gistfile1.rb

#!/usr/local/bin/ruby

require 'rubygems'
require 'hpricot'
require 'open-uri'
require 'pp'

# sudo gem install hpricot

# http://www.katalogiseo.info/str-2-1,60.html
# http://www.katalogiseo.info/str-2-1,120.html
# http://www.katalogiseo.info/str-2-1,180.html

def img_true?(*html)
    html.map { |img|
        img.attributes['src'] =~ /N\.png/ ? false : true
    }
end

dir_count = 0
60.step(5820,60) { |n| puts n
    url = "http://www.katalogiseo.info/str-2-1,#{n}.html"
    puts "#{url}..."

    html = open(url).read
    doc = Hpricot html

    doc.search("table.bodyline > tr > td > table > tr").each { |e|
        id = e.at("td").inner_text.to_s.gsub(/\[(\d+)\]/,'\1')
        next if id !~ /^\d+$/
        dir_count += 1

        type = e.at("td").siblings_at(2).inner_text
        dir = e.at("td > div > div > a.linki")
        dir_name = dir.inner_text
        dir_url = dir.attributes['href']
        back_link = img_true?(e.at("td").siblings_at(3).at('img'))
        free, mod = img_true?(*e.at("td").siblings_at(4).search("img"))

        puts ">>> #{dir_url} - #{dir_name}"
        puts "  >>> #{type}"
        puts "  >>> back_link?/free/mod : #{back_link}/#{free}/#{mod}"
    }
}

puts "count: #{dir_count}"

#  60
#  http://www.katalogiseo.info/str-2-1,60.html...
#  >>> http://www.dir.comoj.com - Katalog dir.comoj.com
#    >>> qlWeb by Jachu
#    >>> back_link?/free/mod : false/true/false
#  >>> http://randki.amor.com.pl - Katalog AMOR
#    >>> ABIX
#    >>> back_link?/free/mod : true/true/true
#  >>> http://maszynykat.pl - Katalog Maszyny Do Metalu
#    >>> qlWeb
#    >>> back_link?/free/mod : false/true/true
#  >>> http://blog.tb3.pl - P.Page To Be Free
#    >>> WordPress
#    >>> back_link?/free/mod : false/true/true
#  >>> http://www.pres.impresa.pl - P.Page Impresa
#    >>> MocneLinki
#    >>> back_link?/free/mod : false/true/false
# etc....