rss_scoring.rb

# -*- coding: utf-8 -*-
require "rexml/document"
require 'open-uri'
require './cityofhypes'
include REXML

class TopArticles
  def self.scoring(rss_url,id)
    # return id,score,url
    # id = RSS Feed ID 
    # [{:id => Feed ID for filtering articles 
    #  :score => aritlce's score determined by CityOfHypes
    #  :url => aritcle's URL}...each aritcle]
    rss = Document.new(open(rss_url).read) rescue return
    link = rss.elements.collect("rss/channel/item/link"){|element|  element.text }
    scores = link.map {|url| {:id =>id,:score =>  CityOfHypes.score_of(url), :url =>url}}
  end
  
  def self.top_articles(urls,each_feed,max)
    #returns array of top articles
    # urls:RSS Feeds
    # each_feed: maximum number of each feed
    # max : maximum number for all articles
    # [
    #   {
    #    :id => Feed ID,
    #    :score => score given by cityofhypes
    #    :url => article's url
    #   }
    # ]
    
    list = urls.collect.with_index{|url,id| self.scoring(url,id)}.flatten
    list.sort!{|a,b| a[:score].to_i <=> b[:score].to_i }
    
    count = Array.new(urls.size,0)
    ta = list.inject([]) do |score,info|      
      return score if score.size >= max
      if count[info[:id].to_i].to_i < each_feed
        count[info[:id].to_i] += 1
        score << info 
      end
      score
    end
    return ta
  end
end




# sample URL 'TechCrunch JP' 'gizmodo jp' engadget jp' 'gigazine' 'lifehacker jp' 'lifehacker us' 'gizmodo us'
urls = ["http://www.pheedo.jp/f/JapaneseTechCrunch","http://www.gizmodo.jp/index.xml",
        "http://japanese.engadget.com/rss.xml","http://gigazine.net/index.php/news/rss_2.0/",
        "http://www.lifehacker.jp/index.xml","http://lifehacker.com/tag/top/index.xml",
        "http://gizmodo.com/tag/top/index.xml"]

p TopArticles.top_articles(urls,2,10)