Class: Metanorma::Standoc::Cleanup::SpansToBibitem

Inherits:
Object
  • Object
show all
Includes:
Utils
Defined in:
lib/metanorma/cleanup/spans_to_bibitem.rb,
lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb

Constant Summary

Constants included from Utils

Utils::SECTION_CONTAINERS, Utils::SUBCLAUSE_XPATH

Instance Attribute Summary collapse

Instance Method Summary collapse

Methods included from Utils

#add_id, #add_id_text, #add_noko_elem, #adoc2xml, #asciimath_key, #attr_code, #complete_and_compare_dates, #complete_iso_date, #complete_year_month, #complete_year_only, #csv_split, #dl_to_attrs, #dl_to_elems, #document_ns_attributes, #grkletters, #insert_before, #isodoc, #kv_parse, #link_unwrap, #noko, #parse_complete_date, #parse_partial_date, #processor, #quoted_csv_split, #refid?, #section_containers, #separate_numbering_footnotes, #term_expr, #textcleanup, #to_xml, #wrap_in_para, #xml_encode

Constructor Details

#initialize(bib) ⇒ SpansToBibitem

Returns a new instance of SpansToBibitem.



11
12
13
14
15
16
17
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 11

def initialize(bib)
  @bib = bib
  @err = []
  @spans = spans_preprocess(extract_spans(bib))
  ids = spans_preprocess(extract_docid(bib))
  @spans[:docid] = override_docids(ids[:docid], @spans[:docid])
end

Instance Attribute Details

#err ⇒ Object (readonly)

Returns the value of attribute err.



9
10
11
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 9

def err
  @err
end

#out ⇒ Object (readonly)

Returns the value of attribute out.



9
10
11
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 9

def out
  @out
end

Instance Method Details

#convert ⇒ Object 



36
37
38
39
40
41
42
43
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 36

def convert
  ret = spans_to_bibitem(@spans)
  @out = Nokogiri::XML("<bibitem>#{ret}</bibitem>").root
  %i(type language script locale).each do |k|
    @spans[k] and @out[k.to_s] = @spans[k]
  end
  self
end

#empty_span_hash ⇒ Object 



32
33
34
35
 
# File 'lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb', line 32

def empty_span_hash
  { contrib: [], docid: [], uri: [], date: [], classification: [],
    keyword: [], image: [], note: [], extent: {}, in: {} }
end

#extract_docid(bib) ⇒ Object 



25
26
27
28
29
30
 
# File 'lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb', line 25

def extract_docid(bib)
  bib.xpath("./docidentifier").each_with_object([]) do |d, m|
    m << { key: "docid", type: d["type"], val: d.text }
    d.remove unless bib.at("./title")
  end
end

#extract_spans(bib) ⇒ Object 



5
6
7
8
9
10
11
12
13
14
15
 
# File 'lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb', line 5

def extract_spans(bib)
  ret = bib.xpath("./formattedref//span").each_with_object([]) do |s, m|
    s.at("./ancestor::span") and next
    extract_spans1(s, m)
  end
  bib.xpath("./formattedref//image").each do |i|
    i.delete("id")
    ret << { key: "image", type: nil, val: i.remove.to_xml }
  end
  ret
end

#extract_spans1(span, acc) ⇒ Object 



17
18
19
20
21
22
23
 
# File 'lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb', line 17

def extract_spans1(span, acc)
  keys = span["class"].split(".", 2)
  acc << { key: keys[0], type: keys[1],
           val: span.children.to_xml }
  (span["class"] == "type" and span.remove) or
    span.replace(span.children)
end

#host_rearrange(ret) ⇒ Object 



99
100
101
102
103
104
105
106
107
108
109
 
# File 'lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb', line 99

def host_rearrange(ret)
  ret[:in][:title] or return ret
  ret[:in].merge!(empty_span_hash, { type: "misc" }) { |_, o, _| o }
  %i(series).each do |k|
    ret[:in][k] = ret[k]
    ret.delete(k)
  end
  /^in/.match?(ret[:type]) and
    ret[:in][:type] = ret[:type].sub(/^in/, "")
  ret
end

#multiple_givennames?(span, contrib) ⇒ Boolean

Returns:

  • (Boolean) 


