Support for Schema 4.5

lib/bolognese/datacite_utils.rb

@@ -106,9 +106,19 @@ def insert_titles(xml)
         end
       end
     end
-
+    
     def insert_publisher(xml)
-      xml.publisher(publisher || container && container["title"])
+      if publisher.is_a?(Hash)
+        attributes = {
+          'publisherIdentifier' => publisher["publisherIdentifier"] || nil,
+          'publisherIdentifierScheme' => publisher["publisherIdentifierScheme"] || nil,
+          'schemeURI' => publisher["schemeUri"] || nil,
+          "xml:lang" => publisher["lang"] || nil
+        }.compact
+        xml.publisher(publisher["name"] || container && container["title"], attributes)
+      else
+        xml.publisher(publisher || container && container["title"])
+      end
     end
 
     def insert_publication_year(xml)
@@ -375,13 +385,11 @@ def insert_geo_locations(xml)
               end
             end
             if geo_location["geoLocationPolygon"]
-              geo_location["geoLocationPolygon"].each do |geo_location_polygon|
-                xml.geoLocationPolygon do
-                  geo_location_polygon.each do |polygon_point|
-                    xml.polygonPoint do
-                      xml.pointLatitude(polygon_point.dig("polygonPoint", "pointLatitude"))
-                      xml.pointLongitude(polygon_point.dig("polygonPoint", "pointLongitude"))
-                    end
+              xml.geoLocationPolygon do
+                Array.wrap(geo_location["geoLocationPolygon"]).each do |polygon_point|
+                  xml.polygonPoint do
+                    xml.pointLatitude(polygon_point.dig("polygonPoint", "pointLatitude"))
+                    xml.pointLongitude(polygon_point.dig("polygonPoint", "pointLongitude"))
                   end
                 end
               end

lib/bolognese/metadata.rb

@@ -217,7 +217,12 @@ def dates
     end
 
     def publisher
-      @publisher ||= meta.fetch("publisher", nil)
+      @publisher ||= 
+        if meta.fetch("publisher", nil).is_a?(Hash)
+          meta.fetch("publisher", nil)
+        elsif meta.fetch("publisher", nil).is_a?(String)
+          { "name" => meta.fetch("publisher", nil) }.compact
+        end
     end
 
     def identifiers

lib/bolognese/metadata_utils.rb

@@ -161,7 +161,7 @@ def citeproc_hsh
         "volume" => container.to_h["volume"],
         "issue" => container.to_h["issue"],
         "page" => page,
-        "publisher" => publisher,
+        "publisher" => publisher["name"],
         "title" => parse_attributes(titles, content: "title", first: true),
         "URL" => url,
         "copyright" => Array.wrap(rights_list).map { |l| l["rights"] }.first,

lib/bolognese/readers/bibtex_reader.rb

@@ -86,7 +86,7 @@ def read_bibtex(string: nil, **options)
           "titles" => meta.try(:title).present? ? [{ "title" => meta.try(:title).to_s }] : [],
           "creators" => creators,
           "container" => container,
-          "publisher" => meta.try(:publisher).to_s.presence,
+          "publisher" => meta.try(:publisher).present? ? { "name" => meta.try(:publisher).to_s }.compact : nil,
           "related_identifiers" => related_identifiers,
           "dates" => dates,
           "publication_year" => publication_year,

lib/bolognese/readers/codemeta_reader.rb

@@ -43,7 +43,7 @@ def read_codemeta(string: nil, **options)
         dates << { "date" => meta.fetch("dateCreated"), "dateType" => "Created" } if meta.fetch("dateCreated", nil).present?
         dates << { "date" => meta.fetch("dateModified"), "dateType" => "Updated" } if meta.fetch("dateModified", nil).present?
         publication_year = meta.fetch("datePublished")[0..3] if meta.fetch("datePublished", nil).present?
-        publisher = meta.fetch("publisher", nil)
+        publisher = { "name" => meta.fetch("publisher", nil) } if meta.fetch("publisher", nil).present?
         state = meta.present? || read_options.present? ? "findable" : "not_found"
         schema_org = meta.fetch("@type", nil)
         types = {

