Repository to present GitHub Actions on PyCon PL 2023
Poetry >= 1.3.0 installation guide Python >= 3.7 git GitHub account
PyPI account (optional)
Clone github-ci-presentation repository:
git clone git@github.com:uoboda-splunk/github-ci-presentation.git
Install project
poetry install
Spawn a poetry shell
poetry shell
jobsim library is designed to be a simple workplace manager. For now it allows to register new employees and schedule introduction trainings for them. Different departments and locations can be defined in jobsim\departments and jobsim\locations directories respectively. After assigning the employee to its deaprtment and location trainings can be scheduled.
The example.py file in jobsim repo contains an example how the library can be used
from jobsim.employee import Employee
from jobsim.location import Location
from jobsim.department import Department
from jobsim.trainings_scheduler import schedule_trainings
IT = "it"
FINANCE = "finance"
SF = "San_francisco"
KRK = "Krakow"
SYD = "Sydney"
employee_1 = Employee("Jackie", "1", Department(IT), Location(SF))
employee_2 = Employee("Molly", "2", Department(FINANCE), Location(SF))
employee_3 = Employee("Mark", "3", Department(IT), Location(KRK))
employee_4 = Employee("John", "4", Department(IT), Location(SYD))
employee_5 = Employee("Bob", "5", Department(FINANCE), Location(SYD))
employee_6 = Employee("Andrew", "6", Department(FINANCE), Location(KRK))
employee_7 = Employee("Tom", "7", Department(FINANCE), Location(SYD))
employee_8 = Employee("Gary", "8", Department(IT), Location(SF))
employee_9 = Employee("Martin", "9", Department(IT), Location(SF))
schedule_trainings(
[
employee_1,
employee_2,
employee_3,
employee_4,
employee_5,
employee_6,
employee_7,
employee_8,
employee_9,
]
)To run example:
python example.py
Output:
Jackie is a new employee in San_francisco 🇺🇸
Jackie is joining IT department 💻
Molly is a new employee in San_francisco 🇺🇸
Molly is joining FINANCE department 💵
Mark is a new employee in Krakow 🇵🇱
Mark is joining IT department 💻
John is a new employee in Sydney 🇦🇺
John is joining IT department 💻
Bob is a new employee in Sydney 🇦🇺
Bob is joining FINANCE department 💵
Andrew is a new employee in Krakow 🇵🇱
Andrew is joining FINANCE department 💵
Tom is a new employee in Sydney 🇦🇺
Tom is joining FINANCE department 💵
Gary is a new employee in San_francisco 🇺🇸
Gary is joining IT department 💻
Martin is a new employee in San_francisco 🇺🇸
Martin is joining IT department 💻
----------------- Required trainings -----------------
Required trainings for Jackie: security_team, hr_intro, health_training,
Required trainings for Molly: law_team, hr_intro, health_training,
Required trainings for Mark: security_team, hr_intro, health_training,
Required trainings for John: security_team, hr_intro, health_training,
Required trainings for Bob: law_team, hr_intro, health_training,
Required trainings for Andrew: law_team, hr_intro, health_training,
Required trainings for Tom: law_team, hr_intro, health_training,
Required trainings for Gary: security_team, hr_intro, health_training,
Required trainings for Martin: security_team, hr_intro, health_training,
----------------- Scheduled trainings -----------------
Trainings for Jackie: security_team miami Tue 9-17, hr_intro las_vegas Fri 8-10, health_training chicago Fri 9-14,
Trainings for Molly: law_team boston Wed 7-12, hr_intro las_vegas Fri 8-10, health_training chicago Fri 9-14,
Trainings for Mark: security_team warsaw Mon 9-17, hr_intro rzeszow Mon 8-10, health_training warsaw Fri 8-13,
Trainings for John: security_team melbourne Wed 9-17, hr_intro melbourne Mon 9-11, health_training perth Mon 10-15,
Trainings for Bob: law_team perth Fri 11-16, hr_intro melbourne Mon 9-11, health_training perth Mon 10-15,
Trainings for Andrew: hr_intro rzeszow Mon 8-10, health_training warsaw Fri 8-13,
Trainings for Tom: law_team perth Fri 11-16, hr_intro melbourne Mon 9-11, health_training perth Mon 10-15,
Trainings for Gary: security_team miami Tue 9-17, hr_intro las_vegas Fri 8-10, health_training chicago Fri 9-14,
Trainings for Martin: security_team miami Tue 9-17, hr_intro las_vegas Fri 8-10, health_training chicago Fri 9-14,
To run small tests:
python -m pytest -v tests/small
To run large tests:
python -m pytest -v tests/large
During the workshop we will use this repository to present GitHub Actions functionality. Please fork this repo to start workshop.
