mechanize_content 0.3.2 → 0.4.0

data/spec/mechanize_content_spec.rb

@@ -3,7 +3,7 @@ require File.expand_path(File.dirname(__FILE__) + '/spec_helper')
 
 describe "MechanizeContent" do
   use_vcr_cassette :record => :new_episodes
-  
+
   describe "#best_title" do
     context "given a page with a title" do
       it "will be that title" do
@@ -11,7 +11,7 @@ describe "MechanizeContent" do
         mc.best_title.should eql("Techmeme")
       end
     end
-    
+
     context "given the page has an incorrect class" do
       it "will be the url" do
         p = MechanizeContent::Page.new("http://techmeme.com/")
@@ -21,7 +21,7 @@ describe "MechanizeContent" do
         mc.best_title.should eql("http://techmeme.com/")
       end
     end
-    
+
     context "given page has no title" do
       it "will be the url" do
         p = mock("page", :title => nil, :url => "http://techmeme.com/")
@@ -30,16 +30,16 @@ describe "MechanizeContent" do
         mc.best_title.should eql("http://techmeme.com/")
       end
     end
-    
+
     context "given multiple urls are passed in" do
       it "will find the best title" do
         mc = MechanizeContent::Parser.new("http://www.facebook.com/RockBand", "http://www.vg247.com/2010/03/09/rock-band-3-out-this-holiday-will-revolutionize-genre/")
         mc.best_title.should eql("Rock Band | Facebook")
       end
     end
-    
+
   end
-  
+
   describe "#best_text" do
     context "page with no text" do
       it "will be nil" do
@@ -47,42 +47,42 @@ describe "MechanizeContent" do
         mc.best_text.should eql(nil)
       end
     end
-    
+
     context "given a page with text" do
       it "will be the text" do
         mc2 = MechanizeContent::Parser.new("http://www.gamesetwatch.com/2010/03/gdc_2010_rounds_off_indie_cove.php")
         mc2.best_text.should eql("Game Developers Conference organizers have confirmed the final set of independent game-specific content, including Ron Carmel on the just-debuted Indie Fund, the Gamma IV party/showcase, and the EGW-replacing Nuovo Sessions game showcase.The newly confirmed details round off a multitude of independent game-specific content at the March 9th-13th event, held at the Moscone Center in San Francisco, including the 12th Annual Independent Games Festival -- featuring over 30 top indie games playable on the GDC Expo floor from Thursday 11th to Saturday 13th, as well as the major IGF Awards on Thursday 11th at 6.30pm.In addition, the 4th Independent Games Summit on Tuesday 9th and Wednesday 10th has added and clarified a number of sessions, with 2D Boy's Ron Carmel kicking off the event with 'Indies and Publishers: Fixing a System That Never Worked', now confirmed to discuss the new Indie Fund organization.Another major new panel, 'Tripping The Art Fantastic', features Spelunky creator Derek Yu, Braid artist David Hellman and Super Meat Boy co-creator Edmund McMillen discussing \"how each one of these figures influences the state of game art, from hand painted epics to short form experimental Flash games.\"")
       end
     end
-    
+
     context "given multiple urls are passed in" do
       it "will find the best text" do
         mc = MechanizeContent::Parser.new("http://www.facebook.com/RockBand", "http://www.vg247.com/2010/03/09/rock-band-3-out-this-holiday-will-revolutionize-genre/")
         mc.best_text.should include("Harmonix just confirmed that Rock Band 3 will release this holiday season.Said the firm on Rock Band’s Facebook page:“Harmonix is developing Rock Band 3 for worldwide release this holiday season! The game, which will be published by MTV Games and distributed by Electronic Arts, will innovate and revolutionize the music genre once again, just as Harmonix did with the original Rock Band, Rock Band 2 and The Beatles: Rock Band. Stay tuned for more details!”There’s no more detail right now, but keep watching for updates from GDC.")
