bolognese 2.0.0 → 2.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 23ccf372b59b7d9c8919ec7ad7137eef6e99512eed488977e847f938f389aa2f
-  data.tar.gz: 390e56023d57e637a52776a98dcd0bea4fa0acd8b8459ced05cbc6785f69fe0a
+  metadata.gz: b97f37a5ebcbedcb33b6937353b9afe721084c52591033d04657c8ef1e3fa4b7
+  data.tar.gz: 119f596d499c6465304b91eb2ea4733a6d83be967d564bf0bc895e69312133d0
 SHA512:
-  metadata.gz: 3a7364a75d69846ada1320bc0e8954b72695216d2807ec39ab1694ac4d120d97806addb9ceed8b9da8c1eb06ae28740cf6d2c118cb76a05e15ca993bf81ad470
-  data.tar.gz: 6281527f6b312d398a96b9f1df9be55de991f6ac939417131633b24ddaa24741c3b86c3b2e4e73408ca1350e0b17addf9439ae30dff998bc6dda07e8c4459db7
+  metadata.gz: 77194ea5ab5fac3555fad930a642081f4cb146dc804baed87b35be72f9c91cf88a72e44a1b2a8ab760f15ce5d49843dd65f2c41b24aa3ce840a940093705d715
+  data.tar.gz: 504322b19079059e7946043c087408237e48e1266f76cc4cb9bbfaf980d47d0f628dda1e1b35d20d79a27b07e18fb220b5a22234ee572b82020a31d98f52c196

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    bolognese (2.0.0)
+    bolognese (2.1.0)
       activesupport (>= 4.2.5)
       benchmark_methods (~> 0.7)
       bibtex-ruby (>= 5.1.0)
@@ -95,10 +95,10 @@ GEM
       concurrent-ruby (~> 1.0)
     iso8601 (0.9.1)
     json (2.6.2)
-    json-canonicalization (0.3.2)
-    json-ld (3.2.5)
+    json-canonicalization (0.3.1)
+    json-ld (3.2.4)
       htmlentities (~> 4.3)
-      json-canonicalization (~> 0.3, >= 0.3.2)
+      json-canonicalization (~> 0.3)
       link_header (~> 0.0, >= 0.0.8)
       multi_json (~> 1.15)
       rack (>= 2.2, < 4)
@@ -225,6 +225,7 @@ DEPENDENCIES
   bundler (>= 1.0)
   byebug
   hashdiff (>= 1.0.0.beta1, < 2.0.0)
+  json-canonicalization (= 0.3.1)
   rack-test (~> 0)
   rake (~> 12.0)
   rspec (~> 3.4)
@@ -234,4 +235,4 @@ DEPENDENCIES
   webmock (~> 3.0, >= 3.0.1)
 
 BUNDLED WITH
-   2.4.20
+   2.4.22

data/bolognese.gemspec

@@ -50,6 +50,7 @@ Gem::Specification.new do |s|
   s.add_development_dependency 'simplecov', '0.17.1'
   s.add_development_dependency 'hashdiff', ['>= 1.0.0.beta1', '< 2.0.0']
   s.add_development_dependency 'byebug'
+  s.add_development_dependency 'json-canonicalization', '0.3.1'
 
   s.require_paths = ["lib"]
   s.files       = `git ls-files`.split($/)

data/lib/bolognese/author_utils.rb

@@ -30,20 +30,20 @@ module Bolognese
       name_type = parse_attributes(author.fetch("creatorName", nil), content: "nameType", first: true) || parse_attributes(author.fetch("contributorName", nil), content: "nameType", first: true)
 
