prettyrepo

This is a cookie-cutter project for the way I like to set up a Python repo.

Components:

• MIT LICENSE.md file and CHANGELOG.md changelog template
• pipenv for dependency management (Pipfile and Pipfile.lock)
• setup.py, setup.cfg, MANFIEST.in for packaging
• flake8 linting
  • Configured in a .flake8 file
  • Extension flake8-quotes to favor single-quoted strings
  • Extension flake8-import-order to force import order
• pytest unit tests
  • .gitignore set up to include runConfigurations to share PyCharm IDE config
  • pytest.ini to configure xunit XML output
  • pytest-flake8 to run flake8 as part of the unit testing and get JUnit XML output
  • Coverage reporting with the help of pytest-cov, configured in a .coveragerc
• .rst README and docs directory for sphinx documentation
• semver.txt to drive the version
• tox to test installation of the package on multiple python versions
  • Also including a Dockerfile to run tox in a dockerized environment rather than manage multiple installations in the test environment
• .gitlab-ci.yml for CI using the GitLab runner
  • Runs pytest --flake8 with some config to get xunit xml output which it uploads so it can be viewed in the pipeline report
  • Runs actual unit tests using the same Docker image and uploads tests
  • In GitLab you can go to Settings → CI/CD → General Pipelines and set the coverage regex to ^TOTAL.+?(\d+.\d+\%)$

Maybe one day part of this repo will be to help automate actually setting things up.

Notes

• readthedocs doesn't support reading requirements from pipenv, so you'll have to run something like pipenv run pip freeze > docs/requirements.txt

<<<<<<< HEAD <<<<<<< HEAD

c814c69... Add badges

Overview

All of the functionality as invoke with a smidge more flexibility (maybe just for me?) than something like invocations.

In general, this project (as well as invoke), give an OS-agnostic solution for automating common commands during development (cleaning build files, running unit tests, etc.). Similar to how a Makefile might be used, prettyrepo makes it so that with a small set of commands you can run any typical task. Furthermore, you can have configurable arguments for each of those tasks being run (see example).

This project introduces a TaskManager so every available task can be tweaked as desired and persistent for a project. Reference tasks.py for this project's own tasks file.

Example

from invoke import task

from prettyrepo import TaskManager
from prettyrepo.utils import run_unittest
from prettyrepo.utils import get_python_cleans

# Create a TaskManager that knows the root of your project
mgr = TaskManager(source_root='prettyrepo', tasks_module=__name__)

# Add tasks from the "task library" or a simple command (by specifying the cmd kwarg)
mgr.add_task('clean', cleans=get_python_cleans())
mgr.add_task('test', lib='unittest', dir_='tests')
mgr.add_task('cov', aliases=['cover'], lib='coverage', func=run_unittest, dir_='tests')
mgr.add_task('lint', aliases=['flake8'], cmd='flake8 prettyrepo', doc='Run flake8 lints.')

# You can use the normal invoke task decorator in conjunction with prettyrepo tasks.
@task(name='bleh', help={'times': 'number of times to say blah'})
def blah(_, times=1):
    """Say blah."""
    print('\n'.join(['blah']*times))


ns = mgr.namespace

With this in your tasks.py file in the root of your repository, you can run

inv <command> [-h] [<command options...>]

Where -h is available to show the help message. For example, running inv clean -h yields:

Docstring:
  Clean unnecessary files.

Options:
  -l INT, --level=INT       the level of "cleanliness" [0-2], default 1
  -v INT, --verbosity=INT   how verbose information to log to the console [0-3], default 1

Where -l and -v can be used to specify what the clean task does (the higher the level, the more files it removes - i.e. venv, dist files; the higher the verbosity, the more info it logs):

Documentation available in the repository wiki.

c77323a... add cookie cutter files