Scrapping book cover images with Scrapy and Python 3

2017-05-20 ·

Let's start by creating a virtual environment:

python3 -m venv ~/.venv/bookspider
source ~/.venv/bookspider/bin/activate

Now we can install scrapy

pip install scrapy

Let's generate the initial directory structure

scrapy startproject bookspider

items.py allows us to define a data object model for the pages crawled by the spider. This class encapsulates each piece of data being scraped.

import scrapy

class Book(scrapy.Item):
  title = scrapy.Field()
  description = scrapy.Field()
  file_urls = scrapy.Field()
  files = scrapy.Field()

file_urls & files are special fields which must be explicitly defined to scrape binary files (images, PDFs, MP3s, etc).

Create a spider inside spiders/ directory. It's a class that inherits from scrapy.Spider and must have name and start_urls fields defined.

import scrapy
from speck.items import BookItem

class BookSpider(scrapy.Spider):
  start_urls = [
    'http://books.toscrape.com/catalogue/category/books/romance_8/page-1.html',
    'http://books.toscrape.com/catalogue/category/books/romance_8/page-2.html'
  ]
  name = 'book'

  def parse(self, response):
    for book in response.css('li article.product_pod'):
      href = book.css('a::attr(href)').extract_first()
      full_url = response.urljoin(href)

      yield scrapy.Request(full_url, callback=self.parse_book)

  def parse_book(self, response):
    title = response.css('h1::text').extract_first()
    description = response.css('.product_page > p::text').extract()
    src = response.css('.product_page .thumbnail img::attr(src)').extract_first()
    cover = response.urljoin(src)

    yield BookItem(title=title, description=description, file_urls=[cover])

start_urls is a list of the seed URLs the spider will crawl first. We are targeting Scrapy playground available at books.toscrape.com which provides a dummy book store.

Every spider must have at least a parse() method that handles the URLs defined in start_urls . This method can either yield other requests which will trigger other pages to be crawled either return the values.

scrapy allows to traverse the DOM using both CSS and XPath selectors.

The last thing is to enable the item pipelines in settings.py so that Scrapy automatically downloads each files put into file_urls

ITEM_PIPELINES = {
  'scrapy.pipelines.files.FilesPipeline': 1
}
FILES_STORE = 'book_cover_images'

Additionally we must also define FILES_STORE setting i.e. the path to the output directory where the download images will be stored.

Finally, let's run the scrapper

scrapy crawl book -o book.jl
Copyright © 2018 All rights reserved.
Powered by Kulfon.