sitemap_generator 0.2.3 → 0.2.4

data/README.md

@@ -6,9 +6,9 @@ This plugin enables ['enterprise-class'][enterprise_class] Google Sitemaps to be
 Foreword
 -------
 
-Unfortunately, Adam Salter passed away recently.  Those who knew him know what an amazing guy he was, and what an excellent Rails programmer he was.  His passing is a great loss to the Rails community.
+Unfortunately, Adam Salter passed away in 2009.  Those who knew him know what an amazing guy he was, and what an excellent Rails programmer he was.  His passing is a great loss to the Rails community.
 
-I will be taking over maintaining this gem from him. -- Karl
+[Karl Varga](http://github.com/kjvarga) has taken over development of SitemapGenerator.  The canonical repository is [http://github.com/kjvarga/sitemap_generator][canonical_repo]
 
 Raison d'être
 -------
@@ -171,20 +171,15 @@ Wishlist
 Thanks (in no particular order)
 ========
 
-- [Karl Varga (aka Bear Grylls)](http://github.com/kjvarga)
 - [Dan Pickett](http://github.com/dpickett)
 - [Rob Biedenharn](http://github.com/rab)
 - [Richie Vos](http://github.com/jerryvos)
+- [Adrian Mugnolo](http://github.com/xymbol)
 
 
-Follow me on:
----------
-
->  Twitter: [twitter.com/adamsalter](http://twitter.com/adamsalter)  
->  Github: [github.com/adamsalter](http://github.com/adamsalter)
-
 Copyright (c) 2009 Adam @ [Codebright.net][cb], released under the MIT license
 
+[canonical_repo]:http://github.com/kjvarga/sitemap_generator
 [enterprise_class]:https://twitter.com/dhh/status/1631034662 "I use enterprise in the same sense the Phusion guys do - i.e. Enterprise Ruby. Please don't look down on my use of the word 'enterprise' to represent being a cut above. It doesn't mean you ever have to work for a company the size of IBM. Or constantly fight inertia, writing crappy software, adhering to change management practices and spending hours in meetings... Not that there's anything wrong with that - Wait, what?"
 [sitemap_engines]:http://en.wikipedia.org/wiki/Sitemap_index "http://en.wikipedia.org/wiki/Sitemap_index"
 [sitemaps_org]:http://www.sitemaps.org/protocol.php "http://www.sitemaps.org/protocol.php"

data/Rakefile

@@ -1,30 +1,81 @@
-require 'rake/testtask'
-require 'find'
+require 'rake'
+require 'rake/rdoctask'
+require 'spec/rake/spectask'
 
 begin
   require 'jeweler'
-  Jeweler::Tasks.new do |s|
-    s.name = "sitemap_generator"
-    s.summary = %Q{Generate 'enterprise-class' Sitemaps for your Rails site using a simple 'Rails Routes'-like DSL and a single Rake task}
-    s.description = %Q{Install as a plugin or Gem to easily generate ['enterprise-class'][enterprise_class] Google Sitemaps for your Rails site, using a simple 'Rails Routes'-like DSL and a single rake task.}
-    s.email = "kjvarga@gmail.com"
-    s.homepage = "http://github.com/kjvarga/sitemap_generator"
-    s.authors = ["Adam Salter", "Karl Varga"]
-    s.files =  FileList["[A-Z]*", "{bin,lib,rails,templates,tasks}/**/*"]
-    # s is a Gem::Specification... see http://www.rubygems.org/read/chapter/20 for additional settings
+  Jeweler::Tasks.new do |gem|
+    gem.name = "sitemap_generator"
+    gem.summary = %Q{Easily generate enterprise class Sitemaps for your Rails site using a simple 'Rails Routes'-like DSL and a single Rake task}
+    gem.description = %Q{Installs as a plugin or Gem to easily generate enterprise class Sitemaps readable by all search engines.  Automatically ping search engines to notify them of new sitemaps, including Google, Yahoo and Bing. Provides rake tasks to easily manage your sitemaps.  Won't clobber your old sitemaps if the new one fails to generate. Setup a cron schedule and never worry about your sitemaps again.}
+    gem.email = "kjvarga@gmail.com"
+    gem.homepage = "http://github.com/kjvarga/sitemap_generator"
+    gem.authors = ["Adam Salter", "Karl Varga"]
+    gem.files =  FileList["[A-Z]*", "{bin,lib,rails,templates,tasks}/**/*"]
+    gem.test_files = []
+    gem.add_development_dependency "rspec"
   end
   Jeweler::GemcutterTasks.new
 rescue LoadError
   puts "Jeweler (or a dependency) not available. Install it with: gem install jeweler"
 end
 
