predictive models to support MLOps.

Table of Contents

Getting Started

Step-by-step instructions for setting up and running the app.

Prerequisites

  • Docker - Make sure Docker and Docker-compose are installed.

Setting Up

git clone https://github.com/DimitrisParaskevopoulos/MLOps-feature-factory.git
cd your-local-path/MLOps-feature-factory

Running

docker-compose up -d --build

Usage

  1. Use the FastAPI endpoint at http://localhost:8000/get_raw_data/{ID} to get raw data for specific player ID.

  2. Use the API endpoint at http://localhost:8000/get_features/{ID} to get features extracted for specific player ID. For example you can get stats for the ID you requested: A) COUNT B) MEAN, SUM, MAX, MIN C) SKEW, STD D) MODE, NUM_UNIQUE

Logger

Write log messages to the /app/logs/app.log file within the container, and you can access the log file on the host machine within the specified logs directory.

Github Actions

This workflow is typical for a continuous integration setup, where code changes pushed to the main branch trigger a series of steps to build a Docker image and push it to Docker hub. This workflow is also designed for continuous deployment, where changes to the repository trigger a series of steps to update and run a Docker container to an EC2 instance. Coming soon...