I'm archiving this repository, I suggest that anybody using it copy the delegateto.py
to their codebase. The license is ok with that.
This tiny library implements two simple ways of method or property delegators, by descriptors and by class decorators.
From wikipedia, what is delegation:
In object-oriented programming, delegation refers to evaluating a member (property or method) of one object (the receiver) in the context of another original object (the sender). Delegation can be done explicitly, by passing the sending object to the receiving object, which can be done in any object-oriented language; or implicitly, by the member lookup rules of the language, which requires language support for the feature.
This is Python 3 only, but it's very simple library with no dependences, it would not be hard to port it to Python 2, contributions are welcome!
pip install delegateto
DelegateTo descriptor let you delegate method or property calls
The argument name is the name of the method or property that you want to delegate, for example.
from delegateto import DelegateTo
class Foo:
upper = DelegateTo('v')
__len__ = DelegateTo('l')
__iter__ = DelegateTo('l')
def __init__(self, v, l):
self.v = v
self.l = l
foo = Foo('hello world', [1, 2, 3])
To call a method just call its delegator
foo.up() # => 'HELLO WORLD'
Magic methods are supported
len(foo) # => 3
[x*2 for x in foo] # => [2, 4, 6]
At class parsing time is not possible for DelegateTo to know to what
attribute you're assigning it to. For example foo = DelegateTo('bar') Pay
attention that DelegateTo doesn't receive any information about foo
attribute, but it will discover this latter. The method name is discovered at
the first call. This is done by iterating over all the object's attributes.
Once found the method is cached and no search is performed in the subsequent
calls.
In fact, in Python +3.6 we can take advantage of __set_name__ to get the
attribute name information not needing to iterate through all the object attributes.
This make it more efficient. If __set_name__ is not available it will resort to
the loop so the above paragraph still applies on Python <=3.5. Please note that
while this should work on Python 3.5, it's not tested on it, only 3.6+ versions
are tested on the CI.
Still, if you need to avoid this iteration (on Python 3.5 and below) you can initialize the method name with the same name of the attribute name.
For example
class Foo:
upper = DelegateTo('v', 'upper')
def __init__(self, v):
self.v = v
This make possible the creation of aliases
class Foo:
up = DelegateTo('v', 'upper')
def __init__(self, v):
self.v = v
Foo('hello').up() # => 'HELLO'
In this context 'self' has a special meaning of delegating a method to another method in the same object. For example
class Foo:
foo = DelegateTo('self', 'bar')
def bar(self):
return 'bar'
Foo().foo() # => 'bar'
There is another way of creating delegators with class decorators, here is how
from delegateto import delegate
@delegate('v', 'upper')
@delegate('v', 'lower')
@delegate('v', 'wrong_method')
@delegate('not_an_attribute', 'wrong_attribute')
class Foo:
def __init__(self, v):
self.v = v
Foo('foo').upper() # => 'FOO'
Foo('FOO').lower() # => 'foo'
Foo('foo').wrong_method() # => raises AttributeError: 'str' object has no attribute 'wrong_method'
Foo('foo').wrong_attribute() # => raises AttributeError: 'Foo' object has no attribute 'not_an_attribute'
As a shortcut you can use pass any number of methods or properties to delegate
@delegate('v', 'upper', 'lower')
class Foo:
def __init__(self, v):
self.v = v
Example using properties instead of methods
class Bar:
def __init__(self):
self._param = 0
@property
def param(self):
return self._param
@param.setter
def param(self, param):
self._param = param
@delegate('v', 'param')
class Foo2:
def __init__(self):
self.v = Bar()
foo2 = Foo2()
foo2.param # => 0
foo2.param = 2
foo2.param # => 2
foo2.v.param # => 2
foo2.v._param # => 2
Simply run the module python -m delegateto