refactoring of xml eml builder

GlobalNamesArchitecture · Dec 30, 2013 · aa40735 · aa40735
1 parent 915be98
commit aa40735
Show file tree

Hide file tree

Showing 4 changed files with 161 additions and 49 deletions.
diff --git a/lib/dwc-archive/generator_eml_xml.rb b/lib/dwc-archive/generator_eml_xml.rb
@@ -22,51 +22,9 @@ def create
             :'xmlns:res' =>        'eml://ecoinformatics.org/resource-2.1.1',
             :'xmlns:dc' =>         'http://purl.org/dc/terms/',
             :'xmlns:xsi' =>        'http://www.w3.org/2001/XMLSchema-instance',
-            :'xsi:schemaLocation' => 'eml_uri')  do
-            xml.dataset(id: @data[:id]) do
-              xml.title(@data[:title])
-              xml.license(@data[:license])
-              contacts = []
-              @data[:authors].each_with_index do |a, i|
-                creator_id = i + 1
-                contacts << creator_id
-                xml.creator(id: creator_id, scope: 'document') do
-                  xml.individualName do
-                    xml.givenName(a[:first_name])
-                    xml.surName(a[:last_name])
-                  end
-                  xml.organizationName(a[:organization]) if a[:organization]
-                  xml.positionName(a[:position]) if a[:position]
-                  xml.onlineUrl(a[:url]) if a[:url]
-                  xml.electronicMailAddress(a[:email])
-                end
-              end
-              @data[:metadata_providers].each_with_index do |a, i|
-                xml.metadataProvider do
-                  xml.individualName do
-                    xml.givenName(a[:first_name])
-                    xml.surName(a[:last_name])
-                  end
-                  xml.organizationName(a[:organization]) if a[:organization]
-                  xml.positionName(a[:position]) if a[:position]
-                  xml.onlineUrl(a[:url]) if a[:url]
-                  xml.electronicMailAddress(a[:email])
-                end
-              end if @data[:metadata_providers]
-              xml.pubDate(Time.now.to_s)
-              xml.abstract() do
-                xml.para(@data[:abstract])
-              end
-              contacts.each do |contact|
-                xml.contact { xml.references(contact) }
-              end
-            end
-            xml.additionalMetadata do
-              xml.metadata do
-                xml.citation(@data[:citation])
-                xml.resourceLogoUrl(@data[:logo_url]) if @data[:logo_url]
-              end
-            end
+            :'xsi:schemaLocation' => 'eml_uri') do
+            build_dataset(xml)
+            build_additional_metadata(xml)
             xml.parent.namespace = xml.parent.namespace_definitions.first
           end
         end
@@ -77,6 +35,85 @@ def create
       end
 
       private
+
+      def build_dataset(xml)
+        xml.dataset(id: @data[:id]) do
+          xml.title(@data[:title])
+          xml.license(@data[:license])
+          contacts = []
+          build_authors(xml, contacts)
+          build_metadata_providers(xml)
+          xml.pubDate(Time.now.to_s)
+          build_abstract(xml)
+          build_contacts(xml, contacts)
+        end
+      end
+
+      def build_abstract(xml)
+        xml.abstract() do
+          xml.para(@data[:abstract])
+        end
+      end
+
+      def build_contacts(xml, contacts)
+        contacts.each do |contact|
+          xml.contact { xml.references(contact) }
+        end
+      end
+
+      def build_metadata_providers(xml)
+        @data[:metadata_providers].each_with_index do |a, i|
+          xml.metadataProvider do
+            build_person(xml, a)
+          end
+        end if @data[:metadata_providers]
+      end
+
+      def build_authors(xml, contacts)
+        @data[:authors].each_with_index do |a, i|
+          creator_id = i + 1
+          contacts << creator_id
+          xml.creator(id: creator_id, scope: 'document') do
+            build_person(xml, a)
+          end
+        end
+      end
+
+      def build_additional_metadata(xml)
+        xml.additionalMetadata do
+          xml.metadata do
+            xml.citation(@data[:citation])
+            xml.resourceLogoUrl(@data[:logo_url]) if @data[:logo_url]
+          end
+        end
+      end
+
+      def build_header(xml)
+        xml.eml(packageId:      "%s/%s" % [@data[:id], timestamp],
+          system:               @data[:system] || 'http://globalnames.org',
+          :'xml:lang' =>         'en',
+          :'xmlns:eml' =>        'eml://ecoinformatics.org/eml-2.1.1',
+          :'xmlns:md' =>         'eml://ecoinformatics.org/methods-2.1.1',
+          :'xmlns:proj' =>       'eml://ecoinformatics.org/project-2.1.1',
+          :'xmlns:d' =>          'eml://ecoinformatics.org/dataset-2.1.1',
+          :'xmlns:res' =>        'eml://ecoinformatics.org/resource-2.1.1',
+          :'xmlns:dc' =>         'http://purl.org/dc/terms/',
+          :'xmlns:xsi' =>        'http://www.w3.org/2001/XMLSchema-instance',
+          :'xsi:schemaLocation' => 'eml_uri') 
+      end
+
+      def build_person(xml, data)
+        a = data
+        xml.individualName do
+          xml.givenName(a[:first_name])
+          xml.surName(a[:last_name])
+        end
+        xml.organizationName(a[:organization]) if a[:organization]
+        xml.positionName(a[:position]) if a[:position]
+        xml.onlineUrl(a[:url]) if a[:url]
+        xml.electronicMailAddress(a[:email])
+      end
+
       def timestamp
         t = Time.now.getutc.to_a[0..5].reverse
         t[0..2].join('-') + '::' + t[-3..-1].join(':')

