check serp of one or more sites on google for a number of keyword queries

getserp.rb

#!/usr/bin/env ruby
#
# check google serp for multiple sites using provided queries
#

require 'rubygems'
require 'mechanize'

SITES = [
  'moneyworks.ca',
  'cognito.co.nz'
]

QUERIES = [
  "moneyworks",
  "moneyworks canada",
  "moneyworks united states",
  "moneyworks us",
  "accounting software",
  "mac accounting software",
  "accounting software canada",
  "mac accounting software canada",
  "accounting software united states",
  "mac accounting software united states"
]

def get_position(a, query, site)
  offset = 0

  0.step(900,100) do |offset|
    a.get("http://google.com/search?pws=0&q=#{query.gsub(/ /,'+')}&num=100&start=#{offset}") do |page|
      # puts ">>>>>"
      # puts page.root.css('a').collect { |l| l.inspect }.join("\n")
      # puts ">>>>>"

      links = page.root.css('h3.r a.l')
      return nil if links.empty?

      links.each_with_index do |link,i|
        href = link['href']
        return "[#{i + offset + 1}] #{href}" if href =~ %r"^http(?:s)?://(?:.+\.)?#{site}/.*$"
      end
    end
  end

  nil
end

a = WWW::Mechanize.new { |agent| agent.user_agent_alias = 'Mac Safari' }

puts "keywords\t" + SITES.join("\t")
QUERIES.each do |q|
  print "#{q}\t"
  puts SITES.collect { |s| get_position(a, q, s) || '-' }.join("\t")
end