129
130
131
132
133
134
135
136
137
138
 
# File 'lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb', line 129

def multiple_givennames?(span, contrib)
  (%w(formatted-initials givenname).include?(span[:key]) &&
    (contrib[-1][:"formatted-initials"] || contrib[-1][:givenname])) or
    return false
  if contrib[-1][:"formatted-initials"]
    contrib[-1][:givenname] = contrib[-1][:"formatted-initials"]
    contrib[-1].delete(:"formatted-initials")
  end
  true
end

#override_docids(old, new) ⇒ Object

override old values with new values if type is the same comparison is case-insensitive if types differ in case, use the old value's type, not the new



22
23
24
25
26
27
28
29
30
31
32
33
34
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 22

def override_docids(old, new)
  ret = new
  keys = new.map { |a| a[:type]&.upcase }
  old.each do |e|
    if keys.include?(e[:type]&.upcase)
      ret.each do |a|
        a[:type]&.upcase == e[:type]&.upcase and a[:type] = e[:type]
      end
    else ret << e
    end
  end
  ret
end

#span_preprocess1(span, ret) ⇒ Object 



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
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
 
# File 'lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb', line 43

def span_preprocess1(span, ret)
  case span[:key]
  when "uri", "docid", "classification", "keyword"
    val = link_unwrap(Nokogiri::XML.fragment(span[:val])).to_xml
    ret[span[:key].to_sym] << { type: span[:type], val: }
  when "date"
    ret[span[:key].to_sym] << { type: span[:type] || "published",
                                val: span[:val] }
  when "pages", "volume", "issue"
    ret[:extent][span[:key].to_sym] ||= []
    ret[:extent][span[:key].to_sym] << span[:val]
  when "pubplace", "title", "type", "series", "edition", "version",
    "abstract", "language", "script", "locale"
    ret[span[:key].to_sym] = span[:val]
  when "image", "note"
    ret[span[:key].to_sym] << { type: span[:type], val: span[:val] }
  when "in_title"
    ret[:in][:title] = span[:val]
  when "publisher"
    ret[:contrib] << { role: "publisher", entity: "organization",
                       name: span[:val] }
  when "surname", "initials", "givenname", "formatted-initials"
    ret[:contrib] = spans_preprocess_contrib(span, ret[:contrib])
  when "fullname"
    ret[:contrib] = spans_preprocess_fullname(span, ret[:contrib])
  when "organization"
    ret[:contrib] = spans_preprocess_org(span, ret[:contrib])
  when "in_surname", "in_initials", "in_givenname",
    "in_formatted-initials"
    ret[:in][:contrib] ||= []
    span[:key].sub!(/^in_/, "")
    ret[:in][:contrib] =
      spans_preprocess_contrib(span, ret[:in][:contrib])
  when "in_fullname"
    ret[:in][:contrib] ||= []
    span[:key].sub!(/^in_/, "")
    ret[:in][:contrib] =
      spans_preprocess_fullname(span, ret[:in][:contrib])
  when "in_organization"
    ret[:in][:contrib] ||= []
    span[:key].sub!(/^in_/, "")
    ret[:in][:contrib] =
      spans_preprocess_org(span, ret[:in][:contrib])
  else
    msg = "unrecognised key '#{span[:key]}' in " \
          "`span:#{span[:key]}[#{span[:val]}]`"
    @err << { msg: }
  end
end

#span_to_contrib(span, title) ⇒ Object 



163
164
165
166
167
168
169
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 163

def span_to_contrib(span, title)
  e = if span[:entity] == "organization"
        "<organization><name>#{span[:name]}</name></organization>"
      else span_to_person(span, title)
      end
  "<contributor><role type='#{span[:role]}'/>#{e}</contributor>"
end

#span_to_date(span) ⇒ Object 



145
146
147
148
149
150
151
152
153
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 145

def span_to_date(span)
  val = if /[-–](?=\d{4})/.match?(span[:val])
          from, to = span[:val].split(/[-–](?=\d{4})/, 2)
          "<from>#{from}</from><to>#{to}</to>"
        else "<on>#{span[:val]}</on>"
        end
  type = span[:type] ? " type='#{span[:type]}'" : ""
  "<date#{type}>#{val}</date>"
end

