cul_hydra 1.9.3 → 1.10.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 8f3619958e270bc6bec9a1e7452032b1859264350705dc0b87be69492fb1ce75
-  data.tar.gz: c8a57028e065ad4dcd862714fe5d1885708ea91705e26f0c11f6f2358d1fd634
+  metadata.gz: c9f2d1672e6b0151d647c2982cb663e138e8f9066ca610926a87a0e993f41e8f
+  data.tar.gz: 4f5c9ded2819a0e2a0456c64c88623de5e2c85fc49388ed511ad674800adf15c
 SHA512:
-  metadata.gz: 328b4fc330baf639835cf543bc1994bdd5fff32d5030b657d48a85b14481bbf894fff1ccb121039d0191a22e3c8b5dbaffa8020969067a52dc0c6f00422de2f8
-  data.tar.gz: 4c142006b25ade25def658155a10ad8ea5ae4d896647f140c06ed7b16d99027b6d1deec68f56340e2e73bb38fcac530098604b886ff3c5c1d83a6f820206035a
+  metadata.gz: e28a9ab016bc62b8ebac7f0ded9481c766db74e74f36426d1eeffd46535a7036ab3a8b80b444188c9385132b7afd67a40dfc86ed8d6a24b43de6036629ffc655
+  data.tar.gz: 4eac117be6bf09d0c59e4222a5c5c5422941826f4cd287b5839d09a38db580edb92277d1c838de80d7a687af472e6bac8eda52c1932b1a44aae852405e9e5f95

data/app/controllers/concerns/cul/hydra/controller.rb

@@ -1,22 +1,10 @@
 module Cul::Hydra::Controller
 
-  def asset_path_from_config(asset)
-    Rails.configuration.assets.paths.each do |dir|
-      result = "#{dir}/#{asset}"
-      return result if File.exists?(result)
-    end
-    return nil
-  end
-
   def asset_url(source)
     URI.join(root_url, ActionController::Base.helpers.asset_path(source))
   end
 
   def http_client
-    unless @http_client
-      @http_client ||= HTTPClient.new
-    end
-    @http_client
+    @http_client ||= HTTPClient.new
   end
-
-end
+end

data/app/models/administrative_set.rb

@@ -30,42 +30,4 @@ class AdministrativeSet < ::ActiveFedora::Base
     r.collect {|hit| members << SolrDocument.new(hit) } unless r.blank?
     members
   end
-
-  def thumbnail_info
-    members = solr_members
-    if members.length == 0
-      thumb = {:asset=>"cul_hydra/crystal/file.png",:mime=>'image/png'}
-    else
-      thumb = thumb_from_members(members)
-    end
-    return thumb || {:asset=>"cul_hydra/crystal/file.png",:mime=>'image/png'}
-  end
-
-  private
-  def thumb_from_members(members)
-    sorted = members.sort do |a,b|
-      c = a['title_si'] <=> b['title_si']
-      if c == 0 && a['identifier_ssim']
-        if b['identifier_ssim']
-          a['identifier_ssim'].delete(a.id) unless a['identifier_ssim'].length == 1
-          b['identifier_ssim'].delete(a.id) unless b['identifier_ssim'].length == 1
-          a['identifier_ssim'][0] <=> b['identifier_ssim'][0]
-        else
-          -1
-        end
-      else
-        c
-      end
-    end
-    thumb_from_solr_doc(sorted[0])
-  end
-
-  def thumb_from_solr_doc(solr_doc)
-    if solr_doc and (member =  ActiveFedora::Base.find(solr_doc.id, :cast=>true)).respond_to? :thumbnail_info
-      member.thumbnail_info
-    else
-      return nil
-    end
-  end
-
 end

data/app/models/concerns/cul/hydra/models/aggregator.rb

@@ -41,21 +41,6 @@ module Cul::Hydra::Models::Aggregator
     solr_members(opts).collect {|hit| hit.id}
   end
 
-  def thumbnail_info
-    members = solr_members
-    if members.length == 0
-      thumb = {:asset=>"cul_hydra/crystal/file.png",:mime=>'image/png'}
-    else
-      thumb = nil
-      unless datastreams['structMetadata'].new?
-        thumb = thumb_from_struct(members)
-      else
-        thumb =  thumb_from_members(members)
-      end
-    end
-    return thumb || {:asset=>"cul_hydra/crystal/file.png",:mime=>'image/png'}
-  end
-
   # a representative image URI for this object
   # defined by the predicate https://schema.org/image
   # @return [String, URI]
@@ -68,48 +53,6 @@ module Cul::Hydra::Models::Aggregator
   end
 