-desc 'Default: run unit tests.'
 task :default => :test
 
-desc 'Test.'
-Rake::TestTask.new(:test) do |t|
-  t.libs << 'lib'
-  t.pattern = 'test/**/*_test.rb'
-  t.verbose = true
+namespace :test do
+  task :gem => ['test:prepare:gem', 'multi_spec']
+  task :plugin => ['test:prepare:plugin', 'multi_spec']
+  
+  task :multi_spec do
+    Rake::Task['spec'].invoke
+    Rake::Task['spec'].reenable
+  end
+  
+  namespace :prepare do
+    task :gem do
+      ENV["SITEMAP_RAILS"] = 'gem'
+      prepare_path(local_path('spec/mock_app_gem/vendor/gems/sitemap_generator-1.2.3'))
+      rm_rf(local_path('spec/mock_app_gem/public/sitemap*'))
+    end
+    
+    task :plugin do
+      ENV["SITEMAP_RAILS"] = 'plugin'
+      prepare_path(local_path('spec/mock_app_plugin/vendor/plugins/sitemap_generator-1.2.3'))
+      rm_rf(local_path('spec/mock_app_plugin/public/sitemap*'))
+    end
+
+    def local_path(path)
+      File.join(File.dirname(__FILE__), path)
+    end
+    
+    def prepare_path(path)
+      rm_rf path
+      mkdir_p path
+      cp_r(FileList["[A-Z]*", "{bin,lib,rails,templates,tasks}"], path)
+    end
+  end
 end
 
+desc "Run all tests both as a plugin and gem"
+task :test => ['test:plugin', 'test:gem']
+
+Spec::Rake::SpecTask.new(:spec) do |spec|
+  spec.libs << 'lib' << 'spec'
+  spec.spec_files = FileList['spec/**/*_spec.rb']
+end
+task :spec => :check_dependencies
+
+Spec::Rake::SpecTask.new(:rcov) do |spec|
+  spec.libs << 'lib' << 'spec'
+  spec.pattern = 'spec/**/*_spec.rb'
+  spec.rcov = true
+end
+
+desc 'Generate documentation'
+Rake::RDocTask.new(:rdoc) do |rdoc|
+  rdoc.rdoc_dir = 'rdoc'
+  rdoc.title    = 'SitemapGenerator'
+  rdoc.options << '--line-numbers' << '--inline-source'
+  rdoc.rdoc_files.include('README.md')
+  rdoc.rdoc_files.include('lib/**/*.rb')
+end

data/VERSION

@@ -1 +1 @@
-0.2.3
+0.2.4

data/lib/sitemap_generator.rb

@@ -4,16 +4,20 @@ require 'sitemap_generator/link_set'
 require 'sitemap_generator/helper'
 
 module SitemapGenerator
-  class <<self
+  silence_warnings do
+    VERSION = File.read(File.dirname(__FILE__) + "/../VERSION").strip
+    MAX_ENTRIES = 50_000
+    Sitemap = LinkSet.new
+  end
+  
+  class << self
     attr_accessor :root, :templates
   end
+
   self.root = File.expand_path(File.join(File.dirname(__FILE__), '../'))
   self.templates = {
-    :sitemap_index => File.join(self.root, 'templates/sitemap_index.builder'),
-    :sitemap_xml   => File.join(self.root, 'templates/xml_sitemap.builder'),
+    :sitemap_index  => File.join(self.root, 'templates/sitemap_index.builder'),
+    :sitemap_xml    => File.join(self.root, 'templates/xml_sitemap.builder'),
     :sitemap_sample => File.join(self.root, 'templates/sitemap.rb'),
   }
-  
-  Sitemap = LinkSet.new
 end
-  

