cul_scv_hydra 0.8.0 → 0.9.0

data/config/fedora.yml

@@ -1,10 +1,10 @@
 development:
-  :url: http://127.0.0.1:8983/fedora
-  :user: fedoraAdmin
-  :password: fedoraAdmin
-#  :url: http://sayers.cul.columbia.edu:8080/fedora
+#  :url: http://127.0.0.1:8983/fedora
 #  :user: fedoraAdmin
-#  :password: f+BULUS*^
+#  :password: fedoraAdmin
+  :url: http://sayers.cul.columbia.edu:8080/fedora
+  :user: fedoraAdmin
+  :password: f+BULUS*^
 test:
   :url: http://127.0.0.1:8983/fedora
   :user: fedoraAdmin

data/config/solr_value_maps.yml

@@ -1,4 +1,5 @@
 :marc_to_facet :
+  'NNC' : 'General Collections'
   'NNC-A' : 'Avery'
   'NNC-ART' : 'Office of Art Properties'
   'NNBa' : 'Barnard College Library'
@@ -10,6 +11,7 @@
   'NyNyCBL' : 'Burke Library'
   'NyNyCOH' : 'CCOH'
 :marc_to_display :
+  'NNC' : 'General Collections'
   'NNC-A' : 'Avery Architectural & Fine Arts Library'
   'NNC-ART' : 'Office of Art Properties'
   'NNBa' : 'Barnard College Library'
@@ -21,9 +23,11 @@
   'NyNyCBL' : 'Burke Library at Union Theological Seminary'
   'NyNyCOH' : 'Columbia Center for Oral History Collections'
 :project_to_facet :
-  'Customer Order Collection' : 'PRD Orders'
+  'Customer Order Collection' : 'Pres Orders'
   "Children's Drawings of the Spanish Civil War (online exhibition)" : 'Spanish Civil War'
   "Jewels in her crown: treasures of Columbia University Libraries special collections" : 'Jewels in her Crown'
   "Russian Imperial Corps of Pages" : 'Russian Corps of Pages'
-  "Osama bin Laden posters" : 'Osama bin Laden posters'
-  "Chinese paper gods" : 'Chinese paper gods'
+  "Preserving Historic Audio Content" : 'Audio Preservation'
+  'Papers of John Jay' : 'John Jay Papers'
+  'Project Facet Mapping Test' : 'Successful Project Mapping'
+  'G.E.E. Lindquist Native American Photographs' : 'Lindquist Photographs'

data/lib/cul_scv_hydra/active_fedora/model.rb

@@ -5,3 +5,5 @@ end
 require 'cul_scv_hydra/active_fedora/model/aggregator'
 require 'cul_scv_hydra/active_fedora/model/common'
 require 'cul_scv_hydra/active_fedora/model/resource'
+require 'cul_scv_hydra/active_fedora/model/nokogiri_datastreams'
+require 'cul_scv_hydra/active_fedora/model/struct_metadata'

data/lib/cul_scv_hydra/active_fedora/model/aggregator.rb

@@ -3,6 +3,7 @@ module Aggregator
   extend ActiveSupport::Concern
 
   included do
+    has_metadata :name => "structMetadata", :type=>Cul::Scv::Hydra::ActiveFedora::Model::StructMetadata, :versionable => true, :controlGroup => 'M'
     has_many :parts, :property => :cul_member_of, :class_name=>'ActiveFedora::Base'
     after_create :aggregator!
   end

data/lib/cul_scv_hydra/active_fedora/model/common.rb

@@ -138,14 +138,15 @@ module Common
       solr_doc["descriptor_ssi"] = ["dublin core"]
     end
     # if no mods, pull some values from DC
-    if (solr_doc["title_ssm"].nil? or solr_doc["title_ssm"].length == 0)
-      if self.dc.term_values(:dc_title).first
-        solr_doc["title_ssm"] = [self.dc.term_values(:dc_title).first]
+    if (solr_doc["title_display_ssm"].nil? or solr_doc["title_display_ssm"].length == 0)
+      if self.datastreams["DC"].term_values(:dc_title).first
+        solr_doc["title_display_ssm"] = self.datastreams["DC"].term_values(:dc_title)
       else
-        solr_doc["title_ssm"] = [self.dc.term_values(:dc_identifier).first]
+        solr_doc["title_display_ssm"] = self.datastreams["DC"].term_values(:dc_identifier).reject {|dcid| dcid.eql? self.id}
       end
-      if self.dc.term_values(:dc_relation).first
-        self.dc.term_values(:dc_relation).each {|val|
+      solr_doc["title_si"] = self.datastreams["DC"].term_values(:dc_title).first
+      if self.datastreams["DC"].term_values(:dc_relation).first
+        self.datastreams["DC"].term_values(:dc_relation).each {|val|
           if val =~ /clio:/
             solr_doc["clio_ssim"] ||= []
             solr_doc["clio_ssim"] << val.split(':')[-1]
@@ -153,8 +154,11 @@ module Common
         }
       end
     end
-    if (solr_doc["title_ssm"].length > 1)
-      solr_doc["title_ssm"].uniq!
+    if (solr_doc["identifier_ssim"].nil? or solr_doc["identifier_ssim"].length == 0)
+        solr_doc["identifier_ssim"] = self.datastreams["DC"].term_values(:dc_identifier).reject {|dcid| dcid.eql? self.id}
+    end
+    if (solr_doc["title_display_ssm"].length > 1)
+      solr_doc["title_display_ssm"].uniq!
     end
     solr_doc["format_ssi"] = [self.route_as]
     solr_doc["index_type_label_ssi"] = [self.index_type_label]
@@ -166,6 +170,7 @@ module Common
         value.strip!
       end
     }
+    solr_doc[:structured_bsi] = 'false' unless solr_doc.has_key? :structured_bsi
     solr_doc
   end
   

