py_cli_template

What?

This is a template package to quickly spin up a python CLI library. There are many like it, but this one is mine.

Why?

Background

I found myself writing the same boilerplate setup.cfg, setup.py, src/, tests/ structure over and over.

I also found that if I start a side project, I spend all my time configuring it, and never actually do the side project.

This is an attempt to alleviate some of that bikeshedding, and give myself a clean place to start future projects, while still retaining some of the good things about that bikeshedding, like having a sensible project structure and tests already configured.

Just clone this repo, plug in the right dependencies, change some names, delete most of this readme, and start actually writing the code you meant to write.

Why not just use cookiecutter?

To put it simply, I don't need it. Cookiecutter is great, and fully featured, but I don't need a lot of the functionality it provides. If someday I do, I will cut this over to be a cookiecutter project instead.

Prerequisites

Python >= 3.6
Git

Structure

+-- .github/workflows
|   +-- ci.yml       # github action for pytest/flake8/black
+-- src              # package containing the main app code
|   +-- __init__.py
|   +-- cli.py       # cli entrypoint
+-- tests            # pytest tests
|   +-- test_cli.py  # single test to get started
+-- .editorconfig    # for standardizing editing across OSs
+-- .gitignore
+-- LICENSE
+-- readme.md
+-- sample.env       # sample .env for secrets
+-- setup.cfg        # app config, dependency definition
+-- setup.py         # basic setup.py - use setup.cfg instead

Usage

Installation

This is how you install the package present in the template locally. You probably want to keep this section around after you use the template, as it won't change.

For installing dev dependencies run pip install -e .[dev]
- Note if you're using zsh you may need to quote any brackets:
  - pip install -e ".[dev]"
- See setup.cfg for what these dependencies are
Copy the sample.env to just .env for envvar loading cp sample.env .env
There is an initial CLI entrypoint set up in setup.cfg to src.cli.main()
- After installation, you can see this by running the app command
  - To change the name - see the "Forking and using" section below
To run tests, you can run pytest for the whole test suite or pytest tests/your_dir for a specific subset of tests
Auto-formatting is handled with black which can be run with black .
Linting is handled with flake8 with can be run with flake8 .
- Custom flake8 config is in setup.cfg
Upon opening a PR or committing to main the GitHub Action present in .github/workflows/ci.yml will test your project against a matrix of OS' and Python versions, and run flake8, black, and pytest.

Using

This is how you use the "template" piece of this repo. You can delete this section of the readme later once you've completed the steps below.

Clone this repo, and then initialize a new repo on top of it. You probably don't want the git history of this project, just the files within it.

git clone SSH_OR_HTTPS_URL your_project_name
cd your_project_name
rm -rf .git
git init

Modify the [metadata] block of setup.cfg with your project's metadata.

diff --git a/setup.cfg b/setup.cfg
[metadata]
+ name = your_project_name
version = 0.0.1
+ description = What the project does
long_description = file: readme.md
+ author = You, unless you're me
+ author_email = you@email.com

Modify the [options] and [options.entry_points] blocks of setup.cfg with your project's dependencies.
The default name what you'll use the run the CLI is app but you probably want to name it something else.

diff --git a/setup.cfg b/setup.cfg
[options.entry_points]
console_scripts =
-    app = src.cli:main
+    your_project_name = src.cli:main

Once you've reconfigured your project, you can install it locally in editable mode with pip install -e .[dev] or see the "Installation" section above
Delete a bunch of this readme referencing the "template" - you don't have a template anymore, just a blank project, and make an initial commit.
Write your project. Easiest part.

Contributing

PRs and issues welcome, with the caveat that this is my own personal setup. I might not change anything.

Todos:

Flesh out the CLI entrypoint further with help, sample arguments, etc.
Add a Dockerfile

verhey/py_cli_template