Module: Jekyll::AgentAudit::Rules::Provenance

Defined in:
lib/jekyll/agent_audit/rules/provenance.rb

Constant Summary collapse

ARTICLE_TYPES = 
%w[Article BlogPosting NewsArticle].freeze
INTENT_KEYS = 
{
  published: [[:date], ["date"]],
  modified: [[:seo, :date_modified], ["seo", "date_modified"], [:last_modified_at], ["last_modified_at"]],
  author: [[:author], ["author"]]
}.freeze

Class Method Summary collapse

Class Method Details

.ambiguous_article?(document, context) ⇒ Boolean

Returns:

  • (Boolean) 


226
227
228
229
230
231
232
233
234
235
236
237
238
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 226

def ambiguous_article?(document, context)
  nodes = jsonld_nodes(document)
  articles = nodes.values.select { |node| article?(node) }
  return false if articles.length < 2
  url = document[:url] || document["url"]
  canonical = canonical_urls(document, context)
  articles.count do |article|
    [article["url"], article["mainEntityOfPage"]].any? do |candidate|
      candidate = candidate["@id"] if candidate.is_a?(Hash)
      candidate && [url, *canonical].include?(candidate)
    end
  end != 1
end

.article?(node) ⇒ Boolean

Returns:

  • (Boolean) 


389
390
391
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 389

def article?(node)
  Array(node["@type"]).any? { |type| ARTICLE_TYPES.include?(type) }
end

.author_identity_kind(value) ⇒ Object 



281
282
283
284
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 281

def author_identity_kind(value)
  authors = value.is_a?(Array) ? value : [value]
  authors.any? { |author| author.is_a?(Hash) && (author["@id"] || author["url"]) } ? :id : :name
end

.author_set(value) ⇒ Object 



269
270
271
272
273
274
275
276
277
278
279
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 269

def author_set(value)
  authors = value.is_a?(Array) ? value : [value]
  authors.map do |author|
    author = author[:value] if author.is_a?(Hash) && author.key?(:value)
    if author.is_a?(Hash)
      author["@id"] || author["url"] || normalize_name(author["name"])
    else
      normalize_name(author)
    end
  end.compact.to_set
end

.author_sources(document, context) ⇒ Object 



160
161
162
163
164
165
166
167
168
169
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 160

def author_sources(document, context)
  sources = explicit_sources(document, :author) + metadata_sources(document, :author)
  graph = primary_article(document, context)
  return sources unless graph
  author_reference = graph["author"]
  author = resolved_value(author_reference, document, context)
  entity = author.is_a?(Hash) ? (author["@id"] || author_reference["@id"] || graph["@id"]) : graph["@id"]
  sources << source(author, :jsonld, entity, graph["__audit_line"]) if author
  sources
end

.canonical_urls(document, context) ⇒ Object 



286
287
288
289
290
291
292
293
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 286

def canonical_urls(document, context)
  Array(document[:canonicals] || document["canonicals"]).map do |item|
    href = item[:url] || item["url"] || item[:href] || item["href"]
    next href unless href && context[:resolver]
    resolved = context[:resolver].resolve(href, document)
    resolved.is_a?(Hash) ? (resolved[:url] || resolved["url"]) : nil
  end.compact
end

.collect_nodes(value, result, line_number = nil) ⇒ Object 



250
251
252
253
254
255
256
257
258
259
260
261
262
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 250

def collect_nodes(value, result, line_number = nil)
  case value
  when Array then value.each { |item| collect_nodes(item, result, line_number) }
  when Hash
    if value["@graph"]
      collect_nodes(value["@graph"], result, line_number)
    elsif value["@id"]
      result[value["@id"]] = value.merge("__audit_line" => line_number)
    elsif value["@type"]
      result["__anonymous_#{result.length}"] = value.merge("__audit_line" => line_number)
    end
  end
end

.comparable_date(values, context) ⇒ Object 



321
322
323
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 321

def comparable_date(values, context)
  values.map { |item| next unless item[:explicit]; parsed = parse_date(item[:value], context); parsed && item.merge(interval: parsed) }.compact.first
end

.date_order(document, context) ⇒ Object 



73
74
75
76
77
78
79
80
81
82
83
84
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 73

def date_order(document, context)
  sources = provenance_sources(document, context)
  published = comparable_date(sources[:published], context)
  modified = comparable_date(sources[:modified], context)
  return [] unless published && modified
  return [] unless interval_before?(modified[:interval], published[:interval])

  [draft(document, "Explicit dateModified precedes datePublished for the primary article.", {
    published: published[:value], modified: modified[:value],
    published_origin: published[:origin], modified_origin: modified[:origin]
  }, modified[:line])]
end

.disjoint?(a, b) ⇒ Boolean

Returns:

  • (Boolean) 


325
326
327
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 325

def disjoint?(a, b)
  !overlaps?(a, b)
end

.draft(document, message, observation, line = nil, related_locations = []) ⇒ Object 



406
407
408
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 406

def draft(document, message, observation, line = nil, related_locations = [])
  {document: document, message: message, observation: observation, line: line, related_locations: related_locations}
end

.evaluate(id, document, context) ⇒ Object 



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
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 21

def evaluate(id, document, context)
  return {status: :not_applicable, findings: [], reason: "unknown rule"} unless %w[
    provenance.jsonld_invalid provenance.date_invalid provenance.date_order provenance.metadata_conflict
  ].include?(id)

  return {status: :not_applicable, findings: [], reason: "no document"} unless document

  if %w[provenance.date_order provenance.metadata_conflict].include?(id) && ambiguous_article?(document, context)
    return {status: :skipped, findings: [], reason: "ambiguous primary Article entity"}
  end
  if id == "provenance.date_order"
    dates = provenance_sources(document, context)
    unless dates[:published].any? { |source| source[:explicit] } && dates[:modified].any? { |source| source[:explicit] }
      return {status: :skipped, findings: [], reason: "no comparable explicit publication and modification dates"}
    end
  end
  if id == "provenance.metadata_conflict" && !metadata_comparable?(document, context)
    return {status: :skipped, findings: [], reason: "no comparable provenance metadata pair"}
  end
  findings = case id
  when "provenance.jsonld_invalid" then invalid_jsonld(document)
  when "provenance.date_invalid" then invalid_dates(document, context)
  when "provenance.date_order" then date_order(document, context)
  when "provenance.metadata_conflict" then metadata_conflict(document, context)
  end
  {status: :evaluated, findings: findings}
end

.explicit_sources(document, field) ⇒ Object 



185
186
187
188
189
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 185

def explicit_sources(document, field)
  value, present = intent_value(document[:explicit_data] || document["explicit_data"], field)
  return [] unless present
  [source(value, :explicit, nil)]
end

.infer_seo_rendered_fallback!(values, document, context) ⇒ Object 



150
151
152
153
154
155
156
157
158
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 150

def infer_seo_rendered_fallback!(values, document, context)
  return unless seo_plugin_active?(context) && explicit_sources(document, :modified).empty?
  published = values[:published].map { |item| parse_date(item[:value], context) }.compact
  return if published.empty?
  values[:modified].each do |modified|
    parsed = parse_date(modified[:value], context)
    modified[:explicit] = false if parsed && published.any? { |interval| overlaps?(interval, parsed) }
  end
end

.intent_value(data, field) ⇒ Object 



344
345
346
347
348
349
350
351
352
353
354
355
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 344

def intent_value(data, field)
  return [nil, false] unless data.is_a?(Hash)
  paths = INTENT_KEYS[field] || []
  paths.each do |path|
    current = data
    path.each do |key|
      current = current[key] || current[key.to_s] if current.is_a?(Hash)
    end
    return [current, true] unless current.nil?
  end
  [nil, false]
end

.interval_before?(a, b) ⇒ Boolean

Returns:

  • (Boolean) 


