starting to clean things up to read real xmp files

bin/avm2avm

@@ -0,0 +1,7 @@
+#!/usr/bin/env ruby
+
+require 'thor'
+require 'avm/cli'
+
+AVM::CLI.start
+

lib/avm/cli.rb

@@ -0,0 +1,17 @@
+require 'thor'
+require 'avm/image'
+require 'pp'
+
+module AVM
+  class CLI < ::Thor
+    default_task :convert
+
+    desc 'convert', "Convert a file from one format to another"
+    def convert
+      data = $stdin.read
+
+      pp AVM::Image.from_xml(data).to_h
+    end
+  end
+end
+

lib/avm/creator.rb

@@ -75,12 +75,12 @@ module AVM
     def from_xml(image, document)
       contacts = []
       document.get_refs do |refs|
-        refs[:dublin_core].search('.//rdf:li').each do |name|
+        refs[:dublin_core].search('.//dc:creator//rdf:li').each do |name|
-          contacts << { :name => name.text }
+          contacts << { :name => name.text.strip }
         end
 
         IPTC_MULTI_FIELD_MAP.each do |key, element_name|
-          if node = refs[:iptc].at_xpath("//Iptc4xmpCore:#{element_name}")
+          if node = refs[:iptc].at_xpath(".//Iptc4xmpCore:#{element_name}")
             node.text.split(',').collect(&:strip).each_with_index do |value, index|
               contacts[index][key] = value
             end

lib/avm/image.rb

@@ -11,12 +11,16 @@ module AVM
   class Image
     DUBLIN_CORE_FIELDS = [ :title, :description ]
 
+    PHOTOSHOP_SINGLES = {
+      'Headline' => :headline,
+      'DateCreated' => :date
+    }
+
     AVM_SINGLE_FIELDS = [ 
       'Distance.Notes',
       'Spectral.Notes',
       'ReferenceURL',
       'Credit',
-      'Date',
       'ID',
       'Type',
       'Image.ProductQuality',
@@ -40,7 +44,6 @@ module AVM
       :spectral_notes,
       :reference_url,
       :credit,
-      :date,
       :id,
       :type,
       :quality,
@@ -64,7 +67,6 @@ module AVM
       :spectral_notes, 
       :reference_url, 
       :credit, 
-      :string_date, 
       :id, 
       :image_type, 
       :image_quality,
@@ -141,6 +143,7 @@ module AVM
         end
 
         refs[:photoshop].add_child(%{<photoshop:Headline>#{headline}</photoshop:Headline>})
+        refs[:photoshop].add_child(%{<photoshop:DateCreated>#{string_date}</photoshop:DateCreated>})
 
         AVM_SINGLES_FOR_MESSAGES.each do |tag, message|
           if value = send(message)
@@ -229,14 +232,18 @@ module AVM
           end
         end
 
-        if node = refs[:photoshop].at_xpath('./photoshop:Headline')
+        PHOTOSHOP_SINGLES.each do |tag, field|
-          options[:headline] = node.text
+          if node = refs[:photoshop].at_xpath("./photoshop:#{tag}")
+            options[field] = node.text
+          end
         end
 
         if node = refs[:avm].at_xpath('./avm:Distance')
           list_values = node.search('.//rdf:li').collect { |li| li.text }
 
           case list_values.length
+          when 0
+            options[:light_years] = node.text
           when 1
             options[:light_years] = list_values.first
           when 2

lib/avm/node.rb

@@ -0,0 +1,27 @@
+require 'delegate'
+
+module AVM
+  class Node < DelegateClass(Nokogiri::XML::Node)
+    def initialize(xmp, node)
+      @xmp, @node = xmp, node
+      super(@node)
+    end
+
+    def at_xpath(path)
+      if node = @node.at_xpath(path, @xmp.namespaces)
+        self.class.new(@xmp, node)
+      else
+        nil
+      end
+    end
+
+    def search(path)
+      self.class.from_nodeset(@xmp, @node.search(path, @xmp.namespaces))
+    end
+
+    def self.from_nodeset(xmp, nodeset)
+      nodeset.collect { |node| new(xmp, node) }
+    end
+  end
+end
+

lib/avm/xmp.rb

@@ -1,4 +1,5 @@
 require 'nokogiri'
+require 'avm/node'
 
 module AVM
   class XMP
@@ -9,6 +10,15 @@ module AVM
       'avm' => 'AVM'
     }
 
