ihub-rss.rb

#!/usr/local/bin/ruby

require 'rubygems'
require 'hpricot'
require 'open-uri'
require 'atom/pub'
require 'htmlentities'

ENV['TZ'] = 'US/Eastern'
BASE_URL = "http://investorshub.advfn.com/boards/"
FORUM_URL = BASE_URL + "board.aspx?board_id=2276"
DESTINATION = "/tmp/ihub.atom"
CODER = HTMLEntities.new

feed = Atom::Feed.new do |f|
  f.title = "Investor's Hub - NEOM"
  f.id = "http://flairly.com/"
  f.links << Atom::Link.new(:href => "http://flairly.com/ihub.atom", :rel => 'self')
  Hpricot(open(FORUM_URL)).search("//table[@id='ctl00_CP1_gv']/tr").each do |row|
    unless row[:style] == 'background-color:PaleGoldenrod;'
      if link = (row/"td:nth(1)/a").first
        f.entries << Atom::Entry.new do |e|
          e.title = "#{CODER.decode((row/"td:nth(1)/a").inner_text)} ...".capitalize
          e.links << Atom::Link.new(:href => "#{BASE_URL}#{link[:href]}", :rel => :alternate)
          e.updated = Time.parse((row/"td:nth(3)").inner_text)
          e.id = "tag:flairly.com,#{e.updated.strftime('%Y-%m-%d')}:/post/#{link[:href].match(/.*=(\d+)/)[1]}"
          author = (row/"td:nth(2)/a").inner_text
          e.authors << Atom::Person.new(:name => author)
          message = Hpricot(open("#{BASE_URL}#{link[:href]}"))
          content = message.at("//span[@id='intelliTXT']").inner_html
          if replied_to = message.at("//a[@id='ctl00_CP1_mh1_hlReplyTo']")
            replied_to = replied_to.inner_text
            content = "<span style='font-size:small;color:gray;font-weight:bold;'>#{author} said to #{replied_to.strip}:</span><br/>" + content
          end
          e.content = Atom::Content::Html.new(content)
        end
      end
    end
  end
  f.updated = f.entries[0].updated
end

File.open(DESTINATION,"w") do |f|
  f.write(feed.to_xml)
end