Contributing to lakeFS-spec¶
Thank you for your interest in contributing to this project!
We appreciate issue reports, pull requests for code and documentation, as well as any project-related communication through GitHub Discussions.
Getting Started¶
To get started with development, you can follow these steps:
-
Clone this repository:
-
Navigate to the directory and install the development dependencies into a virtual environment:
-
After making your changes, verify they adhere to our Python code style by running
pre-commit
:You can also set up Git hooks through
pre-commit
to perform these checks automatically: -
To run the tests against an ephemeral lakeFS instance, you just run
pytest
:To spin up a local lakeFS instance quickly for testing, you can use the Docker Compose file bundled with this repository:
Updating dependencies¶
Dependencies should stay locked for as long as possible, ideally for a whole release. If you have to update a dependency during development, you should do the following:
- If it is a core dependency needed for the package, add it to the
dependencies
section in thepyproject.toml
. - In case of a development dependency, add it to the
dev
section of theproject.optional-dependencies
table instead. - Dependencies needed for documentation generation are found in the
docs
sections ofproject.optional-dependencies
.
After adding the dependency in either of these sections, run the helper script hack/lock-deps.sh
(which in turn uses pip-compile
) to pin all dependencies again:
In addition to these manual steps, we also provide pre-commit
hooks that automatically lock the dependencies whenever pyproject.toml
is changed.
Selective package upgrade for existing dependencies are also handled by the helper script above.
If you want to update the lakefs-sdk
dependency, for example, simply run:
Tip
Since the official development version is Python 3.11, please run the above commands in a virtual environment with Python 3.11.
Working on Documentation¶
Improvements or additions to the project's documentation are highly appreciated.
The documentation is based on the MkDocs and Material for MkDocs (mkdocs-material
) projects, see their homepages for in-depth guides on their features and usage. We use the Numpy documentation style for Python docstrings.
To build the documentation locally, you need to first install the optional docs
dependencies from requirements-docs.txt
,
e.g., with pip install -r requirements-docs.txt
. You can then start a local documentation server with mkdocs serve
, or
build the documentation into its output folder in public/
.
In order to maintain documentation for multiple versions of this library, we use the mike tool, which automatically maintains individual documentation builds per version and publishes them to the gh-pages
branch.
The GitHub CI pipeline automatically invokes mike
as part of the release process with the correct version and updates the GitHub pages branch for the project.