lib/bolognese/readers/crossref_reader.rb

@@ -40,8 +40,7 @@ def read_crossref(string: nil, **options)
         journal_metadata = nil
         journal_issue = {}
         journal_metadata = nil
-        publisher = query.dig("crm_item", 0)
-        publisher = nil unless publisher.is_a?(String)
+        publisher = query.dig("crm_item", 0).is_a?(String) ? { "name" => query.dig("crm_item", 0) } : nil
 
         case model
         when "book"

lib/bolognese/readers/datacite_reader.rb

@@ -94,6 +94,22 @@ def read_datacite(string: nil, **options)
 
         titles = get_titles(meta)
 
+        publisher = Array.wrap(meta.dig("publisher")).map do |r|
+          if r.blank?
+            nil
+          elsif r.is_a?(String)
+            { "name" => r.strip }
+          elsif r.is_a?(Hash)
+            {
+              "name" => r["__content__"].strip,
+              "publisherIdentifier" => r["publisherIdentifierScheme"] == "ROR" ? normalize_ror(r["publisherIdentifier"]) : r["publisherIdentifier"],
+              "publisherIdentifierScheme" => r["publisherIdentifierScheme"],
+              "schemeUri" => r["schemeURI"],
+              "lang" => r["lang"],
+            }.compact
+          end
+        end.compact.first
+
         descriptions = Array.wrap(meta.dig("descriptions", "description")).map do |r|
           if r.blank?
             nil
@@ -287,7 +303,7 @@ def read_datacite(string: nil, **options)
           "creators" => get_authors(Array.wrap(meta.dig("creators", "creator"))),
           "contributors" => get_authors(Array.wrap(meta.dig("contributors", "contributor"))),
           "container" => set_container(meta),
-          "publisher" => parse_attributes(meta.fetch("publisher", nil), first: true).to_s.strip.presence,
+          "publisher" => publisher,
           "agency" => "datacite",
           "funding_references" => funding_references,
           "dates" => dates,

lib/bolognese/readers/ris_reader.rb

@@ -89,7 +89,7 @@ def read_ris(string: nil, **options)
           "url" => meta.fetch("UR", nil),
           "titles" => meta.fetch("T1", nil).present? ? [{ "title" => meta.fetch("T1", nil) }] : nil,
           "creators" => get_authors(author),
-          "publisher" => meta.fetch("PB", "(:unav)"),
+          "publisher" => { "name" => meta.fetch("PB", "(:unav)") },
           "container" => container,
           "related_identifiers" => related_identifiers,
           "dates" => dates,

lib/bolognese/readers/schema_org_reader.rb

@@ -74,7 +74,12 @@ def read_schema_org(string: nil, **options)
           creators = get_authors(from_schema_org_creators(Array.wrap(authors)))
         end
         contributors = get_authors(from_schema_org_contributors(Array.wrap(meta.fetch("editor", nil))))
-        publisher = parse_attributes(meta.fetch("publisher", nil), content: "name", first: true)
+        publisher = 
+          if parse_attributes(meta.fetch("publisher", nil), content: "name", first: true)
+            {
+              "name" => parse_attributes(meta.fetch("publisher", nil), content: "name", first: true),
+            }.compact
+          end
 
         ct = (schema_org == "Dataset") ? "includedInDataCatalog" : "Periodical"
         container = if meta.fetch(ct, nil).present?

