cartographer 1.0.0 → 1.1.0

data/History.txt

@@ -1,3 +1,9 @@
+=== 1.1.0 / 2011-03-28
+
+* from_xml takes sitemap.xml content and integrates it into the current Cartographer object. 
+* uniq! processes URLs through a uniqueness filter. 
+* Nokogiri is now a runtime dependency.
+
 === 1.0.0 / 2011-03-23
 
 * 1 major enhancement

data/Rakefile

@@ -17,12 +17,14 @@ spec = Hoe.spec 'cartographer' do
   self.rubyforge_name = nil
 
   extra_deps << ['haml', '>= 0']
+  extra_deps << ['nokogiri', '>= 0']
 
   extra_dev_deps << ['minitest', '>= 0']
   extra_dev_deps << ['hoe-roodi', '>= 0']
   extra_dev_deps << ['hoe-reek', '>= 0']
   extra_dev_deps << ['hoe-git', '>= 0']
-  extra_dev_deps << ['nokogiri', '>= 0']
+
+  self.testlib = :minitest
 end
 
 task :install => [:gem] do

data/lib/cartographer.rb

@@ -2,6 +2,7 @@ require 'find'
 require 'haml'
 require 'uri'
 require 'time'
+require 'nokogiri'
 
 #
 # Cartographer is a sitemap manipulation library.
@@ -23,7 +24,7 @@ require 'time'
 # See Cartographer and Cartographer::URL RDoc for more information.
 #
 class Cartographer
-  VERSION = '1.0.0'
+  VERSION = '1.1.0'
 
   ##
   # List of Cartographer::URL objects
@@ -97,12 +98,40 @@ class Cartographer
 
   ##
   #
-  # Transforms the given Cartographer::URL list to the sitemap XML format.
+  # Transforms the given Cartographer::URL list to the sitemap XML format. 
+  #
+  # Be aware this does not alter the root URL structure of these elements, so
+  # you may have mixed sitemap content if you do not rewrite the URLs.
   #
   def to_xml
     Haml::Engine.new(File.read(XML_TEMPLATE)).render(self)
   end
 
+  ##
+  #
+  # Given a string containing sitemap.xml content, parses it and adds each item
+  # to the URL list. 
+  #
+  def from_xml(xml)
+    xml = Nokogiri::XML.parse(xml)
+    xml.root.xpath('xmlns:url').each do |url|
+      add Cartographer::URL.new(
+        :location   => URI.parse(url.xpath('xmlns:loc').first.content),
+        :lastmod    => (Time.parse(url.xpath('xmlns:lastmod').first.content) rescue Time.now),
+        :changefreq => (url.xpath('xmlns:changefreq').first.content rescue nil),
+        :priority   => (url.xpath('xmlns:priority').first.content rescue nil)
+      )
+    end
+  end
+
+  ##
+  #
+  # Uniq's the URL list. This operation is destructive.
+  #
+  def uniq!
+    urls.uniq!
+  end
+
   ##
   # Cartographer::URL is a structure to enforce the sitemap format.
   #
@@ -125,7 +154,6 @@ class Cartographer
     # Priority, stored as a Numeric object, from 0 to 1 with decimal values included.
     attr_accessor :priority
 
-
     ##
     #
     # Constructor. Takes a hash of arguments:
@@ -168,5 +196,21 @@ class Cartographer
     def to_xml
       Haml::Engine.new(File.read(XML_TEMPLATE)).render(self)
     end
+
+    ##
+    # Computes a hash based off the URI.
+    def hash
+      location.hash
+    end
+
+    ##
+    #
+    # Compares two locations and returns true/false.
+    #
+    def eql?(url2)
+      location == url2.location
+    end
+
+    alias == eql?
   end
 end

data/test/test_cartographer.rb

@@ -44,6 +44,19 @@ class TestCartographer < MiniTest::Unit::TestCase
       sm.urls.map(&:location), 
       URI.parse("http://example.org/test_cartographer/")
     )
+
+    require 'fileutils'
+    FileUtils.mkdir_p("test/prune_me")
+
+    sm = Cartographer.new("http://example.org")
+    sm.add_tree(File.expand_path("test")) { |x| x =~ /prune_me/ }
+
+    refute_includes(
+      sm.urls.map(&:location),
+      URI.parse("http://example.org/prune_me/")
+    )
+
+    FileUtils.rm_r("test/prune_me")
   end
 
