Utility for easy object composition via delegation. Roughly ported from Ruby's forwardable standard library.

Python 2.7 or 3.3 w/ standard library. Might work on other version of Python, too.

$ pip install forwardable

The @forwardable.forwardable() decorator enables you to use def_delegator() and def_delegators() in a class definition block.

Use def_delegators() to define multiple attr forwarding:

from forwardable import forwardable

@forwardable() # Note the () here, which is required.
class Foo(object):
    def_delegators('bar', 'add, __len__')

    def __init__(self):
        self.bar = set()

foo = Foo()
foo.add(1) # Delegates to foo.bar.add()
assert len(foo) == 1 # Magic methods works, too

Easy, heh?

In case you only need to delegate one method to a delegatee, just use def_delegator:

from forwardable import forwardable

@forwardable()
class Foo(object):
    def_delegator('bar', '__len__')

    def __init__(self):
        self.bar = set()

assert len(Foo()) == 0

And it should work just fine. Actually, def_delegators() calls def_delegator() under the hood.

from forwardable import forwardable

@forwardable()
class MyDict(object):
    def_delegator('dct.get', '__call__')
    def __init__(self):
        self.dct = {'foo', 42}

d = MyDict()
# Equivlant to d.dct.get('foo')
assert d('foo') == 42

The @forwardable() decorator injects def_delegator{,s} into the module scope temporarily, which is why you don't have to import them explicitly. This is admittedly magical but discourages the usage of import *. And it's always nice to type less characters whenever unnecessary.

If you hesitate to utilize this injection magic, just explicitly say from forwardable import def_delegator, def_delegators, use them in a class definition and you'll be fine.

• Source Repository: https://github.com/5long/forwardable
• Feedback: https://github.com/5long/forwardable/issues

MIT license.