Class: HTML::Proofer::UrlValidator

Inherits:
Object
  • Object
show all
Includes:
Utils
Defined in:
lib/html/proofer/url_validator.rb

Instance Attribute Summary collapse

Instance Method Summary collapse

Methods included from Utils

create_nokogiri, swap

Constructor Details

#initialize(logger, external_urls, options, typhoeus_opts, hydra_opts) ⇒ UrlValidator

Returns a new instance of UrlValidator.



12
13
14
15
16
17
18
19
20
21
 
# File 'lib/html/proofer/url_validator.rb', line 12

def initialize(logger, external_urls, options, typhoeus_opts, hydra_opts)
  @logger = logger
  @external_urls = external_urls
  @iterable_external_urls = {}
  @failed_tests = []
  @options = options
  @hydra = Typhoeus::Hydra.new(hydra_opts)
  @typhoeus_opts = typhoeus_opts
  @external_domain_paths_with_queries = {}
end

Instance Attribute Details

#external_urls ⇒ Object

Returns the value of attribute external_urls.



10
11
12
 
# File 'lib/html/proofer/url_validator.rb', line 10

def external_urls
  @external_urls
end

#hydra ⇒ Object

Returns the value of attribute hydra.



10
11
12
 
# File 'lib/html/proofer/url_validator.rb', line 10

def hydra
  @hydra
end

#iterable_external_urls ⇒ Object

Returns the value of attribute iterable_external_urls.



10
11
12
 
# File 'lib/html/proofer/url_validator.rb', line 10

def iterable_external_urls
  @iterable_external_urls
end

#logger ⇒ Object

Returns the value of attribute logger.



10
11
12
 
# File 'lib/html/proofer/url_validator.rb', line 10

def logger
  @logger
end

Instance Method Details

#add_external_issue(filenames, desc, status = nil) ⇒ Object 



164
165
166
167
168
169
170
 
# File 'lib/html/proofer/url_validator.rb', line 164

def add_external_issue(filenames, desc, status = nil)
  if filenames.nil?
    @failed_tests << CheckRunner::Issue.new('', desc, nil, status)
  else
    filenames.each { |f| @failed_tests << CheckRunner::Issue.new(f, desc, nil, status) }
  end
end

#check_hash_in_2xx_response(href, effective_url, response, filenames) ⇒ Object

Even though the response was a success, we may have been asked to check if the hash on the URL exists on the page



141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
 
# File 'lib/html/proofer/url_validator.rb', line 141

def check_hash_in_2xx_response(href, effective_url, response, filenames)
  return if @options[:only_4xx]
  return unless @options[:check_external_hash]
  return unless (hash = hash?(href))

  body_doc = create_nokogiri(response.body)

  # user-content is a special addition by GitHub.
  xpath = %(//*[@name="#{hash}"]|//*[@id="#{hash}"])
  if URI.parse(href).host.match(/github\.com/i)
    xpath << %(|//*[@name="user-content-#{hash}"]|//*[@id="user-content-#{hash}"])
  end

  return unless body_doc.xpath(xpath).empty?

  add_external_issue filenames, "External link #{href} failed: #{effective_url} exists, but the hash '#{hash}' does not", response.code
end

#clean_url(href) ⇒ Object 



106
107
108
 
# File 'lib/html/proofer/url_validator.rb', line 106

def clean_url(href)
  Addressable::URI.parse(href).normalize
end

#external_link_checker(external_urls) ⇒ Object

Proofer runs faster if we pull out all the external URLs and run the checks at the end. Otherwise, we’re halting the consuming process for every file during the check_directory_of_files process.

In addition, sorting the list lets libcurl keep connections to the same hosts alive.

Finally, we’ll first make a HEAD request, rather than GETing all the contents. If the HEAD fails, we’ll fall back to GET, as some servers are not configured for HEAD. If we’ve decided to check for hashes, we must do a GET–HEAD is not an option.



74
75
76
77
78
79
80
81
82
83
84
85
86
87
 
# File 'lib/html/proofer/url_validator.rb', line 74

def external_link_checker(external_urls)
  external_urls = Hash[external_urls.sort]

  count = external_urls.length
  check_text = "#{count} " << (count == 1 ? 'external link' : 'external links')
  logger.log :info, :blue, "Checking #{check_text}..."