#span_to_docid(span, key) ⇒ Object 



138
139
140
141
142
143
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 138

def span_to_docid(span, key)
  if span[:type]
    "<#{key} type='#{span[:type]}'>#{span[:val]}</#{key}>"
  else "<#{key}>#{span[:val]}</#{key}>"
  end
end

#span_to_extent(span, key) ⇒ Object 



129
130
131
132
133
134
135
136
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 129

def span_to_extent(span, key)
  values = span.split(/[-–]/)
  ret = "<locality type='#{key}'>" \
        "<referenceFrom>#{values[0]}</referenceFrom>"
  values[1] and
    ret += "<referenceTo>#{values[1]}</referenceTo>"
  "#{ret}</locality>"
end

#span_to_person(span, title) ⇒ Object 



178
179
180
181
182
183
184
185
186
187
188
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 178

def span_to_person(span, title)
  validate_span_to_person(span, title)
  pre = (span[:"formatted-initials"] and
               "<formatted-initials>" \
               "#{span[:"formatted-initials"]}</formatted-initials>") ||
    Array(span[:givenname]).map do |x|
      "<forename>#{x}</forename>"
    end.join
  "<person><name>#{pre}<surname>#{span[:surname]}</surname></name>" \
    "</person>"
end

#spans_defaults(spans) ⇒ Object 



93
94
95
96
97
 
# File 'lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb', line 93

def spans_defaults(spans)
  spans[:language] && !spans[:script] and
    spans[:script] = ::Metanorma::Utils.default_script(spans[:language])
  spans
end

#spans_preprocess(spans) ⇒ Object 



37
38
39
40
41
 
# File 'lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb', line 37

def spans_preprocess(spans)
  ret = empty_span_hash
  spans.each { |s| span_preprocess1(s, ret) }
  spans_defaults(host_rearrange(ret))
end

#spans_preprocess_contrib(span, contrib) ⇒ Object 



111
112
113
114
115
116
117
118
119
120
121
 
# File 'lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb', line 111

def spans_preprocess_contrib(span, contrib)
  span[:key] == "initials" and span[:key] = "formatted-initials"
  spans_preprocess_new_contrib?(span, contrib) and
    contrib << { role: span[:type] || "author", entity: "person" }
  if multiple_givennames?(span, contrib)
    contrib[-1][:givenname] = [contrib[-1][:givenname],
                               span[:val]].flatten
  else contrib[-1][span[:key].to_sym] = span[:val]
  end
  contrib
end

#spans_preprocess_fullname(span, contrib) ⇒ Object 



140
141
142
143
144
145
146
147
148
149
150
 
# File 'lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb', line 140

def spans_preprocess_fullname(span, contrib)
  name = span[:val].gsub(/\.(?=\p{Alpha})/, ". ").split(/ /)
  out = { role: span[:type] || "author", entity: "person",
          surname: name[-1] }
  if name.size > 1 && name[0..-2].all? { |x| /\.$/.match?(x) }
    out[:"formatted-initials"] = name[0..-2].join(" ")
  else out[:givenname] = name[0..-2]
  end
  contrib << out
  contrib
end

#spans_preprocess_new_contrib?(span, contrib) ⇒ Boolean

Returns:

  • (Boolean) 


123
124
125
126
127
 
# File 'lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb', line 123

def spans_preprocess_new_contrib?(span, contrib)
  contrib.empty? || contrib[-1][:entity] == "organization" ||
    (span[:key] == "surname" && contrib[-1][:surname]) ||
    contrib[-1][:role] != (span[:type] || "author")
end

#spans_preprocess_org(span, contrib) ⇒ Object 



152
153
154
155
156
 
# File 'lib/metanorma/cleanup/spans_to_bibitem_preprocessing.rb', line 152

def spans_preprocess_org(span, contrib)
  contrib << { role: span[:type] || "author", entity: "organization",
               name: span[:val] }
  contrib
end