       end
     end
-    
+
     context "given the first link has no text" do
       it "will pull text from the next link" do
         mc = MechanizeContent::Parser.new("http://www.gog.com/en/gamecard/another_world_15th_anniversary_edition", "http://www.destructoid.com/-nuff-said-good-old-games-gets-another-world-168150.phtml", "http://www.joystiq.com/2010/03/18/another-world-15th-anniversary-edition-now-on-gog-com/")
         mc.best_text.should eql("Éric Chahi masterpiece is looking better than ever\r\r\rPosted on 2010-03-18 09:46:36 by \r\rEclipse:\r\rThe first time I've ever played Another World was on my 486, I remember I booted up the game and wow, I was so shocked at how "cinematic" it felt, it looked and played like an animated sci-fi movie, fluid animations, always unique enemies, action puzzles... I remember it was the first game I've ever played that needed you to actually run away from the beast atread more\r the end of the first level instead of just shooting the hell out of your enemies. The world around Lester felt more alive than every game I've ever played before and Lester himself was a true human, not some sort of pumped superdude with a shotgun.Smooth animations like I was never able to see from Prince of Persia times, glorious and fascinating graphics, rapid action sequences and an hard yet satisfying gameplay.When I played this remade version for the first time I had a similar feeling, It wasn't something new, I've played this game dozen of times during years, but the sharp vector graphics i was able too see as child on my old crt monitor looked pixelly on my new 23" LCD, this restyle reminds me more than the original version how I used to look at this game, I know it may sound strange but it's true.And yay, no more passwords to remember :)This game made my childhood and it's one of the best games EVER.\r\r\r\rWas this helpful?\r\r\r(105 of 119 people found this helpful)")
       end
     end
-    
+
     context "given a page with strange class names" do
       it "should still be able to find the text" do
         mc = MechanizeContent::Parser.new("http://www.destructoid.com/-nuff-said-good-old-games-gets-another-world-168150.phtml")
         mc.best_text.should eql("Another World -- or Out of this World, as many of you will know it by -- is now on DRM-free digital distribution service Good Old Games. It can be had for $9.99. Need I say more?\rI love the game, even though I have never made it more than oh, five minutes in. It's more or less universally loved by the Destructoid staff. Not long after we got an email detailing the good news, the thread soon reached fifteen or so replies full of praise for the game.\rOther, less exciting recent releases include: Empire Earth II Gold, Gabriel Knight 3, and Aquanox. Not to completely s**t on these games, but this is Another World we're talking about here.")
       end
     end
-    
+
     context "given a flash marketing micro site" do
       it "will be nil" do
         mc = MechanizeContent::Parser.new("http://www.sonypictures.co.uk/movies/cloudywithachanceofmeatballs/")
         mc.best_text.should eql(nil)
       end
     end
-    
+
     context "given a first page with text that is too short" do
       it "will pull text from the next suitable page" do
         mc = MechanizeContent::Parser.new("http://www.examiner.com/fps-in-national/playstation-network-up-japan-video-announcement-video",
@@ -92,7 +92,7 @@ describe "MechanizeContent" do
       end
     end
   end
-  
+
   describe "#best_image" do
     context "given a page with a suitable image" do
       it "will be the image url" do
@@ -100,14 +100,14 @@ describe "MechanizeContent" do
         mc.best_image.should eql("http://media.rockstargames.com/rockstargames/img/global/news/upload/lanoire_streetcrimes_cameraobscura.jpg")
       end
     end
-    
+
     context "given a page with no suitable images" do
       it "will be nil" do
         mc3 = MechanizeContent::Parser.new("http://www.gog.com/en/gamecard/another_world_15th_anniversary_edition")
         mc3.best_image.should eql(nil)
       end
     end
-    
+
     context "given multiple urls are passed in" do
       it "will find the best image" do
         mc = MechanizeContent::Parser.new("http://www.facebook.com/RockBand", "http://www.vg247.com/2010/03/09/rock-band-3-out-this-holiday-will-revolutionize-genre/")
@@ -117,12 +117,12 @@ describe "MechanizeContent" do
 
     context "given a url with strange utf8 encoding" do
       it "will find the best image" do
-        mc = MechanizeContent::Parser.new(%w(http://www.computerandvideogames.com/310643/news/lg-dedicated-gaming-handhelds-are-over/ 
+        mc = MechanizeContent::Parser.new(%w(http://www.computerandvideogames.com/310643/news/lg-dedicated-gaming-handhelds-are-over/
                                                    http://www.destructoid.com/lg-age-of-dedicated-gaming-handhelds-is-over--205485.phtml))
         mc.best_image.should eql("http://cdn.medialib.computerandvideogames.com/screens/screenshot_259495_thumb_wide300.jpg")
       end
     end
-    
+
     context "given the first pick does not support hot linking" do
       it "will fall back to the next image" do
