Learning Python - Mark Lutz [562]
To Metaclass or Not to Metaclass
Metaclasses are perhaps the most advanced topic in this book, if not the Python language as a whole. To borrow a quote from the comp.lang.python newsgroup by veteran Python core developer Tim Peters (who is also the author of the famous “import this” Python motto):
[Metaclasses] are deeper magic than 99% of users should ever worry about. If you wonder whether you need them, you don’t (the people who actually need them know with certainty that they need them, and don’t need an explanation about why).
In other words, metaclasses are primarily intended for programmers building APIs and tools for others to use. In many (if not most) cases, they are probably not the best choice in applications work. This is especially true if you’re developing code that other people will use in the future. Coding something “because it seems cool” is not generally a reasonable justification, unless you are experimenting or learning.
Still, metaclasses have a wide variety of potential roles, and it’s important to know when they can be useful. For example, they can be used to enhance classes with features like tracing, object persistence, exception logging, and more. They can also be used to construct portions of a class at runtime based upon configuration files, apply function decorators to every method of a class generically, verify conformance to expected interfaces, and so on.
In their more grandiose incarnations, metaclasses can even be used to implement alternative coding patterns such as aspect-oriented programming, object/relational mappers (ORMs) for databases, and more. Although there are often alternative ways to achieve such results (as we’ll see, the roles of class decorators and metaclasses often intersect), metaclasses provide a formal model tailored to those tasks. We don’t have space to explore all such applications first-hand in this chapter but you should feel free to search the Web for additional use cases after studying the basics here.
Probably the reason for studying metaclasses most relevant to this book is that this topic can help demystify Python’s class mechanics in general. Although you may or may not code or reuse them in your work, a cursory understanding of metaclasses can impart a deeper understanding of Python at large.
Increasing Levels of Magic
Most of this book has focused on straightforward application-coding techniques, as most programmers spend their time writing modules, functions, and classes to achieve real-world goals. They may use classes and make instances, and might even do a bit of operator overloading, but they probably won’t get too deep into the details of how their classes actually work.
However, in this book we’ve also seen a variety of tools that allow us to control Python’s behavior in generic ways, and that often have more to do with Python internals or tool building than with application-programming domains:
Introspection attributes
Special attributes like __class__ and __dict__ allow us to inspect internal implementation aspects of Python objects, in order to process them generically—to list all attributes of an object, display a class’s name, and so on.
Operator overloading methods
Specially named methods such as __str__ and __add__ coded in classes intercept and provide behavior for built-in operations applied to class instances, such as printing, expression operators, and so on. They are run automatically in response to built-in operations and allow classes to conform to expected interfaces.
Attribute interception methods
A special category of operator overloading methods provide a way to intercept attribute accesses on instances generically: __getattr__, __setattr__, and __getattribute__ allow wrapper classes to insert automatically run code that may validate attribute requests and delegate them to embedded