   private
-  def thumb_from_struct(members)
-    puts "thumb thumb_from_struct"
-    sm = datastreams['structMetadata']
-    first = sm.divs_with_attribute(false,'ORDER','1').first
-    if first
-      members.each do |member|
-        puts "looking for #{first["CONTENTIDS"]} in #{member["identifier_ssim"].inspect}"
-        if member["identifier_ssim"].include? first["CONTENTIDS"]
-          return thumb_from_solr_doc(member)
-        end
-      end
-      return thumb_from_solr_doc(members.first) if members.first
-    else
-      return nil
-    end
-  end
-
-  def thumb_from_members(members)
-    sorted = members.sort do |a,b|
-      c = a['title_si'] <=> b['title_si']
-      if c == 0 && a['identifier_ssim']
-        if b['identifier_ssim']
-          a['identifier_ssim'].delete(a.id) unless a['identifier_ssim'].length == 1
-          b['identifier_ssim'].delete(a.id) unless b['identifier_ssim'].length == 1
-          a['identifier_ssim'][0] <=> b['identifier_ssim'][0]
-        else
-          -1
-        end
-      else
-        c
-      end
-    end
-    thumb_from_solr_doc(sorted[0])
-  end
-
-  def thumb_from_solr_doc(solr_doc)
-    if solr_doc and (member =  ActiveFedora::Base.find(solr_doc.id, :cast=>true)).respond_to? :thumbnail_info
-      member.thumbnail_info
-    else
-      return nil
-    end
-  end
 
   def to_uri(obj)
     if obj.respond_to? :internal_uri

data/app/models/concerns/cul/hydra/models/common.rb

@@ -284,10 +284,6 @@ module Cul::Hydra::Models::Common
     return result
   end
 
-  def thumbnail_info
-    {:asset=>("cul_hydra/crystal/kmultiple.png"),:mime_type=>"image/png"}
-  end
-
   def get_singular_rel(predicate)
     property = relationships(predicate).first
     return nil unless property

data/app/models/concerns/cul/hydra/models/image_resource.rb

@@ -8,19 +8,6 @@ module Cul::Hydra::Models::ImageResource
   FORMAT_OF_PREDICATE = ActiveFedora::Predicates.short_predicate("http://purl.org/dc/terms/isFormatOf").to_s
   FORMAT_URI = RDF::URI("http://purl.org/dc/terms/format")
 
-  DJATOKA_THUMBNAIL_PARMS = {
-    "url_ver" => "Z39.88-2004",
-    "svc_id" => "info:lanl-repo/svc/getRegion",
-    "svc_val_fmt" => "info:ofi/fmt:kev:mtx:jpeg2000",
-    "svc.format" => "image/jpeg",
-    "svc.level" => "",
-    "svc.rotate" => "0",
-    "svc.scale" => "200",
-    "svc.clayers" => ""
-  }
-
-  DJATOKA_BASE_URL = "http://iris.cul.columbia.edu:8888/resolve"
-
   EXIF_ORIENTATION_TO_DEGREES = {
     'top-left' => 0,
     'left-bottom' => 90,
@@ -57,33 +44,6 @@ module Cul::Hydra::Models::ImageResource
     end
   end
 
-  def zooming?
-    zoom = rels_int.relationships(datastreams['content'],:foaf_zooming).first
-    return zoom.object.to_s if zoom
-    datastreams.each do |k,v|
-      if v.mimeType =~ /image\/jp2$/i
-        zoom = "info:fedora/#{k.dsid}"
-      end
-    end
-    return zoom
-  end
-
-  def thumbnail_info
-    thumb = rels_int.relationships(datastreams['content'],:foaf_thumbnail).first
-    if thumb
-      t_dsid = thumb.object.to_s.split('/')[-1]
-      t_url = "#{ActiveFedora.fedora_config.credentials[:url]}/objects/#{pid}/datastreams/#{t_dsid}/content"
-      return {:url=>t_url,:mime=>datastreams[t_dsid].mimeType}
-    elsif (zoom = self.zooming?)
-      t_dsid = zoom.split('/')[-1]
-      t_parms = DJATOKA_THUMBNAIL_PARMS.merge({"rft_id" => datastreams[t_dsid].dsLocation})
-      url = "#{DJATOKA_BASE_URL}?#{options.map { |key, value|  "#{CGI::escape(key.to_s)}=#{CGI::escape(value.to_s)}"}.join("&")  }"
-      {:url => url, :mime => t_parms["svc.format"]}
-    else
-      return {:asset=>"cul_hydra/crystal/file.png",:mime=>'image/png'}
-    end
-  end
-
   # The number of rotational degrees required to display this image as upright
   def required_rotation_for_upright_display
     required_rotation_orientation_in_degrees = (360 - self.orientation) % 360
@@ -102,5 +62,4 @@ module Cul::Hydra::Models::ImageResource
   def orientation
     self.relationships(:orientation).present? ? EXIF_ORIENTATION_TO_DEGREES[self.relationships(:orientation).first.to_s] : 0
   end
-
 end

