Update packaging guide (#44)

* Update packaging guide to refer to 'flit'

* Switch from 'flit' to 'poetry'

* Update to pyansys-templates

Co-authored-by: Jorge Martinez <[email protected]>

Author: greschd

doc/source/library_description/packaging.rst

@@ -2,73 +2,229 @@
 
 Packaging
 #########
-A Python package organizes and structures a Python library, which contains several
-modules and assets such as examples or binary extensions. A Python package
-offers an easy, reliable, and comprehensive way to distribute and install
-a Python library on a variety of platforms and environments.
+
+A Python package organizes and structures a Python library, which contains
+several modules and assets such as examples or binary extensions. A Python
+package offers an easy, reliable, and comprehensive way to distribute and
+install a Python library on a variety of platforms and environments.
+
+.. note::
+
+   If you want to create a new PyAnsys project according to the guidelines
+   presented in the following lines, consider using the `ansys-templates tool`_.
+
+
+Python Scripts, Modules, Sub-packages, and Packages
+---------------------------------------------------
+
+It is important to understand the difference between Python scripts, modules,
+sub-packages, and packages:
+
+* ``Script``: Any Python file with logic source code.
+* ``Module``: Any Python script hosted next to an ``__init__.py`` file.
+* ``Sub-package``: Any directory containing various Python modules.
+* ``Package``: Any directory containing Python modules and sub-packages.
+
+The following structure is shown to better explain previous concepts:
+
+.. code:: bash
+
+    .
+    ├── src
+    │   └── package
+    │       ├── subpackage_a
+    │       │   ├── __init__.py
+    │       │   └── module_c.py
+    │       ├── __init__.py
+    │       ├── module_a.py
+    │       └── module_b.py
+    ├── LICENSE
+    ├── README.rst
+    └── pyproject.toml
+
 
 Namespace Packaging
 -------------------
-A PyAnsys library uses `namespace packaging`_.
-Namespace packages allow a user to easily split subpackages from a package into
-a single, independent distribution.
+A PyAnsys library uses `namespace packaging`_.  Namespace packages allow you
+to easily split sub-packages from a package into single, independent
+distributions.
+
+There are different approaches available for creating a namespace package. For
+the ``ansys`` namespace, we use the `PEP 420`_ `native namespace packages`_
+approach.
+
+Therefore, the source directory of any `PyAnsys library` should look like this:
+
+.. code:: bash
 
-Three different approaches are currently available for creating a namespace package:
+    .
+    └── src
+        └── ansys
+            └── product
+                └── library
+                    └── __init__.py
 
-* `native namespace packages`_
-* pkgutil-style namespace packages
-* pkg_resources-style namespace packages
 
 Required Files
 --------------
 
-* README.rst file: Describes the purpose of the package.
+* ``README.rst`` file: Describes the purpose of the package.
   *The format of this file must be reStructuredText.*
 
-* LICENSE file: Specifies copyrights and required authorization.
+* ``LICENSE`` file: Specifies copyrights and required authorization.
 
-* setup.py file: Provides package information.
-  The presence of this file indicate that the package was likely created using ``disutils``,
-  which is the Python standard for building and distributing a Python package.
+* ``pyproject.toml`` file: Provides package metadata and defines how the package
+  is built. There are different build backends available, such as `setuptools`_,
+  `poetry`_, and `flit`_.
 
+* ``src/ansys/product/library/__init__.py`` file: Usually contains the
+  version of the package in a variable named ``__version__``. The value of this
+  variable can be parsed from the ``pyproject.toml`` file so that the version 
+  is only specified in one location.
 
-Setup File
-----------
-The `setup.py`_ file is the build script for ``setuptools``. It exposes dynamic metadata and contains
-package information, such as a description, author, and version.
-In this file, the ``setuptools`` module is used to configure the metadata (as opposed to ``distutils``).
 
