BoilerpipeArticle 0.0.3 → 0.0.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: a2cbf7a8c4c979aecefeb76935ec831f8f7e010b
-  data.tar.gz: 427bb161caf9302c93739c52ffdda87795d08ab1
+  metadata.gz: 86e9801c9b5e6eacf758aa4bbff2ec298313e163
+  data.tar.gz: 58bdb8a6a7624e08f5f28f8bbbc094da2a3fc02a
 SHA512:
-  metadata.gz: d7268bf78155ffba0641a79c2aa83b36953f965587fe404ea96b41fbf7b960b7c70189a22690eedd9993ef2cd0f04d12408718a84e2a3ef5db445636be7f6fe7
-  data.tar.gz: 8366fd8c1ad057b7bb3718109c6e9070e60dd68e06fc4372e9a16f1ed72c9dd7527039e46c841a4f6aafa86d8ae72829f8f1775c1d85d76f1ad9a8ddd40fd24a
+  metadata.gz: f8dcba151737a254fb7f0840d24d187a757575ed87c8b0817ffb24dd3d3e3b28bebacf7ac4ee980bf8bc085317c9f3ae94a651c7c5b77fee43511f1a615077ab
+  data.tar.gz: e3a3ed339a688ae3ed778cd4f4c2e2e426ea070b94e07dcc156d945731b38961540f7a46fa4d421dbdac21aa47022c70df6586f6458bf4f620b95e3628278cd5

data/lib/boilerpipe_article.rb

@@ -5,13 +5,13 @@ class BoilerpipeArticle
     @html = html
   end
 
-  def run
-    html =  Nokogiri::HTML.parse(@html).to_s
+  def getText(html = @html)
+    html =  Nokogiri::HTML.parse(html).to_s
     html.gsub!(/<!-[\s\S]*?->/, '')
     html.gsub!(/\r?\n|\r/, '')
 
     doc = Nokogiri::HTML(html)
-    badHtmlTags = ['li','ol','ul','head','script','style','a','img']
+    badHtmlTags = ['nav','head','script','style','a','img']
     badHtmlTags.each do |tag|
       doc.search(tag).each do |src|
         src.remove
@@ -19,10 +19,7 @@ class BoilerpipeArticle
     end
 
     html = doc.to_html.to_s
-
     selfClosingTags = ['<area','<base','<br','<col','<command','<embed','<hr','<img','<input','<keygen','<link','<meta','<param','<source','<track','<wbr']
-
-
     time = Time.now.to_f
     depth = 1
     i = 0
@@ -74,7 +71,20 @@ class BoilerpipeArticle
     articlesStats.each do |line,stats|
       text = "#{text} #{stats[0]}" if stats[1] == best
     end
-
     return Nokogiri::HTML.parse(text).text
   end
+
+  def getOgMetas(html = @html)
+    metas = Hash.new
+    doc = Nokogiri.parse(html)
+    properties = ['title','type','url','description','image','type','updated_time','locale','url','site_name']
+    properties.each do |prop|
+      if doc.at("meta[property=\"og:#{prop}\"]") != nil
+        metas.store(prop,doc.at("meta[property=\"og:#{prop}\"]")['content'])
+      else
+        metas.store(prop,' ')
+      end
+    end
+    return metas
+  end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: BoilerpipeArticle
 version: !ruby/object:Gem::Version
-  version: 0.0.3
+  version: 0.0.4
 platform: ruby
 authors:
 - David Layer-Reiss
@@ -25,17 +25,17 @@ dependencies:
       - !ruby/object:Gem::Version
         version: 1.6.8
 description: This gem removes the surplus “clutter” (boilerplate, templates) around
-  the main textual content of a web page (pure Ruby implementation). Check GitHub
-  for usage examples.
+  the main textual content of a web page (pure Ruby implementation). BoilerpipeArticle
+  can be also used to parse open graph meta data. Check GitHub for usage examples.
 email: layerreiss@gmail.com
 executables: []
 extensions: []
 extra_rdoc_files: []
 files:
 - lib/boilerpipe_article.rb
-homepage: https://github.com/davidlr99/BoilerpipeArticle
+homepage: http://peppersoft.net/
 licenses:
-- MIT
+- GPL-2.0
 metadata: {}
 post_install_message: 
 rdoc_options: []