#!/usr/bin/env ruby

require 'rubygems'
require 'open-uri'
require 'hpricot'
require "activesupport"
require 'cgi'

@headers = {
  'User-Agent' => 'Mozilla/5.0 (Macintosh; U; Intel Mac OS X; en-US; rv:1.8.1.12) Gecko/20080201 Firefox/2.0.0.12',
  'Accept' => 'image/png,*/*;q=0.5',
  'Accept-Language' => 'en-us,en;q=0.5',
  'Accept-Charset' => 'ISO-8859-1,utf-8;q=0.7,*;q=0.7',
  'Referer' => 'http://www.bbc.co.uk/iplayer/'
}

@programs = []

def scrape_page(url)
  url.gsub!("&", "&")
  # sleep(4 + (rand() * 4)) # Being subtle...
  doc = Hpricot(open(url, @headers))
  begin
    (doc/'#results div.result').each do |program|
      @programs << {
        :src          => 'http://www.bbc.co.uk' + program.at('h3 a.resultlink')['href'], 
        :name         => program.at('h3 a.resultlink').inner_html,
        :img_src      => program.at('a.resultlink img')['src'],
        :title        => program.at('.resultSynopsis .title').inner_text.split('|').last.strip,
        :description  => program.at('.resultSynopsis .description').inner_text,
        :time_left    => program.at('.episode .available').inner_html.strip
      }
    end
  rescue
  end
  if pn = doc.at('#pager-next')
   scrape_page('http://www.bbc.co.uk' + pn['href'])
  end
end

# Changes yesterday:
# 
# date = Time.now.yesterday
# iplayer_date = date.strftime("%d-%m")
# morning   = "http://www.bbc.co.uk/iplayer/last7days/?filter=txdate%3A#{iplayer_date}&filter=txslot%3Amorning&scope=iplayerlast7days"
# afternoon = "http://www.bbc.co.uk/iplayer/last7days/?filter=txdate%3A#{iplayer_date}&filter=txslot%3Aafternoon&scope=iplayerlast7days"
# evening   = "http://www.bbc.co.uk/iplayer/last7days/?filter=txdate%3A#{iplayer_date}&filter=txslot%3Aevening&scope=iplayerlast7days"
# 
# [afternoon, evening].each do |url|
#   scrape_page(url)
# end

urls = [
  'http://www.bbc.co.uk/iplayer/categories/?filter=category%3A200017&order=date&scope=iplayercategories',        # Drama -  Classic & Period
  'http://www.bbc.co.uk/iplayer/categories/?filter=category%3A200018&order=date&scope=iplayercategories',        # Drama -  Crime
  'http://www.bbc.co.uk/iplayer/categories/?filter=category%3A200032&order=date&scope=iplayercategories',        # Drama -  Scifi fantasy
  'http://www.bbc.co.uk/iplayer/categories/?filter=category%3A100002PT004&order=date&scope=iplayercategories',   # Comedy - Chat shows
  'http://www.bbc.co.uk/iplayer/categories/?filter=category%3A200009&order=date&scope=iplayercategories',        # Comedy - Satire
  'http://www.bbc.co.uk/iplayer/categories/?filter=category%3A200011&order=date&scope=iplayercategories',        # Comedy - Sketch
  'http://www.bbc.co.uk/iplayer/categories/?filter=category%3A200047&order=date&scope=iplayercategories',        # Factual - Cars & Motors
  'http://www.bbc.co.uk/iplayer/categories/?filter=category%3A200055&order=date&scope=iplayercategories',        # Factual - History
  'http://www.bbc.co.uk/iplayer/categories/?filter=category%3A200045&order=date&scope=iplayercategories',        # Factual - Life stories
  #'http://www.bbc.co.uk/iplayer/categories/?filter=category%3A200058&order=date&scope=iplayercategories',        # Factual - Politics
  'http://www.bbc.co.uk/iplayer/categories/?filter=category%3A200059&order=date&scope=iplayercategories',        # Factual -  Science, Nature & Environment
  'http://www.bbc.co.uk/iplayer/categories/?filter=category%3A200060&order=date&scope=iplayercategories',        # Factual - Travel
  'http://www.bbc.co.uk/iplayer/categories/?filter=category%3A100003&order=date&start=1&scope=iplayercategories', # Drama
]

urls.each {|u| scrape_page(u) }

@programs.uniq!

@programs.each do |program|
  program[:body] =<<-EOF
    <p><a href="#{program[:src]}"><img src="#{program[:img_src]}" /></a></p>
    <ul>
      <li>Name: #{program[:name]}</li>
      <li>Title: #{program[:title]}</li>
      <li>Description: #{program[:description]}</li>
      <li>Time left: #{program[:time_left]}</li>
    </ul>
  EOF
end

@programs.each do |program|
  program[:published_at] = (7 - program[:time_left].to_i).days.ago
end

xml =  Builder::XmlMarkup.new()
xml.instruct!
xml.rss "version" => "2.0", "xmlns:dc" => "http://purl.org/dc/elements/1.1/" do
  xml.channel do
    xml.pubDate CGI.rfc1123_date(Time.now)
    xml.title "BBC iPlayer Feed"
    xml.link "http://feeds.feedburner.com/iplayer"
    xml.description "Big British Castle's iPlayer's RSS Feeed"
    @programs.each do |program|
      xml.item do
        xml.title("#{program[:name]} - #{program[:title]}")
        xml.description program[:body]
        xml.guid program[:src]
        xml.pubDate CGI.rfc1123_date(program[:published_at])
        xml.link program[:src]
      end
    end
  end
end

File.open(File.expand_path(ARGV[0] || 'iplayer-feed.rss'), 'w') {|f| f.write xml.target! }