data/app/models/concerns/cul/hydra/models/resource.rb

@@ -28,9 +28,6 @@ module Cul::Hydra::Models::Resource
 
       after_create :resource!
     end
-    if self.respond_to? :has_datastream
-      has_datastream :name => "CONTENT", :type=>::ActiveFedora::Datastream, :versionable => true
-    end
   end
 
   def resource!
@@ -39,50 +36,4 @@ module Cul::Hydra::Models::Resource
     @metadata_is_dirty = true
     update
   end
-
-  def add_content_blob(blob, opts)
-    file_name = opts[:file_name]
-    mime = opts[:mime_type].nil? ? mime_type(file_name) : opts[:mime_type]
-    add_file_datastream(blob, :label=>file_name, :mimeType=>mime, :dsid => 'CONTENT')
-    set_title_and_label( file_name, :only_if_blank=>true )
-    if IMAGE_MIME_TYPES.include? mime
-      blob.rewind
-      # retrieve Nokogiri of image property RDF
-      image_properties = Cul::Image::Properties.identify(blob)
-      if image_properties
-        image_prop_nodes = image_properties.nodeset
-        relsext = datastreams['RELS-EXT']
-        image_prop_nodes.each { |node|
-          if node["resource"]
-            is_literal = false
-            object = RDF::URI.new(node["resource"])
-          else
-            is_literal = true
-            object = RDF::Literal(node.text)
-          end
-          subject = RDF::URI(internal_uri)
-          predicate = RDF::URI("#{node.namespace.href}#{node.name}")
-          query = RDF::Query.new({ :subject => {predicate => :object}})
-          relationships(predicate).dup.each { |stmt|
-            relationships.delete(stmt)
-          }
-          add_relationship(predicate,object, is_literal)
-          relationships_are_dirty=true
-        }
-        # add mimetype to DC:format values
-        self.datastreams['DC'].update_values({[:format] => mime})
-      end
-    end
-    blob.rewind
-  end
-
- private
-
-  # Return the mimeType for a given file name
-  # @param [String] file_name The filename to use to get the mimeType
-  # @return [String] mimeType for filename passed in. Default: application/octet-stream if mimeType cannot be determined
-  def mime_type file_name
-    mime_types = MIME::Types.of(file_name)
-    mime_type = mime_types.empty? ? "application/octet-stream" : mime_types.first.content_type
-  end
 end

data/app/models/concerns/cul/hydra/models.rb

@@ -17,7 +17,6 @@ module Hydra
     autoload :Aggregator, 'cul/hydra/models/aggregator'
     autoload :Common, 'cul/hydra/models/common'
     autoload :ImageResource, 'cul/hydra/models/image_resource'
-    autoload :LinkableResources, 'cul/hydra/models/linkable_resources'
     autoload :RelsInt, 'cul/hydra/models/rels_int'
     autoload :Resource, 'cul/hydra/models/resource'
   end

data/app/models/cul/hydra/datastreams/rels_int.rb

@@ -13,11 +13,11 @@ module Hydra
 
       def serialize!
         self.content = to_rels_int() if changed_attributes.include? 'relationships'
-        clear_attribute_changes 'relationships'
+        clear_attribute_changes ['relationships']
       end
 
       def relationships_will_change!
-        changed_attributes['relationships'] = nil
+        attribute_will_change!('relationships')
       end
 
       def content
@@ -114,7 +114,7 @@ module Hydra
           g << stmt
         end
         self.relationships_loaded = true
-        clear_attribute_changes 'relationships'
+        clear_attribute_changes ['relationships']
         @graph = g
       end
 

data/app/models/generic_resource.rb

@@ -1,4 +1,3 @@
-require "cul_image_props"
 require "mime/types"
 require "uri"
 require "open-uri"
@@ -53,15 +52,6 @@ class GenericResource < ::ActiveFedora::Base
       end
     end
 
-    if self.zooming?
-      fz = rels_int.relationships(datastreams['content'], :foaf_zooming).first.object.to_s.split('/')[-1]
-      ds = datastreams[fz]
-      unless ds.nil?
-        rft_id = ds.controlGroup == 'E' ? datastreams[fz].dsLocation : legacy_content_path(ds,'info:fedora/datastreams/')
-        solr_doc['rft_id_ss'] = rft_id
-      end
-    end
-
     if (service_ds = self.service_datastream)
       solr_doc['service_dslocation_ss'] = service_ds.dsLocation
     end
@@ -82,22 +72,6 @@ class GenericResource < ::ActiveFedora::Base
     solr_doc
   end
 
