cookiecutter-dcpypackage

A cookiecutter template, with the author's preferred Python package setup.

Features:

• Does not assume your project will live on GitHub. There are cookiecutter variables for the project homepage and issues page.
• pytest for unit tests, with integration into setup.py.
  • pytest-cov plugin for test coverage reporting.
  • pytest-sugar plugin for pretty test output.
  • HTML coverage reports are enabled by default. See the htmlcov directory.
  • pytest configuration for py.test --runslow option that runs slow tests. There is a @slow decorator defined in tests/common.py for marking slow tests.
  • looping test runner script that will run your tests in a loop. Simply hit any key to re-run tests, or ctrl-c to quit. Takes two arguments: the max number of loops, and the optional word 'slow' to indicate that slow tests should be included.
• pylint, with tests excluded by default.
• sphinx documentation, with shared files between Sphinx and the Python package files, for reduced maintenance.
• bumpversion for version management
• Universal build, targetting Python 2 and 3.
  • python-future included for writing backward-compatible Python 3 code.
• Most build artifacts go to a common build directory.
• Tests directory is underneath the module directory, as this allows easy use of the pkg_resources API for unit test data files.
• Basic tox configuration.
• Utility makefile for common tasks.
• Entry points for both CLI and GUI scripts are generated. Remove either or both if not required. I required one or the other often enough that having them generated suits me.

Inspiration, and some boilerplate text taken from other templates, including:

• cookiecutter-pypackage
• cookiecutter-pylibrary

Usage

Install cookiecutter if you haven't already. Then, create a project in your current working directory using:

$ cookiecutter gh:DC23/cookiecutter-dcpypackage

Once your project is created, the next steps should be to create a virtual environment (using either virtualenv/virtualenvwrapper or the virtual environment functionality built into Python 3):

$ mkvirtualenv -a . my_project

With the virtual environment activated, install the initial development dependencies including installing your package in development mode:

$ make develop

After that, you can start customising the templates to suit your needs. Run make with no target for a list of make options:

$ make

Note that bumpversion and tox do not have make targets. It is easier to run them directly.