xml-sitemap 1.1.3 → 1.2.0

data/README.md

@@ -114,6 +114,11 @@ index.render
 index.render_to('/path/to/file.xml')
 ```
 
+## Authors & Contributors
+
+- [Dan Sosedoff](https://github.com/sosedoff) (author)
+- [Dan Healy](https://github.com/danhealy)
+
 ## License
 
 Copyright © 2010-2011 Dan Sosedoff.

data/lib/xml-sitemap.rb

@@ -2,8 +2,14 @@ require 'time'
 require 'date'
 require 'zlib'
 require 'builder'
+begin
+  require 'nokogiri'
+rescue LoadError
+end
 
 require 'xml-sitemap/options'
+require 'xml-sitemap/render_engine'
+require 'xml-sitemap/item'
 require 'xml-sitemap/map'
 require 'xml-sitemap/index'
 

data/lib/xml-sitemap/item.rb

@@ -0,0 +1,42 @@
+module XmlSitemap
+  class Item
+    DEFAULT_PRIORITY = 0.5
+
+    # ISO8601 regex from here: http://www.pelagodesign.com/blog/2009/05/20/iso-8601-date-validation-that-doesnt-suck/
+    ISO8601_REGEX = /^([\+-]?\d{4}(?!\d{2}\b))((-?)((0[1-9]|1[0-2])(\3([12]\d|0[1-9]|3[01]))?|W([0-4]\d|5[0-2])(-?[1-7])?|(00[1-9]|0[1-9]\d|[12]\d{2}|3([0-5]\d|6[1-6])))([T\s]((([01]\d|2[0-3])((:?)[0-5]\d)?|24\:?00)([\.,]\d+(?!:))?)?(\17[0-5]\d([\.,]\d+)?)?([zZ]|([\+-])([01]\d|2[0-3]):?([0-5]\d)?)?)?)?$/
+
+    attr_reader :target, :updated, :priority, :changefreq, :validate_time
+
+    def initialize(target, opts={})
+      @target         = target.to_s.strip
+      @updated        = opts[:updated]  || Time.now
+      @priority       = opts[:priority] || DEFAULT_PRIORITY
+      @changefreq     = opts[:period]   || :weekly
+      @validate_time  = (opts[:validate_time] != false)
+      
+      unless @updated.kind_of?(Time) || @updated.kind_of?(Date) || @updated.kind_of?(String)
+        raise ArgumentError, "Time, Date, or ISO8601 String required for :updated!"
+      end
+
+      if @validate_time && @updated.kind_of?(String) && !(@updated =~ ISO8601_REGEX)
+        raise ArgumentError, "String provided to :updated did not match ISO8601 standard!"
+      end
+      
+      unless XmlSitemap::PERIODS.include?(@changefreq)
+        raise ArgumentError, "Invalid :period value '#{@changefreq}'"
+      end
+
+      @updated = @updated.to_time if @updated.kind_of?(Date)
+    end
+
+    # Returns the timestamp value for rendere
+    #
+    def lastmod_value
+      if @updated.kind_of?(Time)
+        @updated.utc.iso8601
+      else
+        @updated.to_s
+      end
+    end
+  end
+end

data/lib/xml-sitemap/map.rb

@@ -1,33 +1,25 @@
-module XmlSitemap
-  class Item
-    DEFAULT_PRIORITY = 0.5
-    
-    attr_reader :target, :updated, :priority, :changefreq
-    
-    def initialize(target, opts={})
-      @target     = target.to_s.strip
-      @updated    = opts[:updated]  || Time.now
-      @priority   = opts[:priority] || DEFAULT_PRIORITY
-      @changefreq = opts[:period]   || :weekly
-      
-      # allow only date or time object
-      unless @updated.kind_of?(Time) || @updated.kind_of?(Date)
-        raise ArgumentError, "Time or Date required for :updated!"
-      end
-      
-      # use full time and date only!
-      @updated = @updated.to_time if @updated.kind_of?(Date)
-    end
-  end
-
+module XmlSitemap  
   class Map
+    include XmlSitemap::RenderEngine
+    
     attr_reader   :domain, :items
     attr_reader   :buffer
     attr_reader   :created_at
     attr_reader   :root
     attr_reader   :group
     
-    # Creates new Map class for specified domain
+    # Initializa a new Map instance
+    #
+    # domain - Primary domain for the map (required)
+    # opts   - Map options
+    #
+    # opts[:home]   - Automatic homepage creation. To disable set to false. (default: true)
+    # opts[:secure] - Force HTTPS for all items. (default: false)
+    # opts[:time]   - Set default lastmod timestamp for items (default: current time)
+    # opts[:group]  - Group name for sitemap index. (default: sitemap)
+    # opts[:root]   - Force all links to fall under the main domain.
