omelete 1.0.0 → 2.0.0

data/lib/model/movie.rb

@@ -0,0 +1,8 @@
+class Movie
+  attr_accessor :omelete_id, :name, :runtime, :genre, :age_rating, :image
+  attr_accessor :cast, :directed_by, :release_date, :synopsis
+  
+  def initialize(omelete_id, name)
+    self.omelete_id, self.name = omelete_id, name
+  end
+end

data/lib/model/movie_theater.rb

@@ -0,0 +1,7 @@
+class MovieTheater
+  attr_accessor :omelete_id, :name, :values
+  
+  def initialize(omelete_id, name, values)
+    self.omelete_id, self.name, self.values = omelete_id, name, values
+  end
+end

data/lib/model/showtime.rb

@@ -0,0 +1,17 @@
+class Showtime
+  attr_accessor :omelete_movie_id, :obs, :kind, :time, :theater
+  
+  def initialize(omelete_movie_id, obs, kind, time, theater)
+    self.omelete_movie_id, self.obs, self.kind, self.time, self.theater = omelete_movie_id, pretty_obs(obs), kind, time, theater
+  end
+  
+  def pretty_obs(obs)
+    if obs
+      while obs.match(/\.[A-Z] | feriado[A-Z]/) do
+        obs.insert(obs.index(/\.[A-Z]/) + 1, "\n") if obs.match(/\.[A-Z]/)
+        obs.insert(obs.index(/feriado[A-Z]/) + 7, "\n") if obs.match(/feriado[A-Z]/)
+      end
+    end
+    obs    
+  end
+end

data/lib/model/theater.rb

@@ -0,0 +1,7 @@
+class Theater
+  attr_accessor :name
+  
+  def initialize(name)
+    self.name = name
+  end
+end

data/lib/omelete/movie_agent.rb

@@ -0,0 +1,39 @@
+# -*- encoding : utf-8 -*-
+#!/usr/bin/env ruby
+require 'nokogiri'
+require 'open-uri'
+require File.expand_path("../../../lib/omelete", __FILE__)
+require File.expand_path("../../../lib/model/movie", __FILE__)
+
+module Omelete
+  class MovieAgent
+    
+    def initialize(omelete_id)
+      @client = Client.new
+      @page_doc = @client.page_doc("cinema/#{omelete_id}")
+    end
+    
+    def movies
+      movies=[]
+      omelete_id, name, runtime, genre, age_rating, image = nil, nil, nil, nil, nil, nil
+      @page_doc.css("div[class='programacao_filme']").each do |m|
+        omelete_id = m.css("div[class='programacao_filme_desc'] h2").css("a")[0]
+        omelete_id = omelete_id ? omelete_id["href"][/\d+/] : nil
+        name       = m.css("div[class='programacao_filme_desc'] h2 a").text
+        movie = Movie.new(omelete_id, name)
+        
+        image            = m.css("div[class='programacao_filme_poster']").css("a")[0]
+        genre            = m.css("div[class='programacao_filme_desc'] h4").text.split(" - ")[0]
+        movie.runtime    = m.css("div[class='programacao_filme_desc'] h4").text.split(" - ")[1]
+        age_rating       = m.css("div[class='programacao_filme_desc'] h4").text.split(" - ")[2]
+        
+        movie.image = image ? image["href"] : nil
+        movie.age_rating = age_rating ? age_rating[/\d+/] : nil
+        
+        movies << movie
+      end
+      movies
+    end
+    
+  end
+end

data/lib/omelete/movie_details_agent.rb

@@ -0,0 +1,32 @@
+# -*- encoding : utf-8 -*-
+#!/usr/bin/env ruby
+require 'nokogiri'
+require 'open-uri'
+require File.expand_path("../../../lib/omelete", __FILE__)
+require File.expand_path("../../../lib/model/movie", __FILE__)
+
+module Omelete
+  class MovieDetailsAgent
+    
+    def initialize(movie_id)
+      @client = Client.new
+      @id = movie_id
+      @page_doc = @client.page_doc("filme/#{movie_id}")
+    end
+    
+    def movie
+      synopsis, cast, directed_by = nil, nil, nil
+      movie = Movie.new @id, nil
+      movie.synopsis = @page_doc.css("div#tab_#{@id}_sinopse blockquote").text
+      @page_doc.css("div#tab_#{@id}_ficha dl").children.each do |child|
+        if child.node_name == "dd" && child.previous_element && child.previous_element.content == "Direção:"
+          movie.directed_by = child.content.delete("\r\n")
+        elsif child.node_name == "dd" && child.previous_element && child.previous_element.content == "Elenco:"
+          movie.cast = child.content.delete("\r\n")
+        end
+      end
+      movie
+    end
+    
+  end
+end