data/lib/cul_scv_hydra/active_fedora/model/nokogiri_datastreams.rb

@@ -0,0 +1,142 @@
+module Cul::Scv::Hydra::ActiveFedora::Model::NokogiriDatastreams
+  extend ActiveSupport::Concern
+
+  included do
+    self.extend ClassMethods
+  end
+
+  module ClassMethods
+    # Create an instance of this class based on xml content
+    # @param [String, File, Nokogiri::XML::Node] xml the xml content to build from
+    # @param [ActiveFedora::OmDatastream] tmpl the Datastream object that you are building @default a new instance of this class
+    # Careful! If you call this from a constructor, be sure to provide something 'ie. self' as the @tmpl. Otherwise, you will get an infinite loop!
+    def from_xml(xml, tmpl=nil)
+      tmpl = self.new if tmpl.nil?  ## This path is used only for unit testing (e.g. MarpaDCDatastream.from_xml(fixture("data.xml")) )
+
+      if !xml.present?
+        tmpl.ng_xml = self.xml_template
+      elsif xml.kind_of? Nokogiri::XML::Node || xml.kind_of?(Nokogiri::XML::Document)
+        tmpl.ng_xml = xml
+      else
+        tmpl.ng_xml = Nokogiri::XML::Document.parse(xml)
+      end
+
+      tmpl.ng_xml_doesnt_change!
+
+      return tmpl
+    end
+  end
+
+  def local_or_remote_content(ensure_fetch = true)
+    @content = to_xml if ng_xml_changed? || autocreate?
+    super
+  end
+
+  def datastream_content
+    @datastream_content ||= Nokogiri::XML(super).to_xml  {|config| config.no_declaration}.strip
+  end
+
+  def ng_xml 
+    @ng_xml ||= begin
+      if new?
+        ## Load up the template
+        self.class.xml_template
+      else
+        Nokogiri::XML::Document.parse(datastream_content)
+      end
+    end
+  end
+  
+  def ng_xml=(new_xml)
+    # before we set ng_xml, we load the datastream so we know if the new value differs.
+    local_or_remote_content(true)
+
+    case new_xml 
+    when Nokogiri::XML::Document
+      self.content=new_xml.to_xml
+    when  Nokogiri::XML::Node 
+      ## Cast a fragment to a document
+      self.content=new_xml.to_s
+    when String 
+      self.content=new_xml
+    else
+      raise TypeError, "You passed a #{new_xml.class} into the ng_xml of the #{self.dsid} datastream. OmDatastream.ng_xml= only accepts Nokogiri::XML::Document, Nokogiri::XML::Element, Nokogiri::XML::Node, or raw XML (String) as inputs."
+    end
+  end
+  
+  # don't want content eagerly loaded by proxy, so implementing methods that would be implemented by define_attribute_methods 
+  def ng_xml_will_change!
+    changed_attributes['ng_xml'] = nil
+  end
+
+  def ng_xml_doesnt_change!
+    changed_attributes.delete('ng_xml')
+  end
+  
+  # don't want content eagerly loaded by proxy, so implementing methods that would be implemented by define_attribute_methods 
+  def ng_xml_changed?
+    changed_attributes.has_key? 'ng_xml'
+  end
+
+    def datastream_content
+    @datastream_content ||= Nokogiri::XML(super).to_xml  {|config| config.no_declaration}.strip
+  end
+  
+  def content=(new_content)
+    if inline?
+      # inline datastreams may be transformed by fedora 3, so we test for equivalence instead of equality
+      if !EquivalentXml.equivalent?(datastream_content, new_content)
+        ng_xml_will_change! 
+        @ng_xml = Nokogiri::XML::Document.parse(new_content)
+        super(@ng_xml.to_s)
+      end
+    else
+      if datastream_content != new_content
+        ng_xml_will_change! 
+        @ng_xml = Nokogiri::XML::Document.parse(new_content)
+        super(@ng_xml.to_s)
+      end
+    end
+  end
+
+  def content_changed?
+    return false if !xml_loaded
+    super
+  end
+
+  def xml_loaded
+    instance_variable_defined? :@ng_xml
+  end
+
+  def to_xml(xml = nil)
+    xml = self.ng_xml if xml.nil?
+    ng_xml = self.ng_xml
+    if ng_xml.respond_to?(:root) && ng_xml.root.nil? && self.class.respond_to?(:root_property_ref) && !self.class.root_property_ref.nil?
+      ng_xml = self.class.generate(self.class.root_property_ref, "")
+      if xml.root.nil?
+        xml = ng_xml
+      end
+    end
+
+    unless xml == ng_xml || ng_xml.root.nil?
+      if xml.kind_of?(Nokogiri::XML::Document)
+          xml.root.add_child(ng_xml.root)
+      elsif xml.kind_of?(Nokogiri::XML::Node)
+          xml.add_child(ng_xml.root)
+      else
+          raise "You can only pass instances of Nokogiri::XML::Node into this method.  You passed in #{xml}"
+      end
+    end
+    
+    return xml.to_xml {|config| config.no_declaration}.strip
+  end
+
+  def method_missing method, *args, &block
+    if ng_xml.respond_to? method
+      ng_xml.send(method, *args, &block)
+    else
+      super
+    end
+  end
+
+end

