combomethod 1.0.8

Decorator indicating a method is both a class and an instance method

| |travisci| |version| |downloads| |versions| |impls| |wheel| |coverage| |br-coverage|

.. |travisci| image:: https://api.travis-ci.org/jonathaneunice/combomethod.svg
:target: http://travis-ci.org/jonathaneunice/combomethod

.. |version| image:: http://img.shields.io/pypi/v/combomethod.svg?style=flat
:alt: PyPI Package latest release
:target: https://pypi.python.org/pypi/combomethod

.. |versions| image:: https://img.shields.io/pypi/pyversions/combomethod.svg
:alt: Supported versions
:target: https://pypi.python.org/pypi/combomethod

.. |impls| image:: https://img.shields.io/pypi/implementation/combomethod.svg
:alt: Supported implementations
:target: https://pypi.python.org/pypi/combomethod

.. |wheel| image:: https://img.shields.io/pypi/wheel/combomethod.svg
:alt: Wheel packaging support
:target: https://pypi.python.org/pypi/combomethod

.. |coverage| image:: https://img.shields.io/badge/test_coverage-100%25-6600CC.svg
:alt: Test line coverage
:target: https://pypi.python.org/pypi/combomethod

.. |br-coverage| image:: https://img.shields.io/badge/branch_coverage-100%25-6600CC.svg
:alt: Test branch coverage
:target: https://pypi.python.org/pypi/combomethod

Python has instance methods, class methods (``@classmethod``), and
static methods (``@staticmethod``). But it doesn't have a clear way
to invoke a method on either a class *or*
its instances. With ``combomethod``, it does.

::

from combomethod import combomethod

class A(object):

@combomethod
def either(receiver, x, y):
return x + y

a = A()
assert a.either(1, 3) == 4
assert A.either(1, 3) == 4

*Voila!* You method now takes either the class or the instance--whichever
one you want to call it with.

Discussion
==========

In some cases, you can fake ``@combomethod`` with ``@classmethod``. In
the code above, for example, there is no real reference to the class
or instance, and ``either`` could have been designated a ``@classmethod``,
since they can be called with either classes or instances. But, there's a
problem: Class methods *always* pass the class to the method, even if they're
called with an instance. With this approach, you can never access the
instance variables. Ouch!

Alternatively, ``either`` could have been designated a ``@staticmethod``,
had its ``receiver`` parameter been removed. But while it would then be
callable from either an instance or a class, in neither case would it pass
the object the method was called from. There'd never be a way to access
either the class or instance variables. Ouch again!

As useful as ``@classmethod`` and ``@staticmethod`` are, they don't
handle the important case where you need to call with either the class or
an instance *and* you need genuine access to the object doing the call.
Here's an example that needs this::

class Above(object):

base = 10

def __init__(self, base=100):
self.base = base

@combomethod
def above_base(receiver, x):
return receiver.base + x

a = Above()
assert a.above_base(5) == 105
assert Above.above_base(5) == 15

aa = Above(12)
assert aa.above_base(5) == 17
assert Above.above_base(5) == 15

When you need to call with either an instance or a class, and you also care
about the object doing the calling, ``@combomethod`` rocks and rolls.

Notes
=====

* This module is primarily a convenient packaging, testing,
and documentation of insights and code from Mike Axiak's
`Stack Overflow post <http://stackoverflow.com/questions/2589690/creating-a-method-that-is-simultaneously-an-instance-and-class-method>`_.
Thank you, Mike!

* Automated multi-version testing managed with
`pytest <http://pypi.python.org/pypi/pytest>`_,
`pytest-cov <http://pypi.python.org/pypi/pytest-cov>`_,
`coverage <http://pypi.python.org/pypi/coverage>`_, and
`tox <http://pypi.python.org/pypi/tox>`_.
Continuous integration testing
with `Travis-CI <https://travis-ci.org/jonathaneunice/combomethod>`_.
Packaging linting with `pyroma <https://pypi.python.org/pypi/pyroma>`_.

* Version 2.2.6 updates testing for early 2017 Python
versions. Successfully packaged for, and
tested against, all late-model versions of Python: 2.6, 2.7, 3.3,
3.4, 3.5, and 3.6, as well as PyPy 5.6.0 (based on
2.7.12) and PyPy3 5.5.0 (based on 3.3.5).

* See ``CHANGES.yml`` for the complete Change Log.

* The author, `Jonathan Eunice <mailto:[e-mail]>`_ or
`@jeunice on Twitter <http://twitter.com/jeunice>`_
welcomes your comments and suggestions.

Installation
============

To install or upgrade to the latest version::

pip install -U combomethod

To ``easy_install`` under a specific Python version (3.3 in this example)::

python3.3 -m easy_install --upgrade combomethod

(You may need to prefix these with ``sudo`` to authorize installation. In
environments without super-user privileges, you may want to use ``pip``'s
``--user`` option, to install only for a single user, rather than
system-wide.)