-.. code:: python
+Additional Directories
+----------------------
+
+The following directories may be specified at the same level as the ``src/`` one:
+
+* ``tests/``: Contains all unit tests for the package. It is
+  likely that these tests take advantage of the `pytest`_ framework.
+
+* ``doc/``: Contain all documentation files and examples on
+  how to use the package.
+
+
+Project File and Build System
+------------------------------
+
+The ``pyproject.toml`` file is the standardized build configuration file for Python
+projects. It must contain at least a ``[build-system]`` section, which determines
+how the project is built. Some commonly used packaging tools are `setuptools`_,
+`poetry`_, and `flit`_. All three of these packaging tools are currently supported by
+the ``pyansys-advanced`` template, which is included in the `ansys-templates tool`_.
+
+
+Flit
+^^^^
+
+Flit is a modern and lightweight build system that requires developers
+to manage virtual environments on their own. Developers must:
+
+* Create a virtual environment and activate it.
+* Install the package in editable mode.
+
+Flit is the default tool for creating a new ``pyansys`` project when using the
+`ansys-templates tool`_.
+
+The ``[project]`` section specifies the project's metadata and required
+dependencies. For more information, see `flit pyproject.toml
+guidelines`_.
+
 
-  import setuptools
-  setuptools.setup(...)
+Poetry
+^^^^^^
+
+Because of its ``poetry.lock`` file, Poetry provides strong dependency pinning. When
+installing a package, poetry creates a virtual environment, thus ensuring an isolated
+package development environment.
+
+Nevertheless, it is possible to make Poetry ignore the `poetry.lock` file by running:
+
+.. code:: bash
+
+   poetry config virtualenvs.create false --local
+
+Using `poetry`_ is popular because it:
+
+* Supports pinning dependency versions via a ``poetry.lock`` file that can be 
+  used for testing and CI
+* Allows downstream packages to still consume a loose dependency specification
+* Integrates with `dependabot`_ to update the pinned version
+
+The ``[tool.poetry]`` section contains metadata and defines the project's
+dependencies. For more information, see `poetry pyproject.toml documentation`_.
+
+
+Setuptools
+^^^^^^^^^^
+
+Setuptools is a very well known build system in the Python ecosystem. It is used
+in projects requiring a ``setup.py`` file and can be used in projects with a
+``pyproject.toml`` file, although not all metadata in this second file
+is fully supported yet.
+
+The main advantage of this build system is the ability to create custom build
+steps in the form of Python code.
+
+
+Specifying Package Version
+--------------------------
+
+It is very common for packages to specify their current version in the
+``__version__`` variable. This variable is usually declared in the
+``__init__.py`` file included in the ``library`` directory.
+
+However, it is also required to specify the version in the ``pyproject.toml`` or
+``setup.py`` file. This leads to a duplicate declaration of the project's version,
+which could lead to a potential mismatch between both.
+
+Therefore, a good practice is to take advantage of the `importlib.metadata package`_
+for parsing the version from package metadata. This guarantees that there is no mismatch
+between both version declarations.
 
-This file gathers all namespace packages and files that must be included in the distributed
-package.
 
 .. code:: python
 
-  packages = []
-  for package in setuptools.find_namespace_packages(include='ansys*'):
-      if package.startswith('ansys.tools.example_coverage'):
-          packages.append(package)
+  try:
+      import importlib.metadata as importlib_metadata
+  except ModuleNotFoundError:
+      import importlib_metadata
+
+  __version__ = importlib_metadata.version(__name__.replace(".", "-"))
+
+
+Extra Tools Configuration
+-------------------------
+
+There are plenty of tools in the Python ecosystem that enable developers to
+write clean code according to different coding style guidelines. Some of these
+tools are `black`_, `isort`_, `flake8`_, and `mypy`_.
+
+Some of these tools can be configured. This configuration might be specified in
+custom files required by the tool or in the ``pyproject.toml`` file, thus reducing the
+number of files in the project directory.
 