data/lib/cul_scv_hydra/active_fedora/model/resource.rb

@@ -51,7 +51,6 @@ module Resource
       image_properties = Cul::Image::Properties.identify(blob)
       if image_properties
         image_prop_nodes = image_properties.nodeset
-        puts image_properties.instance_variable_get(:@ng_xml).to_xml
         relsext = datastreams['RELS-EXT']
         image_prop_nodes.each { |node|
           if node["resource"]
@@ -63,7 +62,6 @@ module Resource
           end
           subject = RDF::URI(internal_uri)
           predicate = RDF::URI("#{node.namespace.href}#{node.name}")
-          puts "Adding a resource predicate : #{predicate}"
           query = RDF::Query.new({ :subject => {predicate => :object}})
           relationships(predicate).dup.each { |stmt|
             relationships.delete(stmt)

data/lib/cul_scv_hydra/active_fedora/model/struct_metadata.rb

@@ -0,0 +1,118 @@
+module Cul::Scv::Hydra::ActiveFedora::Model
+class StructMetadata < ::ActiveFedora::Datastream
+  include NokogiriDatastreams
+
+	def self.default_attributes
+		super.merge(:controlGroup => 'M', :mimeType => 'text/xml')
+  end
+
+  def self.xml_template
+    Nokogiri::XML::Document.parse("<mets:structMap xmlns:mets=\"http://www.loc.gov/METS/\">")
+  end
+
+  def self.div_template(prefix="mets")
+    prefix.nil? ? '<div/>' : "<#{prefix}:div/>"
+  end
+
+  def initialize(digital_object=nil, dsid=nil, options={})
+    super
+  end
+
+  # Indicates that this datastream has metadata content. 
+  # @return true 
+  def metadata?
+    true
+  end
+
+  def autocreate?
+    changed_attributes.has_key? :profile
+  end
+
+  def label=(value)
+    struct_map["LABEL"] = value
+    ng_xml_will_change!
+  end
+
+  def label
+    struct_map["LABEL"]
+  end
+
+  def type=(value)
+    struct_map["TYPE"] = value
+    ng_xml_will_change!
+  end
+
+  def type
+    struct_map["TYPE"]
+  end
+
+  def prefix
+    prefix = nil
+    ng_xml.namespaces.each do |p, href|
+      prefix = p.sub(/xmlns:/,'') if href == "http://www.loc.gov/METS/"
+    end
+    prefix
+  end
+
+  def struct_map
+    prefix = self.prefix
+    path = prefix.nil? ? 'xmlns:structMap' : "#{prefix}:structMap"
+    ng_xml.xpath(path, ng_xml.namespaces).first
+  end
+
+  def create_div_node(parent=nil, atts={})
+    if parent.nil?
+      parent = struct_map
+    end
+    divNode = parent.add_child(StructMetadata.div_template(parent.namespace.prefix)).first
+    [:label, :order, :contentids]. each do |key|
+      divNode[key.to_s.upcase] = atts[key].to_s if atts.has_key? key
+    end
+    ng_xml_will_change! if (divNode.document == ng_xml.document) 
+    divNode
+  end
+
+  def divs_with_attribute(descend=true, name=nil, value=nil)
+    prefix = self.prefix || 'xmlns'
+    xpath = descend ? "//#{prefix}:div" : "/#{prefix}:structMap/#{prefix}:div"
+    if !name.nil?
+      xpath << "[@#{name}"
+      if !value.nil?
+        xpath << "='#{value}'"
+      end
+      xpath << ']'
+    end
+    ng_xml.xpath(xpath, ng_xml.namespaces)
+  end
+
+  # a convenience method for setting attributes and creating divs (if necessary) for R/V structure
+  # returns the mets:structMap node
+  def recto_verso!
+    self.type= 'physical' unless self.type == 'physical'
+    self.label= 'Sides' unless self.label == 'Sides'
+    create_div_node struct_map, {:order=>'1'} unless divs_with_attribute(false,'ORDER','1').first
+    create_div_node struct_map, {:order=>'2'} unless divs_with_attribute(false,'ORDER','2').first
+    if (div = divs_with_attribute(false,'ORDER','1').first)
+      div['LABEL'] = 'Recto' unless div['LABEL'] == 'Recto'
+    end
+    if (div = divs_with_attribute(false,'ORDER','2').first)
+      div['LABEL'] = 'Verso' unless div['LABEL'] == 'Verso'
+    end
+    struct_map
+  end
+
+  def recto
+    divs_with_attribute(false, 'LABEL', 'Recto').first
+  end
+
+  def verso
+    divs_with_attribute(false, 'LABEL', 'Verso').first
+  end
+
+  def to_solr(doc={})
+    doc[:structured_bsi] = (has_content? ? 'true' : 'false')
+    doc
+  end
+
+end
+end

data/lib/cul_scv_hydra/om/dc_metadata.rb

@@ -31,7 +31,8 @@ module Om
         :index_as=>[:displayable, :searchable])
       t.dc_identifier(:path=>"identifier",
         :namespace_prefix=>"dc",
-        :index_as=>[:displayable, :searchable])
+        :type=>:string,
+        :index_as=>[:symbol])
       t.dc_language(:path=>"language",
         :namespace_prefix=>"dc",
         :index_as=>[:displayable, :searchable])

