Python wheels are great. Building them across Mac, Linux, Windows, on multiple versions of Python, is not.
cibuildwheel
is here to help. cibuildwheel
runs on your CI server - currently it supports GitHub Actions, Azure Pipelines, Travis CI, AppVeyor, CircleCI, and GitLab CI - and it builds and tests your wheels across all of your platforms.
macOS Intel | macOS Apple Silicon | Windows 64bit | Windows 32bit | Windows Arm64 | manylinux musllinux x86_64 |
manylinux musllinux i686 |
manylinux musllinux aarch64 |
manylinux musllinux ppc64le |
manylinux musllinux s390x |
|
---|---|---|---|---|---|---|---|---|---|---|
CPython 3.6 | โ | N/A | โ | โ | N/A | โ | โ | โ | โ | โ |
CPython 3.7 | โ | N/A | โ | โ | N/A | โ | โ | โ | โ | โ |
CPython 3.8 | โ | โ | โ | โ | N/A | โ | โ | โ | โ | โ |
CPython 3.9 | โ | โ | โ | โ | โ ยฒ | โ ยณ | โ | โ | โ | โ |
CPython 3.10 | โ | โ | โ | โ | โ ยฒ | โ | โ | โ | โ | โ |
CPython 3.11 | โ | โ | โ | โ | โ ยฒ | โ | โ | โ | โ | โ |
PyPy 3.7 v7.3 | โ | N/A | โ | N/A | N/A | โ ยน | โ ยน | โ ยน | N/A | N/A |
PyPy 3.8 v7.3 | โ | โ โด | โ | N/A | N/A | โ ยน | โ ยน | โ ยน | N/A | N/A |
PyPy 3.9 v7.3 | โ | โ โด | โ | N/A | N/A | โ ยน | โ ยน | โ ยน | N/A | N/A |
ยน PyPy is only supported for manylinux wheels.
ยฒ Windows arm64 support is experimental.
ยณ Alpine 3.14 and very briefly 3.15's default python3 was not able to load musllinux wheels. This has been fixed; please upgrade the python package if using Alpine from before the fix.
โด Cross-compilation not supported with PyPy - to build these wheels you need to run cibuildwheel on an Apple Silicon machine.
- Builds manylinux, musllinux, macOS 10.9+, and Windows wheels for CPython and PyPy
- Works on GitHub Actions, Azure Pipelines, Travis CI, AppVeyor, CircleCI, GitLab CI, and Cirrus CI
- Bundles shared library dependencies on Linux and macOS through auditwheel and delocate
- Runs your library's tests against the wheel-installed version of your library
See the cibuildwheel 1 documentation if you need to build unsupported versions of Python, such as Python 2.
cibuildwheel
runs inside a CI service. Supported platforms depend on which service you're using:
Linux | macOS | Windows | Linux ARM | macOS ARM | Windows ARM | |
---|---|---|---|---|---|---|
GitHub Actions | โ | โ | โ | โ ยน | โ ยฒ | โ โด |
Azure Pipelines | โ | โ | โ | โ ยฒ | โ โด | |
Travis CI | โ | โ | โ | |||
AppVeyor | โ | โ | โ | โ ยฒ | โ โด | |
CircleCI | โ | โ | โ | โ ยฒ | ||
Gitlab CI | โ | โ | โ ยน | |||
Cirrus CI | โ | โ ยณ | โ | โ | โ |
ยน Requires emulation, distributed separately. Other services may also support Linux ARM through emulation or third-party build hosts, but these are not tested in our CI.
ยฒ Uses cross-compilation. It is not possible to test arm64
and the arm64
part of a universal2
wheel on this CI platform.
ยณ Uses cross-compilation. Thanks to Rosetta 2 emulation, it is possible to test x86_64
and both parts of a universal2
wheel on this CI platform.
โด Uses cross-compilation. It is not possible to test arm64
on this CI platform.