data/lib/omelete/movie_theater_agent.rb

@@ -0,0 +1,24 @@
+# -*- encoding : utf-8 -*-
+#!/usr/bin/env ruby
+require 'nokogiri'
+require 'open-uri'
+require File.expand_path("../../../lib/omelete", __FILE__)
+require File.expand_path("../../../lib/model/movie_theater", __FILE__)
+
+module Omelete
+  class MovieTheaterAgent
+    
+    def initialize(omelete_id)
+      client = Client.new
+      @id = omelete_id
+      @page_doc = client.page_doc("cinema/#{@id}")
+    end
+    
+    def movie_theater
+      name = @page_doc.css("div[class='programacao_cinema'] h2 a").text
+      values = @page_doc.css("div#tab_#{@id}_ingressos dl dd").text
+      @movie_theater = MovieTheater.new @id, name, values
+    end
+    
+  end
+end

data/lib/omelete/showtime_agent.rb

@@ -0,0 +1,37 @@
+# -*- encoding : utf-8 -*-
+#!/usr/bin/env ruby
+require 'nokogiri'
+require 'open-uri'
+require File.expand_path("../../../lib/omelete", __FILE__)
+require File.expand_path("../../../lib/model/showtime", __FILE__)
+
+module Omelete
+  class ShowtimeAgent
+    
+    def initialize(omelete_id)
+      client = Client.new
+      @id = omelete_id
+      @page_doc = client.page_doc("cinema/#{@id}")
+    end
+    
+    def showtimes
+      showtimes = []
+      obs, kind, time, theater = nil, nil, nil, nil
+      @page_doc.css("div[class='programacao_horarios']").each do |div|
+        omelete_movie_id = div.css("p a")[0]["href"][/\d+/]
+        div.css("table tr td").each_with_index do |td, i|
+          obs = td.text if td.attribute("colspan")
+          unless td.attribute("colspan")
+            theater = td.text if i == 0
+            kind = td.text if i == 1
+            time = td.text if i == 2
+          end
+        end
+        
+        showtimes << Showtime.new(omelete_movie_id, obs, kind, time, theater)
+      end
+      showtimes
+    end
+    
+  end
+end

data/lib/omelete/theater_agent.rb

@@ -0,0 +1,29 @@
+# -*- encoding : utf-8 -*-
+#!/usr/bin/env ruby
+require 'nokogiri'
+require 'open-uri'
+require File.expand_path("../../../lib/omelete", __FILE__)
+require File.expand_path("../../../lib/model/theater", __FILE__)
+
+module Omelete
+  class TheaterAgent
+    
+    def initialize(omelete_id)
+      client = Client.new
+      @id = omelete_id
+      @page_doc = client.page_doc("cinema/#{@id}")
+    end
+    
+    def theaters
+      theaters = []
+      @page_doc.css("div[class='programacao_horarios'] table tr td").each do |td|
+        if td.text.match(/^[0-9]{1,}$/)
+          @theater = Theater.new td.text
+          theaters << @theater
+        end
+      end
+      theaters
+    end
+    
+  end
+end

data/lib/omelete.rb

@@ -2,21 +2,21 @@
 #!/usr/bin/env ruby
 require 'nokogiri'
 require 'open-uri'
-require File.expand_path('../omelete/movie_page', __FILE__)
-
 
 module Omelete
   class Client
     
-    def initialize(state,city)
-      @state=state
-      @city=city
-      @movie_page = MoviePage.new(@state,@city)
-    end
+    HOME_URL = "http://omelete.uol.com.br/filmes-em-cartaz/"
+    
+    def page_doc(complement)
+      begin
+        uri = URI.escape(HOME_URL + complement)
+        @page_doc = Nokogiri::HTML(open(uri))
+        @page_doc
+      rescue SocketError => e
+        raise e.message
+      end
+    end    
     
