First off, we would like to thank you for taking the time to help improve Hikari, it's greatly appreciated. We have some contribution guidelines that you should follow to ensure that your contribution is at its best.
Hikari has a code of conduct that must be followed at all times by all the members of the project. Breaking the code of conduct can lead to a ban from the project and a report to GitHub.
You can read the code of conduct here.
This project follows the versioning scheme stated by PEP 440.
The development version number is increased automatically after each release in the master
branch in the master
repository.
Please also refer to the Semantic Versioning specification for more information.
The removal or renaming of anything facing the public facing API must go through a deprecation process, which should
match that of the versioning scheme. There are utilities under hikari.internal.deprecation
to aid with it.
To aid with the generation of CHANGELOG.md
as well as the releases changelog we use towncrier
.
You will need to install towncrier
and hikari
from source before making changelog additions.
pip install -r dev-requirements/towncrier.txt
pip install -e .
For every pull request made to this project, there should be a short explanation of the change under changes/
with the following format: {pull_request_number}.{type}.md
,
Possible types are:
feature
: Signifying a new feature.bugfix
: Signifying a bugfix.doc
: Signifying a documentation improvement.removal
: Signifying a deprecation or removal of public API.
For changes that do not fall under any of the above cases, please specify the lack of the changelog in the pull request description so that a maintainer can skip the job that checks for newly added fragments.
Best way to create the fragments is to run towncrier create {pull_request_number}.{type}.md
after creating the
pull request, edit the created file and committing the changes.
Multiple fragment types can be created per pull request if it covers multiple areas.
We would like to keep consistency in naming branches in the remote.
To push branches directly to the remote, you will have to name them like this:
feature/issue-number-small-info-on-branch
- This should be used for branches that require more tasks to merge into before going as one MR into
master
.
- This should be used for branches that require more tasks to merge into before going as one MR into
bugfix/issue-number-small-info-on-branch
- This should be used for bugfixes.
task/issue-number-small-info-on-branch
- This should be the default for any commit that doesn't fall in any of the cases above.
issue-number
is optional (only use if issue exists) and can be left out. small-info-on-branch
should be replaced
with a small description of the branch.
We have nox to help out with running pipelines locally and provides some helpful functionality.
You will need to install nox
locally before running any pipelines.
pip install -r dev-requirements.txt
Nox is similar to tox, but uses a pure Python configuration instead of an INI based configuration. Nox and tox are both tools for generating virtual environments and running commands in those environments. Examples of usage include installing, configuring, and running flake8, running pytest, etc.
You can check all the available nox commands by running nox -l
.
Before committing we recommend you to run nox
to run all important pipelines and make sure the pipelines won't fail.
You may run a single pipeline with nox -s name
or multiple pipelines with nox -s name1 name3 name9
.
We have several jobs to ensure that the code is at its best that in can be.
This includes:
test
- Run tests and installation of the package on different OS's and python versions.
linting
- Linting (
flake8
), type checking (mypy
), safety (safety
) and spelling (codespell
).
- Linting (
twemoji
- Force test all discord emojis.
pages
- Generate webpage + documentation.
All jobs will need to succeed before anything gets merged.