2.2.13. Distributing Python software with GitHub, PyPI, Docker Hub, and Read The Docs

The goal of this tutorial is to teach you how to distribute software with GitHub, PyPI, and Read the Docs. GitHub is a code repository that can be used to distribute source code. PyPI is a repository for Python software that makes it easy for users to use pip to install your software. Docker Hub is a repository for Docker containers that makes it easy to distribute entire virtual machines that have your all of your software installed and fully configured. Read The Docs is a repository for documentation that can be use to easily distribute documentation to end users and external developers.

2.2.13.1. Required packages

Execute the following commands to install the packages required for this tutorial:

apt-get install pandoc
pip install \
    pypandoc \
    setuptools \
    twine \
    wheel

2.2.13.2. Prepare your package for distribution

2.2.13.2.1. Annotate the version number of your package in VERSION and __init__.py

Save the following to the VERSION file of your package, e.g. /path/to/intro_to_wc_modeling/intro_to_wc_modeling/VERSION:

0.0.1

Save the following to the __init__.py file of your package, e.g. /path/to/package/intro_to_wc_modeling/__init__.py:

import pkg_resources

with open(pkg_resources.resource_filename('intro_to_wc_modeling', 'VERSION'), 'r') as file:
    __version__ = file.read().strip()
# :obj:`str`: version

2.2.13.2.2. Create a README.md file with an overview of the package

Save a brief description of the package to /path/to/package/README.md. GitHub will display the content of this file on the landing page for the repository. For example:

# intro_to_wc_modeling

The goal of this tutorial is to teach you how to test and document Python code.

2.2.13.2.3. Create a file requirements.txt which lists the required dependencies

The following example illustrates how to use /path/to/package/requirements.txt to specify requirements including how to specify package sources, how to specify version dependencies, and how to specify required options.:

numpy
scipy<=1.2
matplotlib==2.3[option]
git+https://github.com/KarrLab/obj_model.git#egg=obj_model

Packages that should be installed from PyPI should be listed by their names. Packages that should be installed from GitHub should be listed by their GitHub URL.

Version dependencies can be specified with ‘<’, ‘>’, ‘<=’, ‘>=’, and ‘=’.

Required options can be specified by post-pending option names to each dependency.

Similarly, docs/requirements.txt and tests/requirements.txt can be used to specify packages required for testing and documentation.

2.2.13.2.4. Create a file requirements.optional.txt which lists the dependencies

Optional dependencies can be listed in /path/to/package/requirements.optional.txt according this syntax:

[option] dependency_1 dependency_2 …

These optional dependencies can be installed by post-pending the option name(s) during pip and setup.py commands, e.g.:

pip install package_name[option_name]

2.2.13.2.5. Create a license file

Save the following to /path/to/package/LICENSE:

The MIT License (MIT)

Copyright (c) <Year> Karr Lab

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

2.2.13.2.6. Create a setup configuration file

Create a setup configuration file by following this example and saving it to /path/to/package/setup.cfg:

[bdist_wheel]
universal = 1

[coverage:run]
source =
    intro_to_wc_modeling

[sphinx-apidocs]
packages =
    intro_to_wc_modeling

2.2.13.2.7. Create a MANIFEST.in file which describes additional files that should be packaged with your Python code

For example, save the following to /path/to/package/MANIFEST.in:

# documentation
include README.rst

# version
include package/VERSION

# license
include LICENSE

# requirements
include requirements.txt
include requirements.optional.txt

2.2.13.2.8. Create a setup script

You can use the setuptools package to build a install script for your package. Simply edit this template and save it to /path/to/intro_to_wc_modeling/setup.py:

import setuptools
try:
    import setuptools_utils
except ImportError:
    import pip
    pip.main(['install', 'git+https://github.com/KarrLab/setuptools_utils.git#egg=setuptools_utils'])
    import setuptools_utils
import os

name = 'intro_to_wc_modeling'
dirname = os.path.dirname(__file__)

