This utility is aimed at creating the boilerplate necessary when starting a new python package. I've found out that packaging even my own script helps tremendously in the long run, and also improves the general quality of my projects.
Simply open a terminal/prompt and type
pip install pypackinitand a new command, ppinit will be added to your PATH.
If you can't install packages in your system interpreter (you shouldn't anyway), you can add the --user option.
Following the PyPA guide on packaging, this script simply generates the essential files
to get you started on a new project. Suppose you want to start a project named panino.
This means that you have to create by hand the setup.py script, fill it up, then create
README, CHANGELOG, LICENSE, etc. With pypackinit, all of this is done with a single
line (spread on multiple lines for clarity).
ppinit panino 'The best panino on Earth, with even more cheese!' \
-a 'Nino Pa' \
-e ninopa@bread.com \
-u https://panino.bread.com
Now you have a directory structure like this
project_root/
├── panino
│ └── __init__.py
├── CHANGELOG.md
├── LICENSE.txt
├── MANIFEST.in
├── README.md
└── setup.py
ready to go! In practice, this script is an extremely simplified version of cookiecutter (and it's not a fork).
Packaging means that you don't have to set up your PYTHONPATH every time you change machine and
that your virtualenvs can simply "install" your other projects.
Suppose that you have Project A and Project B, where B depends on A. With packaging, you
could install A in B's virtual environment and then you could fix issues and work on additional
features of A without compromising the functionality of B. I have had this issue several times
and by pacakging basically all my scripts and modules I have solved most of my problems when it
comes to re-use my projects!
Also, I found out that sharing code with people that are not comfortable with git and the likes
is way easier if you can just send them a wheel. :)