commonmeta-ruby 3.2.5 → 3.2.7

data/resources/crossref/common5.3.1.xsd

@@ -1,4 +1,9 @@
-<xsd:schema xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:fr="http://www.crossref.org/fundref.xsd" xmlns:ct="http://www.crossref.org/clinicaltrials.xsd" xmlns:ai="http://www.crossref.org/AccessIndicators.xsd" xmlns:mml="http://www.w3.org/1998/Math/MathML"><!-- *************************************************************
+<xsd:schema xmlns:xsd="http://www.w3.org/2001/XMLSchema"
+   xmlns:fr="http://www.crossref.org/fundref.xsd"
+   xmlns:ct="http://www.crossref.org/clinicaltrials.xsd"
+   xmlns:ai="http://www.crossref.org/AccessIndicators.xsd"
+   xmlns:mml="http://www.w3.org/1998/Math/MathML">
+   <!--*************************************************************
 	Crossref Metadata Input Schema, common file v. 5.3.1
 	*************************************************************
 	
@@ -8,7 +13,7 @@ As of version 4.5.0 all changes are noted in the release notes for the schema re
 	
 Prior updates are noted as comments within previous schema versions.
 	
-   -->
+  -->
    <xsd:import namespace="http://www.w3.org/1998/Math/MathML" schemaLocation="http://www.w3.org/Math/XMLSchema/mathml3/mathml3.xsd"/>
    <xsd:import namespace="http://www.crossref.org/fundref.xsd" schemaLocation="fundref.xsd"/>
    <xsd:import namespace="http://www.crossref.org/AccessIndicators.xsd" schemaLocation="AccessIndicators.xsd"/>
@@ -20,7 +25,7 @@ Prior updates are noted as comments within previous schema versions.
    </xsd:complexType>
    <xsd:group name="face_markup">
       <xsd:annotation>
-         <xsd:documentation>The following are basic data types for face markup. Face markup that appears in the title, subtitle, and original_language_title elements should be retained when depositing metadata. Face markup in other elements (e.g. small caps in author names) must be dropped. Face markup support includes bold (b), italic (i), underline (u), over-line (ovl), superscript (sup), subscript (sub), small caps (scp), and typewriter text (tt). See https://support.crossref.org/hc/en-us/articles/214532023 MathML may also be included using the &apos;mml&apos; namespace prefix.</xsd:documentation>
+         <xsd:documentation>The following are basic data types for face markup. Face markup that appears in the title, subtitle, and original_language_title elements should be retained when depositing metadata. Face markup in other elements (e.g. small caps in author names) must be dropped. Face markup support includes bold (b), italic (i), underline (u), over-line (ovl), superscript (sup), subscript (sub), small caps (scp), and typewriter text (tt). See https://support.crossref.org/hc/en-us/articles/214532023 MathML may also be included using the 'mml' namespace prefix.</xsd:documentation>
       </xsd:annotation>
       <xsd:choice>
          <xsd:element ref="b"/>
@@ -50,7 +55,7 @@ Prior updates are noted as comments within previous schema versions.
    <xsd:element name="font" type="xrefFaces"/>
    <xsd:element name="a">
       <xsd:annotation>
-         <xsd:documentation> content is &quot;Inline&quot; except that anchors shouldn&apos;t be nested </xsd:documentation>
+         <xsd:documentation> content is "Inline" except that anchors shouldn't be nested			</xsd:documentation>
       </xsd:annotation>
       <xsd:complexType mixed="true">
          <xsd:simpleContent>
@@ -133,7 +138,7 @@ Prior updates are noted as comments within previous schema versions.
    </xsd:element>
    <xsd:element name="edition_number">
       <xsd:annotation>
-         <xsd:documentation>The edition number of a book. edition_number should include only a number and not additional text such as &quot;edition&quot;. For example, you should submit &quot;3&quot;, not &quot;third edition&quot; or &quot;3rd edition&quot;. Roman numerals are acceptable.</xsd:documentation>
+         <xsd:documentation>The edition number of a book. edition_number should include only a number and not additional text such as "edition". For example, you should submit "3", not "third edition" or "3rd edition". Roman numerals are acceptable.</xsd:documentation>
       </xsd:annotation>
       <xsd:simpleType>
          <xsd:restriction base="xsd:string">
@@ -141,11 +146,12 @@ Prior updates are noted as comments within previous schema versions.
             <xsd:minLength value="1"/>
          </xsd:restriction>
       </xsd:simpleType>
-   </xsd:element><!--  =============================================================
+   </xsd:element>
+   <!-- =============================================================
 
                           5. Journal elements
 
-     =============================================================  -->
+     ============================================================= -->
    <xsd:element name="issue">
       <xsd:annotation>
          <xsd:documentation>The issue number or name in which an article is published. The issue number takes precedence over any other name. For example, if an issue has only a seasonal name, then the season should be listed in issue.</xsd:documentation>
@@ -156,11 +162,12 @@ Prior updates are noted as comments within previous schema versions.
             <xsd:minLength value="1"/>
          </xsd:restriction>
       </xsd:simpleType>
-   </xsd:element><!--  =============================================================
+   </xsd:element>
+   <!-- =============================================================
 
        6. Elements shared by content types
 
-     =============================================================  -->
+     ============================================================= -->
    <xsd:element name="doi_data">
       <xsd:annotation>
          <xsd:documentation>The container for elements related directly to a DOI. </xsd:documentation>
@@ -203,10 +210,9 @@ Prior updates are noted as comments within previous schema versions.
    </xsd:simpleType>
    <xsd:element name="collection">
       <xsd:annotation>
-         <xsd:documentation>
-Container for item elements containing non-primary URIs associated with the item being registered. Collections are supported for the following (defined in the property attribute):
-<ul>
-               <li>list-based: Multiple Resolution, more info: https://www.crossref.org/education/content-registration/creating-and-managing-dois/multiple-resolution/</li>
+         <xsd:documentation>Container for item elements containing non-primary URIs associated with the item being registered. Collections are supported for the following (defined in the property attribute):				
+            <ul>
+               <li>list-based:  Multiple Resolution, more info: https://www.crossref.org/education/content-registration/creating-and-managing-dois/multiple-resolution/</li>
                <li>country-based: more info: https://www.crossref.org/education/content-registration/creating-and-managing-dois/multiple-resolution/#00130</li>
                <li>crawler-based: for Similarity Check URLs, more info: https://www.crossref.org/education/similarity-check/participate/urls-for-new-deposits/</li>
                <li>text-mining: supply specific URLs for text and data mining, more info: https://www.crossref.org/education/retrieve-metadata/rest-api/text-and-data-mining-for-members/</li>
@@ -540,7 +546,7 @@ Container for item elements containing non-primary URIs associated with the item
    </xsd:element>
    <xsd:attributeGroup name="reference_distribution_opts.att">
       <xsd:annotation>
-         <xsd:documentation>NOTE: not currently supported. Use to flag references for distribution. &quot;none&quot; is the default and follows current protocol - references are only distributed to everyone if the prefix level permission is set, otherwise reference distribution is limited to the DOI owner. Setting the value to &quot;query&quot; releases references to anyone making a query request (this overrides any established prefix level permission). Value &quot;any&quot; allows bulk distribution to anyone (using a Crossref query account) using the OAI-PMH protocol, and also releases references to anyone making a query request.</xsd:documentation>
+         <xsd:documentation>NOTE: not currently supported.				Use to flag references for distribution. "none" is the default and follows current protocol - references are only distributed to everyone if the prefix level permission is set, otherwise reference distribution is limited to the DOI owner. Setting the value to "query" releases references to anyone making a query request (this overrides any established prefix level permission). Value "any" allows bulk distribution to anyone (using a Crossref query account) using the OAI-PMH protocol, and also releases references to anyone making a query request.</xsd:documentation>
       </xsd:annotation>
       <xsd:attribute name="reference_distribution_opts" use="optional" default="none">
          <xsd:simpleType>
@@ -551,7 +557,9 @@ Container for item elements containing non-primary URIs associated with the item
             </xsd:restriction>
          </xsd:simpleType>
       </xsd:attribute>
-   </xsd:attributeGroup><!-- END NAME ELEMENTs --><!-- ============   end various dates ==========  -->
+   </xsd:attributeGroup>
+   <!--END NAME ELEMENTs-->
+   <!--============   end various dates ========== -->
    <xsd:simpleType name="doi_t">
       <xsd:restriction base="xsd:string">
          <xsd:maxLength value="2048"/>
@@ -703,7 +711,7 @@ Container for item elements containing non-primary URIs associated with the item
    </xsd:complexType>
    <xsd:element name="citation">
       <xsd:annotation>
-         <xsd:documentation>citation is used to deposit each reference in the reference list of the item for which the DOI is being deposited. For details see: https://www.crossref.org/education/metadata-stewardship/maintaining-your-metadata/add-references/ </xsd:documentation>