# get package metadata
md = setuptools_utils.get_package_metadata(dirname, name)

# install package
setup(
    name=name,
    version=md.version,

    description='Python tutorial',
    long_description=md.long_description,

    # The project's main homepage.
    url='https://github.com/KarrLab/' + name,
    download_url='https://github.com/KarrLab/' + name,

    author='Jonathan Karr',
    author_email='jonrkarr@gmail.com',

    license='MIT',

    # See https://pypi.python.org/pypi?%3Aaction=list_classifiers
    classifiers=[
        'Development Status :: 3 - Alpha',
        'Intended Audience :: Developers',
        'Topic :: Software Development',
        'License :: OSI Approved :: MIT License',
        'Programming Language :: Python',
    ],

    keywords='python, tutorial',

    # packages not prepared yet
    packages=setuptools.find_packages(exclude=['tests', 'tests.*']),
    package_data={
        name: [
            'VERSION',
        ],
    },
    entry_points={
        'console_scripts': [
            'intro_to_wc_modeling = intro_to_wc_modeling.__main__:main',
        ],
    },

    install_requires=md.install_requires,
    extras_require=md.extras_require,
    tests_require=md.tests_require,
    dependency_links=md.dependency_links.
)

Use the entry_points argument to specify the location(s) of command line programs that should be created. Use the install_requires argument to list any dependencies. Use the tests_require argument to specify any additional packages needed to run the tests.

See The Hitchhiker’s Guide to Packaging for a more detailed explanation of the arguments to setup.

You can test the install script by running it locally:

pip install -e .

2.2.13.3. Distributing source code with GitHub

GitHub can be used to distribute source code simply by changing the public/private setting of a repository. The versions of key revisions should be marked using Git tags as illustrated below. See Section 2.2.8 for more information about using Git and GitHub.:

git add <path>
git commit -m "<message>"
git tag 0.0.8
git push --tags

2.2.13.4. Distributing Python packages with PyPI

Follow the steps below to distribute your code via PyPI.

  1. Create an account at https://pypi.python.org

  2. Save your login information to ~/.pypirc:

    [distutils]
    index-servers =
        pypi
    
    [pypi]
    repository: https://upload.pypi.org/legacy/
    username: <username>
    password: <password>
    
  3. Convert your README.md file to .rst format:

    pandoc --from=markdown --to=rst --output=README.rst README.md
    
  4. Compile your package for source code and binary distribution:

    python2 setup.py sdist bdist_wheel
    python3 setup.py sdist bdist_wheel
    
  5. Upload your package to PyPI:

    twine upload dist/*
    

There are also several online tutorials with more information about how to upload packages to PyPI

2.2.13.5. Distributing containers with Docker Hub

Docker Hub can be used to distribute virtual machines simply by changing the public/private setting of a repository. See Section 2.3.2 for more information about using Docker and Docker Hub.

2.2.13.6. Distributing documentation with Read The Docs

After you have configured Sphinx, committed your code to GitHub, and made your repository public, follow these instructions to configure Read The Docs to compile the documentation for your code upon each push to GitHub. Note, your configuration must follow the Sphinx configuration template in karr_lab_build_utils for Read The Docs to properly compile your documentation. Note also, Read The Docs can only be used to compile and distribute documentation for public GitHub repositories.

  1. Create an account at https://readthedocs.org

  2. Log into Read The Docs

  3. Click the “Import a repository” button

  4. Select the repository that you wish to distribute

  5. Create the project

  6. Use the “Settings” and “Advanced Settings” panels to edit the settings for the project.

    • Set the homepage and tags
    • Set the requirements file to docs/requirements.txt
    • Set the Python configuration file to docs/conf.py
    • Set the Python interpreter to CPython 3.x
  7. Optionally, use YAML files to configure the conda environment used to build the documentation within Read the Docs. This is helpful for documenting packages that depend on OS packages. The default Read the Docs conda environment cannot install OS packages, but some of these dependencies can be obtained from conda.:

    • Add the following to /path/to/package/.readthedocs.yml:

      python:
         version: 3
         setup_py_install: true
      requirements_file: docs/requirements.txt
      conda:
          file: docs/conda.environment.yml
      
    • Add the following to /path/to/package/docs/conda.environment.yml:

      name: <package>-docs
      channels:
        - conda-forge
        - defaults
      dependencies:
        - cython
        - pip
        - python
        - sphinx
        - pip:
          - configparser
          - sphinx_rtd_theme
          - robpol86-sphinxcontrib-googleanalytics
          - sphinxcontrib-bibtex
          - sphinxcontrib-spelling
      
  8. Add your email in the “Notifications panel” so that you receive notifications documentation compilation errors

  9. Check for errors

  • Navigate to “Builds”
  • Click on the latest build
  • Browse the tabs for errors and warnings