Contents

# -*- coding: utf-8 -*-
require 'rss'

require 'rubygems'
require 'hpricot' # 2009/12/26 現在 GAE では nokogiri が使えないので hpricot
require 'dm-core'
require 'dm-types'
require 'dm-validations'

module Consadole
  module NikkanSports
    class Entry
      module Parser
        def self.included mod
          mod.extend ClassMethods
        end

        module ClassMethods
          def parse uri, doc
            target = Hpricot(doc)
            entry = self.new
            entry.uri = uri
            entry.title = target.at('h1').inner_text
            entry.text = target.search('div#news p')[0...-1].to_html.gsub(/【.+】<\/p>$/,'</p>')
            entry.auther = target.search('div#news p')[0...-1].to_html.match(/【(.+)】<\/p>$/).to_a[1]
            entry.post_date = Time.strptime(target.search('p.timeStamp').inner_text.match(/\d+年\d+月\d+日\d+時\d+分/).to_s, '%Y年%m月%d日%H時%M分')
            if tmp = target.at('dl.photo img')
              source = IMAGE_BASE_URI + tmp['src'].sub(/\.jpg$/, '-big.jpg')
              entry.image_src = source
              entry.image_binary = Net::HTTP.get(source)
              entry.image_alt = tmp['alt']
            end
            entry
          end
        end
      end

      IMAGE_BASE_URI = URI.parse('http://cache2.nipc.jp/soccer/news/')

      include Parser
      include DataMapper::Resource

      property :id, Serial
      property :uri, URI, :unique => true
      property :title, String
      property :text, Text
      property :auther, String
      property :post_date, Time
      property :image_src, URI
      property :image_alt, String
      property :image_binary, Object
    end

    FEED_URI = URI.parse('http://www.nikkansports.com/rss/soccer/jleague/consadole.rdf')

    def NikkanSports.uris
      rss = RSS::Parser.parse(FEED_URI, false)
      rss.items.map(&:link)
    end
  end
end

Version data entries

1 entries across 1 versions & 1 rubygems

Version	Path
consadole_aggregator-0.0.4	lib/consadole_aggregator/nikkan_sports.rb