To build manylinux, musllinux, macOS, and Windows wheels on GitHub Actions, you could use this .github/workflows/wheels.yml
:
name: Build
on: [push, pull_request]
jobs:
build_wheels:
name: Build wheels on ${{ matrix.os }}
runs-on: ${{ matrix.os }}
strategy:
matrix:
os: [ubuntu-20.04, windows-2019, macOS-11]
steps:
- uses: actions/checkout@v3
# Used to host cibuildwheel
- uses: actions/setup-python@v3
- name: Install cibuildwheel
run: python -m pip install cibuildwheel==2.12.0
- name: Build wheels
run: python -m cibuildwheel --output-dir wheelhouse
# to supply options, put them in 'env', like:
# env:
# CIBW_SOME_OPTION: value
- uses: actions/upload-artifact@v3
with:
path: ./wheelhouse/*.whl
For more information, including PyPI deployment, and the use of other CI services or the dedicated GitHub Action, check out the documentation and the examples.
The following diagram summarises the steps that cibuildwheel takes on each platform.
Explore an interactive version of this diagram in the docs.
Option | Description | |
---|---|---|
Build selection | CIBW_PLATFORM |
Override the auto-detected target platform |
CIBW_BUILD CIBW_SKIP |
Choose the Python versions to build | |
CIBW_ARCHS |
Change the architectures built on your machine by default. | |
CIBW_PROJECT_REQUIRES_PYTHON |
Manually set the Python compatibility of your project | |
CIBW_PRERELEASE_PYTHONS |
Enable building with pre-release versions of Python if available | |
Build customization | CIBW_BUILD_FRONTEND |
Set the tool to use to build, either "pip" (default for now) or "build" |
CIBW_ENVIRONMENT |
Set environment variables needed during the build | |
CIBW_ENVIRONMENT_PASS_LINUX |
Set environment variables on the host to pass-through to the container during the build. | |
CIBW_BEFORE_ALL |
Execute a shell command on the build system before any wheels are built. | |
CIBW_BEFORE_BUILD |
Execute a shell command preparing each wheel's build | |
CIBW_REPAIR_WHEEL_COMMAND |
Execute a shell command to repair each built wheel | |
CIBW_MANYLINUX_*_IMAGE CIBW_MUSLLINUX_*_IMAGE |
Specify alternative manylinux / musllinux Docker images | |
CIBW_CONTAINER_ENGINE |
Specify which container engine to use when building Linux wheels | |
CIBW_DEPENDENCY_VERSIONS |
Specify how cibuildwheel controls the versions of the tools it uses | |
Testing | CIBW_TEST_COMMAND |
Execute a shell command to test each built wheel |
CIBW_BEFORE_TEST |
Execute a shell command before testing each wheel | |
CIBW_TEST_REQUIRES |
Install Python dependencies before running the tests | |
CIBW_TEST_EXTRAS |
Install your wheel for testing using extras_require | |
CIBW_TEST_SKIP |
Skip running tests on some builds | |
Other | CIBW_BUILD_VERBOSITY |
Increase/decrease the output of pip wheel |
These options can be specified in a pyproject.toml file, as well; see configuration.
Here are some repos that use cibuildwheel.
Name | CI | OS | Notes |
---|---|---|---|
scikit-learn | The machine learning library. A complex but clean config using many of cibuildwheel's features to build a large project with Cython and C++ extensions. | ||
NumPy | The fundamental package for scientific computing with Python. | ||
Tornado | Tornado is a Python web framework and asynchronous networking library, originally developed at FriendFeed. | ||
pytorch-fairseq | Facebook AI Research Sequence-to-Sequence Toolkit written in Python. | ||
Matplotlib | The venerable Matplotlib, a Python library with C++ portions | ||
NCNN | ncnn is a high-performance neural network inference framework optimized for the mobile platform | ||
Kivy | Open source UI framework written in Python, running on Windows, Linux, macOS, Android and iOS | ||
Prophet | Tool for producing high quality forecasts for time series data that has multiple seasonality with linear or non-linear growth. | ||
MyPy | The compiled version of MyPy using MyPyC. | ||
pydantic | Data parsing and validation using Python type hints |
โน๏ธ That's just a handful, there are many more! Check out the Working Examples page in the docs.