       name_identifiers = Array.wrap(author.fetch("nameIdentifier", nil)).map do |ni|
-        ni["__content__"] = ni["__content__"].strip
+        name_identifier = ni["__content__"].strip if ni["__content__"].present?
         if ni["nameIdentifierScheme"] == "ORCID"
           {
-            "nameIdentifier" => normalize_orcid(ni["__content__"]),
+            "nameIdentifier" => normalize_orcid(name_identifier),
             "schemeUri" => "https://orcid.org",
             "nameIdentifierScheme" => "ORCID" }.compact
         elsif ni["nameIdentifierScheme"] == "ROR"
           {
-            "nameIdentifier" => normalize_ror(ni["__content__"]),
+            "nameIdentifier" => normalize_ror(name_identifier),
             "schemeUri" => "https://ror.org",
             "nameIdentifierScheme" => "ROR" }.compact
         else
           {
-            "nameIdentifier" => ni["__content__"],
+            "nameIdentifier" => name_identifier,
             "schemeUri" => ni.fetch("schemeURI", nil),
             "nameIdentifierScheme" => ni["nameIdentifierScheme"] }.compact
         end

data/lib/bolognese/metadata_utils.rb

@@ -96,7 +96,7 @@ module Bolognese
       if container.present?
         container["title"]
       elsif types["citeproc"] == "article-journal"
-        publisher
+        publisher["name"] if publisher.present?
       else
         nil
       end

data/lib/bolognese/readers/datacite_reader.rb

@@ -101,7 +101,7 @@ module Bolognese
             { "name" => r.strip }
           elsif r.is_a?(Hash)
             {
-              "name" => r["__content__"].strip,
+              "name" => r["__content__"].present? ? r["__content__"].strip : nil,
               "publisherIdentifier" => r["publisherIdentifierScheme"] == "ROR" ? normalize_ror(r["publisherIdentifier"]) : r["publisherIdentifier"],
               "publisherIdentifierScheme" => r["publisherIdentifierScheme"],
               "schemeUri" => r["schemeURI"],

data/lib/bolognese/readers/schema_org_reader.rb

@@ -74,8 +74,11 @@ module Bolognese
           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_name = parse_attributes(meta.fetch("publisher", nil), content: "name", first: true)
-        publisher = { "name" => publisher_name } if publisher_name.present?
+
+        publisher = {
+          "name" => parse_attributes(meta.fetch("publisher", nil), content: "name", first: true),
+          "publisherIdentifier" => parse_attributes(meta.fetch("publisher", nil), content: "@id", first: true),
+        }.compact if meta.fetch("publisher", nil).present?
 
         ct = (schema_org == "Dataset") ? "includedInDataCatalog" : "Periodical"
         container = if meta.fetch(ct, nil).present?

data/lib/bolognese/version.rb

@@ -1,3 +1,3 @@
 module Bolognese
-  VERSION = "2.0.0"
+  VERSION = "2.1.0"
 end

data/lib/bolognese/writers/jats_writer.rb

@@ -86,7 +86,10 @@ module Bolognese
       end
 
       def insert_publisher_name(xml)
-        xml.send("publisher-name", publisher["name"])
+        attributes = {
+          "xml:lang" => publisher["lang"]
+        }.compact
+        xml.send("publisher-name", attributes, publisher["name"])
       end
 
       def insert_publication_date(xml)

data/lib/bolognese/writers/schema_org_writer.rb

@@ -37,7 +37,7 @@ module Bolognese
           "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["name"] } : nil,
+          "publisher" => publisher.present? ? { "@type" => "Organization", "@id" => publisher["publisherIdentifier"], "name" => publisher["name"] }.compact : nil,
           "funder" => to_schema_org_funder(funding_references),
           "provider" => agency.present? ? { "@type" => "Organization", "name" => agency } : nil
         }.compact.presence

data/spec/fixtures/datacite-xml-lang.xml

@@ -15,7 +15,7 @@
         <title xml:lang="en">DOI Test 2 title content</title>
         <title xml:lang="en">AAPP</title>
     </titles>
-    <publisher>OSI SAF</publisher>
+    <publisher xml:lang="en">OSI SAF</publisher>
     <publicationYear>2004</publicationYear>
     <resourceType resourceTypeGeneral="Collection">Climate Data Record</resourceType>
     <dates>

