python-semanticversion

This small python library provides a few tools to handle SemVer in Python. It follows strictly the 2.0.0 version of the SemVer scheme.

https://secure.travis-ci.org/rbarrois/python-semanticversion.png?branch=master Latest Version Supported Python versions Wheel status License

Getting started

Install the package from PyPI, using pip:

pip install semantic_version

Or from GitHub:

$ git clone git://github.com/rbarrois/python-semanticversion.git

Import it in your code:

import semantic_version

This module provides two classes to handle semantic versions:

  • Version represents a version number (0.1.1-alpha+build.2012-05-15)
  • Spec represents a requirement specification (>=0.1.1,<0.3.0)

Versions

Defining a Version is quite simple:

>>> import semantic_version
>>> v = semantic_version.Version('0.1.1')
>>> v.major
0
>>> v.minor
1
>>> v.patch
1
>>> v.prerelease
[]
>>> v.build
[]
>>> list(v)
[0, 1, 1, [], []]

If the provided version string is invalid, a ValueError will be raised:

>>> semantic_version.Version('0.1')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/rbarrois/dev/semantic_version/src/semantic_version/base.py", line 64, in __init__
    major, minor, patch, prerelease, build = self.parse(version_string, partial)
  File "/Users/rbarrois/dev/semantic_version/src/semantic_version/base.py", line 86, in parse
    raise ValueError('Invalid version string: %r' % version_string)
ValueError: Invalid version string: '0.1'

In order to define “relaxed” version strings, you must pass in partial=True:

>>> v = semantic_version.Version('0.1', partial=True)
>>> list(v)
[0, 1, None, None, None]

Obviously, Versions can be compared:

>>> semantic_version.Version('0.1.1') < semantic_version.Version('0.1.2')
True
>>> semantic_version.Version('0.1.1') > semantic_version.Version('0.1.1-alpha')
True
>>> semantic_version.Version('0.1.1') <= semantic_version.Version('0.1.1-alpha')
False

You can also get a new version that represents a bump in one of the version levels:

>>> v = semantic_version.Version('0.1.1-pre+build')
>>> new_v = v.next_major()
>>> str(new_v)
'1.0.0'
>>> v = semantic_version.Version('1.1.1-pre+build')
>>> new_v = v.next_minor()
>>> str(new_v)
'1.2.0'
>>> v = semantic_version.Version('1.1.1-pre+build')
>>> new_v = v.next_patch()
>>> str(new_v)
'1.1.2'

It is also possible to check whether a given string is a proper semantic version string:

>>> semantic_version.validate('0.1.3')
True
>>> semantic_version.validate('0a2')
False

Requirement specification

The Spec object describes a range of accepted versions:

>>> s = Spec('>=0.1.1')  # At least 0.1.1
>>> s.match(Version('0.1.1'))
True
>>> s.match(Version('0.1.1-alpha1'))  # pre-release satisfy version spec
True
>>> s.match(Version('0.1.0'))
False

Simpler test syntax is also available using the in keyword:

>>> s = Spec('==0.1.1')
>>> Version('0.1.1-alpha1') in s
True
>>> Version('0.1.2') in s
False

Combining specifications can be expressed in two ways:

  • Components separated by commas in a single string:

    >>> Spec('>=0.1.1,<0.3.0')
    
  • Components given as different arguments:

    >>> Spec('>=0.1.1', '<0.3.0')
    
  • A mix of both versions:

    >>> Spec('>=0.1.1', '!=0.2.4-alpha,<0.3.0')
    

Using a specification

The Spec.filter() method filters an iterable of Version:

>>> s = Spec('>=0.1.0,<0.4.0')
>>> versions = (Version('0.%d.0' % i) for i in range(6))
>>> for v in s.filter(versions):
...     print v
0.1.0
0.2.0
0.3.0

It is also possible to select the ‘best’ version from such iterables:

>>> s = Spec('>=0.1.0,<0.4.0')
>>> versions = (Version('0.%d.0' % i) for i in range(6))
>>> s.select(versions)
Version('0.3.0')

Coercing an arbitrary version string

Some user-supplied input might not match the semantic version scheme. For such cases, the Version.coerce() method will try to convert any version-like string into a valid semver version:

>>> Version.coerce('0')
Version('0.0.0')
>>> Version.coerce('0.1.2.3.4')
Version('0.1.2+3.4')
>>> Version.coerce('0.1.2a3')
Version('0.1.2-a3')

Including pre-release identifiers in specifications

When testing a Version against a Spec, comparisons are only performed for components defined in the Spec; thus, a pre-release version (1.0.0-alpha), while not strictly equal to the non pre-release version (1.0.0), satisfies the ==1.0.0 Spec.

Pre-release identifiers will only be compared if included in the Spec definition or (for the empty pre-release number) if a single dash is appended (1.0.0-):

>>> Version('0.1.0-alpha') in Spec('>=0.1.0')  # No pre-release identifier
True
>>> Version('0.1.0-alpha') in Spec('>=0.1.0-')  # Include pre-release in checks
False

Including build metadata in specifications

Build metadata has no ordering; thus, the only meaningful comparison including build metadata is equality.

>>> Version('1.0.0+build2') in Spec('<=1.0.0')   # Build metadata ignored
True
>>> Version('1.0.0+build2') in Spec('==1.0.0+build2')  # Include build in checks
False

Using with Django

The semantic_version.django_fields module provides django fields to store Version or Spec objects.

More documentation is available in the Interaction with Django section.

Contributing

In order to contribute to the source code:

When submitting patches or pull requests, you should respect the following rules:

  • Coding conventions are based on PEP 8
  • The whole test suite must pass after adding the changes
  • The test coverage for a new feature must be 100%
  • New features and methods should be documented in the Reference section and included in the ChangeLog
  • Include your name in the Contributors section

Note

All files should contain the following header:

# -*- encoding: utf-8 -*-
# Copyright (c) The python-semanticversion project

Indices and tables