data/lib/cul_scv_hydra/om/scv_mods_document.rb

@@ -8,6 +8,11 @@ module Om
   class ModsDocument < ::ActiveFedora::OmDatastream
     include ::OM::XML::TerminologyBasedSolrizer
     include Cul::Scv::Hydra::Solrizer::TerminologyBasedSolrizer
+    include Cul::Scv::Hydra::Solrizer::ScvModsFieldable
+
+    map_field("lib_repo_sim", :marc_to_facet)
+    map_field("lib_repo_ssm", :marc_to_display)
+    map_field("lib_project_sim", :project_to_facet)
 
     set_terminology do |t|
       t.root(:path=>"mods",
@@ -16,11 +21,14 @@ module Om
       }
 
       t.main_title_info(:path=>'titleInfo', :index_as=>[], :attributes=>{:type=>:none}){
+        t.non_sort(:path=>"nonSort", :index_as=>[])
         t.main_title(:path=>"title", :index_as=>[])
       }
       
-      t.title(:proxy=>[:mods, :main_title_info, :main_title],
-       :index_as=>[:displayable, :searchable, :sortable, :textable])
+      t.title(:proxy=>[:mods, :main_title_info, :main_title], :type=>:string,
+       :index_as=>[:searchable, :sortable, :textable])
+      t.title_display(:proxy=>[:mods, :main_title_info], :type=>:string,
+       :index_as=>[:displayable])
 
       t.search_title_info(:path=>'titleInfo', :index_as=>[]){
         t.search_title(:path=>'title', :index_as=>[:searchable])
@@ -31,16 +39,16 @@ module Om
         }
       }
       t.collection(:path=>"relatedItem", :attributes=>{:type=>"host", :displayLabel=>"Collection"}, :index_as=>[]){
-        t.collection_title_info(:path=>'titleInfo', :index_as=>[]){
-          t.lib_collection(:path=>'title', :index_as=>[:facetable, :displayable])
+        t.collection_title_info(:path=>'titleInfo', :index_as=>[:facetable, :displayable]){
+          t.lib_collection(:path=>'title', :index_as=>[])
         }
       }
-      t.lib_project(:proxy=>[:project,:project_title_info, :lib_project],
+      t.lib_project(:proxy=>[:project,:project_title_info],
         :index_as=>[:displayable, :searchable, :project_facetable, :textable])
-      t.lib_collection(:proxy=>[:collection,:collection_title_info, :lib_collection])
+      t.lib_collection(:proxy=>[:collection,:collection_title_info])
 # pattern matches
-      t.identifier(:path=>"identifier", :attributes=>{:type=>"local"}, :index_as=>[:symbol, :textable])
-      t.clio(:path=>"identifier", :attributes=>{:type=>"CLIO"}, :data_type=>:symbol, :index_as=>[:searchable, :displayable])
+      t.identifier(:path=>"identifier", :attributes=>{:type=>"local"}, :type=>:string, :index_as=>[:symbol, :textable])
+      t.clio(:path=>"identifier", :attributes=>{:type=>"CLIO"}, :data_type=>:symbol, :index_as=>[:symbol, :textable])
       t.abstract
       t.subject {
         t.topic
@@ -64,9 +72,8 @@ module Om
           :index_as=>[])
         t.shelf_locator(:path=>"shelfLocator", :index_as=>[:textable])
       }
-      t.lib_repo(:proxy=>[:location, :lib_repo],
-       :index_as=>[:marc_code_facetable, :marc_code_displayable,
-        :marc_code_textable])
+      t.lib_repo(:proxy=>[:location, :lib_repo], :type=>:text,
+       :index_as=>[:marc_code_facetable, :marc_code_displayable, :marc_code_textable])
       t.lib_name(
         :path=>'name',:attributes=>{:type=>'personal'},
         :index_as=>[:facetable, :displayable, :searchable, :textable]){
@@ -98,7 +105,8 @@ module Om
 
       t.origin_info(:path=>"originInfo", :index_as=>[]){
         t.date(:path=>"dateIssued", :attributes=>{:encoding=>'w3cdtf'}, :index_as=>[])
-        t.key_date(:path=>"dateIssued", :attributes=>{:encoding=>'w3cdtf',:keyDate=>'yes'}, :index_as=>[])
+        t.lib_date(:path=>"dateIssued", :attributes=>{:encoding=>'w3cdtf',:keyDate=>'yes'},
+         :index_as=>[:date_sortable])
         t.start_date(:path=>"dateIssued", :attributes=>{:encoding=>'w3cdtf',:keyDate=>'yes',:point=>'start'}, :index_as=>[])
         t.end_date(:path=>"dateIssued", :attributes=>{:encoding=>'w3cdtf',:point=>'end'}, :index_as=>[])
       }

data/lib/cul_scv_hydra/solrizer.rb

@@ -2,11 +2,12 @@ module Cul
   module Scv
     module Hydra
       module Solrizer
+      	autoload :Extractor, "cul_scv_hydra/solrizer/extractor"
+      	autoload :TerminologyBasedSolrizer, "cul_scv_hydra/solrizer/terminology_based_solrizer"
+      	autoload :ValueMapper, "cul_scv_hydra/solrizer/value_mapper"
+      	autoload :ScvModsFieldable, "cul_scv_hydra/solrizer/scv_mods_fieldable"
       end
     end
   end
 end
-require "cul_scv_hydra/solrizer/extractor"
-require "cul_scv_hydra/solrizer/field_mapper"
-require "cul_scv_hydra/solrizer/terminology_based_solrizer"
-require "cul_scv_hydra/solrizer/value_mapper"
+require "cul_scv_hydra/solrizer/field_mapper"