+         <xsd:documentation>citation is used to deposit each reference in the reference list of the item for which the DOI is being deposited. For details see:				https://www.crossref.org/education/metadata-stewardship/maintaining-your-metadata/add-references/ </xsd:documentation>
       </xsd:annotation>
       <xsd:complexType>
          <xsd:complexContent>
@@ -729,7 +737,7 @@ Container for item elements containing non-primary URIs associated with the item
    </xsd:attributeGroup>
    <xsd:element name="unstructured_citation">
       <xsd:annotation>
-         <xsd:documentation>A citation to an item that is not structured with the Crossref citation model. &apos;unstructured_citation&apos; supports deposit of references for which no structural information is available. </xsd:documentation>
+         <xsd:documentation>A citation to an item that is not structured with the Crossref citation model.  'unstructured_citation' supports deposit of references for which no structural information is available. </xsd:documentation>
       </xsd:annotation>
       <xsd:complexType mixed="true">
          <xsd:choice minOccurs="0" maxOccurs="unbounded">
@@ -783,7 +791,7 @@ Container for item elements containing non-primary URIs associated with the item
             <xsd:element ref="crossmark_domains" minOccurs="0" maxOccurs="unbounded"/>
             <xsd:element ref="crossmark_domain_exclusive" minOccurs="0">
                <xsd:annotation>
-                  <xsd:documentation>Some publishers encourage broad third party hosting of the publisher&apos;s content. Other publishers do not. And still others vary their policy depending on whether a particular article has been published under an OA policy or not. This boolean flag allows the publisher to indicate whether the Crossmarked content will only legitimately be updated on the Crossmark domain (true) or whether the publisher encourages updating the content on other sites as well (false).</xsd:documentation>
+                  <xsd:documentation>Some publishers encourage broad third party hosting of the publisher's content. Other publishers do not. And still others vary their policy depending on whether a particular article has been published under an OA policy or not. This boolean flag allows the publisher to indicate whether the Crossmarked content will only legitimately be updated on the Crossmark domain (true) or whether the publisher encourages updating the content on other sites as well (false).</xsd:documentation>
                </xsd:annotation>
             </xsd:element>
             <xsd:sequence minOccurs="0">
@@ -795,13 +803,13 @@ Container for item elements containing non-primary URIs associated with the item
    </xsd:element>
    <xsd:element name="crossmark_policy" type="doi_t">
       <xsd:annotation>
-         <xsd:documentation>A DOI which points to a publisher&apos;s CrossMark policy document. Publishers might have different policies for different publications.</xsd:documentation>
+         <xsd:documentation>A DOI which points to a publisher's CrossMark policy document. Publishers might have different policies for different publications.</xsd:documentation>
       </xsd:annotation>
    </xsd:element>
    <xsd:element name="crossmark_version" type="xsd:string"/>
    <xsd:element name="crossmark_domains">
       <xsd:annotation>
-         <xsd:documentation>Container element for crossmark_domain. A list of domains where the publisher maintains updates and corrections to their content. Minimally, one of these should include the Internet domain name of the publisher&apos;s web site(s), but the publisher might also decide to include 3rd party aggregators (e.g. Ebsco, IngentaConnect) or archives with which the publisher has agreements to update the content </xsd:documentation>
+         <xsd:documentation>Container element for crossmark_domain. A list of domains where the publisher maintains updates and corrections to their content. Minimally, one of these should include the Internet domain name of the publisher's web site(s), but the publisher might also decide to include 3rd party aggregators (e.g. Ebsco, IngentaConnect) or archives with which the publisher has agreements to update the content </xsd:documentation>
       </xsd:annotation>
       <xsd:complexType>
          <xsd:sequence>
@@ -811,7 +819,7 @@ Container for item elements containing non-primary URIs associated with the item
    </xsd:element>
    <xsd:element name="crossmark_domain">
       <xsd:annotation>
-         <xsd:documentation>This should be a simple Internet domain name or subdomain name (e.g. www.psychoceramics.org or psychoceramics.org). It is used to identify when a referring URL is coming from a Crossmark domain. A &quot;crossmark_domain&quot; is made up of two subelements; a &quot;domain&quot; and a &quot;filter&quot;. The filter is only needed for use in situations where content from multiple publishers/publications is on the same host with the same domain name (e.g. an aggregator) and one needs to use the referrer&apos;s URI &quot;path&quot; to further determine whether the content in a crossmark domain.</xsd:documentation>
