dwc-archive 0.5.14 → 0.5.15

data/VERSION

@@ -1 +1 @@
-0.5.14
+0.5.15

data/lib/dwc-archive/classification_normalizer.rb

@@ -19,7 +19,7 @@ class DarwinCore
   end
 
   class SynonymNormalized < Struct.new(:name, :canonical_name, :status);end
-  class VernacularNormalized < Struct.new(:name, :language);end
+  class VernacularNormalized < Struct.new(:name, :language, :locality);end
 
   class ClassificationNormalizer
     attr_reader :error_names, :tree, :normalized_data
@@ -210,9 +210,20 @@ class DarwinCore
       ext, fields = *extension
       ext.read do |rows|
         rows[0].each do |r|
+
+          language = nil
+          if fields[:language]
+            language = r[fields[:language]]
+          elsif fields[:languagecode]
+            language = r[fields[:languagecode]] 
+          end
+          
+          locality = fields[:locality] ? r[fields[:locality]] : nil
+          
           vernacular = VernacularNormalized.new(
             r[fields[:vernacularname]],
-            fields[:languagecode] ? r[fields[:languagecode]] : nil)
+            language,
+            locality)
           @normalized_data[r[fields[:id]]].vernacular_names << vernacular
           add_name_string(vernacular.name)
         end

data/spec/files/meta.xml

@@ -1,22 +1,22 @@
 <?xml version="1.0" encoding="UTF-8"?>
 <archive xmlns="http://rs.tdwg.org/dwc/text/">
-  <core encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy="" ignoreHeaderLines="1" rowType="http://rs.tdwg.org/dwc/terms/DarwinCore">
-    <files>
-      <location>DarwinCore.txt</location>
-    </files>
-    <id index="0" term="http://rs.tdwg.org/dwc/terms/TaxonID"/>
-    <field index="1" term="http://purl.org/dc/terms/source"/>
-    <field index="2" term="http://rs.tdwg.org/dwc/terms/ScientificName"/>
-    <field index="3" term="http://rs.tdwg.org/dwc/terms/HigherTaxonID"/>
-    <field index="4" term="http://rs.tdwg.org/dwc/terms/TaxonRank"/>
-    <field index="5" term="http://rs.tdwg.org/dwc/terms/TaxonomicStatus"/>
-  </core>
-  <extension encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy="" ignoreHeaderLines="1" rowType="http://rs.gbif.org/ipt/terms/1.0/VernacularName">
-    <files>
-      <location>VernacularName.txt</location>
-    </files>
-    <coreid index="0"/>
-    <field index="1" term="http://rs.gbif.org/ecat/terms/vernacularName"/>
-    <field index="2" term="http://rs.gbif.org/thesaurus/languageCode"/>
-  </extension>
+<core encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy="" ignoreHeaderLines="1" rowType="http://rs.tdwg.org/dwc/terms/DarwinCore">
+<files>
+<location>DarwinCore.txt</location>
+</files>
+<id index="0" term="http://rs.tdwg.org/dwc/terms/TaxonID"></id>
+<field index="1" term="http://purl.org/dc/terms/source"></field>
+<field index="2" term="http://rs.tdwg.org/dwc/terms/ScientificName"></field>
+<field index="3" term="http://rs.tdwg.org/dwc/terms/HigherTaxonID"></field>
+<field index="4" term="http://rs.tdwg.org/dwc/terms/TaxonRank"></field>
+<field index="5" term="http://rs.tdwg.org/dwc/terms/TaxonomicStatus"></field>
+</core>
+<extension encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy="" ignoreHeaderLines="1" rowType="http://rs.gbif.org/ipt/terms/1.0/VernacularName">
+<files>
+<location>VernacularName.txt</location>
+</files>
+<coreid index="0"></coreid>
+<field index="1" term="http://rs.gbif.org/ecat/terms/vernacularName"></field>
+<field index="2" term="http://rs.gbif.org/thesaurus/languageCode"></field>
+</extension>
 </archive>

data/spec/lib/dwc-archive_spec.rb

@@ -96,7 +96,9 @@ describe DarwinCore do
       file = File.join(@file_dir, 'data.tar.gz')
       dwc = DarwinCore.new(file)
       norm = dwc.normalize_classification
-      norm.select { |k,v| !v.synonyms.empty? }.map { |k,v| v.synonyms }.size.should > 0
+      nodes_with_syn = norm.select { |k,v| !v.synonyms.empty? }
+      nodes_with_syn.map { |k,v| v.synonyms }.size.should > 0
+      nodes_with_syn.first[1].synonyms.first.status.should == 'synonym'
     end
 
     it "should be able work with files which have scientificNameAuthorship" do
@@ -109,6 +111,16 @@ describe DarwinCore do
       syn = norm.select{|k,v| v.synonyms.size > 0}.map {|k,v| v.synonyms}.flatten.select {|s| s.name.split(" ").size  > s.canonical_name.split(" ").size}
       syn.size.should == 50
     end
+
+    it "should be able to get language and locality fields for vernacular names" do
+      file = File.join(@file_dir, 'language_locality.tar.gz')
+      dwc = DarwinCore.new(file)
+      cn = DarwinCore::ClassificationNormalizer.new(dwc)
+      cn.normalize
+      vn = cn.normalized_data['leptogastrinae:tid:42'].vernacular_names.first
+      vn.language.should == 'en'
+      vn.locality.should == 'New England'
+    end
   end
 
 end

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: dwc-archive
 version: !ruby/object:Gem::Version 
-  hash: 23
+  hash: 21
   prerelease: 
   segments: 
   - 0
   - 5
-  - 14
-  version: 0.5.14
+  - 15
+  version: 0.5.15
 platform: ruby
 authors: 
 - Dmitry Mozzherin
@@ -15,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-05-25 00:00:00 -04:00
+date: 2011-05-26 00:00:00 -04:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -107,6 +107,7 @@ files:
 - spec/files/flat_list.tar.gz
 - spec/files/invalid.tar.gz
 - spec/files/junk_dir_inside.zip
+- spec/files/language_locality.tar.gz
 - spec/files/meta.xml
 - spec/files/minimal.tar.gz
 - spec/files/sci_name_authorship.tar.gz