+    REQUIRED_NAMESPACES = {
+      :x => "adobe:ns:meta/",
+      :rdf => "http://www.w3.org/1999/02/22-rdf-syntax-ns#",
+      :dc => "http://purl.org/dc/elements/1.1/",
+      :photoshop => "http://ns.adobe.com/photoshop/1.0/",
+      :avm => "http://www.communicatingastronomy.org/avm/1.0/",
+      :Iptc4xmpCore => "http://iptc.org/std/Iptc4xmpCore/1.0/xmlns/"
+    }
+
     attr_reader :doc
 
     def initialize(doc = nil)
@@ -25,37 +35,73 @@ module AVM
       new(Nokogiri::XML(string))
     end
 
+    def ensure_xmlns(string)
+      string.gsub(%r{([</@])(\w+):}) { |all, matches| $1 + (prefix_map[$2] || $2) + ':' }
+    end
+
+    alias :% :ensure_xmlns
+
+    def ensure_xpath(path)
+      [ ensure_xmlns(path), namespaces ]
+    end
+
+    def search(path, node = doc)
+      node.search(*ensure_xpath(path))
+    end
+
+    def at_xpath(path, node = doc)
+      node.at_xpath(*ensure_xpath(path))
+    end
+
+    def namespaces
+      @namespaces ||= doc.document.collect_namespaces
+    end
+
     private
+      def prefix_map
+        @prefix_map ||= Hash[doc.document.collect_namespaces.collect { |prefix, namespace|
+          prefix = prefix.gsub('xmlns:', '')
+          result = nil
+
+          REQUIRED_NAMESPACES.each do |original_prefix, target_namespace|
+            result = [ original_prefix.to_s, prefix ] if namespace == target_namespace
+          end
+          result
+        }.compact]
+      end
+
       def ensure_namespaces!
-        {
+        existing = doc.document.collect_namespaces
-          :x => "adobe:ns:meta/",
+
-          :rdf => "http://www.w3.org/1999/02/22-rdf-syntax-ns#",
+        REQUIRED_NAMESPACES.each do |namespace, url|
-          :dc => "http://purl.org/dc/elements/1.1/",
+          doc.root.add_namespace_definition(namespace.to_s, url) if !existing.values.include?(url)
-          :photoshop => "http://ns.adobe.com/photoshop/1.0/",
-          :avm => "http://www.communicatingastronomy.org/avm/1.0/",
-          :Iptc4xmpCore => "http://iptc.org/std/Iptc4xmpCore/1.0/xmlns/"
-        }.each do |namespace, url|
-          doc.root.add_namespace_definition(namespace.to_s, url)
         end
       end
 
       def ensure_descriptions_findable!
         added = []
 
-        doc.search('//rdf:Description').each do |description|
+        search('//rdf:Description').each do |description|
           if first_child = description.first_element_child
             if first_child.namespace
               prefix = first_child.namespace.prefix
 
-              description['rdf:about'] = PREFIXES[prefix]
+              if prefix_description = PREFIXES[prefix_map.index(prefix)]
-              added << prefix
+                description[self % 'rdf:about'] = prefix_description
+                added << prefix
+              end
             end
           end
         end
 
+        if !at_xpath('//rdf:RDF')
+          doc.first_element_child.add_child(self % '<rdf:RDF />')
+        end
+
+
         PREFIXES.each do |prefix, about|
           if !added.include?(prefix)
-            doc.at_xpath('//rdf:RDF').add_child(%{<rdf:Description rdf:about="#{about}" />})
+            at_xpath('//rdf:RDF').add_child(self % %{<rdf:Description rdf:about="#{about}" />})
           end
         end
       end
@@ -77,7 +123,7 @@ module AVM
       end
 
       def at_rdf_description(about)
-        @doc.at_xpath(%{//rdf:Description[@rdf:about="#{about}"]})
+        AVM::Node.new(self, at_xpath(%{//rdf:Description[@rdf:about="#{about}"]}))
       end
 
       def empty_xml_doc

ruby-avm-library.gemspec

@@ -23,4 +23,5 @@ Gem::Specification.new do |s|
   s.add_development_dependency 'mocha'
 
   s.add_dependency 'nokogiri'
+  s.add_dependency 'thor'
 end

spec/avm/image_spec.rb

@@ -157,6 +157,15 @@ describe AVM::Image do
     end
 
     context "image in it" do
+      context 'distance as a single value, assume light years' do
+        let(:file_path) { 'spec/sample_files/image/single_value_light_years.xmp' }
+
+        has_most_options
+
+        its(:redshift) { should be_nil }
+        its(:light_years) { should == light_years }
+      end
+
       context "distance in light years" do
         let(:file_path) { 'spec/sample_files/image/light_years.xmp' }
 