data/lib/sitemap_generator/link_set.rb

@@ -1,28 +1,181 @@
+require File.dirname(__FILE__) + '/helper'
+
 module SitemapGenerator
   class LinkSet
+    include SitemapGenerator::Helper
+    include ActionView::Helpers::NumberHelper
+
     attr_accessor :default_host, :yahoo_app_id, :links
-    
-    def initialize
-      @links = []
-    end
+    attr_accessor :sitemaps
+    attr_accessor :max_entries
+    attr_accessor :link_count
 
-    def default_host=(host)
-      @default_host = host
-      add_default_links
+    alias :sitemap_files :sitemaps
+
+    # Create new link set instance.
+    def initialize
+      self.links       = []
+      self.sitemaps    = []
+      self.max_entries = SitemapGenerator::MAX_ENTRIES
+      self.link_count  = 0
     end
 
+    # Add default links to sitemap files.
     def add_default_links
-      # Add default links
-      @links << Link.generate('/', :lastmod => Time.now, :changefreq => 'always', :priority => 1.0)
-      @links << Link.generate('/sitemap_index.xml.gz', :lastmod => Time.now, :changefreq => 'always', :priority => 1.0)
+      links.push Link.generate('/', :lastmod => Time.now, :changefreq => 'always', :priority => 1.0)
+      links.push Link.generate("/#{index_file}", :lastmod => Time.now, :changefreq => 'always', :priority => 1.0)
+      self.link_count += 2
     end
-    
+
+    # Add links to sitemap files passing a block.
     def add_links
+      raise ArgumentError, "Default hostname not set" if default_host.blank?
+      add_default_links if first_link?
       yield Mapper.new(self)
     end
-    
+
+    # Add links from mapper to sitemap files.
     def add_link(link)
-      @links << link
+      write_upcoming if enough_links?
+      links.push link
+      self.link_count += 1
+    end
+
+    # Write links to sitemap file.
+    def write
+      write_pending
+    end
+
+    # Write links to upcoming sitemap file.
+    def write_upcoming
+      write_sitemap(upcoming_file)
+    end
+
+    # Write pending links to sitemap, write index file if needed.
+    def write_pending
+      write_upcoming
+      write_index
+    end
+
+    # Write links to sitemap file.
+    def write_sitemap(file = upcoming_file)
+      buffer = ""
+      xml = Builder::XmlMarkup.new(:target => buffer)
+      eval(File.read(SitemapGenerator.templates[:sitemap_xml]), binding)
+      filename = File.join(RAILS_ROOT, "public", file)
+      write_file(filename, buffer)
+      show_progress("Sitemap", filename, buffer) if verbose
+      links.clear
+      sitemaps.push filename
+    end
+
+    # Write sitemap links to sitemap index file.
+    def write_index
+      buffer = ""
+      xml = Builder::XmlMarkup.new(:target => buffer)
+      eval(File.read(SitemapGenerator.templates[:sitemap_index]), binding)
+      filename = File.join(RAILS_ROOT, "public", index_file)
+      write_file(filename, buffer)
+      show_progress("Sitemap Index", filename, buffer) if verbose
+      links.clear
+      sitemaps.clear
+    end
+
+    # Return sitemap or sitemap index main name.
+    def index_file
+      "sitemap_index.xml.gz"
+    end
+
+    # Return upcoming sitemap name with index.
+    def upcoming_file
+      "sitemap#{upcoming_index}.xml.gz" unless enough_sitemaps?
+    end
+