-  def thumbnail_info
-    thumb = rels_int.relationships(datastreams['content'],:foaf_thumbnail).first
-    if thumb
-      t_dsid = thumb.object.to_s.split('/')[-1]
-      t_url = "#{ActiveFedora.fedora_config.credentials[:url]}/objects/#{pid}/datastreams/#{t_dsid}/content"
-      return {:url=>t_url,:mime=>datastreams[t_dsid].mimeType}
-    elsif self.zooming?
-      t_dsid = rels_int.relationships(dsuri, :foaf_zooming).first.object.to_s.split('/')[-1]
-      t_parms = DJATOKA_THUMBNAIL_PARMS.merge({"rft_id" => datastreams[t_dsid].dsLocation})
-      url = "#{DJATOKA_BASE_URL}?#{options.map { |key, value|  "#{CGI::escape(key.to_s)}=#{CGI::escape(value.to_s)}"}.join("&")  }"
-      {:url => url, :mime => t_parms["svc.format"]}
-    else
-      return {:asset => "cul_hydra/crystal/file.png",:mime=>'image/png'}
-    end
-  end
-
   def linkable_resources
     # let's start with the known DSIDs from lindquist, then work our way back to parsing the solrized relsint
     results = []
@@ -135,17 +109,6 @@ class GenericResource < ::ActiveFedora::Base
     results
   end
 
-  def zooming?
-    content = datastreams['content']
-    return false unless content
-    zr = rels_int.relationships(content, :foaf_zooming)
-    if (zr && zr.first)
-      return !zr.first.blank?
-    else
-      false
-    end
-  end
-
   def closed?
     return access_levels.map(&:downcase).include?('closed')
   end

data/fixtures/spec/CUL_MODS/mods-titles-extended.xml

@@ -0,0 +1,2 @@
+<?xml version='1.0' encoding='UTF-8'?>
+<mods xmlns='http://www.loc.gov/mods/v3' xmlns:xsi='http://www.w3.org/2001/XMLSchema-instance' xsi:schemaLocation='http://www.loc.gov/mods/v3 http://www.loc.gov/standards/mods/v3/mods-3-4.xsd'><titleInfo><nonSort>The </nonSort><title xml:lang="eng">(ġhotogråphs)	ゐ </title></titleInfo></mods>

data/fixtures/spec/CUL_SOLR/mods-001.xml

@@ -1 +1 @@
-<doc xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:mods="http://www.loc.gov/mods/v3" xmlns="http://solr.apache.org" xmlns:index="http://repository.cul.columbia.edu/namespace/index/1_0/" xmlns:local="http://local.org/custom"><field name="format">multipartitem</field><field name="collection_h">Columbia University Libraries Content/Online Projects/PRD Customer Orders/</field><field name="internal_h">ldpd:91655/ldpd:91656/ldpd:101479/ldpd:102275/</field><field name="format_h">multipartitem/</field><field name="id">ldpd:102275@ldpd:103434</field><field name="pid_ssi">ldpd:102275</field><field name="descriptor">mods</field><field name="object_ssm">https://repository2.cul.columbia.edu:8443/fedora/get/ldpd:102275</field><field name="identifier_ssi">prd.custord.070017</field><field name="all_text_teim">AN070017</field><field name="all_text_teim">prd.custord.070017</field><field name="lib_collection_sim">Pulitzer Prizes</field><field name="lib_collection_ssm">Pulitzer Prizes</field><field name="lib_project_sim">Pres Orders</field><field name="lib_project_ssm">Customer Order Collection</field><field name="lib_project_teim">Customer Order Collection</field><field name="title_sort">Photographs</field><field name="title_ssm">Photographs</field><field name="title_teim">Photographs</field><field name="lib_name_sim">Kennerly, David</field><field name="lib_name_teim">Kennerly, David</field><field name="lib_name_ssm">Kennerly, David</field><field name="lib_repo_long_ssim">Rare Book and Manuscript Library</field><field name="lib_repo_sim">RBML</field><field name="lib_repo_teim">RBML</field><field name="lib_format_ssm">photographs</field><field name="lib_format_sim">photographs</field><field name="all_text_teim">Original PRD customer order number: 070017</field><field name="extent_teim">2 items</field><field name="access_condition_ssi">Columbia Libraries Staff Use Only</field><field name="all_text_teim">Box 8, JP2 Feature Photography</field></doc>
+<doc xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:mods="http://www.loc.gov/mods/v3" xmlns="http://solr.apache.org" xmlns:index="http://repository.cul.columbia.edu/namespace/index/1_0/" xmlns:local="http://local.org/custom"><field name="format">multipartitem</field><field name="collection_h">Columbia University Libraries Content/Online Projects/PRD Customer Orders/</field><field name="internal_h">ldpd:91655/ldpd:91656/ldpd:101479/ldpd:102275/</field><field name="format_h">multipartitem/</field><field name="id">ldpd:102275@ldpd:103434</field><field name="pid_ssi">ldpd:102275</field><field name="descriptor">mods</field><field name="object_ssm">https://repository2.cul.columbia.edu:8443/fedora/get/ldpd:102275</field><field name="identifier_ssi">prd.custord.070017</field><field name="all_text_teim">AN070017</field><field name="all_text_teim">prd.custord.070017</field><field name="lib_collection_sim">Pulitzer Prizes</field><field name="lib_collection_ssm">Pulitzer Prizes</field><field name="lib_project_sim">Pres Orders</field><field name="lib_project_ssm">Customer Order Collection</field><field name="lib_project_teim">Customer Order Collection</field><field name="title_sort">PHOTOGRAPHS</field><field name="title_ssm">Photographs</field><field name="title_teim">Photographs</field><field name="lib_name_sim">Kennerly, David</field><field name="lib_name_teim">Kennerly, David</field><field name="lib_name_ssm">Kennerly, David</field><field name="lib_repo_long_ssim">Rare Book and Manuscript Library</field><field name="lib_repo_sim">RBML</field><field name="lib_repo_teim">RBML</field><field name="lib_format_ssm">photographs</field><field name="lib_format_sim">photographs</field><field name="all_text_teim">Original PRD customer order number: 070017</field><field name="extent_teim">2 items</field><field name="access_condition_ssi">Columbia Libraries Staff Use Only</field><field name="all_text_teim">Box 8, JP2 Feature Photography</field></doc>

