Skip to content

Latest commit

 

History

History
246 lines (166 loc) · 7.85 KB

CONTRIBUTING.rst

File metadata and controls

246 lines (166 loc) · 7.85 KB
Raw

Contributing to vineyard

Vineyard is the product of a dedicated team of software engineers and researchers. We warmly welcome contributions from the open-source community to enhance and refine this project!

Vineyard is licensed under the Apache License 2.0.

Install development dependencies

Vineyard requires the following C++ packages for development:

  • apache-arrow >= 0.17.1
  • gflags
  • glog
  • boost
  • protobuf
  • grpc

and the following python packages that can be easily installed using pip:

  • libclang
  • parsec

Developing Vineyard Using Docker

To streamline the dependency installation process, we offer a pre-built Docker image containing all necessary requirements. You can find this image at vineyardcloudnative/vineyard-dev.

docker pull vineyardcloudnative/vineyard-dev:latest

Build the source

You can do an out-of-source build using CMake:

mkdir build
cd build
cmake ..
make -j$(nproc)

The vineyardd target will be generated under the bin directory.

You may find building and installation instructions for other platforms from our CI:

Running Unit Tests

Vineyard incorporates a comprehensive set of unit tests within its continuous integration process. To build these test cases, execute the following command:

cd build
make vineyard_tests -j$(nproc)

Before running the test cases, ensure that etcd is properly installed by executing brew install etcd on macOS or pip3 install etcd_distro on Linux distributions.

A dedicated script is provided to set up the required environments and execute the test cases:

./test/runner.py --help
usage: runner.py [-h] [--with-cpp] [--with-python] [--with-io] [--with-deployment] [--with-migration] [--with-contrib] [--tests [TESTS [TESTS ...]]]

optional arguments:
  -h, --help            show this help message and exit
  --with-cpp            Whether to run C++ tests
  --with-python         Whether to run python tests
  --with-io             Whether to run IO adaptors tests
  --with-deployment     Whether to run deployment and scaling in/out tests
  --with-migration      Whether to run object migration tests
  --with-contrib        Whether to run python contrib tests
  --tests [TESTS [TESTS ...]]
                        Specify tests cases ro run

As shown above, you could run C++ unittests by

./test/runner --with-cpp

You could only run specified test case as well:

./test/runner --with-cpp --tests array_test dataframe_test

Documentation

Vineyard's documentation is generated using Doxygen and Sphinx. To build the documentation locally, navigate to the docs/ directory and execute the following commands:

cd docs/
make html

Upon successful completion, the HTML documentation will be available under the docs/_build/html directory:

open _build/html/index.html

For the most up-to-date version of the documentation, visit https://v6d.io.

Vineyard offers comprehensive documentation that delves into the design and implementation details of the project. The documentation adheres to the syntax conventions of Doxygen and Sphinx markup. If you identify areas for improvement or wish to contribute, feel free to submit a pull request. We appreciate your enthusiasm and support!

Reporting Bugs

Vineyard is hosted on GitHub and utilizes GitHub issues as its bug tracker. If you encounter any issues or unexpected behavior while using Vineyard, please file an issue.

Before creating a new bug report, we recommend that you first search among existing Vineyard bugs to check if the issue has already been addressed.

When submitting a new bug report, kindly provide essential information regarding your problem in the description, such as the operating system version, Vineyard version, and any relevant system configurations. This will greatly assist us in diagnosing and resolving the issue.

Submitting Pull Requests

We greatly appreciate contributions from the community, including bug fixes and new features. To submit a pull request to Vineyard, please follow the guidelines in this section:

Install Pre-commit

Vineyard uses pre-commit to prevent accidental inclusion of secrets in the Git repository. To install pre-commit, run:

pip3 install pre-commit

Next, configure the necessary pre-commit hooks with:

pre-commit install

Sign Off Your Commits

Vineyard has enabled the DCO, which requires you to sign-off your commits included in pull requests. Git provides a -s command line option to sign-off your commit automatically:

git commit -s -m 'This is my commit message'

Code Formatting

Vineyard adheres to the Google C++ Style Guide. When submitting patches, please format your code using clang-format with the Makefile command make vineyard_clformat, and ensure your code complies with the cpplint convention using the CMakefile command make vineyard_cpplint.

Open a Pull Request

When opening issues or submitting pull requests, please prefix the pull request title with the issue number and the type of patch (BUGFIX or FEATURE) in brackets. For example, [BUGFIX-1234] Fix crash in sealing vector to vineyard or [FEATURE-2345] Support seamless operability with PyTorch's tensors.

Git Workflow for Newcomers

Generally, you do NOT need to rebase your pull requests unless there are merge conflicts with the main branch. If GitHub indicates "Can’t automatically merge" on your pull request, you will be asked to rebase your pull request on top of the latest main branch using the following commands:

  • First, rebase to the most recent main:

    git remote add upstream https://github.com/v6d-io/v6d.git
git fetch upstream
git rebase upstream/main

  • If Git shows conflicts, such as in conflict.cpp,you need to: - Manually modify the file to resolve the conflicts - After resolving, mark it as resolved by

    git add conflict.cpp

  • Then, continue rebasing with:

    git rebase --continue

  • Finally, push to your fork, and the pull request will be updated:

    git push --force

Creating a Release

The Vineyard Python package is built using the manylinux2014 environment. To create a release version, we utilize Docker for a consistent and reliable build process. The base image's details can be found in the docker/pypa/Dockerfile.manylinux1 file.