bolognese 1.1.8 → 1.1.11

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7a87dc9083b780cc8c6a1ca24fa0cc68fba665fa632356dec8cafc96b4ff1cb8
-  data.tar.gz: 06efb31ba651cdf2b685871ea91fd7aa19620dad5a5ba7105590ea3d5e8eb3c9
+  metadata.gz: a40e4d003bd824575e53ce801d8579333a01d87c868a9fb5f5d2f7d83a285985
+  data.tar.gz: b69f7a90359bddfd9f7f65882e9faeb20651f91bb71d634e3885b8cc33a5d3e3
 SHA512:
-  metadata.gz: e5d27ea262092d47ab7218268d544e1983cc75c769abcbc7e59beb7b3cdf26b8c18e7063532501acd176131278aa2e37ddb1c9926a002e5aa95c01343ed4034b
-  data.tar.gz: dc3e3348b0b883fe26fb1012ce3d13e27762e2247beebea81b6086ae217fc3e51cbfe27e6a366276d33fad23b86f806c6365eaba21b7d912717f01cca36fb05b
+  metadata.gz: 88b0e6943536624a09a0b65a5ad91baaf9e0c92086e5ed081daaa8d3cd9e9c2e84432a15d8292eeb3015883bc2d74e0183ba6bed4493776b2b035760910d95ba
+  data.tar.gz: 4738c206fd36978d43b57e91363efb0608b497341a22bb76b7825a3ecda213d27252f0d7c0e25fb57b2f91003b565ca086872db6fc33f05788f380af238fc1ec

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    bolognese (1.1.8)
+    bolognese (1.1.11)
       activesupport (>= 4.2.5, < 6)
       benchmark_methods (~> 0.7)
       bibtex-ruby (~> 4.1)
@@ -28,12 +28,12 @@ PATH
 GEM
   remote: https://rubygems.org/
   specs:
-    activesupport (5.2.2)
+    activesupport (5.2.3)
       concurrent-ruby (~> 1.0, >= 1.0.2)
       i18n (>= 0.7, < 2)
       minitest (~> 5.1)
       tzinfo (~> 1.1)
-    addressable (2.5.2)
+    addressable (2.6.0)
       public_suffix (>= 2.0.2, < 4.0)
     benchmark_methods (0.7)
     bibtex-ruby (4.4.7)
@@ -112,12 +112,12 @@ GEM
     nokogiri (1.8.5)
       mini_portile2 (~> 2.3.0)
     oj (2.18.5)
-    postrank-uri (1.0.23)
-      addressable (>= 2.3.0, < 2.6)
-      nokogiri (>= 1.6.1, < 1.9)
+    postrank-uri (1.0.24)
+      addressable (>= 2.4.0)
+      nokogiri (>= 1.8.0)
       public_suffix (>= 2.0.0, < 2.1)
     public_suffix (2.0.5)
-    rack (2.0.6)
+    rack (2.0.7)
     rack-test (0.8.3)
       rack (>= 1.0, < 3)
     rake (12.3.2)

data/lib/bolognese/author_utils.rb

@@ -51,7 +51,7 @@ module Bolognese
                  "givenName" => given_name,
                  "familyName" => family_name,
                  "nameIdentifiers" => name_identifiers,
-                 "affiliation" => parse_attributes(author.fetch("affiliation", nil), first: true),
+                 "affiliation" => parse_attributes(author.fetch("affiliation", nil)),
                  "contributorType" => contributor_type }.compact
 
       return author if family_name.present?
@@ -74,7 +74,7 @@ module Bolognese
           "givenName" => given_name,
           "familyName" => family_name,
           "nameIdentifiers" => name_identifiers,
-          "affiliation" => parse_attributes(author.fetch("affiliation", nil), first: true),
+          "affiliation" => parse_attributes(author.fetch("affiliation", nil)),
           "contributorType" => contributor_type }.compact
       else
         { "nameType" => name_type, "name" => name }.compact

