Support for Cooperative Multi-inheritance

Note

This section is not about sub-classing from more that one Qt class.

Cooperative multi-inheritance is a technique for implementing classes that inherit multiple super-classes - typically a main super-class and one or more mixin classes that add additional behaviour. It makes it easy to add new mixins at a later date to further extend the behavior, without needing to change either the implementation of the class or any existing code that creates an instance of the class.

The technique requires that all the super-class’s __init__ methods follow the same pattern in the way that they handle unrecognised keyword arguments and use super() to invoke their own super-class’s __init__ methods.

PyQt5’s classes follow this pattern.

See Raymond Hettinger’s Python’s super() considered super! blog post for some more background on the subject.

As an example, let’s say we have a class that represents a person, and that a person has a name. The following might be an initial implementation:

class Person(QObject):
    def __init__(self, name, parent=None)
        QObject.__init__(self, parent)

        self.name = name

An instance would normally be created in one of the following ways:

person = Person("Joe")
person = Person("Joe", some_parent)

This approach has some limitations:

• Only a sub-set of the QObject API is exposed. For example you cannot set the value of a Qt property or connect a signal by passing appropriate keyword arguments to Person.__init__.
• Adding another class to Person’s list of super-classes means that its __init__ implementation needs to be changed. If the new mixin takes non-optional arguments then every call to create a Person instance will need changing.

Consider this alternative implementation:

class Person(QObject):
    def __init__(self, name, **kwds):
        super().__init__(**kwds)

        self.name = name

The difference is that we only handle arguments that are used by the Person class itself and we punt all the other arguments to the super-classes by calling super().

With this implementation an instance would normally be created in one of the following ways:

person = Person("Joe")
person = Person("Joe", parent=some_parent)

Here the difference is that we are using keyword arguments to specify any arguments that are not handled by the Person class itself. Note that we could use keyword arguments for all arguments - whether or not you do so is down to personal choice.

The limitations of the first implementation no longer apply. For example, without any further changes we can also do this:

person = Person("Joe", destroyed=some_callable)

Let’s say we now want to extend the behaviour of the Person class by adding a mixin that handles a person’s age. The implementation of the mixin would be as follows:

class Age(object):
    def __init__(self, age=0, **kwds):
        super().__init__(**kwds)

        self.age = age

This follows a similar pattern to our Person implementation, but notice that we have provided the age argument with a default value.

The following is our new Person implementation:

class Person(QObject, Age):
    def __init__(self, name, **kwds):
        super().__init__(**kwds)

        self.name = name

The only change we have had to make is to add Age to Person’s list of super-classes. More importantly we do not need to change any call to create a Person instance.

If we do want to create a Person instance with a non-default age then we simply pass it as a keyword argument as follows:

person = Person("Joe", age=38)

This technique increases the use of keyword arguments - while this means a bit more typing, it significantly increases the readability of application code.