Regression Model

Building and deploying a regression ML model.

This code is used in this blog post.

Requirements

Python 3

Installation

The Makefile included with this project contains targets that help to automate several tasks.

To download the source code execute this command:

git clone https://github.com/schmidtbri/regression-model

Then create a virtual environment and activate it:

# go into the project directory
cd regression-model

make venv

source venv/bin/activate

Install the dependencies:

make dependencies

The requirements.txt file only includes the dependencies needed to make predictions with the model. To train the model you'll need to install the dependencies from the train_requirements.txt file:

make train-dependencies

Running the Unit Tests

To run the unit test suite execute these commands:

# first install the test dependencies
make test-dependencies

# run the test suite
make test

# clean up the unit tests
make clean-test

Running the Service

To start the service locally, execute these commands:

uvicorn rest_model_service.main:app --reload

Generating an OpenAPI Specification

To generate the OpenAPI spec file for the REST service that hosts the model, execute these commands:

export PYTHONPATH=./
generate_openapi --output_file=service_contract.yaml

Docker

To build a docker image for the service, run this command:

docker build -t insurance_charges_model:0.1.0 .

To run the image, execute this command:

docker run -d -p 80:80 insurance_charges_model:0.1.0

To watch the logs coming from the image, execute this command:

docker logs $(docker ps -lq)

To stop the docker image, execute this command:

docker kill $(docker ps -lq)