Class: SitemapGenerator::Tasks

Inherits:
Object
  • Object
show all
Includes:
Rake::DSL
Defined in:
lib/static_sitemap_tasks.rb

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(options = {}) ⇒ Tasks

Returns a new instance of Tasks.



16
17
18
19
20
21
22
23
24
25
26
27
 
# File 'lib/static_sitemap_tasks.rb', line 16

def initialize(options = {})
  # Root of files to crawl
  @public_root = options[:public_root] || Dir.pwd
  # Change frequency - see: http://www.sitemaps.org/protocol.php#changefreqdef
  @change_frequency = options[:change_frequency]
  # Canonical domain of published site
  @base_url = options[:base_url]
  # Index pages
  @index_files = options[:index_files] || [ 'index.html', 'index.htm' ]
  # Compress output to sitemap.xml.gz
  @gzip_output = options[:gzip_output] || true
end

Class Method Details

.install(options = {}) ⇒ Object 



11
12
13
14
 
# File 'lib/static_sitemap_tasks.rb', line 11

def self.install(options = {})
  dir = options.delete(:dir) || Dir.pwd
  self.new(options).install
end

Instance Method Details

uses Hpricot to grab links from a URI adds uri to @pages_crawled loops each link found adds link to pages array if it should be included, unless it already exists



84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
 
# File 'lib/static_sitemap_tasks.rb', line 84

def crawl_for_links(link_path)
  if link_path.include?('http')
    return unless link_path.include?(@base_url)
    link_path.sub!(@base_url,'')
  end
  file_path = resolve_file_path(File.join(@public_root, link_path))

  if file_path.nil?
    puts "Warning: Unable to resolve #{link_path} to a local file"
    return
  end

  puts "Inspecting #{file_path}...\n"
  doc = Hpricot(open(file_path)) rescue nil
  return unless doc
  @pages_crawled << link_path
  (doc/"a").each do |a|
    if a['href'] && should_be_included?(a['href'])
      @pages << a['href'] unless(link_exists?(a['href'],@pages))
    end
  end
end

#generate_sitemap ⇒ Object 



36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
 
# File 'lib/static_sitemap_tasks.rb', line 36

def generate_sitemap
  # holds pages to go into map, and pages crawled
  @pages = []
  @pages_crawled = []

  # start with index pages
  crawl_for_links('/')

  # crawl each page in pages array unless it's already been crawled
  @pages.each {|page|
    crawl_for_links(page) unless @pages_crawled.include?(page)
  }

  # create xml for sitemap
  xml = Builder::XmlMarkup.new( :indent => 2 )
  xml.instruct!
  xml.comment! "Generated on: " + Time.now.to_s
  xml.urlset("xmlns" => "http://www.sitemaps.org/schemas/sitemap/0.9") {
    # loop through array of pages, and build sitemap.xml
    @pages.sort.each {|link|
      xml.url {
        xml.loc URI.join(@base_url, link)
        # TODO - set changefreq dynamically per page
        xml.changefreq @change_frequency unless @change_frequency.nil?
      }
    }
  }

  # convert builder xml to xml string, and save
  xml_string = xml.to_s.gsub("<to_s/>","")
  filename = File.join(@public_root,'sitemap.xml')

  if @gzip_output
    require 'zlib'
    filename << '.gz'
    xml_file = Zlib::GzipWriter.open(filename)
  else
    xml_file = File.open(filename, 'w')
  end

  xml_file << xml_string
  xml_file.close
end

#install ⇒ Object 



29
30
31
32
33
34
 
# File 'lib/static_sitemap_tasks.rb', line 29

def install
  desc "Generate a sitemap based on the contents of #{@public_root}"
  task 'generate_sitemap' do
    generate_sitemap
  end
end

#link_exists?(str, array) ⇒ Boolean

checks each value in a given array for the given string removes ‘/’ character before comparison

Returns:

  • (Boolean) 


141
142
143
 
# File 'lib/static_sitemap_tasks.rb', line 141

def link_exists?(str, array)
  array.detect{|l| strip_slashes(l) == strip_slashes(str)}
end

#resolve_file_path(path) ⇒ Object 



107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
 
# File 'lib/static_sitemap_tasks.rb', line 107

def resolve_file_path(path)
  file_path = nil

  if File.directory?(path)
    @index_files.each do |f|
      index_file = File.join(path,f)
      if File.exists?(index_file)
        file_path = index_file
        break
      end
    end
  else
    if File.exists?(path)
      file_path = path
    end
  end

  file_path
end

#should_be_included?(str) ⇒ Boolean

returns true if any of the following are true:

  • link isn’t external (eg, contains ‘http://’) and doesn’t contain ‘mailto:’

  • is equal to ‘/’

  • link contains @base_url

Returns:

  • (Boolean) 


131
132
133
134
135
136
137
 
# File 'lib/static_sitemap_tasks.rb', line 131

def should_be_included?(str)
  if ((!str.include?('http://') && !str.include?('mailto:')) || str == '/' || str.include?(@base_url))
    unless str.slice(0,1) == "#"
      return true
    end
  end
end

#strip_slashes(str) ⇒ Object

removes ‘/’ character from string



146
147
148
 
# File 'lib/static_sitemap_tasks.rb', line 146

def strip_slashes(str)
  str.gsub('/','')
end