rockobonaparte/observed

Observer pattern in python

observed allows you to to sign up functions or methods to "observe" other functions or methods:

from observed import observable_function

@observable_function
def observed_func(arg):
    print("observed_func: %s"%(arg,))

def observer_func(arg):
    print("observer_func: %s"%(arg,))

observed_func.add_observer(observer_func)
observed_func('banana')

>>> observed_func: banana
>>> observer_func: banana

You can also register observers for bound methods:

from observed import observable_method

class Foo(object):
    def __init__(self, name):
        self.name = name

    @observable_method
    def bar(self, arg):
        print("Object %s invoked bar with arg='%s'"%(self.name, arg))

    def baz(self, arg):
        print("Object %s invoked baz with arg='%s'"%(self.name, arg))

def callback(arg):
    print("callback was invoked with arg='%s'"%(arg,))

a = Foo('a')
b = Foo('b')
a.bar.add_observer(b.bar)
a.bar.add_observer(b.baz)
a.bar.add_observer(callback)
a.bar('banana')

>>> Object a invoked bar with arg='banana'
>>> Object b invoked bar with arg='banana'
>>> Object b invoked baz with arg='banana'
>>> callback was invoked with arg='banana'

You can ask that the observed object pass itself as the first argument whenever it calls observers:

from observed import observable_function

@observable_function
def observed_func():
    print("observed_func: I was called")

def observer_func(observed):
    print("observer_func: %s called me"%(observed.__name__,))

observed_func.add_observer(observer_func, identify_observed=True)
observed_func()

>>> observed_func: I was called
>>> observer_func: observed_func called me

When observed bound methods pass themselves as the observed object, keep in mind that you can always access the associated instance via .__self__:

from observed import observable_method

class Foo(object):
    def __init__(self, name):
        self.name = name

    @observable_method
    def bar(self):
        print("Object %s invoked bar"%(self.name,))

def callback(observed):
    print("callback was invoked by='%s'"%(observed.__self__.name,))

a = Foo('a')
a.bar.add_observer(callback, identify_observed=True)
a.bar()

>>> Object a invoked bar
>>> callback was invoked by a

Notable features include:

• A function or bound method is not kept alive just because it is observing something else. This is because the observed object does not keep any strong references to the observing objects. In CPython this means that your observers are automatically detached whenever the reference count to that observer goes to zero.
• The @observable_method decorator can be used on methods in classes which are unhashable types, and can be used on an arbitrary number of methods in each class...
• ...but the descriptor which handles observable_method does not paste any data onto the instances it manages. There is zero chance of name collision on the observed instances.
• Tests included :)

Installation

observed exists on the python package index, so you can do pip install observed to install it. Alternatively, you can download the source distribution and in the root directory of the distribution do

$ python setup.py install.

News

See the file NEWS for the user-visible changes from previous releases.

License

See LICENSE.txt.

Downloading

observed can be obtained from the python package index

https://pypi.python.org/pypi/observed

or via git

https://github.com/DanielSank/observed.git

Documentation

Basic usage is illustrated at the top of this file. Further examples are given in ./observed/example.py

The source code is documented. Docstrings are required in contributions.

Development

observed development is hosted on github. The current working repository is given in the Downloading section above.

Bug Reporting

Please submit bug tickets on the github tracking system

https://github.com/DanielSank/observed/issues