data/fixtures/spec/CUL_SOLR/mods-001.yml

@@ -19,7 +19,7 @@ lib_collection_sim: ['Pulitzer Prizes']
 lib_collection_ssm: ['Pulitzer Prizes']
 lib_project_short_ssim: ['Pres Orders']
 lib_project_full_ssim: ['Customer Order Collection']
-title_si: 'Photographs'
+title_si: 'PHOTOGRAPHS'
 title_display_ssm: ['The Photographs']
 lib_format_sim: ['photographs']
 lib_format_ssm: ['photographs']

data/lib/cul_hydra/indexer.rb

@@ -1,9 +1,11 @@
+require "active-fedora"
+
 module Cul::Hydra::Indexer
 
   NUM_FEDORA_RETRY_ATTEMPTS = 3
   DELAY_BETWEEN_FEDORA_RETRY_ATTEMPTS = 5.seconds
   DEFAULT_INDEX_OPTS = {
-    skip_resources: false, verbose_output: false, softcommit: true, reraise: false
+    skip_generic_resources: false, verbose_output: false, softcommit: true, reraise: false
   }.freeze
   def self.descend_from(pid, pids_to_omit=nil, verbose_output=false)
     if pid.blank?
@@ -83,7 +85,7 @@ module Cul::Hydra::Indexer
     index_opts = DEFAULT_INDEX_OPTS.merge(index_opts)
     # assign any legacy positional arguments, permitting explicit nils
     unless args.empty?
-      index_opts[:skip_resources] = args[0] if args.length > 0
+      index_opts[:skip_generic_resources] = args[0] if args.length > 0
       index_opts[:verbose_output] = args[1] if args.length > 1
       index_opts[:softcommit] = args[2] if args.length > 2
     end
@@ -99,7 +101,7 @@ module Cul::Hydra::Indexer
       NUM_FEDORA_RETRY_ATTEMPTS.times do |i|
         begin
           active_fedora_object = ActiveFedora::Base.find(pid, :cast => true)
-          if skip_generic_resources && active_fedora_object.is_a?(GenericResource)
+          if index_opts[:skip_generic_resources] && active_fedora_object.is_a?(GenericResource)
             puts 'Object was skipped because GenericResources are being skipped and it is a GenericResource.'
           else
             if index_opts[:softcommit]

data/lib/cul_hydra/risearch_members.rb

@@ -13,8 +13,7 @@ module ClassMethods
       recursive_member_query += ' and $cmodel <mulgara:is> <info:fedora/ldpd:' + cmodel_type + '>'
     end
 
-    puts 'Performing query:' if verbose_output
-    puts recursive_member_query if verbose_output
+    log_risearch_query(recursive_member_query, verbose_output)
 
     search_response = JSON(Cul::Hydra::Fedora.repository.find_by_itql(recursive_member_query, {
       :type => 'tuples',
@@ -35,8 +34,7 @@ module ClassMethods
       'select $pid from <#ri>
       where $pid <http://purl.oclc.org/NET/CUL/memberOf> <fedora:' + pid + '>'
 
-    puts 'Performing query:' if verbose_output
-    puts direct_member_query if verbose_output
+    log_risearch_query(direct_member_query, verbose_output)
 
     search_response = JSON(Cul::Hydra::Fedora.repository.find_by_itql(direct_member_query, {
       :type => 'tuples',
@@ -72,8 +70,7 @@ module ClassMethods
       where $pid <http://purl.oclc.org/NET/CUL/memberOf> <info:fedora/#{pid}>
       having $k0 <http://mulgara.org/mulgara#occursMoreThan> '0.0'^^<http://www.w3.org/2001/XMLSchema#double>;"
 
