Extract Google Search Results

Question

I would like to periodically check what sub-domains are being listed by Google.

To obtain list of sub-domains, I type 'site:example.com' in Google search box - this lists all the sub-domain results (over 20 pages for our domain).

What is the best way to extract only the URL of the addresses returned by the 'site:example.com' search?

I was thinking of writing a little python script that will do the above search and regex the URLs from the search results (repeat on all result pages). Is this a good start? Could there be a better methodology?

Cheers.

Answer 1

The Google Custom Search API can deliver results in ATOM XML format

Getting Started with Google Custom Search

Answer 2

Regex is a bad idea for parsing HTML. It's cryptic to read and relies of well-formed HTML.

Try BeautifulSoup for Python. Here's an example script that returns URLs from the first 10 pages of a site:domain.com Google query.

import sys # Used to add the BeautifulSoup folder the import path
import urllib2 # Used to read the html document

if __name__ == "__main__":
    ### Import Beautiful Soup
    ### Here, I have the BeautifulSoup folder in the level of this Python script
    ### So I need to tell Python where to look.
    sys.path.append("./BeautifulSoup")
    from BeautifulSoup import BeautifulSoup

    ### Create opener with Google-friendly user agent
    opener = urllib2.build_opener()
    opener.addheaders = [('User-agent', 'Mozilla/5.0')]

    ### Open page & generate soup
    ### the "start" variable will be used to iterate through 10 pages.
    for start in range(0,10):
        url = "http://www.google.com/search?q=site:stackoverflow.com&start=" + str(start*10)
        page = opener.open(url)
        soup = BeautifulSoup(page)

        ### Parse and find
        ### Looks like google contains URLs in <cite> tags.
        ### So for each cite tag on each page (10), print its contents (url)
        for cite in soup.findAll('cite'):
            print cite.text

Output:

stackoverflow.com/
stackoverflow.com/questions
stackoverflow.com/unanswered
stackoverflow.com/users
meta.stackoverflow.com/
blog.stackoverflow.com/
chat.meta.stackoverflow.com/
...

Of course, you could append each result to a list so you can parse it for subdomains. I just got into Python and scraping a few days ago, but this should get you started.