「行旅死亡人データベース」のスクレイパ

corpse.rb

require 'rubygems'
require 'uri'
require 'net/http'
require 'nokogiri'

class Corpse
  def initialize(uri=URI.parse('http://theoria.s284.xrea.com/corpse/index.html'))
    @uri=uri
    @contents=nil
  end
  def document
    if @document.nil?
      @document=Nokogiri(Net::HTTP.get(@uri))
    end
    @document
  end
  def contents
    if @contents.nil?
      entries=document/"*[@class='content']"
      @contents=entries.map{|e|parse_contents(e)}.flatten(1)
    end
    return @contents
  end
  def parse_contents(entry)
    permalink=(entry/"*[@class='footer']//a").first['href']
    entry_contents=[]
    current_content={:description=>[],:data=>{},:permalink=>permalink}
    entry.children.each{|node|
      case node.name
      when 'dl'
        current_data_name=''
        node.children.each{|dlnode|
          case dlnode.name
          when 'dt'
            current_data_name=dlnode.inner_text.strip
            current_content[:data][current_data_name]||=[]
          when 'dd'
            current_content[:data][current_data_name].push dlnode.inner_text.strip
          end
        }
      when 'p'
        current_content[:description].push node.inner_text
      when 'hr'
        entry_contents.push current_content
        current_content={:description=>[],:data=>{},:permalink=>permalink}
      end
    }
    entry_contents.push current_content unless current_content[:data].empty?
  end
end