lib/bolognese/writers/bibtex_writer.rb

@@ -21,7 +21,7 @@ def bibtex
           volume: container.to_h["volume"],
           issue: container.to_h["issue"],
           pages: pages,
-          publisher: publisher,
+          publisher: publisher["name"],
           year: publication_year,
           copyright: Array.wrap(rights_list).map { |l| l["rights"] }.first,
         }.compact

lib/bolognese/writers/codemeta_writer.rb

@@ -19,7 +19,7 @@ def codemeta
           "tags" => subjects.present? ? Array.wrap(subjects).map { |k| parse_attributes(k, content: "subject", first: true) } : nil,
           "datePublished" => get_date(dates, "Issued") || publication_year,
           "dateModified" => get_date(dates, "Updated"),
-          "publisher" => publisher,
+          "publisher" => publisher["name"],
           "license" => Array.wrap(rights_list).map { |l| l["rightsUri"] }.compact.unwrap,
         }.compact
         JSON.pretty_generate hsh.presence

lib/bolognese/writers/csv_writer.rb

@@ -15,7 +15,7 @@ def csv
           resource_type: types["resourceType"],
           title: parse_attributes(titles, content: "title", first: true),
           author: authors_as_string(creators),
-          publisher: publisher,
+          publisher: publisher["name"],
           publication_year: publication_year
         }.values
 

lib/bolognese/writers/jats_writer.rb

@@ -77,16 +77,16 @@ def insert_citation_title(xml)
 
       def insert_source(xml)
         if is_chapter?
-          xml.source(publisher)
+          xml.source(publisher["name"])
         elsif is_article? || is_data?
-          xml.source(container && container["title"] || publisher)
+          xml.source(container && container["title"] || publisher["name"])
         else
           xml.source(parse_attributes(titles, content: "title", first: true))
         end
       end
 
       def insert_publisher_name(xml)
-        xml.send("publisher-name", publisher)
+        xml.send("publisher-name", publisher["name"])
       end
 
       def insert_publication_date(xml)

lib/bolognese/writers/ris_writer.rb

@@ -14,7 +14,7 @@ def ris
           "AB" => parse_attributes(abstract_description, content: "description", first: true),
           "KW" => Array.wrap(subjects).map { |k| parse_attributes(k, content: "subject", first: true) }.presence,
           "PY" => publication_year,
-          "PB" => publisher,
+          "PB" => publisher["name"],
           "LA" => language,
           "VL" => container.to_h["volume"],
           "IS" => container.to_h["issue"],

lib/bolognese/writers/schema_org_writer.rb

@@ -37,7 +37,7 @@ def schema_hsh
           "schemaVersion" => schema_version,
           "periodical" => types.present? ? ((types["schemaOrg"] != "Dataset") && container.present? ? to_schema_org(container) : nil) : nil,
           "includedInDataCatalog" => types.present? ? ((types["schemaOrg"] == "Dataset") && container.present? ? to_schema_org_container(container, type: "Dataset") : nil) : nil,
-          "publisher" => publisher.present? ? { "@type" => "Organization", "name" => publisher } : nil,
+          "publisher" => publisher.present? ? { "@type" => "Organization", "name" => publisher["name"] } : nil,
           "funder" => to_schema_org_funder(funding_references),
           "provider" => agency.present? ? { "@type" => "Organization", "name" => agency } : nil
         }.compact.presence

