This is a plugin for the Poetry dependency
management tool that adds a new command to check if any of the dependencies in
the pyproject.toml
file have been yanked.
This plugin was written to learn how to write a Poetry plugin and to scratch an itch I had for a tool like this. I have more ideas for features and improvements, and I welcome any contributions or suggestions.
The easiest way to install this Poetry plugin is via the self add
command of
Poetry.
poetry self add poetry-plugin-check-yanked
If you used pipx
to install Poetry you can add the plugin via the pipx inject
command.
pipx inject poetry poetry-plugin-check-yanked
Otherwise, if you used pip
to install Poetry you can add the plugin packages
via the pip install
command.
pip install poetry-plugin-check-yanked
The plugin adds a new command to Poetry, check-yanked
, which can be used to
check if any of the dependencies in the pyproject.toml
file have been yanked
from PyPI by their maintainers. There is usually a pretty good reason for a
package to be yanked, so it's a good idea to check for yanked packages in your
project.
When you check for yanked packages, the plugin will download the latest
metadata for each package in the poetry.lock
file and check if any of them are
yanked. If any are found, the command will return a non-zero exit code and list
the yanked packages along with the reason for the yank. Once a specific package
verison has been checked, it will be cached for a period of time (default 1 day)
to minimize the number of requests to PyPI (and speed up future runs).
Note
This plugin uses the poetry.lock
file to determine the exact versions of
the dependencies to check, so it will also check for any yanked dependencies
of the dependencies you have specified in the pyproject.toml
file.
poetry check-yanked
The command will return a non-zero exit code if any dependencies have been yanked along with a list of the yanked dependencies and the reason for the yank.
--full
- Check each project dependency package again, even if it is already in the cache.--refresh
- Refesh the entire cache and exit, no not check for yanked packages.--no-progress
- Don't show the progress bar when checking for yanked packages, useful for CI/CD environments.--quiet
- Don't show any output, just return a non-zero exit code if any dependencies are yanked.--verbose
- Show more detailed output, including each dependency and it's yank status. This disables the progress bar.
At this time, the only configuration option is the cache timeout, which is set
to 1 day by default. This can be changed by adding a [tool.check-yanked]
section to the pyproject.toml
file with a cache_expiry
key. This value is
in seconds. and the default is 86400 (1 day). Future versions of the plugin will
offer pre-defined cache times (e.g. 1 hour, 1 day, 1 week, etc.) as well as the
ability to disable the cache entirely.
[tool.check-yanked]
cache_expiry = 3600 # 1 hour
This plugin can be used as a GitHub Action to check for yanked packages automatically as part of your CI. Here is an example workflow:
name: Check for Yanked Packages
on: [push, pull_request]
jobs:
check-yanked:
runs-on: ubuntu-latest
steps:
- name: Run poetry check-yanked
uses: seapagan/check-yanked-packages@v1
There are two optional inputs that can be used to configure the action:
path
- The path to the directory containing thepoetry.lock
file. This defaults to the root of the repository.python-version
- The version of Python to use when running the action. This defaults to the latest version of Python 3.x available on the runner.- If you are using the
actions/setup-python
action, this will be ignored, and the version of Python installed by that will be used instead.
- If you are using the
These can be set in the workflow file like so:
- name: Run poetry check-yanked
uses: seapagan/check-yanked-packages@v1
with:
python-version: '3.10'
path: 'path/to/directory'
See the
check-yanked-packages
action for more information on using this plugin as a GitHub Action. The
cache-expiry
option is not available when using the action, as the cache is
not persisted between runs.
Check CONTRIBUTING.md for full instructions on how to set up the project for development, and CODE_OF_CONDUCT.md for the project code of conduct.
This project is released under the terms of the MIT license.
The original Python boilerplate for this package was created using Pymaker by Grant Ramsay