338
339
340
341
342
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 338

def interval_before?(a, b)
  return a[0] < b[0] if a[0] == a[1]
  return a[1] <= b[0] if b[0] == b[1]
  a[1] <= b[0]
end

.invalid_dates(document, context) ⇒ Object 



59
60
61
62
63
64
65
66
67
68
69
70
71
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 59

def invalid_dates(document, context)
  sources = provenance_sources(document, context)
  sources.map do |field, values|
    values.map do |source|
      next unless source[:explicit]
      next if parse_date(source[:value], context)
      next if timezone_free_timestamp?(source[:value], context)

      draft(document, "#{field} contains an unparseable explicit machine date.",
        {field: field, value: source[:value], origin: source[:origin]}, source[:line])
    end.compact
  end.flatten
end

.invalid_jsonld(document) ⇒ Object 



49
50
51
52
53
54
55
56
57
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 49

def invalid_jsonld(document)
  scripts(document).map do |script|
    JSON.parse(script[:text] || script["text"])
    nil
  rescue JSON::ParserError => e
    draft(document, "Invalid JSON in an application/ld+json script.",
      {error: e.message.to_s.gsub(/[\x00-\x1f]/, " ")[0, 200], source: :jsonld}, line(script))
  end.compact
end

.jsonld_nodes(document) ⇒ Object 



240
241
242
243
244
245
246
247
248
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 240

def jsonld_nodes(document)
  scripts(document).each_with_object({}) do |script, result|
    begin
      collect_nodes(JSON.parse(script[:text] || script["text"]), result, line(script))
    rescue JSON::ParserError
      next
    end
  end
end

.line(script) ⇒ Object 



385
386
387
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 385

def line(script)
  script[:line] || script["line"]
end

.locations_for(sources) ⇒ Object 



377
378
379
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 377

def locations_for(sources)
  sources.map { |source| {line: source[:line], entity: source[:entity], origin: source[:origin]} }
end

.metadata_comparable?(document, context) ⇒ Boolean

Returns:

  • (Boolean) 


126
127
128
129
130
131
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 126

def metadata_comparable?(document, context)
  sources = provenance_sources(document, context)
  sources.values.any? { |values| values.count { |source| source[:explicit] } >= 2 } ||
    author_sources(document, context).count { |source| source[:explicit] } >= 2 ||
    publisher_sources(document, context).count { |source| source[:explicit] } >= 2
end

.metadata_conflict(document, context) ⇒ Object 



86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 86

def metadata_conflict(document, context)
  sources = provenance_sources(document, context)
  findings = []
  %i[published modified].each do |field|
    parsed = sources[field].map do |source|
      next unless source[:explicit]
      value = parse_date(source[:value], context)
      value && source.merge(interval: value)
    end.compact
    next if parsed.length < 2
    next unless parsed.combination(2).any? { |a, b| disjoint?(a[:interval], b[:interval]) }

    findings << draft(document, "Explicit #{field} metadata values conflict.", {
      field: field, values: parsed.map { |value| source_observation(value) }
    }, parsed.last[:line], locations_for(parsed))
  end

  author_sets = author_sources(document, context).map do |source|
    ids = author_set(source[:value])
    ids.empty? ? nil : source.merge(set: ids, identity_kind: author_identity_kind(source[:value]))
  end.compact
  if author_sets.length >= 2 && author_sets.map { |source| source[:identity_kind] }.uniq.length == 1 &&
      author_sets.combination(2).any? { |a, b| a[:set] != b[:set] }
    findings << draft(document, "Explicit author metadata values conflict.", {
      field: :author, values: author_sets.map { |source| source_observation(source).merge(authors: source[:set].to_a.sort) }
    }, author_sets.last[:line], locations_for(author_sets))
  end
  publisher_sets = publisher_sources(document, context).map do |source|
    set = author_set(source[:value])
    set.empty? ? nil : source.merge(set: set, identity_kind: author_identity_kind(source[:value]))
  end.compact
  if publisher_sets.length >= 2 && publisher_sets.map { |source| source[:identity_kind] }.uniq.length == 1 &&
      publisher_sets.combination(2).any? { |a, b| a[:set] != b[:set] }
    findings << draft(document, "Explicit publisher metadata values conflict.", {
      field: :publisher, values: publisher_sets.map { |source| source_observation(source).merge(publishers: source[:set].to_a.sort) }
    }, publisher_sets.last[:line], locations_for(publisher_sets))
  end
  findings
end