resources/kernel-4.5/include/datacite-contributorType-v4.xsd

@@ -0,0 +1,35 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!-- Version 1.0 - Created 2011-01-13 - FZ, TIB, Germany
+   2013-05 v3.0: Addition of ID to simpleType element, added values "ResearchGroup" & "Other"
+   2014-08-20 v3.1: Addition of value "DataCurator"
+   2015-05-14 v4.0 dropped value "Funder", use new "funderReference" -->
+<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns="http://datacite.org/schema/kernel-4" targetNamespace="http://datacite.org/schema/kernel-4" elementFormDefault="qualified">
+  <xs:simpleType name="contributorType" id="contributorType">
+    <xs:annotation>
+      <xs:documentation>The type of contributor of the resource.</xs:documentation>
+    </xs:annotation>
+    <xs:restriction base="xs:string">
+      <xs:enumeration value="ContactPerson" />
+      <xs:enumeration value="DataCollector" />
+      <xs:enumeration value="DataCurator" />
+      <xs:enumeration value="DataManager" />
+      <xs:enumeration value="Distributor" />
+      <xs:enumeration value="Editor" />
+      <xs:enumeration value="HostingInstitution" />
+      <xs:enumeration value="Other" />
+      <xs:enumeration value="Producer" />
+      <xs:enumeration value="ProjectLeader" />
+      <xs:enumeration value="ProjectManager" />
+      <xs:enumeration value="ProjectMember" />
+      <xs:enumeration value="RegistrationAgency" />
+      <xs:enumeration value="RegistrationAuthority" />
+      <xs:enumeration value="RelatedPerson" />
+      <xs:enumeration value="ResearchGroup" />
+      <xs:enumeration value="RightsHolder" />
+      <xs:enumeration value="Researcher" />
+      <xs:enumeration value="Sponsor" />
+      <xs:enumeration value="Supervisor" />
+      <xs:enumeration value="WorkPackageLeader" />
+    </xs:restriction>
+  </xs:simpleType>
+</xs:schema>

resources/kernel-4.5/include/datacite-dateType-v4.xsd

@@ -0,0 +1,25 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!-- Version 1.0 - Created 2011-01-13 - FZ, TIB, Germany
+     2013-05 v3.0: Addition of ID to simpleType element; addition of value "Collected"; deleted "StartDate" & "EndDate"
+     2017-10-23 v4.1: Addition of value "Other"
+     2019-02-14 v4.2: Addition of value "Withdrawn"-->
+<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns="http://datacite.org/schema/kernel-4" targetNamespace="http://datacite.org/schema/kernel-4" elementFormDefault="qualified">
+  <xs:simpleType name="dateType" id="dateType">
+    <xs:annotation>
+      <xs:documentation>The type of date. Use RKMS‐ISO8601 standard for depicting date ranges.To indicate the end of an embargo period, use Available. To indicate the start of an embargo period, use Submitted or Accepted, as appropriate.</xs:documentation>
+    </xs:annotation>
+    <xs:restriction base="xs:string">
+      <xs:enumeration value="Accepted" />
+      <xs:enumeration value="Available" />
+      <xs:enumeration value="Collected" />
+      <xs:enumeration value="Copyrighted" />
+      <xs:enumeration value="Created" />
+      <xs:enumeration value="Issued" />
+      <xs:enumeration value="Other" />
+      <xs:enumeration value="Submitted" />
+      <xs:enumeration value="Updated" />
+      <xs:enumeration value="Valid" />
+      <xs:enumeration value="Withdrawn" />
+    </xs:restriction>
+  </xs:simpleType>
+</xs:schema>

resources/kernel-4.5/include/datacite-descriptionType-v4.xsd

@@ -0,0 +1,19 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!-- Version 1.0 - Created 2011-01-13 - FZ, TIB, Germany
+     2013-05 v3.0: Addition of ID to simpleType element, addition of value "Methods"
+     2015-02-12 v4.0: Addition of value "TechnicalInfo"-->
+<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns="http://datacite.org/schema/kernel-4" targetNamespace="http://datacite.org/schema/kernel-4" elementFormDefault="qualified">
+  <xs:simpleType name="descriptionType" id="descriptionType">
+    <xs:annotation>
+      <xs:documentation>The type of the description.</xs:documentation>
+    </xs:annotation>
+    <xs:restriction base="xs:string">
+      <xs:enumeration value="Abstract" />
+      <xs:enumeration value="Methods" />
+      <xs:enumeration value="SeriesInformation" />
+      <xs:enumeration value="TableOfContents" />
+      <xs:enumeration value="TechnicalInfo" />
+      <xs:enumeration value="Other" />
+    </xs:restriction>
+  </xs:simpleType>
+</xs:schema>

