#!/usr/bin/env ruby

require "db"
require "simple-rss"
require "open-uri"
require "classifier"
require "madeleine"

MIN_CONTENT = 10

while (true)
  m = SnapshotMadeleine.new("bayes_data")
  bayes = m.system

  categories = Category.find(:all)
  threads = []

  for category in categories
    threads << Thread.new(category) { |c|
      contents = Content.find_by_category_id(:all, c.id)

      if contents.nil? or contents.count < MIN_CONTENT
        needed = MIN_CONTENT - contents.count
        feeds = Feed.find_by_category_id(:all, c.id)

        catch :no_feeds do
          throw :no_feeds unless feeds

          feeds.each { |feed|
            begin
              rss = SimpleRSS.parse open(feed.url)
              feed.lastused_at = Time.now
              feed.save

              for item in feed.items
                item_category = bayes.classify item.description

                if item_category == c.name
                  content = Content.new
                  content.title = item.title
                  content.link = item.link
                  content.description = item.description
                  content.category = c

                  content.save

                  needed = needed - 1
                end

                break unless needed > 0
              end
            rescue
               print "b00m!"
            end

            break unless needed > 0
          }
        end
      end
    }
  end

  threads.each { |t| t.join }
  sleep(600)
end