Feel free to use this repository as a template

You can manually create your project as follows.

Creating a package

Create a project root directory and cd into it

$ mkdir MyProject
$ cd !$

Create a README and .gitignore in this root directory

From the root directory:

$ mkdir src/application
$ cd !$
$ touch __init__.py app.py

Using your favorite editor, ensure init.py has a minimum of a project description and version number:

# MyProject/src/application/__init__.py
"""Project Description"""
__version__ = "0.1.0"

Your main python script should have the main wrapper so you can use your code as a module. A main() function can also be your entry point for calling the script (as seen later):

# MyProject/src/application/app.py
def main():
    pass

if __name__ == "__main__":
    main()

In your application directory, you can also add static files and a configuration file if needed. These will be included automatically when you go to package your file.

Go back to the root directory and create a python virtual environment:

$ python -m venv .venv
$ source .venv/bin/activate

Install the flit package which will be used to package your project. :

$ pip install flit
$ flit init

Running flit init will automatically populate your project with a pyproject.toml and LICENSE (follow the on screen prompts to fill out information).

Project Structure

MyProject
├── .venv
├── src
│  └── application
│     ├── data_files
│     │  └── static_file
│     ├── __init__.py
│     ├── app.py
│     └── config.ini
├── .gitignore
├── LICENSE
├── pyproject.toml
└── README.md

pyproject.toml

You will need to edit the pyproject.toml to add your project's dependencies and entry point.

Under [project], add your dependencies as such:

[project]
.
.
dependencies = [
    "package1",
    "package2",
]

To create the command the user will run to run your program, create an entry point:

[project.scripts]
command = "application.app:main"

where application is the module name, app is the script that will be run, and main is the function that exists in the app script. Typically your command should be the same name as your application so as to avoid confusion.

Uploading to PyPi

Follow instructions in the flit docs to learn how to upload your package to PyPi so your package is installable by other users using pip

You will need to register an account on both TestPyPi and PyPi (two separate accounts)

In account settings, generate an API token. Create a .pypirc file in your system's home directory and populate it (note the username should be the keyword __token__), and the password is the generated api token:

# ~/.pypirc
[distutils]
index-servers =
   pypi
   testpypi

[pypi]
repository = https://upload.pypi.org/legacy/
username = yourname

[testpypi]
repository = https://test.pypi.org/legacy/
username = __token__
password = pypi-Someverylongtokenthatwasgenerated

You are now ready to test your package (assuming your code does something useful ;) ). In your project root, run

$ flit publish --repository testpypi

You can test downloading your package in another location using

$ pip install -i https://test.pypi.org/simple/ application

If all runs well, then you can simply publish to the pypi repository instead.