The goal of this project is to collect in one place:
- all operations that react to GitHub events,
- all operations that act on GitHub repos on a scheduled basis.
This will make it easier to review changes, as well as monitor and manage these operations, compared to the current situations where these functions are spread across cron jobs and ad-hoc scripts.
Table of contents
- Auto-delete pull request branch
- When a pull request is merged from a branch in the same repo, the bot deletes the source branch.
- Repo addons table generator in README.md
- For addons repositories, update the addons table in README.md.
- Addon README.rst generator
- For addons repositories, generate README.rst from readme fragments in each addon directory, and push changes back to github.
- Addon icon generator
- For addons repositories, put default OCA icon in each addon that don't have yet any icon, and push changes back to github.
- setup.py generator
- For addons repositories, run setuptools-odoo-make-defaults, and push changes back to github.
These actions are also run nightly on all repos.
Also nightly, wheels are generated for all addons repositories and rsynced to a PEP 503 simple index.
When there are two approvals, set the approved
label.
When the PR is at least 5 days old, set the ready to merge
label.
When the CI in a Pull Request goes green, set the needs review
label,
unless it has wip:
or [wip]
in it's title.
One can ask the bot to perform some tasks by entering special commands as merge request comments.
/ocabot merge
followed by one of major
, minor
, patch
or nobump
can be used to ask the bot to do the following:
- merge the PR onto a temporary branch created off the target branch
- merge when tests on the rebased branch are green
- optionally bump the version number of the addons modified by the PR
- when the version was bumped, udate the changelog with
oca-towncrier
- run the main branch operations (see above) on it
- when the version was bumped, generate a wheel and rsync it to the PEP 503 simple index
See our open issues, pick one and contribute!
The easiest is to look at examples.
New webhooks are added in the webhooks directory. Webhooks execution time must be very short and they should delegate the bulk of their work as delayed tasks, which have the benefit of not overloading the machine and having proper error handling and monitoring.
Tasks are in the tasks directory. They are Celery tasks.
Tasks can be scheduled, in cron.py, using the Celery periodic tasks mechanism.
First create and customize a file named environment
,
based on environment.sample.
Tasks performed by the bot can be specified by setting the BOT_TASKS
variable. This is useful if you want to use this bot for your own GitHub
organisation.
You can also disable a selection of tasks, using BOT_TASKS_DISABLED
.
docker-compose up --build
will start
- the bot, listening for webhooks calls on port 8080
- a celery
worker
to process long running tasks - a celery
beat
to launch scheduled tasks - a
flower
celery monitoring tool on port 5555
The bot URL must be exposed on the internet through a reverse
proxy and configured as a GitHub webhook, using the secret configured
in GITHUB_SECRET
.
This project uses black
as code formatting convention, as well as isort and flake8.
To make sure local coding convention are respected before
you commit, install
pre-commit and
run pre-commit install
after cloning the repository.
To run tests, type tox
. Test are written with pytest.
Here is a recommended procedure to test locally:
- Prepare an
environment
file by cloning and adaptingenvironment.sample
. - Load
environment
in your shell, for instance with bash:
set -o allexport source environment set +o allexport
- Launch the
redis
message queue:
docker run -p 6379:6379 redis
- Install the maintainer tools and add the generated binaries to your path:
PATH=/path/to/maintainer-tools/env/bin/:$PATH
- Create a virtual environment and install the project in it:
python3 -m venv venv source venv/bin/activate pip install -r requirements.txt -e .
- Then you can debug the two processes in your favorite IDE:
- the webhook server:
python -m oca_github_bot
- the task worker:
python -m celery worker --app=oca_github_bot.queue.app --pool=solo --loglevel=INFO
- the webhook server:
- To expose the webhook server on your local machine to internet, you can use ngrok
- Then configure a GitHub webhook in a sandbox project in your organization so you can start receiving webhook calls to your local machine.
To release a new version, follow these steps:
- towncrier --version YYYYMMDD
- git commit the updated HISTORY.rst and removed newfragments
- git tag vYYYYMMDD
- git push --tags
- Stéphane Bidoul <stephane.bidoul@acsone.eu>
- Holger Brunn <hbrunn@therp.nl>
- Miquel Raïch <miquel.raich@eficent.com>
- Florian Kantelberg <florian.kantelberg@initos.com>
- Laurent Mignon <laurent.mignon@acsone.eu>
- Jose Angel Fentanez <joseangel@vauxoo.com>
- Simone Rubino <simone.rubino@agilebg.com>
- Sylvain Le Gal (https://twitter.com/legalsylvain)
This module is maintained by the OCA.
OCA, or the Odoo Community Association, is a nonprofit organization whose mission is to support the collaborative development of Odoo features and promote its widespread use.