Produces Fulltext Feeds for Heise Online

heise-feed.rb

#!/opt/local/bin/ruby
require 'rubygems'
require 'scrubyt'
require 'open-uri'
require 'builder'

LIMIT = 5

res = []

data = Scrubyt::Extractor.define do
  fetch 'http://www.heise.de/newsticker/classic/', :user_agent => 'Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_5_7; de-de) AppleWebKit/530.17 (KHTML, like Gecko) Version/4.0 Safari/530.17'
 
  link_title "//h3/a", :write_text => true do
    link_url
  end
    

end

data.to_hash.each_with_index do |item,index|
  break if index >= LIMIT
  
  content = Scrubyt::Extractor.define do
    next if item[:link_url] =~ /\Ahttp\:/

    fetch 'http://heise.de' + item[:link_url], :user_agent => 'Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_5_7; de-de) AppleWebKit/530.17 (KHTML, like Gecko) Version/4.0 Safari/530.17'
    
    item '//div[@id="mitte_news"]' do
      item_text '//div[@class="meldung_wrapper"]'
      created_at '//p[@class="news_datum"]'
    end
  end
  
  item_text = content.to_hash.first
  next unless item_text

  date = item_text[:created_at].split(/[\.: ]/).collect{|i| i.to_i}
  date = Time.utc(date[2].to_i, date[1].to_i, date[0].to_i, date[-2].to_i, date[-1].to_i)
  res << {:title => item[:link_title], 
          :link => 'http://heise.de' +  item[:link_url],
          :content => item_text[:item_text].gsub(/  +/,' ').gsub('Anzeige','').gsub(/\n\n+/,"\n"),
          :created_at => date}
  
end

builder = Builder::XmlMarkup.new(:target => STDOUT)
builder.feed do |feed|
  feed.title("Heise Online Newsticker")
  feed.updated(res.first[:created_at].strftime("%Y-%m-%dT%H:%M:%SZ"))
 
  for item in res
    feed.entry do |entry|
      entry.title(item[:title])
      entry.link(:href => item[:link])
      entry.id item[:link]
      entry.content(item[:content], :type => 'text')
      entry.updated(item[:created_at].strftime("%Y-%m-%dT%H:%M:%SZ"))
      entry.author do |author|
        author.name('Heise Online')
      end
    end
  end
end