  Ethon.logger = logger # log from Typhoeus/Ethon

  url_processor(external_urls)

  logger.log :debug, :yellow, "Running requests for all #{hydra.queued_requests.size} external URLs..."
  hydra.run
end

#extract_domain_path(uri) ⇒ Object 



60
61
62
 
# File 'lib/html/proofer/url_validator.rb', line 60

def extract_domain_path(uri)
  uri.host + uri.path
end

#handle_timeout(href, filenames, response_code) ⇒ Object 



159
160
161
162
 
# File 'lib/html/proofer/url_validator.rb', line 159

def handle_timeout(href, filenames, response_code)
  return if @options[:only_4xx]
  add_external_issue filenames, "External link #{href} failed: got a time out", response_code
end

#hash?(url) ⇒ Boolean

Returns:

  • (Boolean) 


172
173
174
175
176
 
# File 'lib/html/proofer/url_validator.rb', line 172

def hash?(url)
  URI.parse(url).fragment
rescue URI::InvalidURIError
  nil
end

#new_url_query_values?(uri) ⇒ Boolean

remember queries we’ve seen, ignore future ones

Returns:

  • (Boolean) 


46
47
48
49
50
51
52
53
54
55
56
57
58
 
# File 'lib/html/proofer/url_validator.rb', line 46

def new_url_query_values?(uri)
  queries = uri.query_values.keys.join('-')
  domain_path = extract_domain_path(uri)
  if @external_domain_paths_with_queries[domain_path].nil?
    @external_domain_paths_with_queries[domain_path] = [queries]
    true
  elsif !@external_domain_paths_with_queries[domain_path].include?(queries)
    @external_domain_paths_with_queries[domain_path] << queries
    true
  else
    false
  end
end

#queue_request(method, href, filenames) ⇒ Object 



110
111
112
113
114
 
# File 'lib/html/proofer/url_validator.rb', line 110

def queue_request(method, href, filenames)
  request = Typhoeus::Request.new(href, @typhoeus_opts.merge({ :method => method }))
  request.on_complete { |response| response_handler(response, filenames) }
  hydra.queue request
end

#remove_query_values ⇒ Object 



29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
 
# File 'lib/html/proofer/url_validator.rb', line 29

def remove_query_values
  return if @external_urls.nil?
  iterable_external_urls = @external_urls.dup
  @external_urls.keys.each do |url|
    uri = begin
            Addressable::URI.parse(url)
          rescue URI::Error, Addressable::URI::InvalidURIError
            @logger.log :error, :red, "#{url} is an invalid URL"
            nil
          end
    next if uri.nil? || uri.query.nil?
    iterable_external_urls.delete(url) unless new_url_query_values?(uri)
  end
  iterable_external_urls
end

#response_handler(response, filenames) ⇒ Object 



116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
 
# File 'lib/html/proofer/url_validator.rb', line 116

def response_handler(response, filenames)
  effective_url = response.options[:effective_url]
  href = response.request.base_url.to_s
  method = response.request.options[:method]
  response_code = response.code

  debug_msg = "Received a #{response_code} for #{href}"
  debug_msg << " in #{filenames.join(' ')}" unless filenames.nil?
  logger.log :debug, :yellow, debug_msg

  if response_code.between?(200, 299)
    check_hash_in_2xx_response(href, effective_url, response, filenames)
  elsif response.timed_out?
    handle_timeout(href, filenames, response_code)
  elsif method == :head
    queue_request(:get, href, filenames)
  else
    return if @options[:only_4xx] && !response_code.between?(400, 499)
    # Received a non-successful http response.
    add_external_issue(filenames, "External link #{href} failed: #{response_code} #{response.return_message}", response_code)
  end
end

#run ⇒ Object 



23
24
25
26
27
 
# File 'lib/html/proofer/url_validator.rb', line 23

def run
  @iterable_external_urls = remove_query_values
  external_link_checker(@iterable_external_urls)
  @failed_tests
end

#url_processor(external_urls) ⇒ Object 



89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
 
# File 'lib/html/proofer/url_validator.rb', line 89

def url_processor(external_urls)
  external_urls.each_pair do |href, filenames|
    href = begin
             clean_url(href)
           rescue URI::Error, Addressable::URI::InvalidURIError
             add_external_issue(filenames, "#{href} is an invalid URL")
             next
           end

    if hash?(href) && @options[:check_external_hash]
      queue_request(:get, href, filenames)
    else
      queue_request(:head, href, filenames)
    end
  end
end