data/lib/bolognese/datacite_utils.rb

@@ -31,12 +31,13 @@ module Bolognese
       insert_alternate_identifiers(xml)
       insert_subjects(xml)
       insert_contributors(xml)
-      insert_funding_references(xml)
       insert_dates(xml)
       insert_related_identifiers(xml)
       insert_version(xml)
       insert_rights_list(xml)
       insert_descriptions(xml)
+      insert_geo_locations(xml)
+      insert_funding_references(xml)
     end
 
     def insert_identifier(xml)
@@ -74,7 +75,9 @@ module Bolognese
       Array.wrap(person["nameIdentifiers"]).each do |ni|
         xml.nameIdentifier(ni["nameIdentifier"], 'nameIdentifierScheme' => ni["nameIdentifierScheme"], 'schemeURI' => ni["schemeUri"])
       end
-      xml.affiliation(person["affiliation"]) if person["affiliation"].present?
+      Array.wrap(person["affiliation"]).each do |affiliation|
+        xml.affiliation(affiliation) 
+      end
     end
 
     def insert_titles(xml)
@@ -214,11 +217,10 @@ module Bolognese
 
       xml.descriptions do
         if container && container["title"].present?
-          title = [container["title"]]
-          volume = "#{container["volume"]}" if container["volume"].present?
-          volume += "(#{container["issue"]})" if container["issue"].present?
+          issue = container["issue"].present? ? "(#{container["issue"]})" : nil
+          volume_issue = container["volume"].present? ? [container["volume"], issue].join("") : nil
           pages = [container["firstPage"], container["lastPage"]].compact.join("-") if container["firstPage"].present?
-          series_information = [title, volume, pages].compact.join(", ")
+          series_information = [container["title"], volume_issue, pages].compact.join(", ")
           xml.description(series_information, 'descriptionType' => "SeriesInformation")
         end
 
@@ -238,6 +240,45 @@ module Bolognese
       end
     end
 
+    def insert_geo_locations(xml)
+      return xml unless geo_locations.present?
+
+      xml.geoLocations do
+        geo_locations.each do |geo_location|
+          xml.geoLocation do
+            xml.geoLocationPlace(geo_location["geoLocationPlace"]) if geo_location["geoLocationPlace"]
+
+            if geo_location["geoLocationPoint"]
+              xml.geoLocationPoint do
+                xml.pointLatitude(geo_location.dig("geoLocationPoint", "pointLatitude"))
+                xml.pointLongitude(geo_location.dig("geoLocationPoint", "pointLongitude"))
+              end
+            end
+
+            if geo_location["geoLocationBox"]
+              xml.geoLocationBox do
+                xml.westBoundLongitude(geo_location.dig("geoLocationBox", "westBoundLongitude"))
+                xml.eastBoundLongitude(geo_location.dig("geoLocationBox", "eastBoundLongitude"))
+                xml.southBoundLatitude(geo_location.dig("geoLocationBox", "southBoundLatitude"))
+                xml.northBoundLatitude(geo_location.dig("geoLocationBox", "northBoundLatitude"))
+              end
+            end
+
+            if geo_location["geoLocationPolygon"]
+              xml.geoLocationPolygon do
+                geo_location["geoLocationPolygon"].each do |polygon_point|
+                  xml.polygonPoint do
+                    xml.pointLatitude(polygon_point["pointLatitude"])
+                    xml.pointLongitude(polygon_point["pointLongitude"])
+                  end
+                end
+              end
+            end
+          end
+        end
+      end
+    end
+
     def root_attributes
       { :'xmlns:xsi' => 'http://www.w3.org/2001/XMLSchema-instance',
         :'xsi:schemaLocation' => 'http://datacite.org/schema/kernel-4 http://schema.datacite.org/meta/kernel-4/metadata.xsd',

data/lib/bolognese/metadata.rb

@@ -67,7 +67,6 @@ module Bolognese
       @style = options[:style]
       @locale = options[:locale]
 
-      @regenerate = options[:regenerate]
       @sandbox = options[:sandbox]
 