data/lib/cul_scv_hydra/solrizer/field_mapper.rb

@@ -1,23 +1,32 @@
 require 'solrizer'
 module Solrizer::DefaultDescriptors
+  def self.date_sortable
+    @date_sortable ||= Solrizer::Descriptor.new(:date, :stored, :indexed, converter: date_sortable_converter)
+  end
+
+  # Produces the field name 'all_text_teim'
   def self.textable
-    @textable_type ||= TextableDescriptor.new(:text, :indexed, :multivalued)
+    @textable_type ||= TextableDescriptor.new(:text_en, :indexed, :multivalued)
   end
 
+  # Produces _sim suffix
   def self.project_facetable
     @project_facet_type ||= ProjectFacetDescriptor.new(:string, :indexed, :multivalued)
   end
 
+  # Produces _sim suffix and a value-mapping converter
   def self.marc_code_facetable
     @marc_code_facet_type ||= MarcCodeFacetDescriptor.new(:string, :indexed, :multivalued)
   end
 
+  # Produces _ssm suffix and a value-mapping converter
   def self.marc_code_displayable
     @marc_code_type ||= MarcCodeDisplayDescriptor.new(:string, :stored, :indexed, :multivalued)
   end
 
+  # Produces all_text_timv fieldname and a value-mapping converter
   def self.marc_code_textable
-    @marc_code_map_text_type ||= MarcCodeTextableDescriptor.new(:text, :indexed, :multivalued)
+    @marc_code_map_text_type ||= MarcCodeTextableDescriptor.new(:text_en, :indexed, :multivalued)
   end
 
   def self.load_value_maps(config_path=nil)
@@ -43,52 +52,81 @@ module Solrizer::DefaultDescriptors
     @@value_maps ||= load_value_maps
   end
 
+  def self.date_sortable_converter
+    lambda do |type|
+      lambda do |val| 
+        begin
+          d = val.length < 11 ? Date.new(*(val.split('-').collect {|s| s.to_i})) : Date.parse(val)
+          iso8601_date(d)
+        rescue ArgumentError
+          nil 
+        end
+      end
+    end
+  end
+
+  module Normal
+    def normal!(value)
+      value.gsub!(/\s+/,' ')
+      value.strip!
+      true
+    end
+  end
 
   class TextableDescriptor < Solrizer::Descriptor 
-    def name_and_converter(field_name, field_type)
-      ['all_text_timv']
+    def name_and_converter(field_name, args=nil)
+      super('all_text', args)
     end
   end
 
   class ProjectFacetDescriptor < Solrizer::Descriptor
+    include Normal
     def converter(field_type)
       map = Solrizer::DefaultDescriptors.value_maps[:project_to_facet] || {}
-      lambda {|value| (map.has_key? value) ? map[value] : value}
+      lambda {|value| (normal!(value) and map.has_key? value) ? map[value] : value}
     end 
   end
 
   class MarcCodeFacetDescriptor < Solrizer::Descriptor
+    include Normal
     def converter(field_type)
       map = Solrizer::DefaultDescriptors.value_maps[:marc_to_facet] || {}
-      lambda {|value| (map.has_key? value) ? map[value] : value}
+      lambda {|value| (normal!(value) and map.has_key? value) ? map[value] : value}
     end 
   end
 
   class MarcCodeDisplayDescriptor < Solrizer::Descriptor
+    include Normal
     def converter(field_type)
       map = Solrizer::DefaultDescriptors.value_maps[:marc_to_display] || {}
-      lambda {|value| (map.has_key? value) ? map[value] : value}
-    end 
+      lambda {|value| (normal!(value) and map.has_key? value) ? map[value] : value}
+    end
   end
 
   class MarcCodeTextableDescriptor < Solrizer::Descriptor
-    def name_and_converter(field_name, field_type)
-      ['all_text_timv', converter(field_type)]
+    include Normal
+    def name_and_converter(field_name, args=nil)
+      super('all_text', args)
     end
     def converter(field_type)
       fmap = Solrizer::DefaultDescriptors.value_maps[:marc_to_facet] || {}
       dmap = Solrizer::DefaultDescriptors.value_maps[:marc_to_display] || {}
       lambda do |value|
-        r = (fmap.has_key? value) ? [fmap[value]] : []
-        r << dmap[value] if (dmap.has_key? value)
-        r.join(' ')
+        if value.is_a? String
+          normal!(value)
+          r = (fmap.has_key? value) ? [fmap[value]] : []
+          r << dmap[value] if (dmap.has_key? value)
+          r.uniq!
+          r.join(' ')
+        else
+          value
+        end
       end
     end
   end
   class MarcCodeDisplayTextableDescriptor < MarcCodeDisplayDescriptor
-    def name_and_converter(field_name, field_type)
-      puts "MarcCodeDisplayTextableDescriptor"
-      ['all_text_timv', converter(field_type)]
+    def name_and_converter(field_name, args=nil)
+      super('all_text', args)
     end
   end
 end

