Tech Star

Hi, In this post I give a detailed guide for scraping web pages which contains JavaScript contents and login forms. I’m using the following existing tools based on python.

• Scrapy: A web scrapping framework
• Splash: A lightweight browser
• Selenium: A web application test-automation tool

Scrapy provides a simple framework to scrap web sites. However, it originally does not provide facility to generate dynamic contents of JavaScript code. Therefore, Scrapy-Splash plugin was introduced. With this plugin a ‘SplashRequest’ can be used to send requests to JavaScript enabled sites and receive the dynamic content in response. However, I had trouble using corresponding ‘SplashFormRequst’ to login in to the web site, before scrapping. Hence, I used the selenium to login to the website first, then pass the cookies to ‘SplashRequest’ for further scrapping the web site.

Step 1: Generating a Scrapy Project

You can refer to scrapy documentation on how to install, create project, and create your spider. In brief

• Install Scrapy

pip install Scrapy

• Goto the desired directory ->

scrapy startproject <project name>

• In the generated project, goto ‘spiders/’ subdirectory. Then create a new python file naming your spider.

Refer: https://docs.scrapy.org/en/latest/intro/tutorial.html#our-first-spider to write your first spider.

import scrapy
from scrapy.spiders import CrawlSpider

class LoginSpider(CrawlSpider):
    name = NAME
    start_urls = [LOGIN_URL]

    def parse(self, response):
        print(response)

Step 2: Installing Splash

To install splash I followed an existing tutorial by ScrapingAuthority. It installs splash through docker platform. In brief the steps are to

• Install docker

sudo apt install docker.io

• Install Splash

sudo docker pull scrapinghub/splash

Possible Error and Solution:

If you get following error:
docker: Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?.
See ‘docker run –help’.

# sudo service docker stop
# sudo service docker start 
# sudo docker run -p 8050:8050 scrapinghub/splash

• Install scrapy-splash python plugin

sudo pip install scrapy-splash

• Update scrapy project’s settings.py with splash related configurations.

DOWNLOADER_MIDDLEWARES = {
    'scrapy_splash.SplashCookiesMiddleware': 723,
    'scrapy_splash.SplashMiddleware': 725,
    'scrapy.downloadermiddlewares.httpcompression.HttpCompressionMiddleware': 810,
}
SPLASH_URL = 'http://localhost:8050'
DUPEFILTER_CLASS = 'scrapy_splash.SplashAwareDupeFilter'
HTTPCACHE_STORAGE = 'scrapy_splash.SplashAwareFSCacheStorage'

• Start splash server

sudo docker run -p 8050:8050 scrapinghub/splash

• Update the YourSpider.py

import scrapy
from scrapy.spiders import CrawlSpider

class LoginSpider(CrawlSpider): 
    name = NAME 
    start_urls = [LOGIN_URL] 

    def parse(self, response): 
        print(response)
        yield SplashRequest(url=LOGIN_URL, callback=self.after_login, endpoint='render.html) 
    def after_login(self, response): 
        print(response)

 Step 3: Use Selenium to Login to the Web Page

• Install selenium

pip install selenium

• Update YourSpider.py as follows.

import scrapy
from scrapy.spiders import CrawlSpider

class LoginSpider(CrawlSpider): 
    name = NAME 
    start_urls = [LOGIN_URL] 

    def parse(self, response): 
        print(response)
        browser = webdriver.Chrome('/usr/bin/chromedriver', chrome_options=chrome_options)
        browser.get('https://ifttt.com/login?wp_=1')
        username = browser.find_element_by_id(USERNAME_KEY)
        password = browser.find_element_by_id(PASS_KEY)
        username.send_keys(USERNAME)
        password.send_keys(PASS)
        browser.find_element_by_name("commit").click()
        time.wait(10)
        yield SplashRequest(url=browser.current_url, callback=self.after_login, endpoint='render.html',cookies=self.browser.get_cookies()) 

    def after_login(self, response): 
        print(response)

Cheers ! 🙂