Since cibuildwheel
repairs the wheel with delocate
or auditwheel
, it might automatically bundle dynamically linked libraries from the build machine.
It helps ensure that the library can run without any dependencies outside of the pip toolchain.
This is similar to static linking, so it might have some license implications. Check the license for any code you're pulling in to make sure that's allowed.
16 Jan 2023
- โจ Adds support for PyPy arm64 wheels. This means that you can build PyPy wheels for Apple Silicon machines. Cross-compilation is not supported for these wheels, so you'll have to build on an Apple Silicon machine. (#1372)
- ๐ Pinned version updates, including PyPy to v7.3.11 and setuptools to 66.0.0.
24 Dec 2022
- ๐ Fix a bug that caused missing wheels on Windows when a test was skipped using CIBW_TEST_SKIP (#1377)
- ๐ Updates CPython 3.11 to 3.11.1 (#1371)
- ๐ Updates PyPy to 7.3.10, except on macOS which remains on 7.3.9 due to a bug on that platform. (#1371)
- ๐ Added a reference to abi3audit to the docs (#1347)
5 Dec 2022
- โจ Improves the 'build options' log output that's printed at the start of each run (#1352)
- โจ Added a friendly error message to a common misconfiguration of the
CIBW_TEST_COMMAND
option - not specifying path using the{project}
placeholder (#1336) - ๐ The GitHub Action now uses Powershell on Windows to avoid occasional incompabilities with bash (#1346)
26 October 2022
- ๐ Updates CPython 3.11 to 3.11.0 - final release (#1327)
- ๐ Simplify the default macOS repair command (#1322)
- ๐ Fix the default
MACOSX_DEPLOYMENT_TARGET
on arm64 (#1312) - ๐ Hide irrelevant pip warnings on linux (#1311)
- ๐ Fix a bug that caused the stdout and stderr of commands in containers to be in the wrong order Previously, stdout could appear after stderr. (#1324)
- ๐ Added a FAQ entry describing how to perform native builds of CPython 3.8 wheels on Apple Silicon. (#1323)
- ๐ Other docs improvements
13 October 2022
- ๐ Updates to the latest manylinux images, and updates CPython 3.10 to 3.10.8.
That's the last few versions.
โน๏ธ Want more changelog? Head over to the changelog page in the docs.
For more info on how to contribute to cibuildwheel, see the docs.
Everyone interacting with the cibuildwheel project via codebase, issue tracker, chat rooms, or otherwise is expected to follow the PSF Code of Conduct.
- Joe Rickerby @joerick
- Yannick Jadoul @YannickJadoul
- Matthieu Darbois @mayeut
- Henry Schreiner @henryiii
- Grzegorz Bokota @Czaki
cibuildwheel
stands on the shoulders of giants.
- โญ๏ธ @matthew-brett for multibuild and matthew-brett/delocate
- @PyPA for the manylinux Docker images pypa/manylinux
- @ogrisel for wheelhouse-uploader and
run_with_env.cmd
Massive props also to-
- @zfrenchee for help debugging many issues
- @lelit for some great bug reports and contributions
- @mayeut for a phenomenal PR patching Python itself for better compatibility!
- @czaki for being a super-contributor over many PRs and helping out with countless issues!
- @mattip for his help with adding PyPy support to cibuildwheel
Another very similar tool to consider is matthew-brett/multibuild. multibuild
is a shell script toolbox for building a wheel on various platforms. It is used as a basis to build some of the big data science tools, like SciPy.
If you are building Rust wheels, you can get by without some of the tricks required to make GLIBC work via manylinux; this is especially relevant for cross-compiling, which is easy with Rust. See maturin-action for a tool that is optimized for building Rust wheels and cross-compiling.