Skip to main content

Class based extension/plugin library

Project description

Extend Me - Class based extension/plugin library

This module provides mechanism of extension of your application based on ‘extension via inheritance’. Under this words I mean ability to define new extensions of application objects simply by subclassing of extensible classes of app.

For example we have app with class ‘Worker’ which we would like to make extensible (allowing third party modules to extend or change its behavior). Thinking strait, there are a lot of work to be done, to impelement mechanism of loading, registering, end enabling extension, with lot of glue code, which must define some entry points to connect extension and main app. But why not make it simpler, supposing that any subclass of ‘Worker’ will extend it? And this module provides implementation of this in two ways:

  • Explicit (by using metaclass ExtensibleType directly)
    • When using this way You will heve seperatly Base class to be subclassed by extension classes and class getter which will construct class based on all defined extensions using multiple inhertance
  • Implicit (by using Extensible class which use metaclass magic implicitly)
    • Extensible class takes care of all metaclass magic
      related to generation objects of correct class

How it Works

Metaclass (ExtensibleType) tracks all subclasses of class it is applied to, and provides method to build class based on all subclasses of base class, thus using all functionality of all subclasses. Thus generation of correct class is separate process which should be used everywhere where extensible class is requred.

To simplify this class Extensible was implemented. It has redefined method __new__ which automaticaly creates instances of correct class (class that inherited from base class and all its extensions’)



At the begining we should create a metaclass that will automaticaly gether all information about all extensions, and apply this metaclass to class we would like to enable extensions for:

>>> import six  # Used for Python 2/3 compatability
>>> mc = ExtensibleType._("Object")

>>> @six.add_metaclass(mc)
... class Object(object):
...     pass

Not method _ of ExtensibleType. This method is used to create metaclass for specific object. It receives one argument - string that will be used as name of class generated by this metaclass

Next we may define extension for this class. It is very simple. Just subclass previously defined class:

>>> class ObjectExtension(Object):
...     cool_attribute = 1
...     def method1(self):
...         return "Test"

So… at this momet we have base class and extension. And here is that core magic occures. Metaclass that was created at the begining automaticaly collects all subclasses of base class. So it is posible now to create new class that is subclass of all subclasses of base class using multiple inheritance. And metaclass mc will do it for You:

>>> cls = mc.get_class()

And now You can use cls for Your needs, instead of base class. It can do all that base class can, and all that extensions can:

>>> obj = cls()
>>> obj.method1()
>>> obj.cool_attribute


Same as ExtensibleType, but allows to build tree of classes for diferent names (types). Just look examples below.

First, create metaclass that will specify inheritance rules:

>>> import six  # Used for Python 2/3 compatability
>>> mc = ExtensibleByHashType._("Connector", hashattr='name')

Here we see aditional parametr in _ method: hashattr='name' which describes what meta attribute will be used as key(hash).

Next step - we have to create Base class with this metaclass. As example we will look into connection classes of openerp_proxy project:

>>> @six.add_metaclass(mc)
... class ConnectorBase(object):
...     # Base class for all connectors
...     def __init__(self, host, port, verbose=False):
... = host
...         self.port = port
...         self.verbose = verbose
...     def _get_service(self, name):
...         raise NotImplementedError
...     def get_service(self, name):
...         # Returns service for specified *name*
...         return self._get_service(name)

Base class describes only interface, and may be some part of abstract logic And as next step we will extend it in diferent ways to support different connection types:

>>> class ConnectorXMLRPC(ConnectorBase):
...     # XML-RPC connector
...     class Meta:
...         name = 'xml-rpc' # remember definition of metaclass?
...                          # this attribute is used as hash(key)
...                          # to unique identify each banch of extensions
...                          # of base class
...     def __init__(self, *args, **kwargs):
...         super(ConnectorXMLRPC, self).__init__(*args, **kwargs)
...         self.__services = {}
...     def get_service_url(self, service_name):
...         return 'http://%s:%s/xmlrpc/%s' % (, self.port, service_name)
...     def _get_service(self, name):
...         service = self.__services.get(name, False)
...         if service is False:
...             service = XMLRPCProxy(self.get_service_url(name), verbose=self.verbose)
...             self.__services[name] = service
...         return service
... # Pay attention on base class.
>>> class ConnectorXMLRPCS(ConnectorXMLRPC):
...     # XML-RPCS Connector
...     class Meta:
...         name = 'xml-rpcs'
...     def get_service_url(self, service_name):
...         return 'https://%s:%s/xmlrpc/%s' % (, self.port, service_name)

Code above creates two connectors: one for XML-RPC and one for XML-RPCS. Each of connectors may be extended by simple inheritance. And if required any extension may define new branch(key)(hash) as wee see in example above.

To use this connector mc has method get_class(name[, default=False]) wich will return class generated for hash=*name*:

>>> cls = mc.get_class('xml-rpc')
>>> [b.__name__ for b in cls.__bases__]
['ConnectorXMLRPC', 'ConnectorBase']
>>> cls.__name__

>>> cls = mc.get_class('xml-rpcs')
>>> [b.__name__ for b in cls.__bases__]
['ConnectorXMLRPCS', 'ConnectorBase']
>>> cls.__name__

Example above shows what classes will be generated for specified names. By default, if mc.get_class called with unregistered name (No extension with == name defined) it will raise ValueError

If You want to allow creating of classes with not defined, just pass default=True to mc.get_class:

>>> cls = mc.get_class('unexisting-protocol', default=True)
>>> [b.__name__ for b in cls.__bases__]
>>> cls.__name__


This class provides one more level of abstraction, allowing to hide all metaclass magic behide the scene. So, using it You don’t need to worry about metaclasses and class creation process. Just inherit extensions form base class, and use in Your program instances of base class. Let’s see it in example:

>>> class MyCoolClass(Extensible):
...     my_attr_1 = 25
...     def my_method1(self, arg1):
...         print('Hello, %s' % arg1)

>>> class MyCoolClassExtension1(MyCoolClass):
...     def my_method1(self, arg1):
...         super(MyCoolClassExtension1, self).my_method1(arg1.upper())
...     def my_method2(self, arg1):
...         print("Good by, %s" % arg1)

And now using simply instances of base class You have all abilities that provided by extensions:

>>> my_cool_obj = MyCoolClass()
>>> print(my_cool_obj.my_attr_1)
>>> my_cool_obj.my_method1('World')
Hello, WORLD
>>> my_cool_obj.my_method2('World')
Good by, World

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

extend_me-1.1.4.tar.gz (18.4 kB view hashes)

Uploaded source

Built Distribution

extend_me-1.1.4-py2.py3-none-any.whl (15.8 kB view hashes)

Uploaded 3 6

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page