data/lib/cul_scv_hydra/solrizer/scv_mods_fieldable.rb

@@ -0,0 +1,162 @@
+module Cul::Scv::Hydra::Solrizer
+	module ScvModsFieldable
+    extend ActiveSupport::Concern
+
+    MODS_NS = {'mods'=>'http://www.loc.gov/mods/v3'}
+
+    module ClassMethods
+      def value_mapper(maps=nil)
+        @value_mapper ||= ValueMapper.new(maps)
+      end
+
+      def map_field(field_key, map_key)
+        value_mapper.map_field(field_key, map_key)
+      end
+
+      def map_value(field_key, value_key)
+        value_mapper.map_value(field_key, value_key)
+      end
+
+      def maps_field?(field_key)
+        value_mapper.maps_field? field_key
+      end
+    end
+
+    def mods
+      ng_xml.xpath('/mods:mods', MODS_NS).first
+    end
+
+    def projects
+      mods.xpath("./mods:relatedItem[@type='host' and @displayLabel='Project']", MODS_NS).collect do |p_node|
+        ScvModsFieldable.normalize(main_title(p_node), true)
+      end
+    end
+
+    def collections
+      mods.xpath("./mods:relatedItem[@type='host' and @displayLabel='Collection']", MODS_NS).collect do |p_node|
+        ScvModsFieldable.normalize(main_title(p_node), true)
+      end
+    end
+
+    def sort_title(node=mods)
+      # include only the untyped [!@type] titleInfo, exclude noSort
+      base_text = ''
+      t = node.xpath('./mods:titleInfo[not(@type)]', MODS_NS).first
+      if t
+        t.children.each do |child|
+          base_text << child.text unless child.name == 'nonSort'
+        end
+      end
+      base_text = ScvModsFieldable.normalize(base_text, true)
+      base_text = nil if base_text.empty?
+      base_text
+    end
+
+    def main_title(node=mods)
+      # include only the untyped [!@type] titleInfo
+      t = node.xpath('./mods:titleInfo[not(@type)]', MODS_NS).first
+      if t
+        ScvModsFieldable.normalize(t.text)
+      else
+        nil
+      end
+    end
+
+    def titles(node=mods)
+      # all titles without descending into relatedItems
+      node.xpath('./mods:titleInfo', MODS_NS).collect do |t|
+        ScvModsFieldable.normalize(t.text)
+      end
+    end
+
+    def names(role_authority=nil, role=nil)
+      # get all the name nodes
+      # reject the ones that aren't type 'personal' or 'corporate'
+      # keep all child text except the role terms
+      xpath = "./mods:name[@type = 'personal' or @type = 'corporate']"
+      unless role_authority.nil?
+        xpath << "/mods:role/mods:roleTerm[@authority='#{role_authority.to_s}'"
+        unless role.nil?
+          xpath << " and normalize-space(text()) = '#{role.to_s.strip}'"
+        end
+        xpath << "]/ancestor::mods:name"
+      end
+      names = mods.xpath(xpath, MODS_NS).collect do |node|
+        base_text = node.xpath('./mods:namePart', MODS_NS).collect { |c| c.text }.join(' ')
+        ScvModsFieldable.normalize(base_text, true)
+      end
+      xpath = "./mods:subject" + xpath[1,xpath.length]
+      mods.xpath(xpath, MODS_NS).each do |node|
+        base_text = node.xpath('./mods:namePart', MODS_NS).collect { |c| c.text }.join(' ')
+        names << ScvModsFieldable.normalize(base_text, true)
+      end
+      names
+    end
+
+    def dates(node=mods)
+      # get all the dateIssued with keyDate = 'yes', but not point = 'end'
+    end
+
+    def formats(node=mods)
+      # get all the form values with authority != 'marcform'
+      node.xpath("./mods:physicalDescription/mods:form[@authority != 'marcform']", MODS_NS).collect do |n|
+        ScvModsFieldable.normalize(n.text)
+      end
+    end
+
+    def repositories(node=mods)
+      # get all the location/physicalLocation[@authority = 'marcorg']
+      node.xpath("./mods:location/mods:physicalLocation[@authority = 'marcorg']", MODS_NS).collect do |n|
+        ScvModsFieldable.normalize(n.text)
+      end
+    end
+
+    def shelf_locators(node=mods)
+      node.xpath("./mods:location/mods:shelfLocator", MODS_NS).collect do |n|
+        ScvModsFieldable.normalize(n.text, true)
+      end
+    end
+
+    def to_solr(solr_doc={})
+      solr_doc = (defined? super) ? super : solr_doc
+      solr_doc["title_si"] = sort_title
+      solr_doc["title_ssm"] = titles
+      solr_doc["lib_collection_sim"] = collections
+      solr_doc["lib_project_sim"] = projects
+      solr_doc["lib_name_sim"] = names
+      solr_doc["lib_name_ssm"] = solr_doc["lib_name_sim"]
+      solr_doc["lib_author_sim"] = names(:marcrelator, 'aut')
+      solr_doc["lib_recipient_sim"] = names(:marcrelator, 'rcp')
+      solr_doc["lib_format_sim"] = formats
+      solr_doc["lib_repo_sim"] = repositories
+      solr_doc["lib_shelf_sim"] = shelf_locators
+      solr_doc.each do |k, v|
+        if self.class.maps_field? k
+          solr_doc[k] = self.class.map_value(k, v)
+        end
+      end
+      solr_doc
+    end
+
+    def self.normalize(t, strip_punctuation=false)
+      # strip whitespace
+      n_t = t.dup.strip
+      # collapse intermediate whitespace
+      n_t.gsub!(/\s+/, ' ')
+      # pull off paired punctuation, and any leading punctuation
+      if strip_punctuation
+        n_t = n_t.sub(/^\((.*)\)$/, "\\1")
+        n_t = n_t.sub(/^\{(.*)\}$/, "\\1")
+        n_t = n_t.sub(/^\[(.*)\]$/, "\\1")
+        n_t = n_t.sub(/^"(.*)"$/, "\\1")
+        n_t = n_t.sub(/^'(.*)'$/, "\\1")
+        n_t = n_t.sub(/^<(.*)>$/, "\\1")
+        #n_t = n_t.sub(/^\p{Ps}(.*)\p{Pe}/u, "\\1")
+        n_t = n_t.sub(/^[[:punct:]]+/, '')
+        # this may have 'created' leading/trailing space, so strip
+        n_t.strip!
+      end
+      n_t
+    end
+  end
+end