-    def movies
-      @movie_page.movies if @movie_page
-    end
-
   end
 end

data/lib/version.rb

@@ -1,3 +1,3 @@
 module Omelete
-  VERSION = "1.0.0"
+  VERSION = "2.0.0"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: omelete
 version: !ruby/object:Gem::Version
-  version: 1.0.0
+  version: 2.0.0
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-05-24 00:00:00.000000000 Z
+date: 2012-10-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
@@ -28,14 +28,14 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
-  name: mechanize
+  name: rspec
   requirement: !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
         version: '0'
-  type: :runtime
+  type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     none: false
@@ -44,21 +44,21 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
-  name: rspec
+  name: mocha
   requirement: !ruby/object:Gem::Requirement
     none: false
     requirements:
-    - - ~>
+    - - ! '>='
       - !ruby/object:Gem::Version
-        version: '2.7'
+        version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     none: false
     requirements:
-    - - ~>
+    - - ! '>='
       - !ruby/object:Gem::Version
-        version: '2.7'
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: vcr
   requirement: !ruby/object:Gem::Requirement
@@ -80,17 +80,17 @@ dependencies:
   requirement: !ruby/object:Gem::Requirement
     none: false
     requirements:
-    - - ~>
+    - - ! '>='
       - !ruby/object:Gem::Version
-        version: 1.8.0
+        version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     none: false
     requirements:
-    - - ~>
+    - - ! '>='
       - !ruby/object:Gem::Version
-        version: 1.8.0
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
@@ -114,10 +114,15 @@ executables: []
 extensions: []
 extra_rdoc_files: []
 files:
-- lib/omelete/detailed_page.rb
-- lib/omelete/models/movie.rb
-- lib/omelete/models/showtime.rb
-- lib/omelete/movie_page.rb
+- lib/model/movie.rb
+- lib/model/movie_theater.rb
+- lib/model/showtime.rb
+- lib/model/theater.rb
+- lib/omelete/movie_agent.rb
+- lib/omelete/movie_details_agent.rb
+- lib/omelete/movie_theater_agent.rb
+- lib/omelete/showtime_agent.rb
+- lib/omelete/theater_agent.rb
 - lib/omelete.rb
 - lib/version.rb
 homepage: http://github.com/mvoto/omelete
@@ -140,7 +145,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 1.8.23
+rubygems_version: 1.8.24
 signing_key: 
 specification_version: 3
 summary: Ruby web crawler to access omelete informations

data/lib/omelete/detailed_page.rb

@@ -1,107 +0,0 @@
-# -*- encoding : utf-8 -*-
-#!/usr/bin/env ruby
-$:.unshift File.expand_path('../models', __FILE__)
-require 'rubygems'
-require 'open-uri'
-require 'nokogiri'
-require 'mechanize'
-require 'movie'
-require 'showtime'
-
-module Omelete  
-  class DetailedPage
-    
-    def initialize(page,link)
-      @page = page.link_with(:href => link).click
-    end
-    
-    def movie_poi(movie)
-      movie.movie_theaters = movie_theaters(movie) unless @page.search('//div[@class = "programacao_cinema"]').first.nil?
-      movie.showtimes = showtimes(movie) unless @page.search('//div[@class = "programacao_cinema"]').first.nil?
-      movie.synopsis = synopsis(movie) unless @page.search("//div[@id = \"tab_#{movie.id}_sinopse\"]/blockquote").first.nil?
-      unless @page.search("//div[@id = \"tab_#{movie.id}_ficha\"]/dl/dd").first.nil?
-        movie.cast = cast(movie)
-        movie.directed_by = directed_by(movie)
-      end
-      movie
-    end
-  
-    def synopsis(movie)
-      @page.search("//div[@id = \"tab_#{movie.id}_sinopse\"]/blockquote").first.content
-    end
-  
-    def cast(movie)
-      @page.search("//div[@id = \"tab_#{movie.id}_ficha\"]/dl/dd").first.content.delete("\r\n")
-    end
-  
-    def directed_by(movie)    
-      directed_by = ""
-      @page.search("//div[@id = \"tab_#{movie.id}_ficha\"]/dl").first.children.each do |child|
-        directed_by = child.content if child.node_name == "dd" && child.previous_element && child.previous_element.content == "Direção:"
-      end
-      directed_by.delete("\r\n")
-    end
-    
-    def movie_theaters(movie)
-      movie_theaters = []
-      @page.search('//div[@class = "programacao_cinema"]/h2/a').each do |movie_theater|
-        movie_theaters << movie_theater.content if movie_theater
-      end
-      movie_theaters
-    end
-  
-    def showtimes(movie)
-      showtimes = []
-      mt = ""
-      obs = ""
-      @page.parser.xpath('//div[@id="content-left"]/div[@class="grid_8"]/div').each do |div_child|
-        # if div_child.node_name == "div"
-        mt = div_child.search('h2').first.text.gsub("\n", "").strip if div_child.attribute("class").value == "programacao_cinema"          
-        if div_child.attribute("class").value == "programacao_horarios"
-          div_child.search('tr').each do |tr_doc|
-            unless tr_doc.content.include?("Sala")
-              obs = div_child.search('td[@colspan="3"]').first.content if div_child.search('td[@colspan="3"]').first
-              sat = ShowtimeAndTheater.new tr_doc
-              showtimes << sat.create_showtime_with(mt,movie,obs)
-            end
-          end
-        end
-        # end
-      end
-      showtimes
-    end
-    
-  end
-
-  class ShowtimeAndTheater
-    
-    def initialize(doc)      
-      @doc = doc
-    end
-    
-    def create_showtime_with(movie_theater,movie,observ)
-      unless @doc.search('td').first.next_element.nil?
-        showtime = Showtime.new(nil,nil,nil,movie,nil,movie_theater)
-        showtime.time = time
-        showtime.theater = theater
-        showtime.kind = kind
-        showtime.obs = observ
-        # p "#{showtime.movie_theater} - #{showtime.theater} - #{showtime.movie.name} - #{showtime.time} - #{showtime.obs}"
-      end
-      showtime
-    end
-    
-    def time
-      @doc.search('td').last.content
-    end
-  
-    def kind
-      @doc.search('td').first.next_element.content
-    end
-  
-    def theater      
-      @doc.search('td').first.content
-    end
-  
-  end
-end

