valda-video_scraper 1.0.1 → 1.0.2

data/README

@@ -11,18 +11,19 @@ Web scraping library for video sharing sites.
 
 Supported sites
 
-* YouTube
-* NICO NICO DOUGA
 * AmebaVision
 * Dailymotion
-* Veoh
-* YourFileHost
-* RedTube
-* Pornhub
-* Ura Agesage
 * MoroTube
+* NICO NICO DOUGA
+* Pornhub
 * Pornotube
+* RedTube
+* Tube8
+* Ura Agesage
+* Veoh
 * YouPorn
+* YouTube
+* YourFileHost
 
 == SYNOPSIS:
 

data/Rakefile

@@ -60,6 +60,7 @@ spec = Gem::Specification.new do |s|
        s.test_files        = Dir["test/*_test.rb"]
 
        s.add_dependency('mechanize', '>=0.8.4')
+       s.add_dependency('hpricot', '>=0.6.164')
        #s.required_ruby_version = '>= 1.8.2'
 
        s.files = %w(README ChangeLog Rakefile) +

data/lib/www/video_scraper.rb

@@ -2,24 +2,23 @@
 
 require 'rubygems'
 require 'open-uri'
+require 'hpricot'
 require 'mechanize'
 require 'kconv'
 require 'json'
 require 'uri'
 begin
-  require 'cgialt'
+  require 'cgialt' unless defined? CGI
 rescue LoadError
   require 'cgi'
 end
 
 module WWW
   module VideoScraper
-    class TryAgainLater < RuntimeError; end
-    class FileNotFound < RuntimeError; end
-    VERSION = '1.0.1'
+    VERSION = '1.0.2'
 
     MODULES_NAME = %w(age_sage ameba_vision dailymotion moro_tube
-                      nico_video pornhub pornotube red_tube veoh
+                      nico_video pornhub pornotube red_tube tube8 veoh
                       you_porn you_tube your_file_host)
 
     @@modules = MODULES_NAME.map do |name|
@@ -34,6 +33,10 @@ module WWW
     }
 
     class << self
+      def modules
+        @@nodules
+      end
+
       def options
         @@options
       end

data/lib/www/video_scraper/base.rb

@@ -2,6 +2,9 @@
 
 module WWW
   module VideoScraper
+    class TryAgainLater < RuntimeError; end
+    class FileNotFound < RuntimeError; end
+
     class Base
       attr_reader :page_url, :video_url, :thumb_url, :embed_tag, :title
 

data/lib/www/video_scraper/tube8.rb

@@ -0,0 +1,28 @@
+# -*- mode:ruby; coding:utf-8 -*-
+
+require File.expand_path(File.dirname(__FILE__) + '/base')
+
+module WWW
+  module VideoScraper
+    class Tube8 < Base
+      url_regex %r!\Ahttp://www\.tube8\.com/.*/(\d+)(?:/|$)!
+
+      def initialize(url, opt = nil)
+        super
+        do_query
+      end
+
+      private
+      def do_query
+        html = http_get(@page_url)
+        doc = Hpricot(html.toutf8)
+        raise FileNotFound unless flashvars = doc.at('//object //param[@name="FlashVars"]')
+        flashvars = CGI.parse(flashvars.attributes['value'])
+        @video_url = flashvars['videoUrl'][0]
+        uri = URI.parse(@page_url)
+        @thumb_url = URI.join("#{uri.scheme}://#{uri.host}", flashvars['imageUrl'][0]).to_s
+        @title = doc.at('//h1[@class="text"]').inner_html rescue nil
+      end
+    end
+  end
+end

data/test/www/video_scraper/test_tube8.rb

@@ -0,0 +1,24 @@
+# -*- mode:ruby; coding:utf-8 -*-
+
+require File.dirname(__FILE__) + '/../../test_helper'
+
+class TestTube8 < Test::Unit::TestCase
+  def setup
+    @cache_root = '/tmp/test_video_scraper_cache'
+    WWW::VideoScraper.configure do |conf|
+      conf[:cache] = FileCache.new('TestVideoScraper', @cache_root, 60*60*24)
+    end
+  end
+
+  def teardown
+    # FileUtils.remove_entry_secure(@cache_root, true)
+  end
+
+  def test_scrape
+    vs = WWW::VideoScraper.scrape('http://www.tube8.com/anal/alexis-amore-pov/56983/')
+    assert_equal 'http://www.tube8.com/anal/alexis-amore-pov/56983/', vs.page_url
+    assert_match %r|http://medianl\d+\.tube8\.com/flv/[[:alnum:]]{32}/\d{8}/\d{4}/\d{2}/[[:alnum:]]+/[[:alnum:]]+\.flv|, vs.video_url
+    assert_equal 'http://www.tube8.com/vs/83/56983.jpg', vs.thumb_url
+    assert_equal 'Alexis Amore POV', vs.title
+  end
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: valda-video_scraper
 version: !ruby/object:Gem::Version 
-  version: 1.0.1
+  version: 1.0.2
 platform: ruby
 authors: 
 - YAMAGUCHI Seiji
@@ -9,7 +9,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2009-01-17 00:00:00 -08:00
+date: 2009-01-18 00:00:00 -08:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -21,6 +21,15 @@ dependencies:
       - !ruby/object:Gem::Version 
         version: 0.8.4
     version: 
+- !ruby/object:Gem::Dependency 
+  name: hpricot
+  version_requirement: 
+  version_requirements: !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: 0.6.164
+    version: 
 description: Web scraping library for video sharing sites.
 email: valda@underscore.jp
 executables: []
@@ -41,6 +50,7 @@ files:
 - test/www/video_scraper/test_dailymotion.rb
 - test/www/video_scraper/test_age_sage.rb
 - test/www/video_scraper/test_pornotube.rb
+- test/www/video_scraper/test_tube8.rb
 - test/www/video_scraper/test_your_file_host.rb
 - test/www/video_scraper/test_moro_tube.rb
 - test/www/video_scraper/test_veoh.rb
@@ -65,6 +75,7 @@ files:
 - lib/www/video_scraper/red_tube.rb
 - lib/www/video_scraper/base.rb
 - lib/www/video_scraper/your_file_host.rb
+- lib/www/video_scraper/tube8.rb
 - lib/www/video_scraper.rb
 has_rdoc: true
 homepage: http://github.com/valda/video_scraper