dmoz_sax_doc 0.0.4 → 0.0.5

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    dmoz_sax_doc (0.0.3)
+    dmoz_sax_doc (0.0.4)
       json (~> 1.7)
       nokogiri (~> 1.5)
 

data/lib/dmoz_sax/alias.rb

@@ -4,13 +4,5 @@ module DmozSax
     def initialize path_str
       @path = DmozSax::Path.new path_str
     end
-
-    def title
-      @path.name
-    end
-
-    def title= str
-      @path.name = str
-    end
   end
 end

data/lib/dmoz_sax/path.rb

@@ -9,30 +9,34 @@ module DmozSax
       resource = str.gsub('_', ' ').split(':')
 
       @name = resource.first if resource.length == 2
+      @level = level.to_i
 
       unless resource.empty?
-        @path = resource.last.split('/').reject {|a| a =~ /^[A-Z0-9]$/}
-        @path.shift if 'Top' == @path.first
-      else 
-        @path = []
+        super(resource.last.split('/'))
+      else
+        super([])
       end
-      @level = level.to_i
-      super(@path.freeze)
     end
 
-    def to_a
-      @path.dup
+    def to_s
+      DmozSax::Path.path_str(DmozSax::Path.normalize(self))
     end
 
-    def to_s
-      @path.join('/')
+    def parent_to_s
+      DmozSax::Path.path_str(DmozSax::Path.normalize(self[0...-1]))
+    end
+
+    def self.path_str arr
+      arr.empty? ? '/' : "/#{ arr.join('/') }"
     end
 
-    def parent_path
-      if @path.length == 0
-        nil
+    def self.normalize arr
+      if arr.nil? or arr.length == 0
+        []
       else
-        @path[0...-1].join('/')
+        dup = arr.dup.reject {|a| a =~ /^[A-Z0-9]$/}
+        dup.shift if arr.first == 'Top'
+        dup
       end
     end
   end

data/lib/dmoz_sax/version.rb

@@ -1,3 +1,3 @@
 module DmozSax
-  VERSION = "0.0.4"
+  VERSION = "0.0.5"
 end

data/spec/alias_spec.rb

@@ -1,6 +1,9 @@
 require 'spec_helper'
 
 describe DmozSax::Alias do
-
-
+  it "has a name and path parsed from a string" do
+    a = DmozSax::Alias.new("Publishing:Top/Arts/Business/Publishing")
+    a.path.name.should == 'Publishing'
+    a.path.to_s.should == '/Arts/Business/Publishing'
+  end
 end

data/spec/content_document_spec.rb

@@ -1,11 +1,6 @@
 require 'spec_helper'
 
 describe DmozSax::ContentDocument do
-  #it 'can parse a real content.rdf.u8 document' do
-  #  parser = Nokogiri::XML::SAX::Parser.new(DmozSax::ContentDocument.new)
-  #  parser.parse(File.open('/opt/data/DMOZ/content.rdf.u8'))
-  #end
-  
   it 'can parse a sample content.rdf.u8 document' do
 
     topics = []

data/spec/path_spec.rb

@@ -10,14 +10,14 @@ describe DmozSax::Path do
 
     (('A'..'Z').to_a + (0..9).to_a).each do |char|
       path = DmozSax::Path.new("Top/This/Topic/#{ char }/Path")
-      path.to_a.should == ['This','Topic','Path']
+      path.to_s.should == "/This/Topic/Path"
     end
   end
 
   it "may optionally be preceeded by a name or identifier" do
     path = DmozSax::Path.new("Sample_Directory:Top/This/Topic/Path")
     path.name.should == 'Sample Directory'
-    path.to_a.should == ['This','Topic','Path']
+    path.to_s.should == '/This/Topic/Path'
   end
 
   context "as an immutable array" do
@@ -41,20 +41,20 @@ describe DmozSax::Path do
   context "getting parent path" do
     it "returns parent path" do
       path = DmozSax::Path.new 'This/Topic/Path'
-      path.parent_path.should == 'This/Topic'
+      path.parent_to_s.should == '/This/Topic'
     end
 
     it "has not parent path if top level path" do
       path = DmozSax::Path.new ''
-      path.parent_path.should be_nil
+      path.parent_to_s.should == '/'
 
       path = DmozSax::Path.new 'Top'
-      path.parent_path.should be_nil
+      path.parent_to_s.should == '/'
     end
 
     it "follows the same rules removing index categories" do
       path = DmozSax::Path.new 'Top/Topic/A/Path'
-      path.parent_path.should == 'Topic'
+      path.parent_to_s.should == '/Topic'
     end
   end
 end

data/spec/structure_document_spec.rb

@@ -20,12 +20,14 @@ describe DmozSax::StructureDocument do
     parser.parse(File.open('spec/samples/structure_sample.rdf.u8'))
     topics.count.should == 2
     topics[1].title.should == 'Arts'
-    topics[1].path.should == ['Arts']
+    topics[1].path.should == ['Top', 'Arts']
     topics[1].description.should include 'aesthetic objects'
     topics[1].cid.should == 381773
 
     aliases.count.should == 2
-    aliases[0].title.should == 'Publishers'
-    aliases[0].path.should == ['Business','Publishing and Printing','Publishing','Books','Arts']
+    aliases[0].path.name.should == 'Publishers'
+    aliases[0].path.should == ['Top','Business','Publishing and Printing','Publishing','Books','Arts']
+    aliases[0].path.to_s.should == '/Business/Publishing and Printing/Publishing/Books/Arts'
+    aliases[0].path.parent_to_s.should == '/Business/Publishing and Printing/Publishing/Books'
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: dmoz_sax_doc
 version: !ruby/object:Gem::Version
-  version: 0.0.4
+  version: 0.0.5
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-03-04 00:00:00.000000000 Z
+date: 2013-03-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri