Mastery of the Python calls for felicity with the full depth and breadth of the language. We'll go all the way in this book to show you how to use the most powerful of Python's language features to greatest effect...with occasional reminders that sometimes there is a simpler way.

The Python Master

Preface

In this chapter we’ll be looking at some more advanced, and possibly even obscure, flow-control techniques used in Python programs with which you should be familiar at the advanced level.

Advanced Flow Control

In this chapter we’ll be going low-level and look at byte-oriented programming in Python. At the bottom, everything is bits and bytes, and sometimes it’s necessary to work at this level, particularly when dealing with binary data from other sources.

Byte Oriented Programming

In this chapter we’ll show how objects are represented internally in Python, as a dictionary called __dict__ . We’ll also demonstrate how to directly query and manipulate objects through the internal __dict__ . We’ll then use this knowledge to implement custom attribute access, by overriding the special __getattr__() , __getattrribute__() , __setattr__() , and __delattr__() methods. We’ll round off by seeing how to make objects more memory efficient by defining __slots__ .

Object Internals and Custom Attributes

In this chapter we’ll investigate a feature of Python you’ve been using — perhaps unknowingly — called descriptors . Descriptors are the mechanism used to implement properties in Python. We covered properties thoroughly in The Python Journeyman , but we’ll start here with a short review.

Descriptors

In this chapter module we’ll be taking a deep-dive into exactly what happens when we create a new object. With this knowledge in hand, we’ll be able to exercise fine control over instance creation, which allows us to customize Python objects in powerful ways.

Instance Creation

We’ve mentioned metaclasses several times. We owe you an explanation!

Metaclasses

In the last chapter we looked at metaclasses, and — while powerful — there’s no doubt that metaclasses can be difficult to understand and reason about. Fortunately, Python supports an alternative which is sufficient for many cases where we want to customise classes at the point they are defined. This alternative is the class decorator .

Class decorators

In this chapter we’ll be investigating the abstract base-class mechanism in Python as originally defined in PEP 3119 . In The Python Journeyman we used some abstract base classes, such as collections.abc.Sequence , when implementing a sorted-set collection type. This time, we’ll look at the tools provided in the Python Standard Library abc module for creating abstract base-classes of your own design.

Abstract Base Classes

Python is a large and complicated language with many moving parts. We find it remarkable that much of this complexity is hidden so well in Python. We hope in this book we’ve given you deeper insight into some important mechanisms in Python which, while a bit trickier to understand, can deliver great expressive power. You have reached the end of this book on advanced Python, and now is the time to take what you have learned and apply these powerful techniques in your work and play with Python. No matter where your journey goes, though, remember that, above all else, it’s great fun to write Python software, so enjoy yourselves!

Afterword: Continue the journey

At least Python 3.5, though any version greater than that will work as well (e.g. 3.6). See https://mail.python.org/pipermail/python-ideas/2009-October/006157.html See http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.103.6084&rep=rep1&type=pdf A widespread misconception is that else clauses associated with for-loops are executed if and only if the iterator which the for-loop is consuming is initially empty. This is not the case, the else clause will be executed when the iterator is exhausted, but only if the for-loop was also exited ‘naturally’, without invoking break . As such, the else clause is useful to handle the “not found” case when searching for something: Note we are required to pass the length parameter to specify how many bytes we want in the result and the byteorder parameter to specify whether we want the bytes returned in big-endian order with the most significant byte first, or little-endian order with the least significant byte first We’ve taken additional steps in this book to make the output even more readable by wrapping the interleaved lines of indices and buffer data to make it more readable. The code we present outputs all indices on one line, and all buffer bytes on the following line. The word cast is a synonym for type-conversion used in many different programming languages. The exact meaning is language dependent, but it typically refers to reinterpreting the bit-level data as a different type. We don’t often talk about casting in Python, as Python is strongly typed, but this one place where the notion of casting crops up. From Python 3.7 dictionary insertion order will be preserved, according to dictat from Python’s Benevolent Dictator for Life (BDFL), Guido van Rossum, on the Python mailing list in December 2017. https://mail.python.org/pipermail/python-dev/2017-December/151283.html You’ll notice that in our view, whatever the International Astronomical Union says, Pluto is a planet! Of course your system will likely report different numbers. A lot of factors are involved in memory usage. The real point is that allocating 10,000 board should show a marked increase in memory usage. Recall from chapter 4 that a descriptor is any object supporting any of the __get__() , __set__() , or __delete__() methods. 

Community Questions