data/spec/fixtures/datacite_blank_name_identifier.xml

@@ -0,0 +1,22 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<resource xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://datacite.org/schema/kernel-4" xsi:schemaLocation="http://schema.datacite.org/meta/kernel-4.3/metadata.xsd">
+  <identifier identifierType="DOI">10.15148/3b68dac4-7688-4fbb-ba64-5f9b3bbab954</identifier>
+  <creators>
+    <creator>
+      <creatorName>SNO KARST</creatorName>
+      <nameIdentifier nameIdentifierScheme="ORCID" schemeURI="https://orcid.org"/>
+    </creator>
+  </creators>
+  <publisher>OSU OREME</publisher>
+  <titles>
+    <title>Time series of type chemistry in Le Tarn basin - PARC NATIONAL CEVENNES observatory - KARST observatory network -  OZCAR Critical Zone network Research Infrastructure</title>
+  </titles>
+  <publicationYear>2023</publicationYear>
+  <contributors>
+    <contributor contributorType="ProjectMember">
+      <contributorName>Manche, Yannick</contributorName>
+      <nameIdentifier schemeURI="http://orcid.org/" nameIdentifierScheme="ORCID"></nameIdentifier>
+    </contributor>
+  </contributors>
+  <resourceType resourceTypeGeneral="Dataset">Dataset</resourceType>
+</resource>

data/spec/fixtures/datacite_blank_publisher.xml

@@ -0,0 +1,18 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<resource xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://datacite.org/schema/kernel-4" xsi:schemaLocation="http://datacite.org/schema/kernel-4 http://schema.datacite.org/meta/kernel-4/metadata.xsd">
+  <identifier identifierType="DOI">10.81360/4DVP-KR57</identifier>
+  <creators>
+    <creator>
+      <creatorName nameType="Personal">Møller, Jørgen</creatorName>
+      <givenName>Jørgen</givenName>
+      <familyName>Møller</familyName>
+    </creator>
+  </creators>
+  <titles>
+    <title xml:lang="en">Economic Crisis and Democratic Breakdown in the Interwar Years: A Reassessment</title>
+    <title xml:lang="de">Wirtschaftskrise und demokratischer Zusammenbruch in der Zwischenkriegszeit: Eine Neubewertung</title>
+  </titles>
+  <publisher publisherIdentifier="https://ror.org/04wxnsj81"></publisher>
+  <publicationYear>2015</publicationYear>
+  <resourceType resourceTypeGeneral="JournalArticle"/>
+</resource>