-      puts 'Performing query:' if verbose_output
-      puts direct_members_with_ds_query if verbose_output
+      log_risearch_query(direct_members_with_ds_query, verbose_output)
 
       search_response = JSON.parse(Cul::Hydra::Fedora.repository.find_by_itql(direct_members_with_ds_query, {
         :type => 'tuples',
@@ -99,8 +96,7 @@ module ClassMethods
       'select $pid from <#ri>
       where $pid <info:fedora/fedora-system:def/relations-external#isConstituentOf> <fedora:' + pid + '>'
 
-    puts 'Performing query:' if verbose_output
-    puts project_constituent_query if verbose_output
+    log_risearch_query(project_constituent_query, verbose_output)
 
     search_response = JSON(Cul::Hydra::Fedora.repository.find_by_itql(project_constituent_query, {
       :type => 'tuples',
@@ -130,8 +126,7 @@ module ClassMethods
       'select $pid from <#ri>
       where $pid <http://purl.org/dc/terms/publisher> <fedora:' + pid + '>'
 
-    puts 'Performing query:' if verbose_output
-    puts project_constituent_query if verbose_output
+    log_risearch_query(project_constituent_query, verbose_output)
 
     search_response = JSON(Cul::Hydra::Fedora.repository.find_by_itql(project_constituent_query, {
       :type => 'tuples',
@@ -199,6 +194,9 @@ module ClassMethods
     return pids_to_return
   end
 
+  def log_risearch_query(query, verbose_output = false)
+    puts "Performing query:\n#{query}" if verbose_output
+  end
 end
 extend ClassMethods
 end

data/lib/cul_hydra/solrizer/mods_fieldable.rb

@@ -39,7 +39,11 @@ module Cul::Hydra::Solrizer
           n_t = n_t.sub(/^"(.*)"$/, "\\1")
           n_t = n_t.sub(/^'(.*)'$/, "\\1")
           is_negative_number = n_t =~ /^-\d+$/
-          n_t = n_t.sub(/^[[:punct:]]+/, '')
+          if strip_punctuation == :all
+            n_t = n_t.gsub(/[[:punct:]]/, '')
+          else
+            n_t = n_t.sub(/^[[:punct:]]+/, '')
+          end
           # this may have 'created' leading/trailing space, so strip
           n_t.strip!
           n_t = '-' + n_t if is_negative_number
@@ -92,7 +96,13 @@ module Cul::Hydra::Solrizer
           base_text << child.text unless child.name == 'nonSort'
         end
       end
-      base_text = ModsFieldable.normalize(base_text, true)
+      base_text = ModsFieldable.normalize(base_text, :all)
+      # decompose and strip unicode combining characters
+      base_text = base_text.unicode_normalize(:nfd)
+      base_text.gsub!(/[\u0300-\u036F]/,'')
+      # uppercase per Unicode, for ASCII/Latin
+      # TODO: decide whether to use full Unicode case, other language options (Turkish, Lithuanian, etc.)
+      base_text = base_text.upcase(:ascii)
       base_text = nil if base_text.empty?
       base_text
     end

data/lib/cul_hydra/version.rb

@@ -1,6 +1,6 @@
 module Cul
   module Hydra
-    VERSION = '1.9.3'
+    VERSION = '1.10.0'
     def self.version
       VERSION
     end

