python/reference/concepts/dunder_methods.md

Dunder methods

TODO: ADD MORE

• the exercise relies on the __init__ dunder method to control class instantiation allergies
• student needs to know when to use dunder methods __init__ and __str__ binary-search-tree
• "dunder" -> "double under", referring to the names of these methods being prefixed with two underscores, e.g. __init__. There is no formal privacy in Python, but conventionally a single underscore indicates a private method, or one that the programmer should assume may change at any time; methods without an underscore are considered part of an object's public API. Double underscores are even more special - they are used by Python's builtin functions like len(), for example, to allow objects to implement various interfaces and functionality. They can also be used for operator overloading. If you have a custom class that you would like to be able to compare to other instances of the same class, implementing __lt__, __gt__, __eq__ etc. allow programmers to use the >, <, = operators. Dunder methods allow programmers to build useful objects with simple interfaces, i.e. you can add two instances together using + instead of writing something like instance1.add(instance2). hamming
• the example uses the __init__ magic method as its constructor for the class matrix
• User defined classes can (and generally do) overload the __init__ method, whose first argument is self, because the result of __init__ is a class instance. phone-number
• The example uses __init__ as a constructor for the class, which also calls __new__. In addition, the example uses __call__() via the appending of () to instance method names, and __eq__() (rich_comparison) via the use of == robot-simulator