+    #                 You can add full urls (not paths) if set to false. (default: true)
+    #
     def initialize(domain, opts={})
       @domain     = domain.to_s.strip
       raise ArgumentError, 'Domain required!' if @domain.empty?
@@ -44,15 +36,18 @@ module XmlSitemap
       yield self if block_given?
     end
     
-    # Yields Map class for easier access
-    def generate
-      raise ArgumentError, 'Block required' unless block_given?
-      yield self
-    end
-    
-    # Add new item to sitemap list
+    # Adds a new item to the map
+    #
+    # target - Path or url
+    # opts   - Item options
+    #
+    # opts[:updated]       - Lastmod property of the item
+    # opts[:period]        - Update frequency. (default - :weekly)
+    # opts[:priority]      - Item priority. (default: 0.5)
+    # opts[:validate_time] - Skip time validation if want to insert raw strings.
+    # 
     def add(target, opts={})
-      raise RuntimeError, 'Only less than 50k records allowed!' if @items.size >= 50000
+      raise RuntimeError, 'Only up to 50k records allowed!' if @items.size > 50000
       raise ArgumentError, 'Target required!' if target.nil?
       raise ArgumentError, 'Target is empty!' if target.to_s.strip.empty?
       
@@ -69,42 +64,47 @@ module XmlSitemap
     end
     
     # Get map items count
+    #
     def size
       @items.size
     end
     
     # Returns true if sitemap does not have any items
+    #
     def empty?
       @items.empty?
     end
     
     # Generate full url for path
+    #
     def url(path='')
       "#{@secure ? 'https' : 'http'}://#{@domain}#{path}"
     end
     
     # Get full url for index
+    #
     def index_url(offset)
       "http://#{@domain}/#{@group}-#{offset}.xml"
     end
     
     # Render XML
-    def render
-      xml = Builder::XmlMarkup.new(:indent => 2)
-      xml.instruct!(:xml, :version => '1.0', :encoding => 'UTF-8')
-      xml.urlset(XmlSitemap::MAP_SCHEMA_OPTIONS) { |s|
-        @items.each do |item|
-          s.url do |u|
-            u.loc        item.target
-            u.lastmod    item.updated.utc.iso8601
-            u.changefreq item.changefreq.to_s
-            u.priority   item.priority.to_s
-          end
-        end
-      }.to_s
+    #
+    # method - Pick a render engine (:builder, :nokogiri, :string).
+    #          Default is :string
+    #
+    def render(method = :string)
+      case method
+        when :nokogiri
+          render_nokogiri
+        when :builder
+          render_builder
+        else
+          render_string
+      end
     end
     
     # Render XML sitemap into the file
+    #
     def render_to(path, options={})
       overwrite = options[:overwrite] == true || true
       compress  = options[:gzip]      == true || false
@@ -130,6 +130,7 @@ module XmlSitemap
     protected
   
     # Process target path or url
+    #
     def process_target(str)
       if @root == true
         url(str =~ /^\// ? str : "/#{str}")

data/lib/xml-sitemap/render_engine.rb

@@ -0,0 +1,70 @@
+module XmlSitemap
+  module RenderEngine
+    private
+    
+    # Render with Nokogiri gem
+    #
+    def render_nokogiri
+      unless defined? Nokogiri
+        raise ArgumentError, "Nokogiri not found!"
