Class: Relaton::Bipm::SiBrochureParser

Inherits:
Object
  • Object
show all
Defined in:
lib/relaton/bipm/si_brochure_parser.rb

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(data_fetcher) ⇒ SiBrochureParser

Create new parser

Parameters:



11
12
13
 
# File 'lib/relaton/bipm/si_brochure_parser.rb', line 11

def initialize(data_fetcher)
  @data_fetcher = WeakRef.new data_fetcher
end

Class Method Details

.parse(data_fetcher) ⇒ Object

Parse documents from SI brochure dataset and write thems to YAML files

Parameters:



20
21
22
 
# File 'lib/relaton/bipm/si_brochure_parser.rb', line 20

def self.parse(data_fetcher)
  new(data_fetcher).parse
end

Instance Method Details

#deep_merge(hash1, hash2) ⇒ Hash

Deep merge two hashes

Parameters:

  • hash1 (Hash)
  • hash2 (Hash)

Returns:

  • (Hash) —

    Merged hash



146
147
148
149
150
151
152
153
154
155
156
 
# File 'lib/relaton/bipm/si_brochure_parser.rb', line 146

def deep_merge(hash1, hash2) # rubocop:disable Metrics/PerceivedComplexity, Metrics/CyclomaticComplexity
  hash1.merge(hash2) do |_, oldval, newval|
    if oldval.is_a?(Hash) && newval.is_a?(Hash)
      deep_merge(oldval, newval)
    elsif oldval.is_a?(Array) && newval.is_a?(Array)
      (oldval + newval).uniq { |i| downcase_all i }
    else
      newval || oldval
    end
  end
end

#downcase_all(content) ⇒ Array, ...

Downcase all values in hash or array

Parameters:

  • content (Array, Hash, String) —

    hash, array or string

Returns:

  • (Array, Hash, String) —

    hash, array or string with downcased values



165
166
167
168
169
170
171
172
 
# File 'lib/relaton/bipm/si_brochure_parser.rb', line 165

def downcase_all(content)
  case content
  when Hash then content.transform_values { |v| downcase_all v }
  when Array then content.map { |v| downcase_all v }
  when String then content.downcase
  else content
  end
end

#extract_editorialgroup(xml) ⇒ Object 



117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
 
# File 'lib/relaton/bipm/si_brochure_parser.rb', line 117

def extract_editorialgroup(xml)
  doc = Nokogiri::XML(xml)
  doc.xpath("//editorialgroup/committee").map do |committee|
    acronym = committee["acronym"]
    names = committee.xpath("variant").map do |v|
      Relaton::Bib::TypedLocalizedString.new(
        content: v.text, language: v["language"], script: v["script"],
      )
    end
    subdiv = Relaton::Bib::Subdivision.new(
      type: "committee", name: names,
      abbreviation: Relaton::Bib::LocalizedString.new(content: acronym),
    )
    bipm_name = [Relaton::Bib::TypedLocalizedString.new(content: acronym)]
    org = Relaton::Bib::Organization.new(name: bipm_name, subdivision: [subdiv])
    desc = Relaton::Bib::LocalizedMarkedUpString.new(content: "committee")
    role = Relaton::Bib::Contributor::Role.new(type: "author", description: [desc])
    Relaton::Bib::Contributor.new(organization: org, role: [role])
  end
end

#fix_si_brochure_id(item) ⇒ void

This method returns an undefined value.

