Class: Relaton::Oasis::DataParser

Inherits:
Object
  • Object
show all
Includes:
DataParserUtils
Defined in:
lib/relaton/oasis/data_parser.rb

Overview

Parser for OASIS document.

Constant Summary

Constants included from DataParserUtils

Relaton::Oasis::DataParserUtils::RETRIABLE_PAGE_ERRORS

Instance Method Summary collapse

Methods included from DataParserUtils

#create_contribution_info, #create_ext, #create_person, #decode_cf_email, #page, #parse_chairs, #parse_contributor, #parse_docid, #parse_doctype, #parse_editors, #parse_editors_from_text, #parse_errata, #parse_part, #parse_spec, #person_affiliation, #person_email, #publisher_oasis, #retry_page

Constructor Details

#initialize(node, errors = {}, agent: nil) ⇒ DataParser

Initialize parser.

Parameters:

  • node (Nokogiri::HTML::Element) —

    document node



12
13
14
15
16
 
# File 'lib/relaton/oasis/data_parser.rb', line 12

def initialize(node, errors = {}, agent: nil)
  @node = node
  @errors = errors
  @agent = agent
end

Instance Method Details

#document_part_refs ⇒ Array<String>

Look for “Cite as” references.

Returns:

  • (Array<String>) —

    document part references



178
179
180
181
182
183
 
# File 'lib/relaton/oasis/data_parser.rb', line 178

def document_part_refs
  @node.css(
    ".standard__grid--cite-as > p > strong",
    "span.Refterm", "span.abbrev", "span.citationLabel > strong"
  ).map { |p| p.text.gsub(/^\[{1,2}|\]$/, "").strip }
end

#link_node ⇒ Object 



143
144
145
146
147
148
 
# File 'lib/relaton/oasis/data_parser.rb', line 143

def link_node
  xpath = "./div/div/div[contains(@class, " \
          "'standard__grid--cite-as')]" \
          "/p[strong or span/strong]/a"
  @link_node ||= @node.at(xpath)
end

#links ⇒ Object 



208
209
210
211
212
 
# File 'lib/relaton/oasis/data_parser.rb', line 208

def links
  l = @node.xpath("./div/div/div[1]/p[1]/a[@href]")
  l = @node.xpath("./div/div/div[1]/p[2]/a[@href]") if l.empty?
  l
end

#parse ⇒ ItemData

Parse document.

Returns:

  • (ItemData) —

    bibliographic item data



34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
 
# File 'lib/relaton/oasis/data_parser.rb', line 34

def parse # rubocop:disable Metrics/MethodLength
  ItemData.new(
    type: "standard",
    title: parse_title,
    docidentifier: parse_docid,
    source: parse_link,
    docnumber: parse_docnumber,
    date: parse_date,
    contributor: parse_contributor,
    abstract: parse_abstract,
    language: ["en"],
    script: ["Latn"],
    relation: parse_relation,
    ext: create_ext,
  )
end

#parse_abstract ⇒ Array<Bib::LocalizedMarkedUpString>

Parse abstract.

Returns:

  • (Array<Bib::LocalizedMarkedUpString>) —

    abstract



84
85
86
87
88
89
90
91
92
93
94
95
96
97
 
# File 'lib/relaton/oasis/data_parser.rb', line 84

def parse_abstract
  c = @node.xpath(
    "./summary/div/div[@class='standard__description']/p",
  ).map { |a| a.text.gsub(/[\n\t]+/, " ").strip }.join("\n")
  result = if c.empty?
             []
           else
             [Bib::Abstract.new(
               content: c, language: "en", script: "Latn",
             )]
           end
  @errors[:abstract] &&= result.empty?
  result
end

#parse_authorizer ⇒ Object

rubocop:disable Metrics/MethodLength



128
129
130
131
132
133
134
135
136
137
138
139
140
141
 
# File 'lib/relaton/oasis/data_parser.rb', line 128

def parse_authorizer # rubocop:disable Metrics/MethodLength
  result = @node.xpath("./div[@class='standard__details']/a").map do |a|
    org = Bib::Organization.new(
      name: [Bib::TypedLocalizedString.new(content: a.text.strip)],
      uri: [Bib::Uri.new(type: "uri", content: a[:href])],
    )
    desc = [Bib::LocalizedMarkedUpString.new(content: "Committee")]
    role = Bib::Contributor::Role.new(type: "authorizer",
                                      description: desc)
    Bib::Contributor.new(organization: org, role: [role])
  end
  @errors[:authorizer] &&= result.empty?
  result
end

#parse_date ⇒ Array<Bib::Date>

Parse date.

