Writing High Performance Python Code

Overview

Python is a very popular, easy-to-use language that is widely adopted by developers, admins and data scientists. However, here be dragons! Many people who pick up Python lack proper coding discipline, which results in code that is hard to maintain and that runs painfully slow.

This course teaches good Python coding discipline, how to avoid bottlenecks and to write high performance code in Python

What you will learn:

Well-architected Python project
Understand python execution architecture
Learn to use profilers to spot bottlenecks
Efficient search and retrieval on Collections
Using numpy for fast numeric operations
Compiling to native code

Duration:

Half day to a day

Audience:

Developers, data analysts, data scientists

Skill level

Intermediate

Prerequisites

Basic knowledge of Python language is recommended

Lab environment

Students will need to have a Python development environment like Anaconda
Instructions will be provided on how to set this up

Students will need the following

A reasonably modern laptop
Chrome browser

Detailed Course Outline

Well-structured project

Modeling
Running
Testing

Profiling and optimization

Timing
Profiling with CProfiler
Optimization

Understand Execution Architecture

Python’s virtual machine
Understand CPU/Memory/Cache/Disk

Efficient search and retrieval on collections

Computational complexity
Working with Lists and Deques
Lists vs Arrays
Using Dictionaries
Sets
Heaps and priority queues
Use caching
Memory optimization

Using numpy for fast numeric operations

Introduction to Numpy library
Converting numeric code to Numpy

Compiling to native code

Introduction to python compilers
Using Cython