.metadata_sources(document, field) ⇒ Object 



191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 191

def metadata_sources(document, field)
  nodes = document[:metadata_nodes] || document["metadata_nodes"] || {}
  candidates = Array(nodes[field] || nodes[field.to_s])
  if field == :author && candidates.length > 1
    values = candidates.map { |candidate| candidate.is_a?(Hash) ? (candidate[:value] || candidate["value"]) : candidate }
    first = candidates.first
    return [source(values, first.is_a?(Hash) ? (first[:origin] || first["origin"] || :rendered) : :rendered,
      nil, first.is_a?(Hash) ? (first[:line] || first["line"]) : nil)]
  end
  candidates.map do |candidate|
    candidate = {value: candidate} unless candidate.is_a?(Hash)
    source(candidate[:value] || candidate["value"], candidate[:origin] || candidate["origin"] || :rendered, nil,
      candidate[:line] || candidate["line"]).tap do |entry|
        entry[:entity] = candidate[:entity] || candidate["entity"] if entry
      end
  end.compact
end

.normalize_name(value) ⇒ Object 



393
394
395
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 393

def normalize_name(value)
  value.to_s.strip.gsub(/\s+/, " ") unless value.nil?
end

.overlaps?(a, b) ⇒ Boolean

Returns:

  • (Boolean) 


329
330
331
332
333
334
335
336
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 329

def overlaps?(a, b)
  a_point = a[0] == a[1]
  b_point = b[0] == b[1]
  return a[0] == b[0] if a_point && b_point
  return a[0] >= b[0] && a[0] < b[1] if a_point
  return b[0] >= a[0] && b[0] < a[1] if b_point
  a[0] < b[1] && b[0] < a[1]
end

.parse_date(value, context) ⇒ Object 



295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 295

def parse_date(value, context)
  return nil unless value.is_a?(String) || value.is_a?(Date) || value.is_a?(Time) || value.is_a?(DateTime)
  if value.is_a?(Date) && !value.is_a?(DateTime)
    next_date = value + 1
    return [Time.utc(value.year, value.month, value.day), Time.utc(next_date.year, next_date.month, next_date.day)]
  end
  if value.is_a?(String) && value.match?(/\A\d{4}-\d{2}-\d{2}\z/)
    date = Date.parse(value)
    next_date = date + 1
    return [Time.utc(date.year, date.month, date.day), Time.utc(next_date.year, next_date.month, next_date.day)]
  end
  return [value.utc, value.utc] if value.respond_to?(:utc)
  return nil unless value.is_a?(String) && value.match?(/\A\d{4}-\d{2}-\d{2}T/)
  return nil if value.match?(/T[^Z+\-]*\z/) && timezone(context).nil?
  timestamp = value
  if timestamp.match?(/T[^Z+\-]*\z/)
    offset = timezone(context)
    return nil unless offset.is_a?(String) && offset.match?(/\A[+-]\d{2}:?\d{2}\z/)
    timestamp = "#{timestamp}#{offset}"
  end
  instant = Time.iso8601(timestamp)
  [instant, instant]
rescue ArgumentError, Date::Error
  nil
end

.primary_article(document, context) ⇒ Object 



209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 209

def primary_article(document, context)
  nodes = jsonld_nodes(document)
  articles = nodes.values.select { |node| article?(node) }
  return nil if articles.empty?
  url = document[:url] || document["url"]
  canonical = canonical_urls(document, context)
  matches = articles.select do |article|
    [article["url"], article["mainEntityOfPage"]].any? do |candidate|
      candidate = candidate["@id"] if candidate.is_a?(Hash)
      candidate && [url, *canonical].compact.include?(candidate)
    end
  end
  return matches.first if matches.length == 1
  return nil if articles.length == 1 && (articles.first["url"] || articles.first["mainEntityOfPage"])
  articles.length == 1 ? articles.first : nil
end