#spans_to_bibitem(spans) ⇒ Object 



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
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 45

def spans_to_bibitem(spans)
  ret = ""
  spans[:title] and ret += "<title>#{spans[:title]}</title>"
  ret += spans_to_bibitem_docid(spans)
  ret += spans_to_contribs(spans)
  ret += spans_to_bibitem_edn(spans)
  ret += spans_to_bibitem_i18n(spans)
  spans[:abstract] and ret += "<abstract>#{spans[:abstract]}</abstract>"
  ret += spans_to_series(spans)
  ret += spans_to_pubplace(spans)
  ret += spans_to_bibitem_host(spans)
  ret += spans_to_bibitem_extent(spans[:extent])
  spans[:classification]&.each do |s|
    ret += span_to_docid(s, "classification")
  end
  spans[:keyword]&.each do |s|
    # keyword content is structured per the grammar: an uncontrolled
    # keyword is a vocab entry, not bare text
    ret += "<keyword><vocab>#{s[:val]}</vocab></keyword>"
  end
  spans[:image]&.each do |s|
    ret += "<depiction>#{s[:val]}</depiction>"
  end
  ret
end

#spans_to_bibitem_docid(spans) ⇒ Object 



100
101
102
103
104
105
106
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 100

def spans_to_bibitem_docid(spans)
  ret = ""
  spans[:uri]&.each { |s| ret += span_to_docid(s, "uri") }
  spans[:docid]&.each { |s| ret += span_to_docid(s, "docidentifier") }
  spans[:date]&.each { |s| ret += span_to_date(s) }
  ret
end

#spans_to_bibitem_edn(spans) ⇒ Object 



108
109
110
111
112
113
114
115
116
117
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 108

def spans_to_bibitem_edn(spans)
  ret = ""
  spans[:edition] and ret += "<edition>#{spans[:edition]}</edition>"
  spans[:version] and ret += "<version>#{spans[:version]}</version>"
  spans[:note]&.each do |n|
    ret += "<note type='#{n[:type]}'>#{n[:val]}</note>"
      .sub("<note type=''>", "<note>")
  end
  ret
end

#spans_to_bibitem_extent(spans) ⇒ Object 



119
120
121
122
123
124
125
126
127
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 119

def spans_to_bibitem_extent(spans)
  spans.nil? and return ""
  ret = ""
  { volume: "volume", issue: "issue", pages: "page" }.each do |k, v|
    spans[k]&.each { |s| ret += span_to_extent(s, v) }
  end
  ret.empty? and return ""
  "<extent>#{ret}</extent>"
end

#spans_to_bibitem_host(spans) ⇒ Object 



92
93
94
95
96
97
98
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 92

def spans_to_bibitem_host(spans)
  spans[:in].nil? || spans[:in].empty? and return ""
  ret =
    "<relation type='includedIn'><bibitem type='#{spans[:in][:type]}'>"
  spans[:in].delete(:type)
  ret + "#{spans_to_bibitem(spans[:in])}</bibitem></relation>"
end

#spans_to_bibitem_i18n(spans) ⇒ Object 



79
80
81
82
83
84
85
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 79

def spans_to_bibitem_i18n(spans)
  ret = ""
  spans[:language] and ret += "<language>#{spans[:language]}</language>"
  spans[:script] and ret += "<script>#{spans[:script]}</script>"
  spans[:locale] and ret += "<locale>#{spans[:locale]}</locale>"
  ret
end

#spans_to_contribs(spans) ⇒ Object 



155
156
157
158
159
160
161
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 155

def spans_to_contribs(spans)
  ret = ""
  spans[:contrib]&.each do |s|
    ret += span_to_contrib(s, spans[:title])
  end
  ret
end

#spans_to_pubplace(spans) ⇒ Object 



71
72
73
74
75
76
77
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 71

def spans_to_pubplace(spans)
  ret = ""
  spans[:pubplace] and ret += <<~XML
    <place><formattedPlace>#{spans[:pubplace]}</formattedPlace></place>
  XML
  ret
end

#spans_to_series(spans) ⇒ Object 



87
88
89
90
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 87

def spans_to_series(spans)
  spans[:series] or return ""
  "<series><title>#{spans[:series]}</title></series>"
end

#validate_span_to_person(span, title) ⇒ Object 



171
172
173
174
175
176
 
# File 'lib/metanorma/cleanup/spans_to_bibitem.rb', line 171

def validate_span_to_person(span, title)
  span[:surname] and return
  msg = "Missing surname: issue with bibliographic markup " \
        "in \"#{title}\": #{span}"
  @err << { msg:, fatal: true }
end