+    # Return upcoming sitemap index, first is 1.
+    def upcoming_index
+      sitemaps.length + 1 unless enough_sitemaps?
+    end
+
+    # Return true if upcoming is first sitemap.
+    def first_sitemap?
+      sitemaps.empty?
+    end
+
+    # Return true if sitemap index needed.
+    def multiple_sitemaps?
+      !first_sitemap?
+    end
+
+    # Return true if more sitemaps can be added.
+    def more_sitemaps?
+      sitemaps.length < max_entries
+    end
+
+    # Return true if no sitemaps can be added.
+    def enough_sitemaps?
+      !more_sitemaps?
+    end
+
+    # Return true if this is the first link added.
+    def first_link?
+      links.empty? && first_sitemap?
+    end
+
+    # Return true if more links can be added.
+    def more_links?
+      links.length < max_entries
+    end
+
+    # Return true if no further links can be added.
+    def enough_links?
+      !more_links?
+    end
+
+    # Commit buffer to gzipped file.
+    def write_file(name, buffer)
+      Zlib::GzipWriter.open(name) { |gz| gz.write buffer }
+    end
+
+    # Report progress line.
+    def show_progress(title, filename, buffer)
+      puts "+ #{filename}"
+      puts "** #{title} too big! The uncompressed size exceeds 10Mb" if buffer.size > 10.megabytes
+    end
+
+    # Copy templates/sitemap.rb to config if not there yet.
+    def install_sitemap_rb
+      if File.exist?(File.join(RAILS_ROOT, 'config/sitemap.rb'))
+        puts "already exists: config/sitemap.rb, file not copied"
+      else
+        FileUtils.cp(SitemapGenerator.templates[:sitemap_sample], File.join(RAILS_ROOT, 'config/sitemap.rb'))
+        puts "created: config/sitemap.rb"
+      end
+    end
+
+    # Remove config/sitemap.rb if exists.
+    def uninstall_sitemap_rb
+      if File.exist?(File.join(RAILS_ROOT, 'config/sitemap.rb'))
+        File.rm(File.join(RAILS_ROOT, 'config/sitemap.rb'))
+      end
+    end
+
+    # Clean sitemap files in output directory.
+    def clean_files
+      FileUtils.rm(Dir[File.join(RAILS_ROOT, 'public/sitemap*.xml.gz')])
+    end
+
+    # Ping search engines passing sitemap location.
+    def ping_search_engines
+      super index_file
+    end
+
+    # Create sitemap files in output directory.
+    def create_files(verbose = true)
+      start_time = Time.now
+      load_sitemap_rb
+      write
+      stop_time = Time.now
+      puts "Sitemap stats: #{number_with_delimiter(SitemapGenerator::Sitemap.link_count)} links, " + ("%dm%02ds" % (stop_time - start_time).divmod(60)) if verbose
     end
   end
-end
+end

data/rails/install.rb

@@ -1,12 +1,2 @@
 # Install hook code here
-
-# Copy sitemap_template.rb to config/sitemap.rb
-require 'fileutils'
-sitemap_template = File.join(File.dirname(__FILE__), '../templates/sitemap.rb')
-new_sitemap = File.join(RAILS_ROOT, 'config/sitemap.rb')
-if File.exist?(new_sitemap)
-  puts "already exists: config/sitemap.rb, file not copied"
-else
-  puts "created: config/sitemap.rb"
-  FileUtils.cp(sitemap_template, new_sitemap)
-end
+SitemapGenerator::Sitemap.install_sitemap_rb

data/rails/uninstall.rb

@@ -1,4 +1,2 @@
 # Uninstall hook code here
-
-new_sitemap = File.join(RAILS_ROOT, 'config/sitemap.rb')
-File.rm(new_sitemap) if File.exist?(new_sitemap)
+SitemapGenerator::Sitemap.uninstall_sitemap_rb