resources/kernel-4.5/include/datacite-funderIdentifierType-v4.xsd

@@ -0,0 +1,16 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!-- Version 1.0 - Created 2016-05-14 -->
+<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns="http://datacite.org/schema/kernel-4" targetNamespace="http://datacite.org/schema/kernel-4" elementFormDefault="qualified">
+  <xs:simpleType name="funderIdentifierType" id="funderIdentifierType">
+    <xs:annotation>
+      <xs:documentation>The type of the funderIdentifier.</xs:documentation>
+    </xs:annotation>
+    <xs:restriction base="xs:string">
+      <xs:enumeration value="ISNI" />
+      <xs:enumeration value="GRID" />
+      <xs:enumeration value="ROR" />
+      <xs:enumeration value="Crossref Funder ID" />
+      <xs:enumeration value="Other" />
+    </xs:restriction>
+  </xs:simpleType>
+</xs:schema>

resources/kernel-4.5/include/datacite-nameType-v4.xsd

@@ -0,0 +1,10 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!-- Version 4.1 - Created 2017-10-23 -->
+<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns="http://datacite.org/schema/kernel-4" targetNamespace="http://datacite.org/schema/kernel-4" elementFormDefault="qualified">
+  <xs:simpleType name="nameType" id="nameType">
+    <xs:restriction base="xs:string">
+      <xs:enumeration value="Organizational" />
+      <xs:enumeration value="Personal" />
+    </xs:restriction>
+  </xs:simpleType>
+</xs:schema>

resources/kernel-4.5/include/datacite-numberType-v4.xsd

@@ -0,0 +1,12 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!-- Version 4.4 - Created 2021-03-05 -->
+<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns="http://datacite.org/schema/kernel-4" targetNamespace="http://datacite.org/schema/kernel-4" elementFormDefault="qualified">
+  <xs:simpleType name="numberType" id="numberType">
+    <xs:restriction base="xs:string">
+      <xs:enumeration value="Article" />
+      <xs:enumeration value="Chapter" />
+      <xs:enumeration value="Report" />
+      <xs:enumeration value="Other" />
+    </xs:restriction>
+  </xs:simpleType>
+</xs:schema>

resources/kernel-4.5/include/datacite-relatedIdentifierType-v4.xsd

@@ -0,0 +1,34 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!-- Version 1.0 - Created 2011-01-13 - FZ, TIB, Germany
+  2013-05 v3.0: Addition of ID to simpleType element; addition of value "PMID"
+  2014-08-20 v3.1: Addition of values "arxiv" and "bibcode"
+  2015-02-12 v4.0 Addition of value "IGSN"
+  2019-02-14 v4.2 Addition of value "w3id" -->
+<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns="http://datacite.org/schema/kernel-4" targetNamespace="http://datacite.org/schema/kernel-4" elementFormDefault="qualified">
+  <xs:simpleType name="relatedIdentifierType" id="relatedIdentifierType">
+    <xs:annotation>
+      <xs:documentation>The type of the RelatedIdentifier.</xs:documentation>
+    </xs:annotation>
+    <xs:restriction base="xs:string">
+      <xs:enumeration value="ARK" />
+      <xs:enumeration value="arXiv" />
+      <xs:enumeration value="bibcode" />
+      <xs:enumeration value="DOI" />
+      <xs:enumeration value="EAN13" />
+      <xs:enumeration value="EISSN" />
+      <xs:enumeration value="Handle" />
+      <xs:enumeration value="IGSN" />
+      <xs:enumeration value="ISBN" />
+      <xs:enumeration value="ISSN" />
+      <xs:enumeration value="ISTC" />
+      <xs:enumeration value="LISSN" />
+      <xs:enumeration value="LSID" />
+      <xs:enumeration value="PMID" />
+      <xs:enumeration value="PURL" />
+      <xs:enumeration value="UPC" />
+      <xs:enumeration value="URL" />
+      <xs:enumeration value="URN" />
+      <xs:enumeration value="w3id" />
+    </xs:restriction>
+  </xs:simpleType>
+</xs:schema>