Python Package Project (Example)

Introduction

This package is a simple implementation of linear regression between two variables: housing price against house area. The package contains the implementation of the algorithm and the unit tests for the algorithm. It is intended to be used as a guide for those who are developing a project in python.

Installation

First, clone the repository:

git clone ...

Second, create the environment:

cd Python_Package_Project
make env

Third, activate the environment:

source vLocalEnv/bin/activate

Fourth, install the package:

cd ..
pip install -e Python_Package_Project

where Python_Package_Project is the name of the directory where the package is located and it is installed in development mode (flag -e). For more information about the installation, please visit: https://packaging.python.org/en/latest/guides/distributing-packages-using-setuptools/

Development

Requirements

The package was developed using Python 3.9.16 and make. The following packages are required to run the code:

• jupyter
• matplotlib
• numpy
• scipy

For more information about the requirements, please visit the requirements.txt file.

For the development, the following packages were used:

• pytest
• pytest-mpl
• pytest-mock
• autopep8

For more information about the development requirements, please visit the requirements-dev.txt file.

Format

This package use autoPEP8 to format the code. For more information about autoPEP8, please visit: https://peps.python.org/pep-0008/

Development Commands

The following commands are available for the development:

• make env: Create a virtual environment, if not exist, and install the requirements. The environment is created in the vLocalEnv directory in the root of the project.
• make dev-env: Create venv directory if not exist with dev requirements. The environment is created in the vLocalEnv directory in the root of the project.
• make test: Run tests quickly with the Python in the virtual environment vLocalEnv.
• make clean: Remove all build, test, coverage and Python artifacts
• make test-image: Run tests quickly with the Python in the virtual environment vLocalEnv and generate the images for the unit tests.
• make dist: Builds source package with sdist. For more information about sdist, please visit: https://packaging.python.org/guides/distributing-packages-using-setuptools/#source-distributions

Unit Tests without Make

To generate the images base-line for the unit tests, run the following command:

pytest -k "TestGetPlotForBestFitLine" --mpl-generate-path tests/visualization/baseline

This image will be saved in the tests/visualization/baseline directory. To run the unit tests and compare the generated images with the base-line images, run the following command:

pytest -k "TestGetPlotForBestFitLine" --mpl

To run the unit tests without make, run the following command:

pytest --mpl

Badges

To generate the badges travis and codecov, please visit: https://docs.travis-ci.com/user/status-images/ and https://docs.codecov.io/docs/quick-start

Notes

• If some tests are added and they have the same name but in different files, is necessary to make subpackages of the tests. For more information, please visit: https://docs.pytest.org/en/6.2.x/goodpractices.html#tests-outside-application-code