lpk 0.2.0

learning python

learn-python

module

minimum unit of module is a python file.
a directory containing python files is also a module.
a single file module is a binary __main__ when executed as python3 <path/to/file>.py, also is a library __init__ when imported.
a directory module can be a binary and executed as python3 <path/to/directory> by containing__main__.py, also can be a library and imported by containing__init__.py.
python to module/__init__.py is as rust to module/mod.rs. python to module.py is almost as rust to module.rs but also a binary. top level module/__main__.py or module.py is as rust to src/main.rs top level module/__init__.py or module.py is as rust to src/lib.rs

__main__ and __init__ are parts of meta info of a module. there are a lot of other meta info. e.g. __name__

package

package is also a module or collection of modules. a package can be published to https://pypi.org/. in actual, body of a package uploaded to the pypi registry is just a single module or some modules. modules of a package are located under the project directly or the src directory by convention.

poetry

https://python-poetry.org/docs/

the best package managing tool in Python project. it was insipired by Rust's Cargo.

manifest file

pyproject.toml https://python-poetry.org/docs/pyproject/

initialize a package

poetry init # current directory
poetry new <another> # create another package under current directory

check virtual environments

poetry env info # check executable section and set this custom path when using vscode via `Python: select interpreter`.
poetry env list

clear cache

poetry cache clear pypi --all
find . | grep -E "__pycache__$" | xargs rm -rf

publish package

• https://towardsdatascience.com/how-to-publish-a-python-package-to-pypi-using-poetry-aa804533fc6f
• https://www.digitalocean.com/community/tutorials/how-to-publish-python-packages-to-pypi-using-poetry-on-ubuntu-22-04

make sure there is not same name package on https://pypi.org

poetry search <package>

currently you cannot publish with username and password. you need to use api token.

poetry build
poetry publish -u=__token__ -p=<your pypi api token>

if it's redundant to input token every time.

poetry config pypi-token.pypi <your pypi api token>

and after that

poetry build
poetry publish

process sources in a file only when executed as binary

# processed whenever regardless of as library or binary.
...


if __name__ == __main__:
    # processed only when executed as binary.
    ...

testing

by convention, integration tests are located under top level tests directory. and unittests are included in each module.

• frameworks
  • unittest
    • (std lib)
    • https://docs.python.org/3/library/unittest.html
  • doctest
    • (std lib)
    • https://docs.python.org/3/library/doctest.html
  • pytest
    • https://docs.pytest.org/
    • https://github.com/pytest-dev/pytest
    • https://pypi.org/project/pytest/

documenting

• standard or third-party tools list summary

  • https://realpython.com/documenting-python-code/#documentation-tools-and-resources
  • https://wiki.python.org/moin/DocumentationTools

• hosting

  • readthedocs
    • https://readthedocs.org/
    • https://docs.readthedocs.io/en/stable/
  • github pages
  • custom domain site of yours.

• generating/building

  • sphinx

    • https://www.sphinx-doc.org/en/master/
    • https://github.com/sphinx-doc/sphinx
    • https://pypi.org/project/Sphinx/
    • functionality
      • autodoc rst from codes
      • build static htmls from rst
    • publish
      • usually automatically build & hosted from rst with readthedocs in CI/CD.
      • you can also build html locally or in CI/CD & host on your own website.

    # first, create conf.py and index.rst in <docs_source_path> like this package.
    # then
    ./sphinx_build.sh
    # open generated <docs_path>/sphinx_build/index.html
    

  • mkdocs

    • https://www.mkdocs.org/
    • https://pypi.org/project/mkdocs/
    • https://github.com/mkdocs/mkdocs
    • functionality
      • build static static htmls from markdowns
      • cannot autodoc from codes.

  • pdoc (easy, simple)

    • https://pypi.org/project/pdoc/
    • https://github.com/mitmproxy/pdoc
    • https://pdoc.dev/
    • functionality
      • autodoc & build static htmls from codes
    • publish
      • you can also build html locally or in CI/CD pipelines, then host on your own website, github pages or other hosting services ...
      • automate this processes with CI/CD like github actions.

    ./pdoc_build.sh
    # open generated docs/with_pdoc/index.html
    

linting

• https://code.visualstudio.com/docs/python/linting
• https://code.visualstudio.com/docs/python/linting#_specific-linters
• linter
  • pylint
    • https://github.com/PyCQA/pylint
    • https://www.pylint.org/
    • https://docs.pylint.org/
    • https://pypi.org/project/pylint/
  • flake8 (we recommend this)
    • https://flake8.pycqa.org/en/latest/
    • https://pypi.org/project/flake8/
    • https://github.com/PyCQA/flake8
  • mypy (we recommend)
    • (typing linter)
    • https://mypy-lang.org/
    • https://mypy.readthedocs.io/en/stable/
    • https://github.com/python/mypy
    • https://pypi.org/project/mypy/
  • pycodestyle
    • (pep8 based)
    • https://pypi.org/project/pycodestyle/
    • https://pycodestyle.pycqa.org/en/latest/
    • https://github.com/PyCQA/pycodestyle
  • pydocstyle
    • http://www.pydocstyle.org/en/stable/
    • https://pypi.org/project/pydocstyle/
    • https://github.com/PyCQA/pydocstyle
  • prospector
    • https://prospector.landscape.io/en/master/
    • https://pypi.org/project/prospector/
    • https://github.com/PyCQA/prospector
  • pylama
    • https://klen.github.io/pylama/
    • https://github.com/klen/pylama
    • https://pypi.org/project/pylama/
  • bandit
    • https://bandit.readthedocs.io/en/latest/
    • https://github.com/PyCQA/bandit
    • https://pypi.org/project/bandit/

formatting

• formatter
  • black (we recommend)
    • https://github.com/psf/black
    • https://black.readthedocs.io/en/stable/
    • https://pypi.org/project/black/
  • autopep8
    • https://pypi.org/project/autopep8/
    • https://github.com/peter-evans/autopep8
  • isort (we recommend)
    • sort imports
    • https://github.com/PyCQA/isort
    • https://pycqa.github.io/isort/
    • https://pypi.org/project/isort/

CI/CD & automation supporting

• frameworks
  • tox
    • https://tox.wiki/en/latest/index.html
    • https://pypi.org/project/tox/
    • https://github.com/tox-dev/tox
  • nox
    • https://nox.thea.codes/en/stable/
    • https://github.com/wntrblm/nox
    • https://pypi.org/project/nox/
  • invoke
    • https://www.pyinvoke.org/
    • https://github.com/pyinvoke/invoke/
    • https://pypi.org/project/invoke/

Python public / private syntax convention

there is no private module and file private objects limited by syntax. by convention, to convey that a module or object is private to users, you can make it start with a single underscore _. and objects private too. on the other hand, class members can be public, protected, private protected member start with a single underscore _. private member start with two under scores __. class to protected member is as a file to private object. private member is completely different. cannot accessible from outside normally. to access it, obj._ClassName__private_member also, private members are not inherited, so cannot be accessible anyhow through the childeren which inherited it.

development with vscode

extensions

• Python
• Pylance
• Live Server
• Even Better TOML
• Bash IDE
• autoDocstring - Python
• GitLens
• markdownlint
• YAML
• shell-format