data/spec/fixtures/datacite_journal_article.xml

@@ -0,0 +1,64 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<resource
+    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+    xmlns="http://datacite.org/schema/kernel-4" xsi:schemaLocation="http://datacite.org/schema/kernel-4 http://schema.datacite.org/meta/kernel-4/metadata.xsd">
+    <identifier identifierType="DOI">10.60843/5EGY-VC42</identifier>
+    <creators>
+        <creator>
+            <creatorName nameType="Personal">Feldman, Heidi Li</creatorName>
+            <givenName>Heidi Li</givenName>
+            <familyName>Feldman</familyName>
+            <nameIdentifier nameIdentifierScheme="ORCID" schemeURI="https://orcid.org">https://orcid.org/0000-0003-1905-2698</nameIdentifier>
+        </creator>
+    </creators>
+    <titles>
+        <title xml:lang="en">Science and Uncertainty in Mass Exposure Litigation</title>
+        <title xml:lang="en">Science and Uncertainty in Mass Exposure Litigation, 74 Tex. L. Rev. 1 (1995)</title>
+    </titles>
+    <publisher>Texas Law Review</publisher>
+    <publicationYear>1995</publicationYear>
+    <resourceType resourceTypeGeneral="JournalArticle">Article</resourceType>
+    <subjects>
+        <subject subjectScheme="Fields of Science and Technology (FOS)" schemeURI="http://www.oecd.org/science/inno/38235147.pdf" classificationCode="5.5">FOS: Law</subject>
+    </subjects>
+    <contributors>
+        <contributor contributorType="HostingInstitution">
+            <contributorName nameType="Organizational">Georgetown Law Library</contributorName>
+            <nameIdentifier nameIdentifierScheme="" schemeURI=""/>
+            <affiliation affiliationIdentifier="https://ror.org/05vzafd60" affiliationIdentifierScheme="ROR" schemeURI="https://ror.org">Georgetown University</affiliation>
+        </contributor>
+    </contributors>
+    <dates>
+        <date dateType="Issued">1995-11</date>
+        <date dateType="Collected">2016-04-25</date>
+    </dates>
+    <language>en</language>
+    <relatedIdentifiers>
+        <relatedIdentifier relatedIdentifierType="ISSN" relationType="IsPublishedIn" resourceTypeGeneral="Journal">0040-4411</relatedIdentifier>
+    </relatedIdentifiers>
+    <relatedItems>
+        <relatedItem relatedItemType="Journal" relationType="IsSourceOf">
+            <relatedItemIdentifier relatedItemIdentifierType="ISSN">0040-4411</relatedItemIdentifier>
+            <creators>
+                <creator>
+                    <creatorName>University of Texas School of Law</creatorName>
+                </creator>
+            </creators>
+            <titles>
+                <title>Texas Law Review</title>
+            </titles>
+            <publicationYear>1995</publicationYear>
+            <volume>74</volume>
+            <issue>1</issue>
+        </relatedItem>
+    </relatedItems>
+    <sizes/>
+    <formats>
+        <format>text/html</format>
+        <format>application/pdf</format>
+    </formats>
+    <version/>
+    <descriptions>
+        <description xml:lang="en" descriptionType="Abstract">Critics of the tort system have condemned courts for their alleged leniency in admitting scientific expert testimony, especially in mass exposure litigation. Claiming that this has resulted in an epidemic of "junk science" in the courtroom, these commentators have urged a more scientific approach to admissibility, intimating that employing more scientific standards would exclude scientific evidence favorable to plaintiffs, thereby demonstrating to factfinders that litigated substances are in fact safe or at least not unsafe.</description>
+    </descriptions>
+</resource>

data/spec/fixtures/schema_org.json

@@ -14,6 +14,7 @@
     }],
     "publisher": {
         "@type": "Organization",
+        "@id": "https://ror.org/04wxnsj81",
         "name": "DataCite"
     },
     "dateCreated": "2016-12-20",

data/spec/readers/datacite_reader_spec.rb

@@ -1759,4 +1759,54 @@ describe Bolognese::Metadata, vcr: true do
       )
     end
   end
+
+  it "blank nameIdentifier" do
+    input = fixture_path + "datacite_blank_name_identifier.xml"
+    subject = Bolognese::Metadata.new(input: input)
+    expect(subject.creators).to eq(
+      [
+        {
+          "nameType" => "Personal",
+          "name" => "KARST, SNO",
+          "givenName" => "SNO",
+          "familyName" => "KARST",
+          "nameIdentifiers" =>
+            [
+              {
+                "schemeUri" => "https://orcid.org",
+                "nameIdentifierScheme" => "ORCID"
+              }
+            ],
+          "affiliation" => []
+        }
+      ]
+    )
+    expect(subject.contributors).to eq(
+      [
+        {
+          "nameType" => "Personal",
+          "name" => "Manche, Yannick",
+          "givenName" => "Yannick",
+          "familyName" => "Manche",
+          "nameIdentifiers" =>
+            [
+              {
+                "schemeUri" => "https://orcid.org",
+                "nameIdentifierScheme" => "ORCID"
+              }
+            ],
+          "affiliation" => [],
+          "contributorType" => "ProjectMember"
+        }
+      ]
+    )
+  end
+
+  it "blank publisher" do
+    input = fixture_path + "datacite_blank_publisher.xml"
+    subject = Bolognese::Metadata.new(input: input)
+    expect(subject.publisher).to eq(
+      { "publisherIdentifier" => "https://ror.org/04wxnsj81" }
+    )
+  end
 end

data/spec/readers/schema_org_reader_spec.rb

@@ -169,7 +169,7 @@ describe Bolognese::Metadata, vcr: true do
       expect(subject.publication_year).to eq("2016")
       expect(subject.related_identifiers.length).to eq(3)
       expect(subject.related_identifiers.last).to eq("relatedIdentifier"=>"10.5438/55e5-t5c0", "relatedIdentifierType"=>"DOI", "relationType"=>"References")
-      expect(subject.publisher).to eq({"name"=>"DataCite"})
+      expect(subject.publisher).to eq({"name"=>"DataCite", "publisherIdentifier"=>"https://ror.org/04wxnsj81"})
     end
 