metadata

@@ -1,44 +1,50 @@
 --- !ruby/object:Gem::Specification
 name: cul_hydra
 version: !ruby/object:Gem::Version
-  version: 1.9.3
+  version: 1.10.0
 platform: ruby
 authors:
 - Benjamin Armintor
 - Eric O'Hanlon
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2020-04-15 00:00:00.000000000 Z
+date: 2021-10-27 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rails
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '5.2'
+    - - "<"
       - !ruby/object:Gem::Version
-        version: 4.2.10
+        version: '6.0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '5.2'
+    - - "<"
       - !ruby/object:Gem::Version
-        version: 4.2.10
+        version: '6.0'
 - !ruby/object:Gem::Dependency
   name: bootstrap-sass
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 3.3.7
+        version: 3.4.1
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 3.3.7
+        version: 3.4.1
 - !ruby/object:Gem::Dependency
   name: blacklight
   requirement: !ruby/object:Gem::Requirement
@@ -115,20 +121,6 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: 0.5.0
-- !ruby/object:Gem::Dependency
-  name: cul_image_props
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
 - !ruby/object:Gem::Dependency
   name: mods
   requirement: !ruby/object:Gem::Requirement
@@ -219,14 +211,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.0.0
+        version: 2.1.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.0.0
+        version: 2.1.0
 - !ruby/object:Gem::Dependency
   name: sprockets
   requirement: !ruby/object:Gem::Requirement
@@ -359,7 +351,6 @@ files:
 - app/controllers/concerns/cul/hydra/application_id_behavior.rb
 - app/controllers/concerns/cul/hydra/controller.rb
 - app/controllers/concerns/cul/hydra/resolver.rb
-- app/controllers/concerns/cul/hydra/thumbnails.rb
 - app/helpers/cul/hydra/ore_proxies_helper_behavior.rb
 - app/helpers/cul/hydra/struct_metadata_helper_behavior.rb
 - app/models/administrative_set.rb
@@ -370,7 +361,6 @@ files:
 - app/models/concerns/cul/hydra/models/aggregator.rb
 - app/models/concerns/cul/hydra/models/common.rb
 - app/models/concerns/cul/hydra/models/image_resource.rb
-- app/models/concerns/cul/hydra/models/linkable_resources.rb
 - app/models/concerns/cul/hydra/models/rels_int.rb
 - app/models/concerns/cul/hydra/models/resource.rb
 - app/models/concerns/nfo/common.rb
@@ -492,6 +482,7 @@ files:
 - fixtures/spec/CUL_MODS/mods-subjects.xml
 - fixtures/spec/CUL_MODS/mods-textual-date.xml
 - fixtures/spec/CUL_MODS/mods-textual-dates-with-unusual-chars.xml
+- fixtures/spec/CUL_MODS/mods-titles-extended.xml
 - fixtures/spec/CUL_MODS/mods-titles.xml
 - fixtures/spec/CUL_MODS/mods-top-level-location-vs-relateditem-location.xml
 - fixtures/spec/CUL_MODS/mods-unmapped-project.xml
@@ -547,7 +538,7 @@ files:
 homepage: https://github.com/cul/cul_hydra
 licenses: []
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - app
@@ -565,8 +556,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.1.2
-signing_key: 
+rubygems_version: 3.0.9
+signing_key:
 specification_version: 4
 summary: ActiveFedora, OM, and Solrizer implementations for CUL repository apps
 test_files: []

data/app/controllers/concerns/cul/hydra/thumbnails.rb

@@ -1,62 +0,0 @@
-require 'httpclient'
-module Cul::Hydra::Thumbnails
-  # some thumbnail urls
-  NO_THUMB = "cul_hydra/crystal/file.png"
-  BROKEN_THUMB = "cul_hydra/crystal/file_broken.png"
-  COLLECTION_THUMB = "cul_hydra/crystal/kmultiple.png"
-  AUDIO_THUMB = "cul_hydra/crystal/sound.png"
-  # some rel predicates
-  FORMAT = "http://purl.org/dc/elements/1.1/format"
-  MEMBER_OF = "http://purl.oclc.org/NET/CUL/memberOf"
-  HAS_MODEL = "info:fedora/fedora-system:def/model#hasModel"
-  IMAGE_WIDTH = "http://purl.oclc.org/NET/CUL/RESOURCE/STILLIMAGE/BASIC/imageWidth"
-  IMAGE_LENGTH = "http://purl.oclc.org/NET/CUL/RESOURCE/STILLIMAGE/BASIC/imageLength"
-
-  def show
-    pid = params[:id].split(/@/)[0]
-    get_by_pid(pid)
-  end
-
-  def get_by_pid(pid)
-    r_obj = ActiveFedora::Base.find(pid, :cast=>true)
-    
-    if r_obj.respond_to? :thumbnail_info
-      url = r_obj.thumbnail_info
-    else
-      url = {:asset=>(COLLECTION_THUMB),:mime=>'image/png'}
-    end
-    if url[:asset]
-      #url = {:url=>asset_url(COLLECTION_THUMB),:mime=>'image/png'}
-      #redirect_to asset_url(url[:asset]).to_s, status: 302
-      #return
-    end
-    Rails.logger.debug "thumbnail #{url[:url] || url[:asset]} #{url[:mime]}"
-    filename = pid + '.' + url[:mime].split('/')[1].downcase
-    h_cd = "filename=""#{CGI.escapeHTML(filename)}"""
-    headers.delete "Cache-Control"
-    headers["Content-Disposition"] = h_cd
-    headers["Content-Type"] = url[:mime]
-   
-    if url[:asset]
-      full_path = Rails.application.assets.resolve(url[:asset]).to_path
-      render :status => 200, :text => File.read(full_path)
-      return
-    elsif url[:url].to_s =~ /^https?:/
-      cl = http_client
-      render :status => 200, :text => cl.get_content(url[:url])
-      return
-    else
-      render :status => 200, :text => File.read(url[:url])
-      return
-    end
-  end
-
-  def jp2_thumbnail(pid)
-    {:url => ActiveFedora.fedora_config.credentials[:url] + '/objects/' + pid + '/methods/ldpd:sdef.Image/getView?max=250', :mime => 'image/jpeg'}
-  end
-
-  def pid_from_uri(uri)
-    return uri.sub(/info:fedora\//,'')
-  end
-  
-end

