Skip to main content

Classes and setuptools plugin for Mercurial and Git repositories

Project description Automated Tests Code style: Black


hgtools is defunct. It has been largely superseded by the setuptools_scm project.


hgtools builds on the setuptools_hg plugin for setuptools. hgtools provides classes for inspecting and working with repositories in the Mercurial and Git version control systems (VCS).

hgtools provides a plugin for setuptools that enables setuptools to find files managed by the VCS (currently only Mercurial support is implemented).

The classes provided by hgtools are designed to use subprocess invocation to leverage the command-line interfaces of the VCS tools hg and git. An in-process RepoManager for Mercurial exists but has been disabled due to issues that arise when run in certain environments (namely setuptools sandboxing).


The setuptools feature

You can read about the setuptools plugin provided by hgtools in the setuptools documentation. It basically returns a list of files that are under VCS when running the setup function, e.g. if you create a source and binary distribution. It’s a simple yet effective way of not having to define package data (non-Python files) manually in MANIFEST templates (


Here’s a simple example of a that uses hgtools:

from setuptools import setup, find_packages

If you run the above, setuptools will automatically download hgtools to the directory where the is located at (and won’t install it anywhere else) to get all package data files from the sourec code repository.

You should not need to, and I recommend you don’t, install hgtools in your site-packages directory. Let setuptools grab it on demand. Also, try not to specify an upper bound for the requirement. Usually, simply specifying ‘hgtools’ will get the latest version, which is likely to remain compatible (as a plugin) for the life of the project. Specifying an upper bound (i.e. hgtools<1.1) will only prevent you from getting bug fixes. Only specify an upper bound if you require support for older versions of Python.

Auto Version Numbering

With the 0.4 release, hgtools adds support for automatically generating project version numbers from the repository in which the project is developed.

To use this feature, your project must follow the following assumptions:

  • Repo tags are used to indicate released versions.
  • Tag names are specified as the version only (i.e. 0.1 and not v0.1 or release-0.1)
  • Released versions currently must conform to the StrictVersion in distutils. Any tags that don’t match this scheme will be ignored. Future releases may relax this restriction.

Thereafter, you may use the RepoManager.get_current_version to determine the version of your product. If the current revision is tagged with a valid version, that version will be used. Otherwise, the tags in the repo will be searched, the latest release will be found, and hgtools will infer the upcoming release version.

For example, if the repo contains the tags 0.1, 0.2, and 0.3 and the repo is not on any of those tags, get_current_version will return ‘0.3.1dev’ and get_current_version(increment=’0.1’) will return ‘0.4dev’.

A distutils hook has been created to hack setuptools to use this version information automatically. To use this functionality, just use the use_vcs_version parameter to setup. For example:

from setuptools import setup, find_packages

If the value supplied to use_vcs_version resolves to True, hgtools will use the tagged version to determine the version of the package (based on get_current_version). If an sdist is created, hgtools will store the calculated version in the tag_build of the setup.cfg and will use that version when deploying remotely. Therefore, if you are using auto-versioning, you should not use setuptools tags explicitly.

See the jaraco.util for an example of this technique.

Versioning Parameters

It’s also possible to pass keyword parameters to use_vcs_version to tweak how it generates version numbers. To pass parameters, instead of setting use_vcs_version = True, set it to a non-empty dictionary with one or more of the following parameters:

  • increment: A string indicating the default version increment for the project. By default, this value is ‘0.1’, meaning hgtools will use the version ‘1.1dev’ for builds following the 1.0 release and ‘1.10dev’ for builds following a 1.9.3 release. Set this value to ‘1.0’ or ‘0.0.1’ for the current tree to help hgtools guess the target version.

  • version_handler: A Python function with the following signature:

    def calc_version(mgr, options):
        return str('1.0')

    hgtools will use this function instead of its default implementation to customize the version number calculation. The mgr object is the hgtools.managers.base.RepoManager object referencing the local repo and the options is the dictionary passed to use_vcs_version.

    Use this option, for example, to include the commit hash or local revision ID in the version:

    def id_as_version(mgr, options):
        "Always return the Mercurial revision ID as the version"
        id_n = mgr._invoke(['id', '-n']).strip()
        return id_n
        use_vcs_version={'version_handler': id_as_version},

    The first thing to note is the mgr does not yet provide a nice interface for getting anything but the tags for a revision, so the example digs into the underlying API to extract the ID. hgtools should provide better support in the HGRepoManager classes in future releases.

    Use this feature with caution. If you have not already read the setuptools documentation on specifying a project version, the author recommends you do read that.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for hgtools, version 8.2.1
Filename, size File type Python version Upload date Hashes
Filename, size hgtools-8.2.1-py3-none-any.whl (21.4 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size hgtools-8.2.1.tar.gz (29.7 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page