+      end
+      builder = Nokogiri::XML::Builder.new(:encoding => "UTF-8") do |xml|
+        xml.urlset(XmlSitemap::MAP_SCHEMA_OPTIONS) { |s|
+          @items.each do |item|
+            s.url do |u|
+              u.loc        item.target
+              u.lastmod    item.lastmod_value
+              u.changefreq item.changefreq.to_s
+              u.priority   item.priority.to_s
+            end
+          end
+        }
+      end
+      builder.to_xml
+    end
+    
+    # Render with Builder gem
+    #
+    def render_bulder
+      xml = Builder::XmlMarkup.new(:indent => 2)
+      xml.instruct!(:xml, :version => '1.0', :encoding => 'UTF-8')
+      xml.urlset(XmlSitemap::MAP_SCHEMA_OPTIONS) { |s|
+        @items.each do |item|
+          s.url do |u|
+            u.loc        item.target
+            u.lastmod    item.lastmod_value
+            u.changefreq item.changefreq.to_s
+            u.priority   item.priority.to_s
+          end
+        end
+      }.to_s
+    end
+    
+    # Render with plain strings
+    #
+    def render_string
+      result = '<?xml version="1.0" encoding="UTF-8"?>' + "\n<urlset"
+        
+      XmlSitemap::MAP_SCHEMA_OPTIONS.each do |key, val|
+        result += ' ' + key + '="' + val + '"'
+      end
+      
+      result += ">\n"
+      
+      item_results = []
+      @items.each do |item|
+        item_string  = "  <url>\n"
+        item_string += "    <loc>#{CGI::escapeHTML(item.target)}</loc>\n"
+        item_string += "    <lastmod>#{item.lastmod_value}</lastmod>\n"
+        item_string += "    <changefreq>#{item.changefreq}</changefreq>\n"
+        item_string += "    <priority>#{item.priority}</priority>\n"
+        item_string += "  </url>\n"
+        
+        item_results << item_string
+      end
+      
+      result = result + item_results.join("") + "</urlset>\n"
+      result
+    end
+  end
+end

data/lib/xml-sitemap/version.rb

@@ -1,3 +1,3 @@
 module XmlSitemap
-  VERSION = '1.1.3'
+  VERSION = '1.2.0'
 end

data/spec/item_spec.rb

@@ -0,0 +1,8 @@
+require 'spec_helper'
+
+describe 'XmlSitemap::Item' do
+  it 'should raise ArgumentError if invalid :period value was passed' do
+    proc { XmlSitemap::Item.new('hello', :period => :foobar) }.
+      should raise_error ArgumentError, "Invalid :period value 'foobar'"
+  end
+end

data/spec/map_spec.rb

@@ -1,3 +1,4 @@
+require 'benchmark'
 require 'spec_helper'
 
 describe XmlSitemap::Map do
@@ -52,10 +53,43 @@ describe XmlSitemap::Map do
     map.add('world', :updated => @extra_time).updated.should == Time.gm(2011, 7, 1, 0, 0, 1)
   end
   
+  it 'should help me test performance' do
+    pending "comment this line to run benchmarks, takes roughly 30 seconds"
+    map = XmlSitemap::Map.new('foobar.com', :time => @base_time)
+    50000.times do |i|
+      map.add("hello#{i}")
+    end
+    
+    Benchmark.bm do |x|
+      x.report("render(:builder)")  { map.render(:builder)  }
+      x.report("render(:nokogiri)") { map.render(:nokogiri) }
+      x.report("render(:string)")   { map.render(:string)   }
+    end
+  end
+  
   it 'should raise Argument error if no time or date were provided' do
+    map = XmlSitemap::Map.new('foobar.com', :time => @base_time)
+    proc { map.add('hello', :updated => 5) }.
+      should raise_error ArgumentError, "Time, Date, or ISO8601 String required for :updated!"
+  end
+  
+  it 'should not raise Argument error if a iso8601 string is provided' do
+    map = XmlSitemap::Map.new('foobar.com', :time => @base_time)
+    proc { map.add('hello', :updated => "2011-09-12T23:18:49Z") }.
+      should_not raise_error
+    map.add('world', :updated => @extra_time.utc.iso8601).updated.should == Time.gm(2011, 7, 1, 0, 0, 1).utc.iso8601
+  end
+  
+  it 'should not raise Argument error if a string is provided with :validate_time => false' do
+    map = XmlSitemap::Map.new('foobar.com', :time => @base_time)
+    proc { map.add('hello', :validate_time => false, :updated => 'invalid data') }.
+      should_not raise_error
+  end
+  
+  it 'should raise Argument error if an invalid string is provided' do
     map = XmlSitemap::Map.new('foobar.com', :time => @base_time)
     proc { map.add('hello', :updated => 'invalid data') }.
-      should raise_error ArgumentError, "Time or Date required for :updated!"
+      should raise_error ArgumentError, "String provided to :updated did not match ISO8601 standard!"