@@ -197,7 +206,7 @@ describe AVM::Image do
       it "should have basic tags" do
         xml.at_xpath('//rdf:RDF').should_not be_nil 
         xml.search('//rdf:RDF/rdf:Description').should_not be_empty 
-        avm.at_xpath('./avm:Date').should be_nil
+        photoshop.at_xpath('./photoshop:DateCreated').should_not be_nil
       end
     end
 
@@ -221,7 +230,7 @@ describe AVM::Image do
         xpath_text(avm, './avm:Spectral.Notes').should == spectral_notes
         xpath_text(avm, './avm:ReferenceURL').should == reference_url
         xpath_text(avm, './avm:Credit').should == credit
-        xpath_text(avm, './avm:Date').should == date
+        xpath_text(photoshop, './photoshop:DateCreated').should == date
         xpath_text(avm, './avm:ID').should == id
         xpath_text(avm, './avm:Type').should == type
         xpath_text(avm, './avm:Image.ProductQuality').should == image_quality

spec/avm/xmp_spec.rb

@@ -24,11 +24,63 @@ describe AVM::XMP do
     end
   end
 
-  describe '.from_string' do
+  describe 'xml from string' do
     let(:xmp) { self.class.describes.from_string(string) }
-    let(:string) { '<xml xmlns:rdf="cats"><rdf:RDF><node /></rdf:RDF></xml>' }
+    let(:doc) { xmp.doc }
 
-    specify { xmp.doc.at_xpath('//node').should_not be_nil }
+    describe '.from_string' do
+      let(:string) { '<xml xmlns:rdf="cats"><rdf:RDF><node /></rdf:RDF></xml>' }
+
+      specify { xmp.doc.at_xpath('//node').should_not be_nil }
+    end
+
+    describe '#ensure_namespaces! and #ensure_xmlns' do
+      let(:rdf_namespace) { AVM::XMP::REQUIRED_NAMESPACES[:rdf] }
+
+      def self.all_default_namespaces
+        it "should have all the namespaces with the default prefixes" do
+          namespaces = doc.document.collect_namespaces
+
+          namespaces_to_test = AVM::XMP::REQUIRED_NAMESPACES.dup
+          yield namespaces_to_test if block_given?
+
+          namespaces_to_test.each do |prefix, namespace|
+            if namespace
+              namespaces["xmlns:#{prefix}"].should == namespace
+            end
+          end
+        end
+      end
+
+      before { doc }
+
+      context 'none of the namespaces exist' do
+        let(:string) { '<xml><node /></xml>' }
+
+        all_default_namespaces
+
+        specify { xmp.ensure_xmlns('.//rdf:what').should == './/rdf:what' }
+      end
+
+      context 'one namespace exists with the same prefix' do
+        let(:string) { %{<xml xmlns:rdf="#{rdf_namespace}"><node /></xml>} }
+
+        all_default_namespaces
+
+        specify { xmp.ensure_xmlns('.//rdf:what').should == './/rdf:what' }
+      end
+
+      context 'one namespace exists with a different prefix' do
+        let(:string) { %{<xml xmlns:r="#{rdf_namespace}"><node /></xml>} }
+
+        all_default_namespaces { |namespaces|
+          namespaces.delete(:rdf)
+          namespaces[:r] = AVM::XMP::REQUIRED_NAMESPACES[:rdf]
+        }
+
+        specify { xmp.ensure_xmlns('.//rdf:what').should == './/r:what' }
+      end
+    end
   end
 
   describe '#ensure_descriptions_findable!' do
@@ -46,7 +98,7 @@ describe AVM::XMP do
       let(:content) { '' }
 
       [ 'Dublin Core', 'IPTC', 'Photoshop', 'AVM' ].each do |which|
-        specify { xmp.doc.at_xpath(%{//rdf:Description[@rdf:about="#{which}"]}).children.should be_empty }
+        specify { xmp.at_xpath(%{//rdf:Description[@rdf:about="#{which}"]}).children.should be_empty }
       end
     end
 
