Book Image

Python Web Scraping - Second Edition

By : Katharine Jarmul
Book Image

Python Web Scraping - Second Edition

By: Katharine Jarmul

Overview of this book

The Internet contains the most useful set of data ever assembled, most of which is publicly accessible for free. However, this data is not easily usable. It is embedded within the structure and style of websites and needs to be carefully extracted. Web scraping is becoming increasingly useful as a means to gather and make sense of the wealth of information available online. This book is the ultimate guide to using the latest features of Python 3.x to scrape data from websites. In the early chapters, you'll see how to extract data from static web pages. You'll learn to use caching with databases and files to save time and manage the load on servers. After covering the basics, you'll get hands-on practice building a more sophisticated crawler using browsers, crawlers, and concurrent scrapers. You'll determine when and how to scrape data from a JavaScript-dependent website using PyQt and Selenium. You'll get a better understanding of how to submit forms on complex websites protected by CAPTCHA. You'll find out how to automate these actions with Python packages such as mechanize. You'll also learn how to create class-based scrapers with Scrapy libraries and implement your learning on real websites. By the end of the book, you will have explored testing websites with scrapers, remote scraping, best practices, working with images, and many other relevant topics.
Table of Contents (10 chapters)

Automating forms with Selenium

The examples built so far work, but each form requires a fair amount of work and testing. This effort can be minimized by using Selenium as we did in Chapter 5, Dynamic Content. Because it is a browser-based solution, Selenium can mock many user interactions including clicks, scrolling and typing. If you are using it with a headless browser like PhantomJS, you will also be able to parallelize and scale your processes because it has less overhead than running a full browser.

Using a complete browser can also be a good solution for "humanizing" your interactions, particularly if you are using a well-known browser or other browser-like headers which can set you apart from other more robot-like identifiers.

Rewriting our login and editing scripts to use Selenium is fairly straightforward, but we must first investigate the page to pick out the CSS or XPath identifiers to use....