+         <xsd:documentation>This should be a simple Internet domain name or subdomain name (e.g. www.psychoceramics.org or psychoceramics.org). It is used to identify when a referring URL is coming from a Crossmark domain. A "crossmark_domain" is made up of two subelements; a "domain" and a "filter". The filter is only needed for use in situations where content from multiple publishers/publications is on the same host with the same domain name (e.g. an aggregator) and one needs to use the referrer's URI "path" to further determine whether the content in a crossmark domain.</xsd:documentation>
       </xsd:annotation>
       <xsd:complexType>
          <xsd:sequence>
@@ -827,7 +835,7 @@ Container for item elements containing non-primary URIs associated with the item
    </xsd:element>
    <xsd:element name="filter" type="xsd:string">
       <xsd:annotation>
-         <xsd:documentation>The filter element is used to disambiguate content in situations where multiple publishers share the same host (e.g. when on an aggregated platform). It should contain a substring of the path that can be used to uniquely identify a publisher&apos;s or publication&apos;s content. For instance, using the string &quot;alpsp&quot; here would help the CrossMark system distinguish between ALPSP publications on the ingentaconnect host and other publications on the same host.</xsd:documentation>
+         <xsd:documentation>The filter element is used to disambiguate content in situations where multiple publishers share the same host (e.g. when on an aggregated platform). It should contain a substring of the path that can be used to uniquely identify a publisher's or publication's content. For instance, using the string "alpsp" here would help the CrossMark system distinguish between ALPSP publications on the ingentaconnect host and other publications on the same host.</xsd:documentation>
       </xsd:annotation>
    </xsd:element>
    <xsd:element name="crossmark_domain_exclusive" type="xsd:boolean"/>
@@ -843,16 +851,15 @@ Container for item elements containing non-primary URIs associated with the item
    </xsd:element>
    <xsd:element name="update">
       <xsd:annotation>
-         <xsd:documentation>The DOI of the content being updated (e.g. corrected, retracted, etc.) In the CrossMark Terms and Conditions &quot;updates&quot; are defined as changes that are likely to &quot;change the reader’s interpretation or crediting of the work.&quot; That is, *editorially significant* changes. &quot;Updates&quot; should not include minor changes to spelling, punctuation, formatting, etc. </xsd:documentation>
+         <xsd:documentation>The DOI of the content being updated (e.g. corrected, retracted, etc.) In the CrossMark Terms and Conditions "updates" are defined as changes that are likely to "change the reader’s interpretation or crediting of the work." That is, *editorially significant* changes. "Updates" should not include minor changes to spelling, punctuation, formatting, etc. </xsd:documentation>
       </xsd:annotation>
       <xsd:complexType mixed="true">
          <xsd:simpleContent>
             <xsd:extension base="doi_t">
                <xsd:attribute name="type" use="required" type="cm_update_type">
                   <xsd:annotation>
-                     <xsd:documentation>
-This attribute should be used to list the update type. Allowed update types are:
-<ul>
+                     <xsd:documentation>This attribute should be used to list the update type. Allowed update types are: 
+                        <ul>
                            <li>addendum</li>
                            <li>clarification</li>
                            <li>correction</li>
@@ -913,27 +920,27 @@ This attribute should be used to list the update type. Allowed update types are:
          </xsd:choice>
          <xsd:attribute name="explanation" type="xsd:anyURI">
             <xsd:annotation>
-               <xsd:documentation>If the publisher wants to provide a further explanation of what the particular &quot;assertion&quot; means, they can link to such an explanation by providing an appropriate url on the &quot;explanation&quot; attribute.</xsd:documentation>
+               <xsd:documentation>If the publisher wants to provide a further explanation of what the particular "assertion" means, they can link to such an explanation by providing an appropriate url on the "explanation" attribute.</xsd:documentation>
             </xsd:annotation>
          </xsd:attribute>
          <xsd:attribute name="group_label" type="cm_assertion_group_label">
             <xsd:annotation>
-               <xsd:documentation>This is the human-readable form of the &quot;group_name&quot; attribute. This is what will be displayed in the group headings on the Crossmark metadata record dialog.</xsd:documentation>
+               <xsd:documentation>This is the human-readable form of the "group_name" attribute. This is what will be displayed in the group headings on the Crossmark metadata record dialog.</xsd:documentation>
             </xsd:annotation>
          </xsd:attribute>
          <xsd:attribute name="group_name" type="cm_assertion_group_name">