@@ -67,8 +119,28 @@ describe AVM::XMP do
       XML
 
       [ 'Dublin Core', 'IPTC', 'Photoshop', 'AVM' ].each do |which|
-        specify { xmp.doc.at_xpath(%{//rdf:Description[@rdf:about="#{which}"]}).should_not be_nil }
+        specify { xmp.at_xpath(%{//rdf:Description[@rdf:about="#{which}"]}).should_not be_nil }
       end
     end
+
+    context 'has a namespace it should know about with a different prefix' do
+      let(:content) { <<-XML }
+<rdf:Description rdf:about="" xmlns:whatever="http://purl.org/dc/elements/1.1/">
+  <whatever:creator />
+</rdf:Description>
+      XML
+
+      specify { xmp.at_xpath(%{//rdf:Description[@rdf:about="Dublin Core"]}).should_not be_nil }
+    end
+
+    context 'has a namespace it knows nothing about' do
+      let(:content) { <<-XML }
+<rdf:Description rdf:about="" xmlns:whatever="http://example.com">
+  <whatever:creator />
+</rdf:Description>
+      XML
+
+      it { expect { xmp }.to_not raise_error }
+    end
   end
 end

spec/sample_files/creator/one_creator.xmp

@@ -5,6 +5,11 @@
       <dc:creator>
         <rdf:Seq><rdf:li>John Bintz</rdf:li></rdf:Seq>
       </dc:creator>
+      <dc:title>
+        <rdf:Alt>
+          <rdf:li xml:lang="x-default">This is an image</rdf:li>
+        </rdf:Alt>
+      </dc:title>
     </rdf:Description>
     <rdf:Description rdf:about=""
       xmlns:Iptc4xmpCore="http://iptc.org/std/Iptc4xmpCore/1.0/xmlns/">

spec/sample_files/image/both.xmp

@@ -22,6 +22,7 @@
     <rdf:Description rdf:about=""
       xmlns:photoshop="http://ns.adobe.com/photoshop/1.0/">
       <photoshop:Headline>Headline</photoshop:Headline>
+      <photoshop:DateCreated>2010-01-01</photoshop:DateCreated>
     </rdf:Description>
     <rdf:Description rdf:about=""
       xmlns:avm="http://www.communicatingastronomy.org/avm/1.0/">
@@ -29,7 +30,6 @@
       <avm:Distance.Notes>Distance Notes</avm:Distance.Notes>
       <avm:ReferenceURL>Reference URL</avm:ReferenceURL>
       <avm:Credit>Credit</avm:Credit>
-      <avm:Date>2010-01-01</avm:Date>
       <avm:ID>ID</avm:ID>
       <avm:Type>Observation</avm:Type>
       <avm:Image.ProductQuality>Good</avm:Image.ProductQuality>

spec/sample_files/image/light_years.xmp

@@ -21,6 +21,7 @@
     <rdf:Description rdf:about=""
       xmlns:photoshop="http://ns.adobe.com/photoshop/1.0/">
       <photoshop:Headline>Headline</photoshop:Headline>
+      <photoshop:DateCreated>2010-01-01</photoshop:DateCreated>
     </rdf:Description>
     <rdf:Description rdf:about=""
       xmlns:avm="http://www.communicatingastronomy.org/avm/1.0/">
@@ -28,7 +29,6 @@
       <avm:Distance.Notes>Distance Notes</avm:Distance.Notes>
       <avm:ReferenceURL>Reference URL</avm:ReferenceURL>
       <avm:Credit>Credit</avm:Credit>
-      <avm:Date>2010-01-01</avm:Date>
       <avm:ID>ID</avm:ID>
       <avm:Type>Observation</avm:Type>
       <avm:Image.ProductQuality>Good</avm:Image.ProductQuality>

spec/sample_files/image/redshift.xmp

@@ -22,6 +22,7 @@
     <rdf:Description rdf:about=""
       xmlns:photoshop="http://ns.adobe.com/photoshop/1.0/">
       <photoshop:Headline>Headline</photoshop:Headline>
+      <photoshop:DateCreated>2010-01-01</photoshop:DateCreated>
     </rdf:Description>
     <rdf:Description rdf:about=""
       xmlns:avm="http://www.communicatingastronomy.org/avm/1.0/">
@@ -29,7 +30,6 @@
       <avm:Distance.Notes>Distance Notes</avm:Distance.Notes>
       <avm:ReferenceURL>Reference URL</avm:ReferenceURL>
       <avm:Credit>Credit</avm:Credit>
-      <avm:Date>2010-01-01</avm:Date>
       <avm:ID>ID</avm:ID>
       <avm:Type>Observation</avm:Type>
       <avm:Image.ProductQuality>Good</avm:Image.ProductQuality>

