Automatically deploy a package to 42maru pypi server by release actions that have a 'created', 'deleted'.
Install this GitHub action by creating a file in your repo at .github/workflows/deploy_package.yml.
A minimal example could be:
on:
release:
types: [created,deleted]
jobs:
deploys:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- name: Set up Python ${{ matrix.python-version }}
uses: actions/setup-python@v2
with:
python-version: "3.8"
- name: Deploy pacakge to 42maru pypi
uses: sunny4249/autodeploy@v0.0.2
with:
token: ${{ secrets.PYPI_SECRET }}
package_name: <package name>
env:
GITHUB_CONTEXT: ${{ toJSON(github) }}
If that package has been not registered in 42maru pypi, create 42maru pypi register templete and deploy the package to pypi. But already the package has been registered, create 42maru pypi update templete.
- Release a package with specific tags(v0.0.1)
- Go to pypi repo pull request tab and just merge that register or update request.
If specific release version package is deleted, only delete the package anchor in the package index.hml(<package_name>/index.hml). But if nothing remains in pypi index related the package, then delete the anchor in main index.html(./index.html).
- Delete a specific version package.
- Go to pypi repo pull request tab and just merge that delete request.
Because this action is fitted in 42maru pypi deploy management policy, there are not many settings.
There are Just two settings.
The PYPI_SECRET is just a token string that have workflow and release authorization.
- Generate token in your developer settings.
- Register that tokens in your repository's settings secrets tab.
The name of secret must be PYPI_SECRET
THe GITHUB_CONTEXT is environment variable that is provided by github and contains the event information to be used in actions for deploying.
PYPI_SECRET and GITHUB_CONTEXT is configured different syntax level.
(PYPI_SECRET is in with syntax and GITHUB_CONTEXT is in env syntax).
Because GITHUB_CONTEXT is provided input by actions rather than custom input, but
PYPI_SECRET is the other way around.