+.. note::
 
-It also extracts the version number from the ``_version.py`` file located in the 
-``ansys/<product>/library`` directory of the source code.
+  When using `setuptools`_ as a build backend, providing the metadata in
+  the ``pyproject.toml`` file is not yet fully supported.  Instead, it also 
+  requires a ``setup.cfg`` file, ``setup.py`` file, or both files.
+
+In the `pyansys template`, all these configurations are included by default in
+the ``.pre-commit-config.yaml`` file because ``pre-commit`` is not able to parse the
+``pyproject.toml`` file nor the ``setup.py`` file.
 
 
 Generate the Package and Upload It on PyPI
 ------------------------------------------
 
-The first time that you want to upload a package on PyPI under the `ansys <https://pypi.org/user/ansys/>`_ 
+The first time that you want to upload a package on PyPI under `ansys <https://pypi.org/user/ansys/>`_
 account, you must perform the following process manually.
 
 Create the python package.
 
 .. code::
 
-  python setup.py sdist
+  pip install build
+  python -m build
+
+If using flit or poetry, you can also run:
+
+.. code::
+
+   flit build
+   poetry build
 
 Verify the distribution's long description rendering with ``twine``.
 
@@ -77,8 +233,14 @@ Verify the distribution's long description rendering with ``twine``.
   pip install twine
   twine check dist/*
 
-Upload the package to PyPI using ``twine`` and the upload token generated for the ``ansys`` PyPI account. 
-Contact [email protected] for the token.
+
+Upload the package to PyPI using ``twine`` and the upload token generated for
+the ``ansys`` PyPI account.  As soon as the package has been released for the
+first time, it is possible to create an independent token dedicated to this
+package.  This way the token stored in the GitHub secrets and used in the
+release's workflow is only related to that specific package.  This limits the
+exposure to any potential token security flaws.  Contact
+[email protected] for the token.
 
 .. code::
 
@@ -127,7 +289,7 @@ Install a package with:
 
 .. code::
 
-  pip install ansys.<product>.<library>
+  pip install ansys-<product>-<library>
 
 To create a package complying with the above standards, here is the minimal content of your PyAnsys library:
 
@@ -136,12 +298,26 @@ To create a package complying with the above standards, here is the minimal cont
    ansys/<product>/<library>/__init__.py
    LICENSE
    README.rst
-   setup.py
+   pyproject.toml
    tests/
 
 
 .. _namespace packaging: https://packaging.python.org/guides/packaging-namespace-packages/
 .. _native namespace packages: https://packaging.python.org/guides/packaging-namespace-packages/#native-namespace-packages
+.. _PEP 420: https://www.python.org/dev/peps/pep-0420/
+.. _setuptools: https://setuptools.pypa.io
+.. _poetry: https://python-poetry.org/docs/
+.. _flit pyproject.toml guidelines: https://flit.readthedocs.io/en/latest/pyproject_toml.html
+.. _flit: https://flit.readthedocs.io
+.. _dependabot: https://docs.github.com/en/code-security/supply-chain-security/keeping-your-dependencies-updated-automatically/about-dependabot-version-updates
+.. _ansys-templates tool: https://github.com/pyansys/pyansys-templates
+.. _poetry pyproject.toml documentation: https://python-poetry.org/docs/pyproject/
+.. _black: https://black.readthedocs.io/en/stable/usage_and_configuration/the_basics.html#configuration-via-a-file
+.. _mypy: https://mypy.readthedocs.io/en/stable/config_file.html#the-mypy-configuration-file
 .. _trunk-based development: https://trunkbaseddevelopment.com/
 .. _secret: https://docs.github.com/en/actions/reference/encrypted-secrets
 .. _setup.py: https://packaging.python.org/tutorials/packaging-projects/#configuring-metadata
+.. _importlib.metadata package: https://docs.python.org/3/library/importlib.metadata.html
+.. _isort: https://github.com/PyCQA/isort
+.. _flake8: https://flake8.pycqa.org/en/latest/
+.. _pytest: https://docs.pytest.org/en/latest/