         mc = MechanizeContent::Parser.new("http://www.andriasang.com/e/blog/2011/05/27/kojima_questions/",
@@ -134,7 +134,30 @@ describe "MechanizeContent" do
       end
     end
   end
-  
+
+  describe "#best_image_iphone" do
+    context "given a page with an iphone fav icon" do
+      it "will be the image url" do
+        mc = MechanizeContent::Parser.new("http://www.officialplaystationmagazine.co.uk/2011/11/16/historic-issue-65-cover-features-hideo-kojima/")
+        mc.best_image_iphone.should eql("http://static.officialplaystationmagazine.co.uk/wp-content/themes/opm/style/images/icon/apple-touch-icon.png")
+      end
+    end
+
+    context "given a page with an iphone fav icon" do
+      it "will be the image url" do
+        mc = MechanizeContent::Parser.new("http://www.eurogamer.net/articles/2011-11-17-oddword-strangers-wrath-hd-ps3-complete")
+        mc.best_image_iphone.should eql("http://www.eurogamer.net/img/EurogamerPage/apple-touch-icon.png")
+      end
+    end
+
+    context "given a page with no iphone icon" do
+      it "will return the normal image url" do
+        mc = MechanizeContent::Parser.new("http://www.rockstargames.com/newswire/article/16021/street-crimes-of-la-noire-new-screenshots-info.html")
+        mc.best_image_iphone.should eql("http://media.rockstargames.com/rockstargames/img/global/news/upload/lanoire_streetcrimes_cameraobscura.jpg")
+      end
+    end
+  end
+
   describe ".initialize" do
     context "given an array of urls are passed in" do
       it "will still parse content" do
@@ -144,7 +167,7 @@ describe "MechanizeContent" do
         mc.best_image.should eql("http://media.rockstargames.com/rockstargames/img/global/news/upload/lanoire_streetcrimes_cameraobscura.jpg")
       end
     end
-    
+
     context "given urls are passed in as seperate arguements" do
       it "will parse arguements" do
         mc = MechanizeContent::Parser.new("http://www.rockstargames.com/newswire/article/16021/street-crimes-of-la-noire-new-screenshots-info.html", "http://www.vg247.com/2010/03/18/gta-iv-episodes-from-liberty-city-sees-slight-delay-on-pc-and-ps3/")
@@ -153,5 +176,5 @@ describe "MechanizeContent" do
         mc.best_image.should eql("http://media.rockstargames.com/rockstargames/img/global/news/upload/lanoire_streetcrimes_cameraobscura.jpg")
       end
     end
-  end  
+  end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: mechanize_content
 version: !ruby/object:Gem::Version
-  version: 0.3.2
+  version: 0.4.0
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2011-10-09 00:00:00.000000000Z
+date: 2011-11-17 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: mechanize
-  requirement: &70365206410400 !ruby/object:Gem::Requirement
+  requirement: &70212194376380 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -21,10 +21,10 @@ dependencies:
         version: 2.0.1
   type: :runtime
   prerelease: false
-  version_requirements: *70365206410400
+  version_requirements: *70212194376380
 - !ruby/object:Gem::Dependency
   name: imagesize
-  requirement: &70365206409620 !ruby/object:Gem::Requirement
+  requirement: &70212194375580 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -32,10 +32,10 @@ dependencies:
         version: 0.1.1
   type: :runtime
   prerelease: false
-  version_requirements: *70365206409620
+  version_requirements: *70212194375580
 - !ruby/object:Gem::Dependency
   name: rspec
-  requirement: &70365206408940 !ruby/object:Gem::Requirement
+  requirement: &70212194236180 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -43,21 +43,21 @@ dependencies:
         version: 2.6.0
   type: :development
   prerelease: false
-  version_requirements: *70365206408940
+  version_requirements: *70212194236180
 - !ruby/object:Gem::Dependency
   name: vcr
-  requirement: &70365206408160 !ruby/object:Gem::Requirement
+  requirement: &70212194235100 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 1.10.0
+        version: 1.11.3
   type: :development
   prerelease: false
-  version_requirements: *70365206408160
+  version_requirements: *70212194235100
 - !ruby/object:Gem::Dependency
   name: fakeweb
-  requirement: &70365206407500 !ruby/object:Gem::Requirement
+  requirement: &70212194233980 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -65,7 +65,7 @@ dependencies:
         version: 1.3.0
   type: :development
   prerelease: false
-  version_requirements: *70365206407500
+  version_requirements: *70212194233980
 description: pass in a url or urls and mechanize-content will select the best block
   of text, image and title by analysing the page content
 email:
@@ -105,21 +105,15 @@ required_ruby_version: !ruby/object:Gem::Requirement
   - - ! '>='
     - !ruby/object:Gem::Version
       version: '0'
-      segments:
-      - 0
-      hash: -160408397260513251
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:
   - - ! '>='
     - !ruby/object:Gem::Version
       version: '0'
-      segments:
-      - 0
-      hash: -160408397260513251
 requirements: []
 rubyforge_project: mechanize_content
-rubygems_version: 1.8.7
+rubygems_version: 1.8.11
 signing_key: 
 specification_version: 3
 summary: scrape the best content from a page