diff --git a/spec/files/generator_eml.xml b/spec/files/generator_eml.xml
@@ -0,0 +1,47 @@
+<?xml version="1.0"?>
+<eml:eml xmlns:eml="eml://ecoinformatics.org/eml-2.1.1" xmlns:md="eml://ecoinformatics.org/methods-2.1.1" xmlns:proj="eml://ecoinformatics.org/project-2.1.1" xmlns:d="eml://ecoinformatics.org/dataset-2.1.1" xmlns:res="eml://ecoinformatics.org/resource-2.1.1" xmlns:dc="http://purl.org/dc/terms/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" packageId="1234/2013-12-30::19:45:33" system="http://globalnames.org" xml:lang="en" xsi:schemaLocation="eml_uri">
+  <dataset id="1234">
+    <title>Test Classification</title>
+    <license>http://creativecommons.org/licenses/by-sa/3.0/</license>
+    <creator id="1" scope="document">
+      <individualName>
+        <givenName>John</givenName>
+        <surName>Doe</surName>
+      </individualName>
+      <organizationName>Example</organizationName>
+      <positionName>Assistant Professor</positionName>
+      <onlineUrl>http://example.org</onlineUrl>
+      <electronicMailAddress>jdoe@example.com</electronicMailAddress>
+    </creator>
+    <creator id="2" scope="document">
+      <individualName>
+        <givenName>Jane</givenName>
+        <surName>Doe</surName>
+      </individualName>
+      <electronicMailAddress>jane@example.com</electronicMailAddress>
+    </creator>
+    <metadataProvider>
+      <individualName>
+        <givenName>Jim</givenName>
+        <surName>Doe</surName>
+      </individualName>
+      <onlineUrl>http://aggregator.example.org</onlineUrl>
+      <electronicMailAddress>jimdoe@example.com</electronicMailAddress>
+    </metadataProvider>
+    <pubDate>2013-12-30 14:45:33 -0500</pubDate>
+    <abstract>
+      <para>test classification</para>
+    </abstract>
+    <contact>
+      <references>1</references>
+    </contact>
+    <contact>
+      <references>2</references>
+    </contact>
+  </dataset>
+  <additionalMetadata>
+    <metadata>
+      <citation>Test classification: Doe John, Doe Jane, Taxnonmy, 10, 1, 2010</citation>
+    </metadata>
+  </additionalMetadata>
+</eml:eml>
diff --git a/spec/files/generator_meta.xml b/spec/files/generator_meta.xml
@@ -0,0 +1,19 @@
+<?xml version="1.0"?>
+<archive xmlns="http://rs.tdwg.org/dwc/text/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://rs.tdwg.org/dwc/terms/xsd/archive/ http://darwincore.googlecode.com/svn/trunk/text/tdwg_dwc_text.xsd">
+  <core encoding="UTF-8" fieldsTerminatedBy="," fieldsEnclosedBy="&quot;" linesTerminatedBy="&#10;" rowType="http://rs.tdwg.org/dwc/terms/Taxon" ignoreHeaderLines="1">
+    <files>
+      <location>core.csv</location>
+    </files>
+    <id index="0"/>
+    <field term="http://rs.tdwg.org/dwc/terms/parentNameUsageID" index="1"/>
+    <field term="http://rs.tdwg.org/dwc/terms/scientificName" index="2"/>
+    <field term="http://rs.tdwg.org/dwc/terms/taxonRank" index="3"/>
+  </core>
+  <extension encoding="UTF-8" fieldsTerminatedBy="," fieldsEnclosedBy="&quot;" linesTerminatedBy="&#10;" rowType="http://rs.gbif.org/terms/1.0/VernacularName" ignoreHeaderLines="1">
+    <files>
+      <location>vern.csv</location>
+    </files>
+    <coreid index="0"/>
+    <field term="http://rs.tdwg.org/dwc/terms/vernacularName" index="1"/>
+  </extension>
+</archive>
diff --git a/spec/lib/generator_spec.rb b/spec/lib/generator_spec.rb
@@ -50,16 +50,25 @@
                         'http://rs.gbif.org/terms/1.0/VernacularName')
 
       gen.add_meta_xml
-      meta = File.read(File.join(gen.path, 'meta.xml'))
-      expect(meta).to match %r|<location>core.csv</location>|
+      meta = File.read(File.join(gen.path, 'meta.xml')).strip
+      meta_from_file= File.read(File.expand_path(
+        '../../files/generator_meta.xml', 
+        __FILE__)).strip
+      expect(meta).to eq meta_from_file
     end
   end
 
   describe '#add_eml_data' do
     it 'adds eml data' do
       gen.add_eml_xml(EML_DATA)
-      eml = File.read(File.join(gen.path, 'eml.xml'))
-      expect(eml).to match /jdoe@example.com/
+      eml = File.read(File.join(gen.path, 'eml.xml')).strip
+      eml.gsub!(%r|(<pubDate>).*?(</pubDate>)|, '\12013-12-30 14:45:33 -0500\2')
+      eml.gsub!(/(packageId=").*?"/, '\11234/2013-12-30::19:45:33"')
+
+      eml_from_file = File.read(File.expand_path(
+        '../../files/generator_eml.xml', 
+        __FILE__)).strip
+      expect(eml.strip).to eq eml_from_file.strip
     end
   end