Project Overview

Given a pre-trained, sklearn model that has been trained to predict housing prices in Boston according to several features, such as average rooms in a home and data about highway access, teacher-to-pupil ratios, and so on. This project is about operationalizing a Python flask app—in a provided file, app.py—that serves out predictions (inference) about housing prices through API calls. This project could be extended to any pre-trained machine learning model, such as those for image recognition and data labeling.

Steps to build and test the project

The files in this project are :

• .crcleci/config.yml : CI configuration file used in circleci for auromatic source code integration
• model_data : data used to train the prediction model
• app.py : the flask application that serve the prediction
• Dockerfile : Used to build a docker image for this project
• make_prediction.sh : posts the features to the flask API to make a prediction
• Makefile : steps to setup the project(includes setting up a python venv, intalling dependencies and linting source code)
• run_docker.sh : builds project's docker image and runs it in a container
• upload_docker.sh: tags and uploads the created image to a docker hub registry
• run_kubernetes.sh : deploys the image into a local kubernetes cluster (use minikube and kubectl)

Setup the Environment

• Create a virtualenv and activate it
• Run make install to install the necessary dependencies

Running app.py

1. Standalone: python app.py
2. Run in Docker: ./run_docker.sh
3. Run in Kubernetes: ./run_kubernetes.sh
4. Make prediction : ./make_prediction.sh

Kubernetes Steps

• Setup and Configure Docker locally
• Setup and Configure Kubernetes locally
• Create Flask app in Container
• Run via kubectl