Module: Bolognese::Readers::CrossrefReader

Included in:
Metadata
Defined in:
lib/bolognese/readers/crossref_reader.rb

Constant Summary collapse

CR_TO_SO_TRANSLATIONS =

CrossRef types from api.crossref.org/types

{
  "Proceedings" => nil,
  "ReferenceBook" => "Book",
  "JournalIssue" => "PublicationIssue",
  "ProceedingsArticle" => nil,
  "Other" => "CreativeWork",
  "Dissertation" => "Thesis",
  "Dataset" => "Dataset",
  "EditedBook" => "Book",
  "JournalArticle" => "ScholarlyArticle",
  "Journal" => nil,
  "Report" => nil,
  "BookSeries" => nil,
  "ReportSeries" => nil,
  "BookTrack" => nil,
  "Standard" => nil,
  "BookSection" => nil,
  "BookPart" => nil,
  "Book" => "Book",
  "BookChapter" => "Chapter",
  "StandardSeries" => nil,
  "Monograph" => "Book",
  "Component" => "CreativeWork",
  "ReferenceEntry" => nil,
  "JournalVolume" => "PublicationVolume",
  "BookSet" => nil,
  "PostedContent" => "ScholarlyArticle"
}
CR_TO_BIB_TRANSLATIONS = 
{
  "Proceedings" => "proceedings",
  "ReferenceBook" => "book",
  "JournalIssue" => nil,
  "ProceedingsArticle" => nil,
  "Other" => nil,
  "Dissertation" => "phdthesis",
  "Dataset" => nil,
  "EditedBook" => "book",
  "JournalArticle" => "article",
  "Journal" => nil,
  "Report" => nil,
  "BookSeries" => nil,
  "ReportSeries" => nil,
  "BookTrack" => nil,
  "Standard" => nil,
  "BookSection" => "inbook",
  "BookPart" => nil,
  "Book" => "book",
  "BookChapter" => "inbook",
  "StandardSeries" => nil,
  "Monograph" => "book",
  "Component" => nil,
  "ReferenceEntry" => nil,
  "JournalVolume" => nil,
  "BookSet" => nil,
  "PostedContent" => "article"
}
CONTACT_EMAIL = 
"[email protected]"

Instance Method Summary collapse

Instance Method Details

#crossref_alternate_name(bibliographic_metadata) ⇒ Object 



134
135
136
137
138
139
140
 
# File 'lib/bolognese/readers/crossref_reader.rb', line 134

def crossref_alternate_name(bibliographic_metadata)
  if bibliographic_metadata.fetch("publisher_item", nil).present?
    parse_attributes(bibliographic_metadata.dig("publisher_item", "item_number"))
  else
    parse_attributes(bibliographic_metadata.fetch("item_number", nil))
  end
end

#crossref_date_published(bibliographic_metadata) ⇒ Object

def datacite_funding_reference(meta)

Array.wrap(meta.dig("fundingReferences", "fundingReference")).map do |f|
  funder_id = parse_attributes(f["funderIdentifier"])
  funder = { "type" => "Organization",
             "id" => normalize_id(funder_id),
             "name" => f["funderName"] }.compact
  if f["awardNumber"].present? || f["awardTitle"].present?
    { "type" => "Award",
      "name" => f.fetch("awardTitle", nil),
      "identifier" => f.dig("awardNumber", "__content__"),
      "url" => f.dig("awardNumber", "awardURI"),
      "funder" => funder }
  else
    funder
  end
end.uniq

end



208
209
210
211
212
213
214
215
216
 
# File 'lib/bolognese/readers/crossref_reader.rb', line 208

def crossref_date_published(bibliographic_metadata)
  pub_date = Array.wrap(bibliographic_metadata.fetch("publication_date", nil)).presence ||
    Array.wrap(bibliographic_metadata.fetch("acceptance_date", nil))
  if pub_date.present?
    get_date_from_parts(pub_date.first["year"], pub_date.first["month"], pub_date.first["day"])
  else
    nil
  end
end

#crossref_description(bibliographic_metadata) ⇒ Object 



142
143
144
145
146
147
148
149
 
# File 'lib/bolognese/readers/crossref_reader.rb', line 142

def crossref_description(bibliographic_metadata)
  des = bibliographic_metadata.fetch("abstract", {}).values.first
  if des.is_a?(Hash)
    sanitize(des.fetch("__content__", nil))
  elsif des.is_a?(String)
    sanitize(des)
  end
end

#crossref_funding_reference(program_metadata) ⇒ Object 



171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
 
# File 'lib/bolognese/readers/crossref_reader.rb', line 171

def crossref_funding_reference(program_metadata)
  fundref = Array.wrap(program_metadata).find { |a| a["name"] == "fundref" } || {}
  Array.wrap(fundref.fetch("assertion", [])).select { |a| a["name"] == "fundgroup" }.map do |f|
    f = Array.wrap(f.fetch("assertion", nil)).first
    funder = { "type" => "Organization",
               "id" => normalize_id(f.dig("assertion", "__content__")),
               "name" => f.dig("__content__").strip }.compact
     if f["awardNumber"].present? || f["awardTitle"].present?
       { "type" => "Award",
         "name" => f.fetch("awardTitle", nil),
         "identifier" => f.dig("awardNumber", "__content__"),
         "url" => f.dig("awardNumber", "awardURI"),
         "funder" => funder }
     else
       funder
     end
  end.unwrap
end

#crossref_is_part_of(journal_metadata) ⇒ Object 



218
219
220
221
222
223
224
225
226
 
# File 'lib/bolognese/readers/crossref_reader.rb', line 218

def crossref_is_part_of(journal_metadata)
  if journal_metadata.present?
    { "type" => "Periodical",
      "title" => journal_metadata["full_title"],
      "issn" => parse_attributes(journal_metadata.fetch("issn", nil)) }.compact
  else
    nil
  end
end