             <xsd:annotation>
-               <xsd:documentation>Some assertions could be logically &quot;grouped&quot; together in the CrossMark dialog. For instance, if the publisher is recording several pieces of metadata related to funding sources (source name, percentage, grant number), they may want to make sure that these three assertions are grouped next to each-other in the CrossMark dialog. The group_name attribute is the machine-readable value that will be used for grouping such assertions.</xsd:documentation>
+               <xsd:documentation>Some assertions could be logically "grouped" together in the CrossMark dialog. For instance, if the publisher is recording several pieces of metadata related to funding sources (source name, percentage, grant number), they may want to make sure that these three assertions are grouped next to each-other in the CrossMark dialog. The group_name attribute is the machine-readable value that will be used for grouping such assertions.</xsd:documentation>
             </xsd:annotation>
          </xsd:attribute>
          <xsd:attribute name="label" type="cm_assertion_label">
             <xsd:annotation>
-               <xsd:documentation>This is the human-readable version of the name attribute which will be displayed in the CrossMark dialog. If this attribute is missing, then the value of the assertion will *not* be displayed in the dialog. Publishers may want to &quot;hide&quot; assertions this way in cases where the assertion value is too large or too complex to display in the dialog, but where the assertion is nonetheless valuable enough to include in API queries and metadata dumps (e.g. detailed licensing terms)</xsd:documentation>
+               <xsd:documentation>This is the human-readable version of the name attribute which will be displayed in the CrossMark dialog. If this attribute is missing, then the value of the assertion will *not* be displayed in the dialog. Publishers may want to "hide" assertions this way in cases where the assertion value is too large or too complex to display in the dialog, but where the assertion is nonetheless valuable enough to include in API queries and metadata dumps (e.g. detailed licensing terms)</xsd:documentation>
             </xsd:annotation>
          </xsd:attribute>
          <xsd:attribute name="name" use="required" type="cm_assertion_name">
             <xsd:annotation>
-               <xsd:documentation>This is the machine-readable name of the assertion. Use the &quot;label&quot; attribute to provide a human-readable version of the name.</xsd:documentation>
+               <xsd:documentation>This is the machine-readable name of the assertion. Use the "label" attribute to provide a human-readable version of the name.</xsd:documentation>
             </xsd:annotation>
          </xsd:attribute>
          <xsd:attribute name="order" type="xsd:integer">
@@ -1024,7 +1031,7 @@ This attribute should be used to list the update type. Allowed update types are:
    </xsd:element>
    <xsd:element name="std_family_designator" type="std_designator_t">
       <xsd:annotation>
-         <xsd:documentation>Provides for defining a DOI for a broad grouping of standards. </xsd:documentation>
+         <xsd:documentation>Provides for defining a DOI for a broad grouping of standards.			</xsd:documentation>
       </xsd:annotation>
    </xsd:element>
    <xsd:element name="std_set_designator">
@@ -1059,12 +1066,12 @@ This attribute should be used to list the update type. Allowed update types are:
    </xsd:element>
    <xsd:element name="std_adopted_from" type="std_designatorvalue_t">
       <xsd:annotation>
-         <xsd:documentation>Designator for standard from which the current deposit is adopted. </xsd:documentation>
+         <xsd:documentation>Designator for standard from which the current deposit is adopted.			</xsd:documentation>
       </xsd:annotation>
    </xsd:element>
    <xsd:element name="std_revision_of" type="std_designatorvalue_t">
       <xsd:annotation>
-         <xsd:documentation>Designator for the previous revision of the standard being deposited. (note: use alt_as_published for revisions within designators having common stem) </xsd:documentation>
+         <xsd:documentation>Designator for the previous revision of the standard being deposited. (note: use alt_as_published for revisions within designators having common stem)			</xsd:documentation>
       </xsd:annotation>
    </xsd:element>
    <xsd:element name="standards_body">
@@ -1103,7 +1110,8 @@ This attribute should be used to list the update type. Allowed update types are:
    </xsd:complexType>
    <xsd:element name="std_designator" type="std_designatorvalue_t"/>
    <xsd:element name="std_alt_script" type="std_designatorvalue_t"/>
-   <xsd:element name="std_variant_form" type="std_designatorvalue_t"/><!--  ================SCN policy =============================  -->
+   <xsd:element name="std_variant_form" type="std_designatorvalue_t"/>
+   <!-- ================SCN policy ============================= -->
    <xsd:element name="scn_policies">
       <xsd:annotation>
          <xsd:documentation>A wrapper for Scholarly Sharing Network (SCN) policy information</xsd:documentation>