       # options that come from the datacite database
@@ -102,6 +101,8 @@ module Bolognese
         :sizes
       ).compact
 
+      @regenerate = options[:regenerate] || read_options.present?
+
       # generate name for method to call dynamically
       @meta = @from.present? ? send("read_" + @from, { string: string, sandbox: options[:sandbox], doi: options[:doi], id: id }.merge(read_options)) : {}
     end

data/lib/bolognese/readers/datacite_reader.rb

@@ -187,7 +187,7 @@ module Bolognese
             }.compact
           end
         end.compact
-        container = set_container(meta)
+
         state = doi.present? || read_options.present? ? "findable" : "not_found"
 
         { "id" => id,
@@ -198,7 +198,7 @@ module Bolognese
           "titles" => titles,
           "creators" => get_authors(Array.wrap(meta.dig("creators", "creator"))),
           "contributors" => get_authors(Array.wrap(meta.dig("contributors", "contributor"))),
-          "container" => container,
+          "container" => set_container(meta),
           "publisher" => parse_attributes(meta.fetch("publisher", nil), first: true).to_s.strip.presence,
           "agency" => "DataCite",
           "funding_references" => funding_references,

data/lib/bolognese/utils.rb

@@ -550,7 +550,7 @@ module Bolognese
 
     def to_schema_org_creators(element)
       element = Array.wrap(element).map do |c|
-        c["affiliation"] = { "@type" => "Organization", "name" => c["affiliation"] } if c["affiliation"].present?
+        c["affiliation"] = { "@type" => "Organization", "name" => Array.wrap(c["affiliation"]).first } if c["affiliation"].present?
         c["@type"] = c["nameType"].present? ? c["nameType"][0..-3] : nil
         c["@id"] = Array.wrap(c["nameIdentifiers"]).first.to_h.fetch("nameIdentifier", nil)
         c["name"] = c["familyName"].present? ? [c["givenName"], c["familyName"]].join(" ") : c["name"]
@@ -560,7 +560,7 @@ module Bolognese
 
     def to_schema_org_contributors(element)
       element = Array.wrap(element).map do |c|
-        c["affiliation"] = { "@type" => "Organization", "name" => c["affiliation"] } if c["affiliation"].present?
+        c["affiliation"] = { "@type" => "Organization", "name" => Array.wrap(c["affiliation"]).first } if c["affiliation"].present?
         c["@type"] = c["nameType"].present? ? c["nameType"][0..-3] : nil
         c["@id"] = Array.wrap(c["nameIdentifiers"]).first.to_h.fetch("nameIdentifier", nil)
         c["name"] = c["familyName"].present? ? [c["givenName"], c["familyName"]].join(" ") : c["name"]

data/lib/bolognese/version.rb

@@ -1,3 +1,3 @@
 module Bolognese
-  VERSION = "1.1.8"
+  VERSION = "1.1.11"
 end

data/resources/kernel-4.2/metadata.xsd

@@ -329,11 +329,7 @@ Use the complete title of a license and include version information if applicabl
                 <xs:complexType mixed="true">
                   <xs:choice>
                     <xs:element name="br" minOccurs="0" maxOccurs="unbounded">
-                      <xs:simpleType>
-                        <xs:restriction base="xs:string">
-                          <xs:length value="0"/>
-                        </xs:restriction>
-                      </xs:simpleType>
+                      <xs:complexType/>
                     </xs:element>
                   </xs:choice>
                   <xs:attribute name="descriptionType" type="descriptionType" use="required"/>

data/resources/kernel-4/metadata.xsd

@@ -329,11 +329,7 @@ Use the complete title of a license and include version information if applicabl
                 <xs:complexType mixed="true">
                   <xs:choice>
                     <xs:element name="br" minOccurs="0" maxOccurs="unbounded">
-                      <xs:simpleType>
-                        <xs:restriction base="xs:string">
-                          <xs:length value="0"/>
-                        </xs:restriction>
-                      </xs:simpleType>
+                      <xs:complexType/>
                     </xs:element>
                   </xs:choice>
                   <xs:attribute name="descriptionType" type="descriptionType" use="required"/>