   def test_to_xml
@@ -71,4 +84,36 @@ class TestCartographer < MiniTest::Unit::TestCase
     assert_equal [ "http://example.org/", "http://example.org/test_cartographer/" ],
       xml.root.xpath('//xmlns:loc', xml.root.namespaces).map(&:content)
   end
+
+  def test_from_xml
+    sm = Cartographer.new("http://example.org")
+    sm.add_tree(File.expand_path("test"))
+
+    assert_includes sm.urls.map { |x| x.location.path }, "/test_cartographer.rb"
+    assert_equal 1, sm.urls.select { |x| x.location.path == "/test_cartographer.rb" }.length
+
+    sm2 = Cartographer.new("http://example.org")
+    sm2.from_xml(sm.to_xml)
+
+    assert_includes sm2.urls.map { |x| x.location.path }, "/test_cartographer.rb"
+    assert_equal 1, sm2.urls.select { |x| x.location.path == "/test_cartographer.rb" }.length
+  end
+
+  def test_uniq!
+    sm = Cartographer.new("http://example.org")
+    sm.add_tree(File.expand_path("test"))
+
+    assert_includes sm.urls.map { |x| x.location.path }, "/test_cartographer.rb"
+    assert_equal 1, sm.urls.select { |x| x.location.path == "/test_cartographer.rb" }.length
+
+    sm.from_xml(sm.to_xml)
+
+    assert_includes sm.urls.map { |x| x.location.path }, "/test_cartographer.rb"
+    assert_equal 2, sm.urls.select { |x| x.location.path == "/test_cartographer.rb" }.length
+
+    sm.uniq!
+
+    assert_includes sm.urls.map { |x| x.location.path }, "/test_cartographer.rb"
+    assert_equal 1, sm.urls.select { |x| x.location.path == "/test_cartographer.rb" }.length
+  end
 end

metadata

@@ -2,7 +2,7 @@
 name: cartographer
 version: !ruby/object:Gem::Version 
   prerelease: 
-  version: 1.0.0
+  version: 1.1.0
 platform: ruby
 authors: 
 - Erik Hollensbe
@@ -11,7 +11,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-03-26 00:00:00 -04:00
+date: 2011-03-28 00:00:00 -04:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -26,7 +26,7 @@ dependencies:
   type: :runtime
   version_requirements: *id001
 - !ruby/object:Gem::Dependency 
-  name: minitest
+  name: nokogiri
   prerelease: false
   requirement: &id002 !ruby/object:Gem::Requirement 
     none: false
@@ -34,10 +34,10 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version 
         version: "0"
-  type: :development
+  type: :runtime
   version_requirements: *id002
 - !ruby/object:Gem::Dependency 
-  name: hoe-roodi
+  name: minitest
   prerelease: false
   requirement: &id003 !ruby/object:Gem::Requirement 
     none: false
@@ -48,7 +48,7 @@ dependencies:
   type: :development
   version_requirements: *id003
 - !ruby/object:Gem::Dependency 
-  name: hoe-reek
+  name: hoe-roodi
   prerelease: false
   requirement: &id004 !ruby/object:Gem::Requirement 
     none: false
@@ -59,7 +59,7 @@ dependencies:
   type: :development
   version_requirements: *id004
 - !ruby/object:Gem::Dependency 
-  name: hoe-git
+  name: hoe-reek
   prerelease: false
   requirement: &id005 !ruby/object:Gem::Requirement 
     none: false
@@ -70,7 +70,7 @@ dependencies:
   type: :development
   version_requirements: *id005
 - !ruby/object:Gem::Dependency 
-  name: nokogiri
+  name: hoe-git
   prerelease: false
   requirement: &id006 !ruby/object:Gem::Requirement 
     none: false