Single-sourcing the package version#


Update this page for build backends other than setuptools.

There are many techniques to maintain a single source of truth for the version number of your project:

  1. Read the file in and get the version. Example (from pip

    import codecs
    import os.path
    def read(rel_path):
        here = os.path.abspath(os.path.dirname(__file__))
        with, rel_path), 'r') as fp:
    def get_version(rel_path):
        for line in read(rel_path).splitlines():
            if line.startswith('__version__'):
                delim = '"' if '"' in line else "'"
                return line.split(delim)[1]
            raise RuntimeError("Unable to find version string.")


    As of the release of setuptools 46.4.0, one can accomplish the same thing by instead placing the following in the project’s setup.cfg file (replacing “package” with the import name of the package):

    version = attr: package.__version__

    As of the release of setuptools 61.0.0, one can specify the version dynamically in the project’s pyproject.toml file.

    name = "package"
    dynamic = ["version"]
    version = {attr = "package.__version__"}

    Please be aware that declarative config indicators, including the attr: directive, are not supported in parameters to

  2. Use an external build tool that either manages updating both locations, or offers an API that both locations can use.

    Few tools you could use, in no particular order, and not necessarily complete: bump2version, changes, commitizen, zest.releaser.

  3. Set the value to a __version__ global variable in a dedicated module in your project (e.g., then have read and exec the value into a variable.

    version = {}
    with open("...sample/") as fp:
        exec(, version)
    # later on we use: version['__version__']

    Example using this technique: warehouse.

  4. Place the value in a simple VERSION text file and have both and the project code read it.

    with open(os.path.join(mypackage_root_dir, 'VERSION')) as version_file:
        version =

    An advantage with this technique is that it’s not specific to Python. Any tool can read the version.


    With this approach you must make sure that the VERSION file is included in all your source and binary distributions (e.g. add include VERSION to your

  5. Set the value in, and have the project code use the importlib.metadata API to fetch the value at runtime. (importlib.metadata was introduced in Python 3.8 and is available to older versions as the importlib-metadata project.) An installed project’s version can be fetched with the API as follows:

    import sys
    if sys.version_info >= (3, 8):
        from importlib import metadata
        import importlib_metadata as metadata
    assert metadata.version('pip') == '1.2.0'

    Be aware that the importlib.metadata API only knows about what’s in the installation metadata, which is not necessarily the code that’s currently imported.

    If a project uses this method to fetch its version at runtime, then its install_requires value needs to be edited to install importlib-metadata on pre-3.8 versions of Python like so:

            'importlib-metadata >= 1.0 ; python_version < "3.8"',

    An older (and less efficient) alternative to importlib.metadata is the pkg_resources API provided by setuptools:

    import pkg_resources
    assert pkg_resources.get_distribution('pip').version == '1.2.0'

    If a project uses pkg_resources to fetch its own version at runtime, then setuptools must be added to the project’s install_requires list.

    Example using this technique: setuptools.

  6. Set the value to __version__ in sample/ and import sample in

    import sample


    Although this technique is common, beware that it will fail if sample/ imports packages from install_requires dependencies, which will very likely not be installed yet when is run.

  7. Keep the version number in the tags of a version control system (Git, Mercurial, etc) instead of in the code, and automatically extract it from there using setuptools_scm.