.provenance_sources(document, context) ⇒ Object 



133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 133

def provenance_sources(document, context)
  graph = primary_article(document, context)
  values = {published: [], modified: []}
  values[:published].concat(explicit_sources(document, :published))
  values[:modified].concat(explicit_sources(document, :modified))
  values[:published].concat(metadata_sources(document, :published))
  values[:modified].concat(metadata_sources(document, :modified))
  if graph
    values[:published] << source(graph["datePublished"], :jsonld, graph["@id"], graph["__audit_line"])
    modified = source(graph["dateModified"], :jsonld, graph["@id"], graph["__audit_line"])
    modified[:explicit] = false if modified && seo_fallback_date?(graph, document, context)
    values[:modified] << modified
  end
  infer_seo_rendered_fallback!(values, document, context)
  values.transform_values { |items| items.compact }
end

.publisher_sources(document, context) ⇒ Object 



171
172
173
174
175
176
177
178
179
180
181
182
183
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 171

def publisher_sources(document, context)
  sources = metadata_sources(document, :publisher)
  graph = primary_article(document, context)
  return sources unless graph
  publisher_reference = graph["publisher"]
  publisher = resolved_value(publisher_reference, document, context)
  if publisher
    value = publisher.is_a?(Hash) && publisher["name"] ? publisher["name"] : publisher
    entity = publisher.is_a?(Hash) ? (publisher["@id"] || publisher_reference["@id"] || graph["@id"]) : graph["@id"]
    sources << source(value, :jsonld, entity, graph["__audit_line"])
  end
  sources
end

.resolved_value(value, document, context) ⇒ Object 



264
265
266
267
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 264

def resolved_value(value, document, context)
  return value unless value.is_a?(Hash) && value["@id"]
  jsonld_nodes(document)[value["@id"]]
end

.scripts(document) ⇒ Object 



381
382
383
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 381

def scripts(document)
  Array(document[:jsonld_scripts] || document["jsonld_scripts"])
end

.seo_fallback_date?(graph, document, context) ⇒ Boolean

Returns:

  • (Boolean) 


362
363
364
365
366
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 362

def seo_fallback_date?(graph, document, context)
  return false unless seo_plugin_active?(context)
  return false if explicit_sources(document, :modified).any?
  graph["datePublished"] && graph["dateModified"] && graph["datePublished"] == graph["dateModified"]
end

.seo_plugin_active?(context) ⇒ Boolean

Returns:

  • (Boolean) 


368
369
370
371
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 368

def seo_plugin_active?(context)
  plugins = context[:plugins] || context.dig(:site_config, :plugins) || context.dig(:site_config, "plugins") || []
  Array(plugins).any? { |plugin| plugin.to_s == "jekyll-seo-tag" }
end

.source(value, origin, entity, line = nil) ⇒ Object 



357
358
359
360
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 357

def source(value, origin, entity, line = nil)
  return nil if value.nil?
  {value: value, origin: origin, entity: entity, line: line, explicit: true}
end

.source_observation(source) ⇒ Object 



373
374
375
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 373

def source_observation(source)
  {value: source[:value], origin: source[:origin], entity: source[:entity], line: source[:line]}
end

.timezone(context) ⇒ Object 



397
398
399
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 397

def timezone(context)
  context.dig(:site_config, :timezone) || context.dig(:site_config, "timezone")
end

.timezone_free_timestamp?(value, context) ⇒ Boolean

Returns:

  • (Boolean) 


401
402
403
404
 
# File 'lib/jekyll/agent_audit/rules/provenance.rb', line 401

def timezone_free_timestamp?(value, context)
  value.is_a?(String) && value.match?(/\A\d{4}-\d{2}-\d{2}T/) &&
    value.match?(/T[^Z+\-]*\z/) && timezone(context).nil?
end