Class: Blinkr::Extensions::Links

Inherits:
Object
  • Object
show all
Includes:
HttpUtils
Defined in:
lib/blinkr/extensions/links.rb

Instance Method Summary collapse

Methods included from HttpUtils

#retry?, #sanitize

Constructor Details

#initialize(config) ⇒ Links

Returns a new instance of Links.



10
11
12
13
 
# File 'lib/blinkr/extensions/links.rb', line 10

def initialize(config)
  @config = config
  @links = {}
end

Instance Method Details

#analyze(context, browser) ⇒ Object 



30
31
32
33
34
35
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
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
 
# File 'lib/blinkr/extensions/links.rb', line 30

def analyze(context, browser)
  puts '----------------------'
  puts " #{@links.length} links to check "
  puts '----------------------'
  start = DateTime.now

  processed = 0

  # Find the internal links
  @links.select{|k| k.start_with? @config.base_url}.each do |url, locations|
    # TODO figure out what to do about relative links
    link = URI.parse(url)

    # fix up links so they're proper, also drop fragments and queries as they won't be in the sitemap that way
    link.fragment = nil
    link.query = nil
    link.path = link.path.gsub(/\/+/, '/') if link.path

    unless context.pages.keys.include?(link.to_s) || context.pages.keys.include?((link.to_s + '/'))
      locations.each do |location|
        location[:page].errors << Blinkr::Error.new({:severity => :warning,
                                                     :category => 'Resource missing from sitemap',
                                                     :type => '<a href=""> target missing from sitemap',
                                                     :url => url, :title => "#{url} (line #{location[:line]})",
                                                     :code => nil,
                                                     :message => 'Missing from sitemap',
                                                     :detail => 'Checked with Typheous',
                                                     :snippet => location[:snippet],
                                                     :icon => 'fa-bookmark-o'
                                                    })
        # It wasn't in the sitemap, so we'll add it to the "external_links" to still be checked
      end
    end
  end
  @links.each do |url, metadata|
    # if link start_with? @config.base_url check to see if it's in the sitemap.xml
    browser.process(url, @config.max_retrys, :method => :get, :followlocation => true, :timeout => 60,
                                             :cookiefile => '_tmp/cookies', :cookiejar => '_tmp/cookies',
                    :connecttimeout => 30, :maxredirs => 3) do |resp|
      puts "Loaded #{url} via #{browser.name} #{'(cached)' if resp.cached?}" if @config.verbose

      resp_code = resp.code.to_i
      if ((resp_code > 300 && resp_code < 400) && @config.warning_on_300s) || resp_code > 400
        response = resp

        detail = nil
        if response.status_message.nil?
          message = response.return_message
        else
          message = response.status_message
          detail = response.return_message unless resp.return_message == 'No error'
        end

        severity = :danger
        if response.code.to_i >= 300 && response.code.to_i < 400
          severity = :warning
        end
        metadata.each do |src|
          src[:page].errors << Blinkr::Error.new({:severity => severity,
                                                  :category => 'Resources missing',
                                                  :type => '<a href=""> target cannot be loaded',
                                                  :url => url, :title => "#{url} (line #{src[:line]})",
                                                  :code => response.code.to_i, :message => message,
                                                  :detail => detail, :snippet => src[:snippet],
                                                  :icon => 'fa-bookmark-o'}) unless response.success?
        end
      end
      processed += 1
      puts "Processed #{processed} of #{@links.size}" if @config.verbose
    end
  end
  browser.hydra.run if browser.is_a? Blinkr::TyphoeusWrapper
  puts "Total time in links: #{(DateTime.now.to_time - start.to_time).duration}" if @config.verbose
end

#collect(page) ⇒ Object 



15
16
17
18
19
20
21
22
23
24
25
26
27
28
 
# File 'lib/blinkr/extensions/links.rb', line 15

def collect(page)
  page.body.css('a[href]').each do |a|
    attr = a.attribute('href')
    src = page.response.effective_url
    url = attr.value
    unless @config.skipped?(url)
      url = sanitize url, src
      unless url.nil?
        @links[url] ||= []
        @links[url] << {:page => page, :line => attr.line, :snippet => attr.parent.to_s}
      end
    end
  end
end