data/app/models/concerns/cul/hydra/models/linkable_resources.rb

@@ -1,108 +0,0 @@
-module Cul::Hydra::Models::LinkableResources
-  include Cul::Hydra::Fedora::UrlHelperBehavior
-  # so as to avoid including all the url hepers via:
-  ## include Rails.application.routes.url_helpers
-  # we are just going to delegate
-  delegate :fedora_content_path,  :to => 'Rails.application.routes.url_helpers'
-  delegate :cache_path,  :to => 'Rails.application.routes.url_helpers'
-
-  def http_client
-    unless @http_client
-      @http_client ||= HTTPClient.new
-      @http_client.ssl_config.verify_mode = OpenSSL::SSL::VERIFY_NONE
-      uname = Cul::Hydra::Fedora.repository.config[:user]
-      pwd = Cul::Hydra::Fedora.repository.config[:password]
-      @http_client.set_auth(nil, uname, pwd)
-    end
-    @http_client
-  end
-
-  def linkable_resources
-    r = self.parts(:response_format => :solr)
-    return [] if r.blank?
-    members = r.collect {|hit|
-      SolrDocument.new(hit)
-    }
-    members.delete_if { |sd| (sd[:has_model_ssim] & ["info:fedora/ldpd:Resource"]).blank? }
-    case self.route_as
-    when "zoomingimage"
-      results = members.collect {|doc| image_resource(doc)}
-      base_id = self.pid
-      url = fedora_ds_url(base_id, 'SOURCE') + '/content'
-      head_req = http_client().head(url)
-      file_size = head_req.header["Content-Length"].first.to_i
-      results << {
-        :dimensions => "Original",
-        :mime_type => "image/jp2",
-        :uri=>base_id, :block=>"SOURCE", :filename=>base_id + "_source.jp2",
-        :content_models=>[]}
-    when "audio"
-      results = members.collect {|doc| audio_resource(doc)}
-    when "image"
-      results = members.collect {|doc| image_resource(doc)}
-    else
-      raise "Unknown format #{self.route_as}"
-    end
-    return results
-  end
-
-  def basic_resource(document)
-    res = {}
-    res[:pid] = document["id"]
-    res[:dsid] = "CONTENT"
-    res[:mime_type] = document["dc_format_ssm"] ? document["dc_format_ssm"].first : "application/octect-stream"
-    res[:content_models] = document["has_model_ssim"]
-    res[:file_size] = document["extent_ssim"].first.to_i
-    res[:size] = (document["extent_ssim"].first.to_i / 1024).to_s + " Kb"
-    res
-  end
-
-  def image_resource(document)
-    res = basic_resource(document)
-    if document["image_width_ssim"]
-      res[:dimensions] = document["image_width_ssim"].first + " x " + document["image_length_ssim"].first
-      res[:width] = document["image_width_ssim"].first
-      res[:height] = document["image_length_ssim"].first
-    else
-      res[:dimensions] = "? x ?"
-      res[:width] = "0"
-      res[:height] = "0"
-    end
-    base_id = document["id"]
-    base_filename = base_id.gsub(/\:/,"")
-    img_filename = base_filename + "." + document["dc_format_ssm"].first.gsub(/^[^\/]+\//,"")
-    res[:filename] = img_filename
-    res[:block] = "CONTENT"
-    res[:mime_type] = document["dc_format_ssm"] ? document["dc_format_ssm"].first : "application/octect-stream"
-    res[:content_models] = document["has_model_ssim"]
-    res[:file_size] = document["extent_ssim"].first.to_i
-    res[:size] = (document["extent_ssim"].first.to_i / 1024).to_s + " Kb"
-    res[:uri] = base_id
-    res
-  end
-
-  def audio_resource(document)
-    res = basic_resource(document)
-    base_id = document["id"]
-    base_filename = base_id.gsub(/\:/,"")
-    if res[:mime_type] =~ /wav/
-      ext = 'wav'
-    elsif res[:mime_type] =~ /mpeg/
-      ext = 'mp3'
-    else
-      ext = 'bin'
-    end
-    filename = base_filename + "." + ext
-    dc_filename = base_filename + "_dc.xml"
-    res[:uri] = base_id
-    res[:block] = "CONTENT"
-    res[:filename] = filename
-    res[:dc_path] = fedora_content_path(:download_method=>"show_pretty", :uri=>base_id, :block=>"DC", :filename=>dc_filename)
-    res[:mime_type] = document["dc_format_ssm"] ? document["dc_format_ssm"].first : "application/octect-stream"
-    res[:content_models] = document["has_model_ssim"]
-    res[:file_size] = document["extent_ssim"].first.to_i
-    res[:size] = (document["extent_ssim"].first.to_i / 1024).to_s + " Kb"
-    res
-  end
-
-end