data/lib/cul_scv_hydra/solrizer/terminology_based_solrizer.rb

@@ -9,18 +9,6 @@ module Cul::Scv::Hydra::Solrizer::TerminologyBasedSolrizer
   # Module Methods
       
   module ClassMethods
-    # Build a solr document from +doc+ based on its terminology
-    # @param [OM::XML::Document] doc  
-    # @param [Hash] solr_doc (optional) solr_doc (values hash) to populate
-    def solrize(doc, solr_doc=Hash.new, field_mapper = nil)
-      unless doc.class.terminology.nil?
-        doc.class.terminology.terms.each_pair do |term_name,term|
-          doc.solrize_term(term, solr_doc, field_mapper) unless term.is_root_term?
-        end
-      end
-
-      return solr_doc
-    end
   # Populate a solr document with solr fields corresponding to the given xml node
   # Field names are generated using settings from the term in the +doc+'s terminology corresponding to +term_pointer+
   # @param [Nokogiri::XML::Node] node to solrize
@@ -31,7 +19,6 @@ module Cul::Scv::Hydra::Solrizer::TerminologyBasedSolrizer
       return solr_doc unless term.index_as && !term.index_as.empty?
       generic_field_name_base = OM::XML::Terminology.term_generic_name(*term_pointer)
       create_and_insert_terms(generic_field_name_base, node_value, term.index_as, solr_doc)
-      
       if term_pointer.length > 1
         #hierarchical_field_name_base = OM::XML::Terminology.term_hierarchical_name(*term_pointer)
         #create_and_insert_terms(hierarchical_field_name_base, node_value, term.index_as, solr_doc)

data/lib/cul_scv_hydra/solrizer/value_mapper.rb

@@ -21,10 +21,21 @@ class ValueMapper
   # Instance methods
 
   def initialize(value_maps=nil)
-    @value_maps = value_maps || Cul::Scv::Hydra::Solrizer::ValueMapper.default_value_maps
+    @value_maps = (value_maps || Cul::Scv::Hydra::Solrizer::ValueMapper.default_value_maps).with_indifferent_access
+    @mapped_fields = {}.with_indifferent_access
   end
-  def solr_value(map_key, value_key)
-    return value_key unless @value_maps.has_key? map_key
+
+  def maps_field?(field_key)
+    @mapped_fields.has_key? field_key
+  end
+
+  def map_field(field_key, map_key)
+    @mapped_fields[field_key] = map_key
+  end
+
+  def map_value(field_key, value_key)
+    map_key = @mapped_fields[field_key]
+    return value_key unless @mapped_fields.has_key? field_key and @value_maps.has_key? map_key
     if value_key.is_a? Array
       value_key.collect{ |val| @value_maps[map_key].fetch(val, val) }
     else

data/lib/cul_scv_hydra/version.rb

@@ -1,7 +1,7 @@
 module Cul
   module Scv
     module Hydra
-        VERSION = '0.8.0'
+        VERSION = '0.9.0'
         def self.version
           VERSION
         end

data/lib/tasks/cmodel.rake

@@ -73,27 +73,27 @@ end
 
 namespace :cul_scv_hydra do
   namespace :cmodel do
-    task :test => :environment do
+    task :test do #=> :environment do
       pid = ENV["PID"]
       puts content_for(pid)
     end
 
-    task :load => :environment do
+    task :load do #=> :environment do
       pid = ENV["PID"]
       load_content(content_for(pid),pid)
     end
 
-    task :purge => :environment do
+    task :purge do #=> :environment do
       pid = ENV["PID"]
       purge(pid)
     end
 
-    task :reload => :environment do
+    task :reload do #=> :environment do
       pid = ENV["PID"]
       reload(pid)
     end
 