Returns:

  • (Array<Bib::Date>) —

    date



68
69
70
71
72
73
74
75
76
77
 
# File 'lib/relaton/oasis/data_parser.rb', line 68

def parse_date
  xpath = "./summary/div/time[@class='standard__date']"
  result = @node.xpath(xpath).map do |d|
    date_str = d.text.match(/\d{2}\s\w+\s\d{4}/).to_s
    date = Date.parse(date_str).to_s
    Bib::Date.new(at: date, type: "issued")
  end
  @errors[:date] &&= result.empty?
  result
end

#parse_docnumber ⇒ String

Parse document number.

Returns:

  • (String) —

    document number



219
220
221
222
223
224
225
226
227
228
229
230
 
# File 'lib/relaton/oasis/data_parser.rb', line 219

def parse_docnumber
  parts = document_part_refs
  result = case parts.size
           when 0
             txt = @node.at("./summary/div/h2").text
             parse_spec title_to_docid(txt)
           when 1 then parse_part parse_spec(parts[0])
           else parts_to_docid parts
           end
  @errors[:docnumber] &&= result.nil?
  result
end

#parse_editorialgroup_contributor ⇒ Array<Bib::Contributor>

Parse editorial group as contributors.

Returns:

  • (Array<Bib::Contributor>) —

    editorial group contributors



104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
 
# File 'lib/relaton/oasis/data_parser.rb', line 104

def parse_editorialgroup_contributor # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
  tcs = @node.xpath("./div[@class='standard__details']/a")
  if tcs.empty?
    result = []
  else
    subdivisions = tcs.map do |a|
      name = [Bib::TypedLocalizedString.new(content: a.text.strip)]
      Bib::Subdivision.new(type: "technical-committee",
                           name: name)
    end
    org = Bib::Organization.new(
      name: [Bib::TypedLocalizedString.new(content: "OASIS")],
      subdivision: subdivisions,
    )
    desc = [Bib::LocalizedMarkedUpString.new(content: "committee")]
    role = Bib::Contributor::Role.new(
      type: "author", description: desc,
    )
    result = [Bib::Contributor.new(organization: org, role: [role])]
  end
  @errors[:editorialgroup_contributor] &&= result.empty?
  result
end

#parse_link ⇒ Object

rubocop:disable Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/MethodLength, Metrics/PerceivedComplexity



185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
 
# File 'lib/relaton/oasis/data_parser.rb', line 185

def parse_link # rubocop:disable Metrics/AbcSize, Metrics/CyclomaticComplexity, Metrics/MethodLength, Metrics/PerceivedComplexity
  result = if parts.size > 1
             []
           else
             links.map do |l|
               type = l[:href].match(/\.(\w+)$/)&.captures&.first
               type ||= "src"
               type.sub!("docx", "doc")
               type.sub!("html", "src")
               Bib::Uri.new(type: type, content: l[:href])
             end
           end
  @errors[:link] &&= result.empty?
  result
end

#parse_relation ⇒ Array<Bib::Relation>

Parse relation.

Returns:

  • (Array<Bib::Relation>) —

    relation



155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
 
# File 'lib/relaton/oasis/data_parser.rb', line 155

def parse_relation # rubocop:disable Metrics/MethodLength
  xpath = "./div/div/div[contains(@class, " \
          "'standard__grid--cite-as')]" \
          "/p[strong or span/strong or b/span]"
  rels = @node.xpath(xpath)
  result = if rels.size > 1
             rels.map do |r|
               docid = DataPartParser.new(r, @errors, agent: @agent).parse_docid
               bib = ItemData.new(formattedref: Bib::Formattedref.new(content: docid[0].content))
               Bib::Relation.new(type: "hasPart", bibitem: bib)
             end
           else
             []
           end
  @errors[:relation] &&= result.empty?
  result
end

#parse_technology_area ⇒ Array<String>

Parse technology areas.

