Welcome to squelette’s documentation!

This package is a template package for my new python projects.

Features

Binary

  • There are several ways to call the binary.

    • The package contains a __main__ module, so you can use python -m squelette [ARGS].

    • There is an entry point in the setup file. That way, the binary is correctly installed, whatever the platform, and once installed, you can use squelette [ARGS].

    • In the bin directory, not shipped when distributing the program, there is a squelette binary. In development environment, you can use ./bin/squelette [ARGS].

    • The package is translated using gettext.

  • My usual logging configuration is set.

Tests

Several parts are tested; everything is configured in tox.ini.

  • First of all, tox is used to aggregate all of the following tests. To perform all the tests (with different python versions), in the root project directory, run:

    $ tox

  • To run unittest, in the root directory, run:

    $ python -m unittest

  • Unittest call also run doctest present accross modules. Thus, you can freely add doctests anywhere in your program, and they will be performed along with unittests or tox. This part is performed in squelette/test.py.

  • It can be a bit cumbersome to add tests for a program or library acting on files. File test/test_add/__init__.py scans module directories for couples of .in and .out, and creates one unittest.TestCase per couple. Using one test case per couple, when a test fails, it is easier to track which couple of files caused the problem than if tests where performed in a loop in a single test case.

    Note that for simple cases (like this one), this could also have been achieved by using unittest.TestCase.subTest().

  • Tox also checks that Code formatting runs without any error.

  • Tox also checks that Documentation is compiled without any warning or error.

Release and Distribution

  • Hatch is used with pyproject.toml file, with my usual configuration. Among others:

    • The README.md file content is used as the long description (don’t repeat yourself).

    • Entry points are used, so that binaries are automatically installed the right way, disregarding specific platform.

  • My usual release process is described in file release.rst.

  • Stdeb is used to build quick and dirty debian packages. It is configured in stdeb.cfg.

  • To access packages data, no matter what release are, they are placed in package sub directories, and accessed using importlib.resources. An example is available in file squelette/__main__.py.

  • A build hook is used to automatically compile gettext message catalogs while building the python package: hatch_build_babel.py.

Documentation

  • My usual README.rst is provided, with links, basic documentation, and badges.

  • The documentation is built using sphinx.

  • Sphinx compilation is tested by tox.

  • Readthedocs is configured using a YAML file.

Code formatting

  • The code is Pylint compliant. Configuration is done in the pylintrc file.

  • The code is formatted using black. If pre-commit is properly installed and configured (i.e. pip install pre-commit to install, and pre-commit install to set up pre-commit hooks), black is automatically run each time you git commit.

  • Tox tests that pylint and black raises no error.

Download and install

See the main project page for instructions, and changelog.

Usage

Here are the command line options for squelette.

Python package template

usage: squelette [-h] [--version] [NUMBER ...]

Positional Arguments

NUMBER

List of numbers to add

Named Arguments

--version

Show version

As a template, this program sums numbers given in argument.