raev 0.1.10 → 0.1.11

data/README.md

@@ -3,6 +3,19 @@ Raev
 
 Raev is a Ruby gem for fetching, parsing and normalizing meta data from websites. It was extracted from http://promoterapp.com.
 
+Install
+--------
+
+```shell
+gem install raev
+```
+or add the following line to Gemfile:
+
+```ruby
+gem 'raev'
+```
+and run `bundle install` from your shell.
+
 Usage
 -----
 
@@ -48,6 +61,13 @@ Raev.url("http://www.polygon.com").feed
 # => "http://www.polygon.com/rss/index.xml"
 ```
 
+Fetch headline from url. Removes double spaces.
+
+```ruby
+Raev.url("http://www.polygon.com/e3-2013/2013/6/14/4429126/the-indie-eight-ps4").headline
+# => "The Indie Eight: Polygon talks with the showcase indies launching on PS4"
+```
+
 Normalize author name. Capitalizes name, strips whitespace, ignores email addresses and removes silly nicknames in quotes. Returns nil for empty strings or non-names like *Editor* or *Staff*.
 
 ```

data/VERSION

@@ -1 +1 @@
-0.1.10
+0.1.11

data/lib/raev/url.rb

@@ -81,6 +81,40 @@ module Raev
       feed_url
     end
     
+    def headline
+      page_title = nil
+      
+      node = document.css(".twitter-share-button")
+      
+      if node.first
+        if node.first['data-text']
+          page_title = node.first['data-text']
+        end
+      end
+
+      if page_title.nil?
+        document.css("head meta").each do |meta|
+          if meta['property'] == 'og:title' || meta['property'] == 'twitter:title'
+            page_title = meta['content']
+          end
+        end
+      end
+            
+      if page_title.nil?
+        node = document.css("#article h1, a[rel=\"bookmark\"], h2[itemprop=\"name\"]")
+                
+        if node.first
+          page_title = node.first.content
+        end
+      end
+      
+      unless page_title.nil?
+        page_title.gsub!(/ +/, ' ')
+      end
+      
+      page_title
+    end
+    
     private
     
     def document

data/raev.gemspec

@@ -5,11 +5,11 @@
 
 Gem::Specification.new do |s|
   s.name = %q{raev}
-  s.version = "0.1.10"
+  s.version = "0.1.11"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["Andreas Zecher"]
-  s.date = %q{2013-06-15}
+  s.date = %q{2013-06-16}
   s.description = %q{Fetch, parse and normalize meta data from websites.}
   s.email = %q{andreas@madebypixelate.com}
   s.extra_rdoc_files = [

data/test/test_raev.rb

@@ -56,6 +56,38 @@ class TestRaev < Test::Unit::TestCase
     assert_equal "http://www.edge-online.com/feed/", url.feed
   end
   
+  should "get headline from url" do
+    url = Raev.url("http://www.polygon.com/e3-2013/2013/6/14/4429126/the-indie-eight-ps4")
+    assert_equal "The Indie Eight: Polygon talks with the showcase indies launching on PS4", url.headline
+
+    url = Raev.url("http://penny-arcade.com/report/article/the-future-improved-hands-on-with-the-high-definition-oculus-rift-dev-kit")
+    assert_equal "The future, improved: hands on with the high-definition Oculus Rift dev kit", url.headline
+
+    url = Raev.url("http://kotaku.com/the-world-of-a-link-to-the-past-has-changed-in-the-new-513424187")
+    assert_equal "The World of A Link To The Past Has Changed in the New 3DS Zelda", url.headline
+    
+    url = Raev.url("http://arstechnica.com/gaming/2012/03/journey-a-hauntingly-beautiful-art-house-film-disguised-as-a-game/")
+    assert_equal "Journey: A hauntingly beautiful art house film disguised as a game", url.headline
+
+    url = Raev.url("http://www.creativeapplications.net/games/below-new-from-the-creators-of-sword-sworcery/")
+    assert_equal "Below – New from the creators of Sword & Sworcery", url.headline
+
+    url = Raev.url("http://www.edge-online.com/news/e3-2013-watch-the-first-hohokum-gameplay-footage-here/")
+    assert_equal "E3 2013: watch the first Hohokum gameplay footage here", url.headline
+
+    url = Raev.url("http://www.giantbomb.com/videos/e3-2013-fez-ii-announcement-teaser/2300-7606/")
+    assert_equal "E3 2013: Fez II Announcement Teaser", url.headline
+
+    url = Raev.url("http://indiegames.com/2013/06/indie_fund_backing_for_two_new.html")
+    assert_equal "Indie Fund backing two new titles for Double Fine", url.headline
+
+    url = Raev.url("http://killscreendaily.com/articles/news/cheat-sheet-614/")
+    assert_equal "Pixels on canvas, Spielberg's predictions, and Polytron's glorious tease", url.headline
+
+    url = Raev.url("http://www.rockpapershotgun.com/2013/06/05/i-spy-an-open-beta-for-spy-party/")
+    assert_equal "I Spy An Open Beta For Spy Party", url.headline
+  end
+    
   should "get twitter and rss feed" do
     url = Raev.url("http://www.polygon.com")
     assert_equal "http://www.polygon.com/rss/index.xml", url.feed
@@ -73,5 +105,5 @@ class TestRaev < Test::Unit::TestCase
     assert_equal nil, Raev.normalize_author("Staff")
     assert_equal nil, Raev.normalize_author(" ")
     assert_equal nil, Raev.normalize_author(nil)
-  end
+  end  
 end

metadata

@@ -5,8 +5,8 @@ version: !ruby/object:Gem::Version
   segments: 
   - 0
   - 1
-  - 10
-  version: 0.1.10
+  - 11
+  version: 0.1.11
 platform: ruby
 authors: 
 - Andreas Zecher
@@ -14,7 +14,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2013-06-15 00:00:00 +02:00
+date: 2013-06-16 00:00:00 +02:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency