url_scrubber 0.7.6 → 0.7.7

data/lib/url_scrubber.rb

@@ -16,7 +16,6 @@ module UrlScrubber
     url.sub!('#!/', '')
     url = downcase_domain(url)
     remove_www!(url)
-    Rails.logger.debug "URLSCRUBBER - url1 = #{url}"
     drop_anchor!(special_cases(url))
   end
 
@@ -69,7 +68,7 @@ module UrlScrubber
     when :facebook
       !!url.match(%r{^http://facebook\.com/(profile\.php?id=\d+|[\w_\.-]+)$})
     when :linkedin
-      !!url.match(%r{^http://linkedin\.com/(company/[\w_-]+|profile/view\?id=\d+)$})
+      !!url.match(%r{^http://linkedin\.com/(company/[\w_-]+|profile/view\?id=\d+)$}) || !!url.match(%r{^http://linkedin\.com/(groups\?gid=[0-9]+)$}) || !!url.match(%r{^http://linkedin\.com/(groups/[\w_-]+)$})
     when :google
       !!url.match(%r{^http://plus\.google\.com/(\+[\w_-]+|\d+)$})
     when :slideshare
@@ -91,6 +90,25 @@ module UrlScrubber
       true
     end
   end
+  
+  def self.linkedin_company_url?(url)
+    url = scrub(url)
+    return false unless url
+    return url.include?('http://linkedin.com/company/')
+  end
+  
+  def self.find_linkedin_identity_from_url(url)
+    return nil if url.nil?
+    scrubbed_url = scrub(url) 
+    if scrubbed_url && linkedin_company_url?(scrubbed_url)
+      scrubbed_url.split("/").last
+    elsif scrubbed_url.include?('linkedin.com/groups/')
+      scrubbed_url.split("/").last
+    elsif scrubbed_url.include?('linkedin.com/groups?gid=')
+      id_partition = scrubbed_url.partition('linkedin.com/groups?gid=')
+      drop_url_ampersand!(id_partition[2])
+    end
+  end
 
   private
 
@@ -119,6 +137,11 @@ module UrlScrubber
     url
   end
 
+  def self.drop_url_ampersand!(url)
+    url.sub!(/\&.*$/, '')
+    url
+  end
+
   def self.drop_url_query!(url)
     url.sub!(/\?.*$/, '')
     url
@@ -168,7 +191,16 @@ module UrlScrubber
   def self.linkedin(url)
 
     url.sub!('linkedin.com/companies/', 'linkedin.com/company/')
-    drop_url_query!(url) if !!url.match(%r{com/company/})
+    if !!url.match(%r{com/company/})
+      drop_url_query!(url) 
+    elsif url.include?('linkedin.com/groups/')
+      drop_url_query!(url)
+    elsif url.include?('linkedin.com/groups?gid=')
+      drop_url_ampersand!(url)
+    elsif url.include?('linkedin.com/groups?home=&gid=')
+      id_partition = url.partition('linkedin.com/groups?home=&gid=')
+      url = "http://linkedin.com/groups?gid=" + drop_url_ampersand!(id_partition[2])
+    end
     url
   end
 
@@ -179,13 +211,10 @@ module UrlScrubber
     url.sub!('/photos', '')
     url.sub!('/of', '')
     url.sub!('/albums', '')
-    Rails.logger.debug "URLSCRUBBER - url3 = #{url}"
 
     path_match = url.match(/^http:\/\/plus\.google\.com\/([^\/]+)/)
-    Rails.logger.debug "URLSCRUBBER - path_match = #{path_match}"
     return url unless path_match
 
-    Rails.logger.debug "URLSCRUBBER - no match url = http://plus.google.com/#{path_match[1]}"
     "http://plus.google.com/#{path_match[1]}"
   end
 

data/lib/url_scrubber/version.rb

@@ -1,3 +1,3 @@
 module UrlScrubber
-  VERSION = "0.7.6"
+  VERSION = "0.7.7"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: url_scrubber
 version: !ruby/object:Gem::Version
-  version: 0.7.6
+  version: 0.7.7
   prerelease: 
 platform: ruby
 authors:
@@ -11,7 +11,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-09-11 00:00:00.000000000 Z
+date: 2013-09-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec