Popular recipes tagged "pattern"

Create on-the-fly class adapters with functools.partial (Python)

2014-05-29T18:48:28-07:00

Python recipe 578884 by Christoph Schueler (adapter, namedtuple, partial, pattern).

functools.partial could not only applied to functions it also works with classes. This opens some interesting perspectives, like on-the-fly creation of class adapters, as the following code illustrates.

Observer Pattern (Python)

2013-03-09T10:03:10-08:00

Python recipe 578484 by Mauro B. Bianc (cascade, observer, oop, pattern, patterns, setattr, __setattr__). Revision 2.

This is a Python implementation of the observer pattern described by Gamma et. al. It defines a one-to many dependency between objects so that when one object changes state, all its dependents (i.e. observers) are notified and updated automatically.

My adaptation gets rid of the need to use specific functions to set the data (and to call Notify) and allows you to be notified for ANY attribute you set. It is possible to specify a list of attributes which should not trigger a notification. In case you need the opposite, it is very easy to invert the behavior of the code.

The example should output: Creating data1 without notification for attrs name & surname
Creating data2 without notification for attr age
Setting data1.name=Heather - Notification unnecessary
Setting data1.num=333 - Notification expected
Observer1: Subject Heather has updated attr num to 333
Setting data2.name=Molly - Notification expected
Observer2: Subject Molly has updated attr name to Molly
Setting data2.age=28 - Notification unnecessary
Setting data2.eyecolor=blue - Notification expected
Observer2: Subject Molly has updated attr eyecolor to blue

Multiple unique class instances - a tentative design pattern (Python)

2013-02-23T14:11:44-08:00

Python recipe 578471 by Moritz Beber (design_pattern, pattern, unique).

Coming from Bioinformatics and having to deal with multiple objects with unique properties, like genes, proteins, and many more, I felt the need for a class that would always yield a unique instance based on a chosen identifier. This is because I always wanted the same instances whose attributes were filled with information and track them in various storage classes, like dictionaries, lists, etc. The code for the class lives over on github. Recently, I've added a lot of parallel-processing code so I adapted the pickling behaviour for this class to also yield existing instances. What follows is an example of how to use it and some discussion questions.

Singleton? We don't need no stinkin' singleton: the Borg design pattern (Python)

2011-09-18T18:47:59-07:00

Python recipe 577870 by Anler Hernández Peral (oop, pattern).

The Singleton design pattern (DP) has a catchy name, but the wrong focus -- on identity rather than on state. The Borg design pattern has all instances share state instead, and Python makes it, literally, a snap.

Concrete Class Finder (Python)

2011-08-24T05:29:15-07:00

Python recipe 577858 by Lucio Santi (case, class, design, finder, match, oop, pattern, suitable, switch).

This recipe implements a design pattern useful for performing an object-oriented case analysis for a particular object (or a collection of objects as well). Essentially, it is an alternative to complex if-then-else or switches. Modelling each case with a particular class, the Concrete Class Finder searches for an appropriate case/class that applies to the given object/s. Once found, this class can be used in an homogeneous way, independently of the object/s previously considered.

Maybe pattern (Python)

2010-07-21T11:08:11-07:00

Python recipe 577248 by Alan Franzoni (design, maybe, pattern). Revision 2.

An example Maybe pattern implementation for Python.

Flexible observer pattern implementation (Python)

2012-12-06T19:23:11-08:00

Python recipe 576962 by Glenn Eychaner (event, observer, pattern, publish, subscribe, threadsafe). Revision 16.

A simple, flexible, general-purpose observer pattern.

Observers can be callable objects or objects with a particular named method (handle_notify() by default). Events can be any object, and observers can select which events they are interested in receiving. Support for a number of different types of lightweight event objects is included.

Multicontext (e.g. asynchronous) inline execution framework using coroutines (Python)

2012-12-06T19:32:20-08:00

Python recipe 576965 by Glenn Eychaner (asynchronous, coroutine, decorator, generator, inline, nonblocking, pattern, thread). Revision 14.

A framework for executing inline code, contained in a generator, across multiple execution contexts, by pairing it with an executor that handles the context switching at each yield. An example of a generator which executes some iterations synchronously and some asynchronously is provided. The framework is general enough to be applied to many different coroutine situations.

Maintenance free Factory design pattern (Python)

2009-03-12T08:54:39-07:00

Python recipe 576687 by Akira Fora (design, factory, free, maintenance, pattern). Revision 3.

An implementation of the Factory design pattern that doesn't require maintenance when a new specialized subclass is added.