GitHub documentation: https://docs.github.com/en/get-started/quickstart/fork-a-repo
After forking the repository you may notice that Actions history is unavailable. First you have to enable Actions itself, after clicking
Actions tab. To create first workflow run please add another trigger to tha main workflow. Use workflow_dispatch trigger to make manual runs possible.
You have to push changes to main branch to make workflow_dispatch active
GitHub documentation: https://docs.github.com/en/actions/using-workflows/events-that-trigger-workflows
Please create manual workflow run
GitHub documentation: https://docs.github.com/en/actions/using-workflows/manually-running-a-workflow
As you may notice report_jobs_results job is failing after forking the repo. This is because missing actions secrets,
which are used for job execution. Please set secrets in Settings -> Secrets and variables -> Actions You need to specify two secrets:
SPLUNK HOST
SPLUNK_TOKEN
Values for those will be provided during the workshop
GitHub documentation: https://docs.github.com/en/actions/security-guides/encrypted-secrets
Examine logs from GitHub Actions console, try to use search box and rerun options
Create another job in main workflow, which runs large tests and name it large-tests. For now, it should run in parallel with
small-tests job. Make sure that results from GitHub Actions run are the same as local runs of large tests.
Modify needs parameter for report_jobs_results, as this job must be ran as the last one int the workflow.
GitHub documentation: https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions#jobsjob_idneeds
In context of needs parameter think about different job hierarchy in our project. You can also add needs to pre-commit, small-tests, large-tests jobs. What would be the best job dependency tree?
Modify your test jobs to create JUnit logs, store those logs as artifacts.
GitHub documentation: https://docs.github.com/en/actions/using-workflows/storing-workflow-data-as-artifacts pytest documentation: https://docs.pytest.org/en/6.2.x/usage.html#creating-junitxml-format-files
Please notice that large-tests are failing, and you need to add some logic because of this. By default, job stops execution after failed step
GitHub documentation: https://docs.pytest.org/en/6.2.x/usage.html#creating-junitxml-format-files
In our workflow we are using action actions/setup-python to setup Python 3.9 for test jobs. We would like to check if our library is also
working without problems with Python versions 3.10 and 3.11. Please modify workflow to run tests also with these versions.
Hint: you don't need to create new jobs. Try parametrization with matrix strategy
GitHub documentation: https://docs.github.com/en/actions/using-jobs/using-a-matrix-for-your-jobs
Please remember that we need to have JUnit logs from every execution (3.9, 3.10, 3.11 Python versions) Some parametrization here might also be needed
Can you combine small and large tests in one job definition with parametrization? Or even further one job definition, but small tests should be run with all Python versions, but large tests only with 3.9 and 3.10?
Please familiarize with .releaserc file as it is crucial for release process. Use action https://github.com/splunk/semantic-release-action to incorporates release process in your pipeline. Try to use conventional commits syntax to create few releases.
Conventional Commits: https://www.conventionalcommits.org/en/v1.0.0/
Try to use action haya14busa/action-update-semver to update shorter tags like v1 v1.1 https://github.com/haya14busa/action-update-semver
This task is optional. We will present how to make a PyPI release. If you have a PyPI account you can do it as well, but it is not required by next tasks. Can we upload package with the same name?
GitHub allows to create workflows which can be reused in many repositories. We prepared workflow which lists PRs and issues for repository - https://github.com/kkania-splunk/gh-monitor-workflow Use this worfklow in your repository. Examine logs from GitHub runs
GitHub documentation: https://docs.github.com/en/actions/using-workflows/reusing-workflows
Make a fork of reusable workflow and modify action to fail if there are any open issues. In the next step try to make this behaviour configurable. This can be done by adding input to the workflow, which determines if we should fail or proceed always