     it "GTEx dataset" do

data/spec/writers/citation_writer_spec.rb

@@ -48,5 +48,14 @@ describe Bolognese::Metadata, vcr: true do
       expect(subject.locale).to eq("en-US")
       expect(subject.citation).to eq("Clark, D. (2019). <i>Exploring the \"Many analysts, one dataset\" project from COS</i>. Gigantum, Inc. https://doi.org/10.34747/g6yb-3412")
     end
+
+    it "journal article with container title" do
+      input = fixture_path + "datacite_journal_article.xml"
+      subject = Bolognese::Metadata.new(input: input, from: "datacite")
+      expect(subject.types["citeproc"]).to eq("article-journal")
+      expect(subject.style).to eq("apa")
+      expect(subject.locale).to eq("en-US")
+      expect(subject.citation).to eq('Feldman, H. L. (1995). Science and Uncertainty in Mass Exposure Litigation. <i>Texas Law Review</i>. https://doi.org/10.60843/5egy-vc42')
+    end
   end
 end

data/spec/writers/jats_writer_spec.rb

@@ -179,6 +179,13 @@ describe Bolognese::Metadata, vcr: true do
       expect(jats.dig("day")).to be_nil
       expect(jats.dig("pub_id")).to eq("pub_id_type"=>"doi", "__content__"=>"10.34747/g6yb-3412")
     end
+
+    it "with publisher with language" do
+      input = fixture_path + 'datacite-xml-lang.xml'
+      subject = Bolognese::Metadata.new(input: input)
+      jats = Maremma.from_xml(subject.jats).fetch("element_citation", {})
+      expect(jats.dig("publisher_name")).to eq({"xml:lang"=>"en", "__content__"=>"OSI SAF"})
+    end
   end
 
   context "change metadata as datacite xml" do

data/spec/writers/schema_org_writer_spec.rb

@@ -365,5 +365,18 @@ describe Bolognese::Metadata, vcr: true do
       expect(json["@type"]).to eq("CreativeWork")
       expect(json["datePublished"]).to eq("2019")
     end
+
+    it "from Schema 4.5 with publisher attributes" do
+      input = fixture_path + 'datacite-example-full-v4.5.xml'
+      subject = Bolognese::Metadata.new(input: input)
+      json = JSON.parse(subject.schema_org)
+      expect(json["publisher"]).to eq(
+         {
+          "@type" => "Organization",
+          "@id" => "https://ror.org/04z8jg394",
+          "name" => "Example Publisher"
+         }
+      )
+    end
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: bolognese
 version: !ruby/object:Gem::Version
-  version: 2.0.0
+  version: 2.1.0
 platform: ruby
 authors:
 - Martin Fenner
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2023-12-04 00:00:00.000000000 Z
+date: 2024-01-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: maremma
@@ -556,6 +556,20 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
+- !ruby/object:Gem::Dependency
+  name: json-canonicalization
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '='
+      - !ruby/object:Gem::Version
+        version: 0.3.1
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '='
+      - !ruby/object:Gem::Version
+        version: 0.3.1
 description: Ruby gem and command-line utility for conversion of DOI metadata from
   and to different metadata formats, including schema.org.
 email: mfenner@datacite.org
@@ -814,7 +828,10 @@ files:
 - spec/fixtures/datacite-xml-lang.xml
 - spec/fixtures/datacite.json
 - spec/fixtures/datacite.xml
+- spec/fixtures/datacite_blank_name_identifier.xml
+- spec/fixtures/datacite_blank_publisher.xml
 - spec/fixtures/datacite_dataset.xml
+- spec/fixtures/datacite_journal_article.xml
 - spec/fixtures/datacite_malformed_creator.xml
 - spec/fixtures/datacite_missing_creator.xml
 - spec/fixtures/datacite_schema_3.xml