data/lib/omelete/models/movie.rb

@@ -1,13 +0,0 @@
-# -*- encoding : utf-8 -*-
-module Omelete
-  class Movie
-
-    attr_accessor :id, :name, :status, :runtime, :cast, :genre, :directed_by, :city, :state
-    attr_accessor :age_rating, :synopsis, :image, :movie_theaters, :showtimes
-  
-    def initialize(id, name)
-      @id = id
-      @name = name
-    end
-  end
-end

data/lib/omelete/models/showtime.rb

@@ -1,16 +0,0 @@
-# -*- encoding : utf-8 -*-
-module Omelete
-  class Showtime
-
-    attr_accessor :time, :obs, :movie, :theater, :kind, :movie_theater
-
-    def initialize(time, obs, kind, movie, theater, movie_theater)
-      @time = time
-      @obs = obs
-      @kind = kind
-      @movie = movie
-      @theater = theater
-      @movie_theater = movie_theater
-    end
-  end
-end

data/lib/omelete/movie_page.rb

@@ -1,50 +0,0 @@
-# -*- encoding : utf-8 -*-
-#!/usr/bin/env ruby
-$:.unshift File.expand_path('../', __FILE__)
-$:.unshift File.expand_path('../models', __FILE__)
-require 'rubygems'
-require 'open-uri'
-require 'nokogiri'
-require 'mechanize'
-require 'movie'
-require 'detailed_page'
-
-module Omelete
-  class MoviePage
-    URL = "http://omelete.uol.com.br/filmes-em-cartaz"
-    def initialize(state,city)
-      @state = state
-      @city = city
-      @agent = Mechanize.new{|agent| agent.user_agent_alias = 'Mac Safari'}
-      url=URL + "?uf=#{@state}&cidade=#{@city}"
-      @agent.get(URI.escape(url)){ |page| @page = page }
-    end
-
-    def movies      
-      @movies = []
-      @page.search('//div[@class = "programacao_filme"]').each do |movie_doc|
-        link = movie_doc.search('p').first.search('a').first.attr('href')
-        id = id_from(link)
-        return @movies if movie_doc.search('h2').first.search('a').first.nil?
-        name = movie_doc.search('h2').first.search('a').first.content.strip
-        movie = Movie.new(id,name)
-        
-        return @movies if movie_doc.search('h4').first.nil?
-        info = movie_doc.search('h4').first.content.strip.split('-')
-        movie.genre = info[0]
-        movie.runtime = info[1]
-        movie.age_rating = info[2]        
-        movie.image = movie_doc.search('div').first.search('a').first.attr('href').strip unless movie_doc.search('div').first.search('a').first.nil?
-      
-        dp = DetailedPage.new(@page, link)        
-        @movies << dp.movie_poi(movie)
-      end
-      @movies
-    end
-
-    def id_from(link)
-      link.match(/\d+/)[0]
-    end
-
-  end
-end