spec/sample_files/image/single_value_light_years.xmp

@@ -0,0 +1,77 @@
+<x:xmpmeta xmlns:x="adobe:ns:meta/" x:xmptk="Adobe XMP Core 4.2.2-c063 53.352624, 2008/07/30-18:05:41        ">
+  <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
+    <rdf:Description rdf:about=""
+      xmlns:dc="http://purl.org/dc/elements/1.1/">
+      <dc:title>:
+        <rdf:Alt>
+          <rdf:li xml:lang="x-default">My title</rdf:li>
+        </rdf:Alt>
+      </dc:title>
+      <dc:description>
+        <rdf:Alt>
+          <rdf:li xml:lang="x-default">Description</rdf:li>
+        </rdf:Alt>
+      </dc:description>
+    </rdf:Description>
+    <rdf:Description rdf:about=""
+      xmlns:Iptc4xmpCore="http://iptc.org/std/Iptc4xmpCore/1.0/xmlns/">
+      <Iptc4xmpCore:CreatorContactInfo rdf:parseType="Resource">
+      </Iptc4xmpCore:CreatorContactInfo>
+    </rdf:Description>
+    <rdf:Description rdf:about=""
+      xmlns:photoshop="http://ns.adobe.com/photoshop/1.0/">
+      <photoshop:Headline>Headline</photoshop:Headline>
+      <photoshop:DateCreated>2010-01-01</photoshop:DateCreated>
+    </rdf:Description>
+    <rdf:Description rdf:about=""
+      xmlns:avm="http://www.communicatingastronomy.org/avm/1.0/">
+      <avm:Spectral.Notes>Spectral Notes</avm:Spectral.Notes>
+      <avm:Distance.Notes>Distance Notes</avm:Distance.Notes>
+      <avm:ReferenceURL>Reference URL</avm:ReferenceURL>
+      <avm:Credit>Credit</avm:Credit>
+      <avm:ID>ID</avm:ID>
+      <avm:Type>Observation</avm:Type>
+      <avm:Image.ProductQuality>Good</avm:Image.ProductQuality>
+      <avm:Distance>Light years</avm:Distance>
+      <avm:Spatial.CoordinateFrame>ICRS</avm:Spatial.CoordinateFrame>
+      <avm:Spatial.Equinox>100</avm:Spatial.Equinox>
+      <avm:Spatial.ReferenceValue>
+        <rdf:Seq>
+          <rdf:li>100</rdf:li>
+          <rdf:li>50</rdf:li>
+        </rdf:Seq>
+      </avm:Spatial.ReferenceValue>
+      <avm:Spatial.ReferenceDimension>
+        <rdf:Seq>
+          <rdf:li>200</rdf:li>
+          <rdf:li>150</rdf:li>
+        </rdf:Seq>
+      </avm:Spatial.ReferenceDimension>
+      <avm:Spatial.ReferencePixel>
+        <rdf:Seq>
+          <rdf:li>25</rdf:li>
+          <rdf:li>15</rdf:li>
+        </rdf:Seq>
+      </avm:Spatial.ReferencePixel>
+      <avm:Spatial.Scale>
+        <rdf:Seq>
+          <rdf:li>40</rdf:li>
+          <rdf:li>35</rdf:li>
+        </rdf:Seq>
+      </avm:Spatial.Scale>
+      <avm:Spatial.Rotation>10</avm:Spatial.Rotation>
+      <avm:Spatial.CoordsystemProjection>TAN</avm:Spatial.CoordsystemProjection>
+      <avm:Spatial.Quality>Full</avm:Spatial.Quality>
+      <avm:Spatial.Notes>Spatial Notes</avm:Spatial.Notes>
+      <avm:Spatial.FITSheader>FITS header</avm:Spatial.FITSheader>
+      <avm:Spatial.CDMatrix>
+        <rdf:Seq>
+          <rdf:li>1</rdf:li>
+          <rdf:li>2</rdf:li>
+          <rdf:li>3</rdf:li>
+          <rdf:li>4</rdf:li>
+        </rdf:Seq>
+      </avm:Spatial.CDMatrix>
+    </rdf:Description>
+  </rdf:RDF>
+</x:xmpmeta>