model.rb

class PixivImage
  include Mongoid::Document
  field :image_id
  field :pixiv_user_id
  field :title
  field :tags, :type => Array
  field :url
  field :image_url
  field :description
  field :saved, :type => Boolean, :default => false
  field :created_at, :type => Time
  validates_uniqueness_of :image_id

  def fetch
    agent = Pixiv.agent
    self.url = "http://www.pixiv.net/member_illust.php?mode=medium&illust_id=#{self.image_id}"
    page = agent.get URI.parse(self.url)
    self.title = page.root.xpath("//div[@class='works_data']/h3").inner_text
    self.tags = page.root.xpath("//*[@id='tags']/a").to_ary.delete_if{|x| x["href"] == "_blank"}.collect{|x| x.inner_text}.delete_if{|x| x.empty?}
    self.image_url = page.root.xpath('//div[@class="works_display"]/a/img').attr("src").to_s
    self.description = page.root.xpath('//div[@class="works_area"]/p').inner_text
  end

  def user
    PixivUser.where(:pixiv_id => self.pixiv_user_id).first
  end
end


class PixivUser
  include Mongoid::Document
  field :name
  field :pixiv_id
  validates_uniqueness_of :pixiv_id

  def images
    agent = Pixiv.agent
    page = agent.get URI.parse("http://www.pixiv.net/member_illust.php?id=#{self.pixiv_id}")
    illusts = page.root.xpath '//div[@class="display_works linkStyleWorks"]'
    links = illusts.xpath '//ul/li/a'
    image_ids = []
    links.each do |l|
      link_to = l.attr("href")
      if /medium/ =~ link_to
        link_to.scan(/[0-9]*/).each do |ids|
          if ids and ids != ""
            image_ids << ids
          end
        end
      end
    end
    image_ids.reverse!
    image_ids.each{|i|
      p = PixivImage.new unless p = PixivImage.where(:image_id => i).first
      p.image_id = i
      p.fetch
    }
    PixivImage.where(:pixiv_user_id => self.pixiv_id).order("created_at desc")
  end



  def self.find_from_pixiv pixiv_id, agent
    unless p = where(:pixiv_id => pixiv_id).first
      p = new
    end
    page = agent.get "http://www.pixiv.net/member.php?id=#{pixiv_id}"
    p.name = page.root.xpath('//div[@class="profile_area"]/h2/span').inner_text.split("\n").pop
    p.pixiv_id = pixiv_id
    p.save!
    p
  end
end

# Service
class Pixiv
  def self.agent
    count = 0
    begin
      agent = Mechanize.new
      page = agent.get URI.parse("http://www.pixiv.net/")
      form = page.forms[0]
      form.pixiv_id = "ID"
      form.pass = "PASSWORD"
      res = agent.submit(form)
      return agent
    rescue
      count += 1
      if count < 10
        retry
      end
    end
  end
end

model_spec.rb

#coding:utf-8
require File.dirname(__FILE__) + '/spec_helper'

describe 'Application' do
  describe Pixiv do
    it "ログイン済みの Agent を取得出来る" do
      agent = Pixiv.agent
      agent.class.should == Mechanize
      agent.page.root.xpath("//div[@class='profile_area']/h2").inner_text.should == "ssig33"
    end
  end
  describe PixivUser do
    it "ユーザー ID を指定して情報が Pixiv から取得出来る" do
      p = PixivUser.find_from_pixiv "1048706", Pixiv.agent
      p.pixiv_id.should == "1048706"
      p.name.should == "ファーボはオワコン"
    end
    it "一度 Pixiv から取得された情報は MongoDB に保存される" do
      p = PixivUser.find_from_pixiv "1048706", Pixiv.agent
      p.pixiv_id.should == "1048706"
      p.name.should == "ファーボはオワコン"
      p = PixivUser.where(:name => "ファーボはオワコン").first
      p.pixiv_id.should == "1048706"
      p.name.should == "ファーボはオワコン"
    end
    it "画像一覧の取得が出来る"
  end
  describe PixivImage do
    it "画像の詳細を取得出来る" do
      p = PixivImage.new
      p.image_id = "9827118"
      p.fetch
      p.url.should == "http://www.pixiv.net/member_illust.php?mode=medium&illust_id=9827118"
      p.title.should == "ボゴ・ザ・イケブクロシティボーイ"
      p.description.should == "サンリオダブルパロ第三弾です！キラリと光るナイフでバイオレンスにキメたボゴ原ボゴ也さんです！"
      p.user.name.should == "ファーボはオワコン"
    end
  end
end