data/spec/fixtures/aida.json

@@ -0,0 +1,82 @@
+{
+  "@context": "http://schema.org",
+  "@type": "Dataset",
+  "@id": "https://doi.org/10.23698/aida/drov",
+  "name": "Ovary data from the Visual Sweden project DROID",
+  "about": "Pathology",
+  "url": "https://doi.aida.medtech4health.se/10.23698/aida/drov",
+  "author": [{
+    "@type": "Person",
+    "@id": "https://orcid.org/0000-0003-1298-517X",
+    "givenName": "Karin",
+    "familyName": "Lindman",
+    "name": "Karin Lindman"
+  }, {
+    "@type": "Person",
+    "@id": "https://orcid.org/",
+    "givenName": "Jerónimo",
+    "familyName": "Rose",
+    "name": "Jerónimo F. Rose"
+  }, {
+    "@type": "Person",
+    "@id": "https://orcid.org/",
+    "givenName": "Martin",
+    "familyName": "Lindvall",
+    "name": "Martin Lindvall"
+  }, {
+    "@type": "Person",
+    "@id": "https://orcid.org/0000-0001-7250-234X",
+    "givenName": "Caroline",
+    "familyName": "Bivik Stadler",
+    "name": "Caroline Bivik Stadler"
+  }],
+  "publisher": {
+    "@type": "Organization",
+    "name": "AIDA"
+  },
+  "copyrightYear": 2019,
+  "copyrightHolder": [{
+    "@type": "Organization",
+    "name": "Linköping University",
+    "url": "https://liu.se/"
+  }, {
+    "@type": "Person",
+    "@id": "https://orcid.org/0000-0002-9368-0177",
+    "givenName": "Claes",
+    "familyName": "Lundström",
+    "name": "Claes Lundström"
+  }],
+  "provider": [{
+    "@type": "Person",
+    "@id": "https://orcid.org/0000-0002-0128-870X",
+    "name": "Karin Lindman",
+    "email": "Karin.Lindman@regionostergotland.se"
+  }, {
+    "@type": "Person",
+    "@id": "https://orcid.org/0000-0002-9368-0177",
+    "name": "Claes Lundstrom",
+    "email": "claes.lundstrom@liu.se"
+  }, {
+    "@type": "Person",
+    "@id": "https://orcid.org/0000-0001-7250-234X",
+    "givenName": "Caroline",
+    "familyName": "Bivik Stadler",
+    "name": "Caroline Bivik Stadler"
+  }, {
+    "@type": "Person",
+    "@id": "https://orcid.org/0000-0001-6443-3604",
+    "name": "Joel Hedlund",
+    "email": "joel.hedlund@liu.se"
+  }],
+  "dateCreated": "2019-01-09",
+  "datePublished": "2019-01-09",
+  "dateModified": "2019-01-09",
+  "keywords": "pathology, whole slide imaging, annotated",
+  "version": "1.0",
+  "description": "This dataset consists of 174 WSI ovary whole slide images (WSI): 158\nmalignant and 16 benign. Eight of the most common, histological definable\ntumour types were annotated: high grade serous carcinoma (HGSC), low grade\nserous carcinoma (LGSC), clear cell carcinoma (CC), endometrioid\nadenocarcinoma (EN), metastastic serous carcinoma (MS), metastatic other\n(MO), serous borderline tumor (SB) and mucinous borderline tumor (MB). Also\nnormal ovarian tissue were annotated. 11258 separate annotations were made.\nFor the benign structures only the epithelial structures, stroma and support\ntissue were annotated.\n",
+  "license": {
+    "name": "Restricted access",
+    "id": "https://datasets.aida.medtech4health.se/10.23698/aida/drov#license"
+  },
+  "citation": null
+}

data/spec/fixtures/datacite-example-geolocation-2.xml

@@ -5,6 +5,7 @@
     <creator>
       <creatorName>Bales, Roger</creatorName>
       <affiliation>UC Merced</affiliation>
+      <affiliation>NSF</affiliation>
     </creator>
     <creator>
       <creatorName>Meadows, Matt</creatorName>

data/spec/readers/datacite_reader_spec.rb

@@ -236,6 +236,18 @@ describe Bolognese::Metadata, vcr: true do
       expect(subject.geo_locations).to eq([{"geoLocationPoint"=>{"pointLatitude"=>"-11.64583333", "pointLongitude"=>"-68.2975"}}])
     end
 
