This repository contains a template for building kani extensions that use the kani.ext.* namespace.
See https://packaging.python.org/en/latest/tutorials/packaging-projects/ for more information.
The first step is to create your own repo using this repo as a template!
Follow the instructions
at https://docs.github.com/en/repositories/creating-and-managing-repositories/creating-a-repository-from-a-template
in order to create your repo. We recommend naming your extension repo like kani-ext-myextension, but you are free to
choose whatever name you like.
An important next step is to choose your code's license. kani is licensed under the MIT license, which allows extension developers to license their own code with a license of their own choice.
We (the kani developers) recommend making your extension available under the MIT license as well. You are, however, free to choose your own license: https://choosealicense.com/
Once you've chosen a license, add it to your repository in a file named LICENSE.
Now that you've created your repo, the next step is to change the template names to your own package's names. The files you'll need to change are:
pyproject.toml: Set your package name and metadatakani/ext/my_extension: Rename themy_extensiondirectory to your own name.github/workflows/pythonpublish.yml(optional): Set the environment variable to the right PyPI URL
Finally, you can delete the contents of this README and replace them with your own! Write your code in the kani/ext/*
package you renamed.
To publish your package to PyPI, this repo comes with a GitHub Action that will automatically build and upload new releases. Alternatively, you can build and publish the package manually.
To use the GitHub Action, you must configure it as a publisher for your project on PyPI: https://pypi.org/manage/account/publishing/
The workflow is configured with the following settings:
- workflow name:
pythonpublish.yml - environment name:
pypi
Once you've configured this, each release you publish on GitHub will automatically be built and uploaded to PyPI. You can also manually trigger the workflow.
Make sure to update the version number in pyproject.toml before releasing a new version!