-    task :reload_all => :environment do
+    task :reload_all do #=> :environment do
       pattern = ENV["PATTERN"]
       pattern = Regexp.compile(pattern) if pattern
       reload("ldpd:nullbind")

metadata

@@ -1,18 +1,20 @@
 --- !ruby/object:Gem::Specification
 name: cul_scv_hydra
 version: !ruby/object:Gem::Version
-  version: 0.8.0
+  version: 0.9.0
+  prerelease: 
 platform: ruby
 authors:
 - Benjamin Armintor
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-09-09 00:00:00.000000000 Z
+date: 2014-02-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: blacklight
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -20,6 +22,7 @@ dependencies:
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -27,34 +30,39 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: active-fedora
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 6.1.1
+        version: '6.1'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 6.1.1
+        version: '6.1'
 - !ruby/object:Gem::Dependency
   name: hydra-head
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - '='
       - !ruby/object:Gem::Version
-        version: 6.2.2
+        version: 6.4.1
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - '='
       - !ruby/object:Gem::Version
-        version: 6.2.2
+        version: 6.4.1
 - !ruby/object:Gem::Dependency
   name: active_fedora_finders
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -62,6 +70,7 @@ dependencies:
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -69,6 +78,7 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: active_fedora_relsint
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -76,6 +86,7 @@ dependencies:
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -83,6 +94,7 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: cul_image_props
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -90,6 +102,23 @@ dependencies:
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: mods
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -97,62 +126,71 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: om
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
-    - - ~>
+    - - ! '>='
       - !ruby/object:Gem::Version
-        version: 2.1.0
+        version: '0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
-    - - ~>
+    - - ! '>='
       - !ruby/object:Gem::Version
-        version: 2.1.0
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: solrizer
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 3.0.0
+        version: '3.0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 3.0.0
+        version: '3.0'
 - !ruby/object:Gem::Dependency
   name: rspec
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 2.10.0
+        version: '2.10'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 2.10.0
+        version: '2.10'
 - !ruby/object:Gem::Dependency
   name: rspec-rails
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 2.10.0
+        version: '2.10'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 2.10.0
+        version: '2.10'
 - !ruby/object:Gem::Dependency
   name: mocha
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - '='
       - !ruby/object:Gem::Version
@@ -160,6 +198,7 @@ dependencies:
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - '='
       - !ruby/object:Gem::Version
@@ -167,6 +206,7 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: equivalent-xml
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -174,6 +214,7 @@ dependencies:
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -181,6 +222,7 @@ dependencies:
 - !ruby/object:Gem::Dependency
   name: rbx-require-relative
   requirement: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -188,6 +230,7 @@ dependencies:
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
+    none: false
     requirements:
     - - ! '>='
       - !ruby/object:Gem::Version
@@ -202,7 +245,9 @@ files:
 - lib/cul_scv_hydra/access_controls_enforcement.rb
 - lib/cul_scv_hydra/active_fedora/model/aggregator.rb
 - lib/cul_scv_hydra/active_fedora/model/common.rb
+- lib/cul_scv_hydra/active_fedora/model/nokogiri_datastreams.rb
 - lib/cul_scv_hydra/active_fedora/model/resource.rb
+- lib/cul_scv_hydra/active_fedora/model/struct_metadata.rb
 - lib/cul_scv_hydra/active_fedora/model.rb
 - lib/cul_scv_hydra/active_fedora.rb
 - lib/cul_scv_hydra/controllers/aggregates.rb
@@ -230,6 +275,7 @@ files:
 - lib/cul_scv_hydra/om.rb
 - lib/cul_scv_hydra/solrizer/extractor.rb
 - lib/cul_scv_hydra/solrizer/field_mapper.rb
+- lib/cul_scv_hydra/solrizer/scv_mods_fieldable.rb
 - lib/cul_scv_hydra/solrizer/terminology_based_solrizer.rb
 - lib/cul_scv_hydra/solrizer/value_mapper.rb
 - lib/cul_scv_hydra/solrizer.rb
@@ -268,25 +314,26 @@ files:
 - config/subs.yml
 homepage: https://github.com/cul/cul_scv_hydra
 licenses: []
-metadata: {}
 post_install_message: 
 rdoc_options: []
 require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
   requirements:
   - - ! '>='
     - !ruby/object:Gem::Version
       version: 1.9.3
 required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
   requirements:
   - - ! '>='
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.0.7
+rubygems_version: 1.8.24
 signing_key: 
-specification_version: 4
+specification_version: 3
 summary: ActiveFedora, OM, and Solrizer implementations for CUL Staff Collection Viewer
 test_files: []

checksums.yaml

@@ -1,15 +0,0 @@
----
-!binary "U0hBMQ==":
-  metadata.gz: !binary |-
-    YTI5YWY3ZWFjMWVhYjdjNjI4Mzc4MGRjYWNkNTg5MTk0MmNiZjY3Mg==
-  data.tar.gz: !binary |-
-    YzU5YzY4YjY3OWFjYWM5NzllOWRiNDU2ZThiYTk0ZTgwMjg4ZWE4MQ==
-!binary "U0hBNTEy":
-  metadata.gz: !binary |-
-    YWU1Y2U2ZGQ5Zjg1ZDlhNmUxNzFmNDBjYTdhZjlmNDdjYjQ1MjhjYjc2YzVj
-    ZjBjYTcwMTY5NTViNTZiZTNhY2U3OTk1Yjc2YWFkODYxYTQyMzE4MjkwZDUz
-    YzBkNWM4ZTA4ZWNmNjk5NDA3M2YyY2JlNjNmNzNmZjk5NTYwNTM=
-  data.tar.gz: !binary |-
-    ZTI1ZjY5ZDRmYzM0MWVhZDQ5MmVmNWQwZDVlZWVjODJmNTVjZDBmMzA3ZWQ2
-    YTRhODNhYTI1Yjc5MmJkNGJmN2Y2OTBjOWM2NzRiM2FhODM1NzhjZjdkZTFi
-    ZmNlNzA2ZWM0NTFjODRjYWYwYWViMWRlNTRlZGE3NTM0ZGJjZjA=