Returns:

  • (Array<String>) —

    technology areas



298
299
300
301
302
 
# File 'lib/relaton/oasis/data_parser.rb', line 298

def parse_technology_area
  result = super(@node)
  @errors[:technology_area] &&= result.empty?
  result
end

#parse_title ⇒ Array<Bib::Title>

Parse title.

Returns:

  • (Array<Bib::Title>) —

    title



56
57
58
59
60
61
 
# File 'lib/relaton/oasis/data_parser.rb', line 56

def parse_title
  result = [Bib::Title.new(type: "main", content: title, language: "en",
                           script: "Latn")]
  @errors[:title] &&= result.empty?
  result
end

#parts ⇒ Object 



201
202
203
204
205
206
 
# File 'lib/relaton/oasis/data_parser.rb', line 201

def parts
  xpath = "./div/div/div[contains(@class, " \
          "'standard__grid--cite-as')]" \
          "/p[strong or span/strong]"
  @parts ||= @node.xpath(xpath)
end

#parts_to_docid(parts) ⇒ String

Create document identifier from parts references.

Parameters:

  • parts (Array<String>) —

    parts references

Returns:

  • (String) —

    document identifier



239
240
241
242
243
244
245
246
247
248
249
250
 
# File 'lib/relaton/oasis/data_parser.rb', line 239

def parts_to_docid(parts) # rubocop:disable Metrics/AbcSize
  id = parts[1..].each_with_object(parts[0].split("-")) do |part, acc|
    chunks = part.split "-"
    chunks.each.with_index do |chunk, idx|
      unless chunk.casecmp(acc[idx])&.zero?
        acc.slice!(idx..-1)
        break
      end
    end
  end.join("-")
  parse_part parse_spec(id)
end

#text ⇒ Object 



22
23
24
25
26
27
 
# File 'lib/relaton/oasis/data_parser.rb', line 22

def text
  xpath = "./div/div/div[contains(@class, " \
          "'standard__grid--cite-as')]" \
          "/p[em or i or a or span]"
  @text ||= @node.at(xpath)&.text&.strip
end

#title ⇒ Object 



18
19
20
 
# File 'lib/relaton/oasis/data_parser.rb', line 18

def title
  @title ||= @node.at("./summary/div/h2").text
end

#title_to_docid(title) ⇒ String

Create document identifier from title.

Parameters:

  • title (String) —

    title

Returns:

  • (String) —

    document identifier



259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
 
# File 'lib/relaton/oasis/data_parser.rb', line 259

def title_to_docid(title) # rubocop:disable Metrics/AbcSize, Metrics/MethodLength, Metrics/CyclomaticComplexity, Metrics/PerceivedComplexity
  abbrs = title.scan(/(?<=\()[^)]+(?=\))/)
  if abbrs.any?
    id = abbrs.map { |abbr| abbr.split.join("-") }.join "-"
    /(?:Version\s|v)(?<ver>[\d.]+)/ =~ title
    id += "-v#{ver}" if ver
    /(?<eb>ebXML|ebMS)/ =~ title
    id = "#{eb}-#{id}" if eb
    id
  else
    series_end = false
    title.sub(/\s\[OASIS\s\d+\]$/,
              "").split(/[,:]?\s|-|(?<=[a-z])(?=[A-Z][a-z])/)
      .each_with_object([""]) do |word, acc|
        if word =~ /^v[\d.]+/
          acc << $MATCH.to_s
          series_end = true
        elsif word.match?(/^Version/)
          acc << "v"
          series_end = false
        elsif word.match?(/^\d|ebXML|ebMS/)
          series_end ? acc << word : acc[-1] += word
          series_end = true
        elsif word.match?(/^\w+$/) && word == word.upcase
          series_end ? acc << word : acc[-1] = word
          series_end = true
        elsif word.match?(/[A-Z]+[a-z]+/)
          series_end ? acc << word[0] : acc[-1] += word[0]
          series_end = false
        end
    end.join "-"
  end
end