+    it "geo_location in separate input" do
+      input = fixture_path + 'datacite-geolocation-empty.xml'
+      geo_locations = [{"geoLocationPoint"=>{"pointLatitude"=>"49.0850736", "pointLongitude"=>"-123.3300992"}}]
+      subject = Bolognese::Metadata.new(input: input, geo_locations: geo_locations)
+      expect(subject.valid?).to be true
+      expect(subject.types["schemaOrg"]).to eq("Dataset")
+      expect(subject.geo_locations).to eq(geo_locations)
+
+      datacite = Maremma.from_xml(subject.datacite).fetch("resource", {})
+      expect(datacite.dig("geoLocations")).to eq("geoLocation"=>{"geoLocationPoint"=>{"pointLatitude"=>"49.0850736", "pointLongitude"=>"-123.3300992"}})
+    end
+
     it "xml:lang attribute" do
       input = fixture_path + 'datacite-xml-lang.xml'
       subject = Bolognese::Metadata.new(input: input)
@@ -321,7 +333,7 @@ describe Bolognese::Metadata, vcr: true do
       expect(subject.identifiers).to eq([{"identifier"=>"https://doi.org/10.6071/z7wc73", "identifierType"=>"DOI"}])
       expect(subject.doi).to eq("10.6071/z7wc73")
       expect(subject.creators.length).to eq(6)
-      expect(subject.creators.first).to eq("familyName"=>"Bales", "givenName"=>"Roger", "name"=>"Bales, Roger", "nameType"=>"Personal", "affiliation" => "UC Merced")
+      expect(subject.creators.first).to eq("familyName"=>"Bales", "givenName"=>"Roger", "name"=>"Bales, Roger", "nameType"=>"Personal", "affiliation" => ["UC Merced", "NSF"])
       expect(subject.titles).to eq([{"title"=>"Southern Sierra Critical Zone Observatory (SSCZO), Providence Creek\n      meteorological data, soil moisture and temperature, snow depth and air\n      temperature"}])
       expect(subject.publisher).to eq("UC Merced")
       expect(subject.dates).to eq([{"date"=>"2014-10-17", "dateType"=>"Updated"}, {"date"=>"2016-03-14T17:02:02Z", "dateType"=>"Available"}, {"date"=>"2013", "dateType"=>"Issued"}])

data/spec/readers/schema_org_reader_spec.rb

@@ -240,6 +240,28 @@ describe Bolognese::Metadata, vcr: true do
       expect(subject.funding_references).to eq([{"funderIdentifier"=>"https://doi.org/10.13039/100000050", "funderIdentifierType"=>"Crossref Funder ID", "funderName"=>"National Heart, Lung, and Blood Institute (NHLBI)"}])
     end
 