   end
   
   it 'should have properly encoded entities' do
@@ -64,11 +98,27 @@ describe XmlSitemap::Map do
     map.render.should == fixture('encoded_map.xml')
   end
   
+  it 'should have properly encoded entities using nokogiri render' do
+    map = XmlSitemap::Map.new('foobar.com', :time => @base_time)
+    map.add('/path?a=b&c=d&e=sample string')
+    s = map.render(:nokogiri)
+    # ignore ordering of urlset attributes by dropping first two lines
+    s.split("\n")[2..-1].join("\n").should == fixture('encoded_map.xml').split("\n")[2..-1].join("\n")
+  end
+  
+  it 'should have properly encoded entities using string render' do
+    map = XmlSitemap::Map.new('foobar.com', :time => @base_time)
+    map.add('/path?a=b&c=d&e=sample string')
+    s = map.render(:string)
+    # ignore ordering of urlset attributes by dropping first two lines
+    s.split("\n")[2..-1].join("\n").should == fixture('encoded_map.xml').split("\n")[2..-1].join("\n")
+  end
+  
   it 'should not allow more than 50k records' do
     map = XmlSitemap::Map.new('foobar.com')
     proc {
-      1.upto(50000) { |i| map.add("url#{i}") }
-    }.should raise_error RuntimeError, 'Only less than 50k records allowed!'
+      1.upto(50001) { |i| map.add("url#{i}") }
+    }.should raise_error RuntimeError, 'Only up to 50k records allowed!'
   end
   
   it 'should not allow urls longer than 2048 characters' do
@@ -120,4 +170,4 @@ describe XmlSitemap::Map do
     File.delete(path) if File.exists?(path)
     File.delete(path_gzip) if File.exists?(path_gzip)
   end
-end
+end

data/xml-sitemap.gemspec

@@ -12,6 +12,7 @@ Gem::Specification.new do |s|
   s.add_development_dependency 'rake',      '~> 0.8'
   s.add_development_dependency 'rspec',     '~> 2.6'
   s.add_development_dependency 'simplecov', '~> 0.4'
+  s.add_development_dependency 'nokogiri',  '~> 1.5.0'
   
   s.add_runtime_dependency      'builder',  '>= 2.0'
   

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: xml-sitemap
 version: !ruby/object:Gem::Version
-  version: 1.1.3
+  version: 1.2.0
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2011-09-08 00:00:00.000000000Z
+date: 2011-09-14 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
-  requirement: &2154904800 !ruby/object:Gem::Requirement
+  requirement: &2162591460 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -21,10 +21,10 @@ dependencies:
         version: '0.8'
   type: :development
   prerelease: false
-  version_requirements: *2154904800
+  version_requirements: *2162591460
 - !ruby/object:Gem::Dependency
   name: rspec
-  requirement: &2154904300 !ruby/object:Gem::Requirement
+  requirement: &2162590440 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -32,10 +32,10 @@ dependencies:
         version: '2.6'
   type: :development
   prerelease: false
-  version_requirements: *2154904300
+  version_requirements: *2162590440
 - !ruby/object:Gem::Dependency
   name: simplecov
-  requirement: &2154903840 !ruby/object:Gem::Requirement
+  requirement: &2162589300 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -43,10 +43,21 @@ dependencies:
         version: '0.4'
   type: :development
   prerelease: false
-  version_requirements: *2154903840
+  version_requirements: *2162589300
+- !ruby/object:Gem::Dependency
+  name: nokogiri
+  requirement: &2162588020 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 1.5.0
+  type: :development
+  prerelease: false
+  version_requirements: *2162588020
 - !ruby/object:Gem::Dependency
   name: builder
-  requirement: &2154903380 !ruby/object:Gem::Requirement
+  requirement: &2162586640 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -54,7 +65,7 @@ dependencies:
         version: '2.0'
   type: :runtime
   prerelease: false
-  version_requirements: *2154903380
+  version_requirements: *2162586640
 description: Provides a wrapper to generate XML sitemaps and sitemap indexes.
 email:
 - dan.sosedoff@gmail.com
@@ -70,8 +81,10 @@ files:
 - Rakefile
 - lib/xml-sitemap.rb
 - lib/xml-sitemap/index.rb
+- lib/xml-sitemap/item.rb
 - lib/xml-sitemap/map.rb
 - lib/xml-sitemap/options.rb
+- lib/xml-sitemap/render_engine.rb
 - lib/xml-sitemap/version.rb
 - spec/fixtures/empty_index.xml
 - spec/fixtures/encoded_map.xml
@@ -80,6 +93,7 @@ files:
 - spec/fixtures/saved_map.xml
 - spec/fixtures/simple_map.xml
 - spec/index_spec.rb
+- spec/item_spec.rb
 - spec/map_spec.rb
 - spec/spec_helper.rb
 - spec/xmlsitemap_spec.rb
@@ -116,6 +130,7 @@ test_files:
 - spec/fixtures/saved_map.xml
 - spec/fixtures/simple_map.xml
 - spec/index_spec.rb
+- spec/item_spec.rb
 - spec/map_spec.rb
 - spec/spec_helper.rb
 - spec/xmlsitemap_spec.rb