data/spec/author_utils_spec.rb

@@ -44,10 +44,42 @@ describe Commonmeta::Metadata, vcr: true do
       expect(subject.is_personal_name?(name: author['name'])).to be true
     end
 
+    it 'has unknown given name and middle initial' do
+      author = { 'name' => 'Tejas S. Sathe' }
+      expect(subject.is_personal_name?(name: author['name'])).to be true
+    end
+
     it 'has no info' do
       author = { 'name' => 'M Fenner' }
       expect(subject.is_personal_name?(name: author['name'])).to be true
     end
+
+    it 'name with title' do
+      author = { 'name' => 'Tejas S. Sathe, MD' }
+      expect(subject.is_personal_name?(name: author['name'])).to be true
+    end
+  end
+
+  context 'cleanup_author' do
+    it 'Smith J.' do
+      author = 'Smith J.'
+      expect(subject.cleanup_author(author)).to eq('Smith, J.')
+    end
+
+    it 'Smith, John' do
+      author = 'Smith, John'
+      expect(subject.cleanup_author(author)).to eq('Smith, John')
+    end
+
+    it 'John Smith' do
+      author = 'John Smith'
+      expect(subject.cleanup_author(author)).to eq('John Smith')
+    end
+
+    it 'with email' do
+      author = 'noreply@blogger.com (Roderic Page)'
+      expect(subject.cleanup_author(author)).to eq('Roderic Page')
+    end
   end
 
   context 'get_one_author' do
@@ -69,6 +101,12 @@ describe Commonmeta::Metadata, vcr: true do
       )
     end
 
+    it 'has name with title' do
+      author = { 'name' => "Tejas S. Sathe, MD" }
+      response = subject.get_one_author(author)
+      expect(response).to eq('givenName' => 'Tejas S.', 'familyName' => 'Sathe', 'type' => 'Person')
+    end
+
     it 'has name in display-order with ORCID' do
       input = 'https://doi.org/10.6084/M9.FIGSHARE.4700788'
       subject = described_class.new(input: input)

data/spec/cli_spec.rb

@@ -341,11 +341,9 @@ describe Commonmeta::CLI do
     end
   end
 
-  describe "json_feed", vcr: true do
-    let(:input) { "prmb582" }
-
-    it "blog post" do
-      expect { subject.json_feed input }.to output(/89a326c6-f5d1-443a-99b8-ea75e705ce11\n/).to_stdout
-    end
-  end
+  # describe "json_feed_unregistered", vcr: true do
+  #   it "blog post uuid" do
+  #     expect { subject.json_feed_unregistered }.to output(/e4872b3e-cd15-407d-8406-33642c1a98b0\n/).to_stdout
+  #   end
+  # end
 end