+    it "aida dataset" do
+      input = fixture_path + 'aida.json'
+      subject = Bolognese::Metadata.new(input: input)
+
+      expect(subject.valid?).to be true
+      expect(subject.identifiers).to eq([{"identifier"=>"https://doi.org/10.23698/aida/drov", "identifierType"=>"DOI"}])
+      expect(subject.url).to eq("https://doi.aida.medtech4health.se/10.23698/aida/drov")
+      expect(subject.types).to eq("bibtex"=>"misc", "citeproc"=>"dataset", "resourceTypeGeneral"=>"Dataset", "ris"=>"DATA", "schemaOrg"=>"Dataset")
+      # expect(subject.creators).to eq([{"familyName"=>"Lindman", "givenName"=>"Karin", "name"=>"Lindman, Karin", "nameIdentifiers"=>[{"nameIdentifier"=> "https://orcid.org/0000-0003-1298-517X", "nameIdentifierScheme"=>"ORCID", "schemeUri"=>"https://orcid.org"}], "nameType"=>"Personal"}])
+      expect(subject.titles).to eq([{"title"=>"Ovary data from the Visual Sweden project DROID"}])
+      expect(subject.version_info).to eq("1.0")
+      expect(subject.subjects).to eq([{"subject"=>"pathology"}, {"subject"=>"whole slide imaging"}, {"subject"=>"annotated"}])
+      expect(subject.dates).to eq([{"date"=>"2019-01-09", "dateType"=>"Issued"}, {"date"=>"2019-01-09", "dateType"=>"Created"}, {"date"=>"2019-01-09", "dateType"=>"Updated"}])
+      expect(subject.publication_year).to eq("2019")
+      expect(subject.id).to eq("https://doi.org/10.23698/aida/drov")
+      expect(subject.publisher).to eq("AIDA")
+      puts subject.rights_list
+      expect(subject.rights_list).to eq({"rights"=>"Restricted access", "rightsUri"=>"https://datasets.aida.medtech4health.se/10.23698/aida/drov#license"})
+      expect(subject.id).to eq("https://doi.org/10.23698/aida/drov")
+    end
+
+
     it "from attributes" do
       subject = Bolognese::Metadata.new(input: nil,
         from: "schema_org",

data/spec/writers/datacite_writer_spec.rb

@@ -190,7 +190,7 @@ describe Bolognese::Metadata, vcr: true do
       expect(subject.valid?).to be true
       expect(subject.types).to eq("bibtex"=>"misc", "citeproc"=>"dataset", "resourceType"=>"dataset", "resourceTypeGeneral"=>"Dataset", "ris"=>"DATA", "schemaOrg"=>"Dataset")
       expect(subject.creators.length).to eq(6)
-      expect(subject.creators.first).to eq("affiliation"=>"UC Merced", "familyName"=>"Bales", "givenName"=>"Roger", "name"=>"Bales, Roger", "nameType"=>"Personal")
+      expect(subject.creators.first).to eq("affiliation"=>["UC Merced", "NSF"], "familyName"=>"Bales", "givenName"=>"Roger", "name"=>"Bales, Roger", "nameType"=>"Personal")
       expect(subject.titles).to eq([{"title"=>"Southern Sierra Critical Zone Observatory (SSCZO), Providence Creek\n      meteorological data, soil moisture and temperature, snow depth and air\n      temperature"}])
       expect(subject.identifiers).to eq([{"identifier"=>"https://doi.org/10.6071/z7wc73", "identifierType"=>"DOI"}])
       expect(subject.rights_list).to eq([{"rights"=>"Creative Commons Attribution 4.0 International (CC BY 4.0)", "rightsUri"=>"https://creativecommons.org/licenses/by/4.0"}])
@@ -202,7 +202,7 @@ describe Bolognese::Metadata, vcr: true do
 
       datacite = Maremma.from_xml(subject.datacite).fetch("resource", {})
       expect(datacite.fetch("xsi:schemaLocation", "").split(" ").first).to eq("http://datacite.org/schema/kernel-4")
-      expect(datacite.dig("creators", "creator", 0, "affiliation")).to eq("UC Merced")
+      expect(datacite.dig("creators", "creator", 0, "affiliation")).to eq(["UC Merced", "NSF"])
     end
 
     it "with data citation schema.org" do

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: bolognese
 version: !ruby/object:Gem::Version
-  version: 1.1.8
+  version: 1.1.11
 platform: ruby
 authors:
 - Martin Fenner
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-03-12 00:00:00.000000000 Z
+date: 2019-04-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: maremma
@@ -672,6 +672,7 @@ files:
 - spec/datacite_utils_spec.rb
 - spec/doi_utils_spec.rb
 - spec/find_from_format_spec.rb
+- spec/fixtures/aida.json
 - spec/fixtures/citeproc-no-categories.json
 - spec/fixtures/citeproc.json
 - spec/fixtures/codemeta.json