Hey, here’s a fun one. Just last week cdcarter needed to scrape and Rubyify the SciFi channel’s listings. (Wow, those guys really like tables, huh? And nondescript markup. And PHP3. (PHP3 was sooo the best.))

Quickly carter and I dusted off Hpricot and, with it, scraped the hell out of the listing page. We then turned each listing into a Show object, easy. With OpenStruct.

%w[open-uri rubygems hpricot ostruct].each { |f| require f }

class Show < OpenStruct
  LISTINGS = 'http://www.scifi.com/schedulebot/index.php3?feed_req=US:Central:E'

  def to_s
    "#{time}: #{title}" << (program ? " [#{program}]" : '')
  end  

  def self.find_all_from_today
    shows = []
    doc = Hpricot open(LISTINGS)
    tds = (doc/:td).select { |td| td.respond_to?(:[]) && td['class'] == 'text' }
    tds.each_with_index do |td, i|
      next unless td.innerHTML =~ /:.+(AM|PM)/
      time    = td.innerHTML
      program = tds[i+1].innerHTML.gsub(/<a.+>(.+)<\/a>/, '\1')
      title   = tds[i+2].innerHTML
      shows << new(:time => time, :program => program, :title => title)
    end
    shows
  end
end

# print all found shows for today
Show.find_all_from_today.each { |show| puts show.to_s }

Run it. I get something like this:

5:00 AM: [PAID PROGRAMMING]
7:00 AM: SHADOW PLAY [TWILIGHT ZONE, THE]
7:30 AM: BLACK MARKET [BATTLESTAR GALACTICA (SEASON 2)]
8:30 AM: SCAR [BATTLESTAR GALACTICA (SEASON 2)]
9:30 AM: SACRIFICE [BATTLESTAR GALACTICA (SEASON 2)]
...

Way cool (even though I’m dying to slip in some returning action). You can imagine how this might be expanded into a nice little pirate RSS feed or something.

Any more cool Struct or OpenStruct uses floating around out there? Jay Fields has done lots of messin’ with OpenStruct and kindly sprinkles a few write-ups throughout his blog. How’s about yous?