#crossref_license(program_metadata) ⇒ Object 



151
152
153
154
155
156
157
158
 
# File 'lib/bolognese/readers/crossref_reader.rb', line 151

def crossref_license(program_metadata)
  access_indicator = Array.wrap(program_metadata).find { |m| m["name"] == "AccessIndicators" }
  if access_indicator.present?
    { "id" => normalize_url(parse_attributes(access_indicator["license_ref"])) }
  else
    nil
  end
end

#crossref_people(bibliographic_metadata, contributor_role) ⇒ Object 



160
161
162
163
164
165
166
167
168
169
 
# File 'lib/bolognese/readers/crossref_reader.rb', line 160

def crossref_people(bibliographic_metadata, contributor_role)
  person = bibliographic_metadata.dig("contributors", "person_name")
  Array.wrap(person).select { |a| a["contributor_role"] == contributor_role }.map do |a|
    { "type" => "Person",
      "id" => parse_attributes(a["ORCID"]),
      "name" => [a["given_name"], a["surname"]].join(" "),
      "givenName" => a["given_name"],
      "familyName" => a["surname"] }.compact
  end.unwrap
end

#crossref_references(bibliographic_metadata) ⇒ Object 



228
229
230
231
232
233
234
235
 
# File 'lib/bolognese/readers/crossref_reader.rb', line 228

def crossref_references(bibliographic_metadata)
   refs = bibliographic_metadata.dig("citation_list", "citation")
   Array.wrap(refs).select { |a| a["doi"].present? }.map do |c|
     { "type" => "CreativeWork",
       "id" => normalize_id(c["doi"]),
       "title" => c["article_title"] }.compact
   end.unwrap
end

#get_crossref(id: nil) ⇒ Object 



65
66
67
68
69
70
71
72
73
 
# File 'lib/bolognese/readers/crossref_reader.rb', line 65

def get_crossref(id: nil)
  return nil unless id.present?

  doi = doi_from_url(id)
  url = "http://www.crossref.org/openurl/?id=doi:#{doi}&noredirect=true&pid=#{CONTACT_EMAIL}&format=unixref"
  response = Maremma.get(url, accept: "text/xml", raw: true)
  string = response.body.fetch("data", nil)
  Nokogiri::XML(string, nil, 'UTF-8', &:noblanks).to_s if string.present?
end

#read_crossref(string: nil) ⇒ Object 



75
76
77
78
79
80
81
82
83
84
85
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
125
126
127
128
129
130
131
132
 
# File 'lib/bolognese/readers/crossref_reader.rb', line 75

def read_crossref(string: nil)
  if string.present?
    m = Maremma.from_xml(string).dig("doi_records", "doi_record") || {}
    meta = m.dig("crossref", "error").nil? ? m : {}
  else
    meta = {}
  end

  return meta unless meta["crossref"].present?

  journal_metadata = meta.dig("crossref", "journal", "journal_metadata").presence || {}
  bibliographic_metadata = meta.dig("crossref", "journal", "journal_article").presence ||
                           meta.dig("crossref", "conference", "conference_paper").presence ||
                           meta.dig("crossref", meta.fetch("crossref", {}).keys.last).presence || {}
  program_metadata = bibliographic_metadata.dig("program") ||
                     bibliographic_metadata.dig("crossmark", "custom_metadata", "program") || {}
  journal_issue = meta.dig("crossref", "journal", "journal_issue").presence || {}

  additional_type = if meta.dig("crossref", "journal").present?
                      meta.dig("crossref", "journal").keys.last.camelize
                    else
                      meta.dig("crossref").keys.last.camelize
                    end
  type = CR_TO_SO_TRANSLATIONS[additional_type] || "ScholarlyArticle"
  doi = bibliographic_metadata.dig("doi_data", "doi")

  { "id" => normalize_doi(doi),
    "type" => type,
    "additional_type" => additional_type,
    "citeproc_type" => Bolognese::Utils::CR_TO_CP_TRANSLATIONS[additional_type] || "article-journal",
    "bibtex_type" => CR_TO_BIB_TRANSLATIONS[additional_type] || "misc",
    "ris_type" => Bolognese::Utils::CR_TO_RIS_TRANSLATIONS[additional_type] || "JOUR",
    "resource_type_general" => Bolognese::Utils::SO_TO_DC_TRANSLATIONS[type],
    "doi" => doi,
    "url" => bibliographic_metadata.dig("doi_data", "resource"),
    "title" => parse_attributes(bibliographic_metadata.dig("titles", "title")),
    "alternate_name" => crossref_alternate_name(bibliographic_metadata),
    "author" => crossref_people(bibliographic_metadata, "author"),
    "editor" => crossref_people(bibliographic_metadata, "editor"),
    "funding" => crossref_funding_reference(program_metadata),
    "publisher" => nil,
    "provider" => "Crossref",
    "is_part_of" => crossref_is_part_of(journal_metadata),
    "references" => crossref_references(bibliographic_metadata),
    "date_published" => crossref_date_published(bibliographic_metadata),
    "date_modified" => Time.parse(meta.fetch("timestamp", "")).utc.iso8601,
    "volume" => journal_issue.dig("journal_volume", "volume"),
    "issue" => journal_issue.dig("issue"),
    "pagination" => [bibliographic_metadata.dig("pages", "first_page"), bibliographic_metadata.dig("pages", "last_page")].compact.join("-").presence,
    "description" => crossref_description(bibliographic_metadata),
    "license" => crossref_license(program_metadata),
    "version" => nil,
    "keywords" => nil,
    "language" => nil,
    "content_size" => nil,
    "schema_version" => nil
  }
end