Created
January 23, 2011 06:59
-
-
Save danneu/791887 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
def scrape | |
require 'nokogiri' | |
require 'open-uri' | |
# keyword = "nokogiri is my favorite" | |
# domain = "tenderlovemaking.com" | |
keyword = self.text.split.join("+") | |
domain = self.website.url | |
user_agent = get_random_user_agent | |
log = Logger.new("scraping.log") | |
log.debug "logger created for '#{keyword}' on '#{domain}'" | |
start = 0 # page (crawling 10 pages: 0,10,20..90) | |
i = 1 # position (1-100) | |
position = nil | |
until position or start==100 # until position gets set or reaches 10th page | |
page = (start.to_i+10)/10 | |
log.debug "page: #{page}... " | |
url = "http://www.google.com/search?q=#{keyword}&start=#{start}" | |
log.debug "url: #{url}" | |
#url = Rails.root + "test/google-search/search#{page}.html" | |
doc = Nokogiri::HTML(open(url, "User-Agent" => user_agent)) | |
links = doc.xpath('//h3/a[contains(@class, "l")]') | |
#links = doc.search('//h3/a[@class="l"]') | |
#if links empty, try a more general scrape (sometimes <a> tags didn't have a class!) | |
if links.empty? | |
links = doc.xpath('//h3[@class = "r"]/a]') | |
end | |
if links.empty? | |
if doc.at_xpath('//div/p').to_s =~ /did not match any documents/ | |
# If no results returned, | |
log.debug "Page contains no results" | |
break | |
else | |
# No links, but also no "found no results" page. throw error | |
log.debug doc.to_s | |
log.debug "Raising: PageContainsNoLinks" | |
raise "PageContainsNoLinks" | |
end | |
end | |
links.each do |link| | |
# Remove protocol prefix | |
to_remove = ["http://", "https://"] | |
reg = Regexp.new(to_remove.map{ |s| "(#{s})" }.join('|')) | |
link = link['href'].gsub(reg, '') | |
# If link start with '/url?q=', remove it | |
if link[0..6] == '/url?q=' | |
link = link[7..-1] | |
end | |
# If link starts with 'www.', remove it | |
if link[0..3] == 'www.' | |
link = link[4..-1] | |
end | |
log.debug "link #{i}. '#{link[0..domain.length-1]}' against '#{domain}'" | |
# If match, set position and stop looping | |
if domain == link[0..domain.length-1] | |
position = i | |
break | |
end | |
i += 1 | |
end | |
sleep 2.seconds # reduce request spam | |
start += 10 | |
end | |
position ||= -1 # -1 if not found | |
log.debug "position: #{position} | measured_at: #{Time.now.to_s(:short)}" | |
log.debug "=============================================================" | |
log.debug "" | |
log.debug "" | |
return {:position => position, :measured_at => Time.now, :engine => "Google"} | |
end |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment