Class: Jekyll::AgentAudit::Extractor

Inherits:
Object
  • Object
show all
Defined in:
lib/jekyll/agent_audit/extractor.rb

Constant Summary collapse

MAX_NAME_BYTES = 
512
MAX_EXCERPT_BYTES = 
4_096

Instance Method Summary collapse

Constructor Details

#initialize(configuration) ⇒ Extractor

Returns a new instance of Extractor.



14
15
16
 
# File 'lib/jekyll/agent_audit/extractor.rb', line 14

def initialize(configuration)
  @configuration = configuration
end

Instance Method Details

#extract(entry) ⇒ Object 



18
19
20
21
22
23
24
25
26
27
28
29
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
 
# File 'lib/jekyll/agent_audit/extractor.rb', line 18

def extract(entry)
  html = entry[:html] || read_artifact(entry[:absolute_path], entry[:destination_root])
  raise InputError, "HTML exceeds max_html_bytes" if html.bytesize > limit(:max_html_bytes, 5 * 1024 * 1024)

  doc = Nokogiri::HTML5.parse(html, nil, "UTF-8")
  content = content_node(doc, entry)
  id_nodes = doc.css("[id]")
  id_index = id_nodes.each_with_object({}) { |node, index| index[node["id"].to_s] ||= node unless node["id"].to_s.empty? }
  result = entry.dup.merge(
    title: compact_text(node_text(doc.at_css("title"))),
    ids: Set.new(id_index.keys), duplicate_ids: duplicate_ids(id_nodes), duplicate_id_locations: duplicate_id_locations(id_nodes),
    links: [], canonicals: [], headings: [], base_href: doc.at_css("head base[href]")&.[]("href"),
    content_basis: content[:basis], content_confident: content[:confident], redirect: redirect?(doc),
    jsonld_scripts: [], metadata_nodes: Hash.new { |h, k| h[k] = [] }, diagnostics: content[:diagnostics]
  )

  doc.css("a[href]").each do |node|
    result[:links] << occurrence(node, accessible_name(node, id_index), :navigation, content[:basis])
  end
  doc.css("[cite]").each do |node|
    href = node["cite"].to_s.strip
    next if href.empty?
    result[:links] << {href: href, line: node.line, name: compact_text(node_text(node)), name_supported: true,
      hidden: hidden?(node), kind: :citation, region: :other}
  end
  doc.css("head link[href]").each do |node|
    rels = node["rel"].to_s.split.map(&:downcase)
    next if rels.include?("canonical")
    kind = rels.include?("citation") ? :citation : (rels.include?("alternate") ? :alternate : nil)
    next unless kind
    result[:links] << {href: node["href"].to_s, line: node.line, name: compact_text(node["title"]),
      name_supported: true, hidden: false, kind: kind, region: :other}
  end

  content[:node]&.css("h1,h2,h3,h4,h5,h6,[role='heading']")&.each do |node|
    next if excluded?(node) || decorative?(node)
    level = node.name =~ /h([1-6])/ ? Regexp.last_match(1).to_i : node["aria-level"].to_i
    next unless (1..6).cover?(level)
    name = accessible_name(node, id_index)
    result[:headings] << {level: level, name: name[:name], name_supported: name[:supported], hidden: hidden?(node), line: node.line}
  end
  doc.css("head link[rel]").each do |node|
    rels = node["rel"].to_s.split.map(&:downcase)
    result[:canonicals] << {href: node["href"], line: node.line} if rels.include?("canonical")
  end
  doc.css("script[type]").each do |node|
    next unless node["type"].to_s.downcase == "application/ld+json"
    raise InputError, "JSON-LD exceeds max_jsonld_bytes" if node.text.bytesize > limit(:max_jsonld_bytes, 1024 * 1024)
    result[:jsonld_scripts] << {text: node.text, line: node.line}
  end
  result[:content_text] = compact_text(content_text(content[:node]), MAX_EXCERPT_BYTES)
  extract_metadata(doc, result)
  result
rescue Errno::ENOENT, Errno::EACCES => e
  raise InputError, e.message
rescue Nokogiri::XML::SyntaxError, ArgumentError => e
  raise InputError, "HTML could not be parsed: #{e.message}"
end