data/tasks/sitemap_generator_tasks.rake

@@ -1,96 +1,26 @@
 require 'zlib'
-require 'sitemap_generator/helper'
-
-class SiteMapRefreshTask < Rake::Task
-  include SitemapGenerator::Helper
-
-  def execute(*)
-    super
-    ping_search_engines("sitemap_index.xml.gz")
-  end
-end
-
-class SiteMapCreateTask < Rake::Task
-  include SitemapGenerator::Helper
-  include ActionView::Helpers::NumberHelper
-
-  def execute(*)
-    super
-    build_files
-  end
-
-  private
-  def build_files
-    start_time = Time.now
-
-    # update links from config/sitemap.rb
-    load_sitemap_rb
-
-    raise(ArgumentError, "Default hostname not defined") if SitemapGenerator::Sitemap.default_host.blank?
-
-    links_grps = SitemapGenerator::Sitemap.links.in_groups_of(50_000, false)
-    raise(ArgumentError, "TOO MANY LINKS!! I really thought 2,500,000,000 links would be enough for anybody!") if links_grps.length > 50000
-
-    Rake::Task['sitemap:clean'].invoke
-
-    # render individual sitemaps
-    sitemap_files = render_sitemap(links_grps)
-
-    # render index
-    render_index(sitemap_files)
-    
-    stop_time = Time.now
-    puts "Sitemap stats: #{number_with_delimiter(SitemapGenerator::Sitemap.links.length)} links, " + ("%dm%02ds" % (stop_time - start_time).divmod(60)) if verbose
-  end
-
-  def render_sitemap(links_grps)
-    sitemap_files = []
-    links_grps.each_with_index do |links, index|
-      buffer = ''
-      xml = Builder::XmlMarkup.new(:target=>buffer)
-      eval(open(SitemapGenerator.templates[:sitemap_xml]).read, binding)
-      filename = File.join(RAILS_ROOT, "public/sitemap#{index+1}.xml.gz")
-      Zlib::GzipWriter.open(filename) do |gz|
-        gz.write buffer
-      end
-      sitemap_files << filename
-      puts "+ #{filename}" if verbose
-      puts "** Sitemap too big! The uncompressed size exceeds 10Mb" if (buffer.size > 10 * 1024 * 1024) && verbose
-    end
-    sitemap_files
-  end
-
-  def render_index(sitemap_files)
-    buffer = ''
-    xml = Builder::XmlMarkup.new(:target=>buffer)
-    eval(open(SitemapGenerator.templates[:sitemap_index]).read, binding)
-    filename = File.join(RAILS_ROOT, "public/sitemap_index.xml.gz")
-    Zlib::GzipWriter.open(filename) do |gz|
-      gz.write buffer
-    end
-    
-    puts "+ #{filename}" if verbose
-    puts "** Sitemap Index too big! The uncompressed size exceeds 10Mb" if (buffer.size > 10 * 1024 * 1024) && verbose
-  end
-end
+require 'sitemap_generator'
 
 namespace :sitemap do
   desc "Install a default config/sitemap.rb file"
   task :install do
-    load File.expand_path(File.join(File.dirname(__FILE__), "../rails/install.rb"))
+    SitemapGenerator::Sitemap.install_sitemap_rb
   end
 
   desc "Delete all Sitemap files in public/ directory"
   task :clean do
-    sitemap_files = Dir[File.join(RAILS_ROOT, 'public/sitemap*.xml.gz')]
-    FileUtils.rm sitemap_files
+    SitemapGenerator::Sitemap.clean_files
   end
 
   desc "Create Sitemap XML files in public/ directory (rake -s for no output)"
-  SiteMapRefreshTask.define_task :refresh => ['sitemap:create']
+  task :refresh => ['sitemap:create'] do
+    SitemapGenerator::Sitemap.ping_search_engines
+  end
 
   desc "Create Sitemap XML files (don't ping search engines)"
   task 'refresh:no_ping' => ['sitemap:create']
 
-  SiteMapCreateTask.define_task :create => [:environment]
+  task :create => [:environment] do
+    SitemapGenerator::Sitemap.create_files
+  end
 end