Update ID of SI brochure

Parameters:

  • item (ItemData) —

    bibliographic item



81
82
83
84
85
86
87
88
89
90
91
92
93
94
 
# File 'lib/relaton/bipm/si_brochure_parser.rb', line 81

def fix_si_brochure_id(item)
  # isbn = hash["docid"].detect { |id| id["type"] == "ISBN" }
  # num = isbn && isbn["id"] == "978-92-822-2272-0" ?  "SI Brochure" : "SI Brochure, Appendix 4"

  update_id item

  prid = primary_id item
  if item.docnumber
    item.docnumber.sub!(/^Brochure(?:\sConcise|\sFAQ)?$/i, prid.sub(/^BIPM\s/, ""))
  else
    item.docnumber = prid.sub(/^BIPM\s/, "")
  end
  item.id = prid.gsub(/[,\s-]/, "")
end

#has_committee_contributor?(item) ⇒ Boolean

Returns:

  • (Boolean) 


111
112
113
114
115
 
# File 'lib/relaton/bipm/si_brochure_parser.rb', line 111

def has_committee_contributor?(item)
  item.contributor.any? do |c|
    c.role.any? { |r| r.description.any? { |d| d.content == "committee" } }
  end
end

#parse ⇒ Object

Parse SI brochure and write them to YAML files



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
 
# File 'lib/relaton/bipm/si_brochure_parser.rb', line 27

def parse # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
  # puts "Parsing SI brochure..."
  # puts "Ls #{Dir['*']}"
  # puts "Ls #{Dir['bipm-si-brochure/*']}"
  # puts "Ls #{Dir['bipm-si-brochure/site/*']}"
  # puts "Ls #{Dir['bipm-si-brochure/site/documents/*']}"
  Dir["bipm-si-brochure/_site/documents/*.rxl"].each do |f|
    puts "Parsing #{f}"
    xml = File.read(f, encoding: "UTF-8")
    xml = xml.force_encoding("UTF-8") if xml.encoding != Encoding::UTF_8
    item1 = Bibdata.from_xml(xml)
    # Workaround for relaton-bib Version#content bug: whitespace between
    # legacy <revision-date>/<draft> children gets captured as @content,
    # blocking the legacy-fold path. Clear it so the getter recomputes.
    item1.version.each do |v|
      c = v.instance_variable_get(:@content)
      next unless c.is_a?(Array) || (c.is_a?(String) && c.strip.empty?)
      v.instance_variable_set(:@content, nil)
    end
    @data_fetcher.errors[:si_brochure_title] &&= item1.title.empty?
    @data_fetcher.errors[:si_brochure_docidentifier] &&= item1.docidentifier.empty?
    unless has_committee_contributor?(item1)
      contribs = extract_editorialgroup(xml)
      contribs.each { |c| item1.contributor << c }
    end
    fix_si_brochure_id item1
    basename = File.join @data_fetcher.output, File.basename(f).sub(/(?:-(?:en|fr))?\.rxl$/, "")
    outfile = "#{basename}.#{@data_fetcher.ext}"
    key = item1.docnumber || basename
    @data_fetcher.index.add_or_update Id.new.parse(key).to_hash, outfile
    item =
      if File.exist? outfile
        warn_duplicate = false
        item2 = Item.from_yaml File.read(outfile, encoding: "UTF-8")
        fix_si_brochure_id item2
        hash1 = YAML.safe_load item1.to_yaml
        hash2 = YAML.safe_load item2.to_yaml
        Item.from_yaml deep_merge(hash1, hash2).to_yaml
      else
        warn_duplicate = true
        item1
      end
    @data_fetcher.write_file outfile, item, warn_duplicate: warn_duplicate
    puts "Saved to #{outfile}"
  end
end

#primary_id(item) ⇒ Object 



105
106
107
108
109
 
# File 'lib/relaton/bipm/si_brochure_parser.rb', line 105

def primary_id(item)
  item.docidentifier.detect do |id|
    id.primary && (id.language == "en" || id.language.nil?)
  end.content
end

#update_id(item) ⇒ Object 



96
97
98
99
100
101
102
103
 
# File 'lib/relaton/bipm/si_brochure_parser.rb', line 96

def update_id(item)
  item.docidentifier.each do |id|
    next unless id.type == "BIPM" && id.content&.match?(/BIPM Brochure/i)

    id.primary = true
    id.content.sub!(/(?<=^BIPM\s)(Brochure)/i, "SI \\1")
  end
end