ddr-models 2.4.7 → 2.4.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: f6f4046f421e1eff0e98a72a91752fcefdc2d58b
-  data.tar.gz: 19170b14ef451f3f41f5dc265462e54e86da74a8
+  metadata.gz: 529d7cf0ef14124eeff5d9aca162a22d99f9f882
+  data.tar.gz: 309f4bdb1800c6f94ab8e92d2fbbace193031a54
 SHA512:
-  metadata.gz: 087dc599aead9cb898234e9b20e16e44fe8cd61f74914cf77573ad92a32f6043d0543401c1631e350d12bacdd68970edc474a9f9589deb1e90ec95a10848bcba
-  data.tar.gz: cc83cf6ed16414010004621c928da3768dc06c3669cabb37f8d28198549650101b31034965830c8ba9f8b1514813cc659894177b7acdc058ddda17b488acb03a
+  metadata.gz: f67fae61c9e4eb4802095b3761df7d77f65ae13bedfc44b042c82f4a245b046f814023bd395a638c9a9316650f1231ac1e440c78bb4119baf38d014d3e8f36f9
+  data.tar.gz: f92dc9427534c10ebee195eaaa629e2371ef61b35eda6550acd1dd73b6cddc85b9f9b230b1b570c1356596cd650277dff1e2719df09b45e4d9204bad4ad281de

data/app/models/component.rb

@@ -12,6 +12,8 @@ class Component < Ddr::Models::Base
   belongs_to :parent, :property => :is_part_of, :class_name => 'Item'
   belongs_to :target, :property => :has_external_target, :class_name => 'Target'
 
+  after_save :index_parent, if: :has_extracted_text?, unless: "parent.nil?"
+
   alias_method :item, :parent
   alias_method :item=, :parent=
 
@@ -23,4 +25,8 @@ class Component < Ddr::Models::Base
     self.collection.internal_uri rescue nil
   end
 
+  def index_parent
+    Resque.enqueue(Ddr::Jobs::UpdateIndex, parent_id)
+  end
+
 end

data/app/models/item.rb

@@ -21,4 +21,17 @@ class Item < Ddr::Models::Base
   alias_method :collection_id, :parent_id
   alias_method :collection=, :parent=
 
+  def children_having_extracted_text
+    item = self
+    Ddr::Index::Query.new do
+      is_part_of item
+      where attached_files_having_content: "extractedText"
+      fields :id, :extracted_text
+    end
+  end
+
+  def all_text
+    children_having_extracted_text.docs.map(&:extracted_text).flatten
+  end
+
 end

data/ddr-models.gemspec

@@ -26,7 +26,7 @@ Gem::Specification.new do |s|
   s.add_dependency "devise", "~> 3.4"
   s.add_dependency "omniauth-shibboleth", "~> 1.2.0"
   s.add_dependency "grouper-rest-client"
-  s.add_dependency "ezid-client", "~> 1.1", ">= 1.1.1"
+  s.add_dependency "ezid-client", "~> 1.4.2"
   s.add_dependency "resque", "~> 1.25"
   s.add_dependency "rdf-vocab", "~> 0.8"
   s.add_dependency "net-ldap", "~> 0.11"

data/lib/ddr/index/fields.rb

@@ -8,7 +8,10 @@ module Ddr::Index
     ACTIVE_FEDORA_MODEL         = Field.new :active_fedora_model, :stored_sortable
     ADMIN_SET                   = Field.new :admin_set, :stored_sortable
     ADMIN_SET_FACET             = Field.new :admin_set_facet, :facetable
+    ALL_TEXT                    = Field.new :all_text, solr_name: "all_text_timv"
     ASPACE_ID                   = Field.new :aspace_id, :stored_sortable
+    ATTACHED_FILES_HAVING_CONTENT =
+      Field.new :attached_files_having_content, :symbol
     BOX_NUMBER_FACET            = Field.new :box_number_facet, :facetable
     COLLECTION_FACET            = Field.new :collection_facet, :facetable
     COLLECTION_URI              = Field.new :collection_uri, :symbol
@@ -24,7 +27,7 @@ module Ddr::Index
     DISPLAY_FORMAT              = Field.new :display_format, :stored_sortable
     DOI                         = Field.new :doi, :symbol
     EAD_ID                      = Field.new :ead_id, :stored_sortable
-    EXTRACTED_TEXT              = Field.new :extracted_text, :searchable, type: :text
+    EXTRACTED_TEXT              = Field.new :extracted_text, solr_name: "extracted_text_tsm"
     HAS_MODEL                   = Field.new :has_model, :symbol
     IDENTIFIER_ALL              = Field.new :identifier_all, :symbol
     INTERNAL_URI                = Field.new :internal_uri, :stored_sortable

data/lib/ddr/index/filter.rb

@@ -76,6 +76,10 @@ module Ddr::Index
         term is_member_of_collection: internal_uri(object_or_id)
       end
 
+      def is_part_of(object_or_id)
+        term is_part_of: internal_uri(object_or_id)
+      end
+
       def model(*models)
         where active_fedora_model: models
       end

data/lib/ddr/models/fixity_checkable.rb

@@ -5,6 +5,7 @@ module Ddr
       def datastreams_to_validate
         datastreams.select { |dsid, ds| ds.has_content? }
       end
+      alias_method :attached_files_having_content, :datastreams_to_validate
 
       def fixity_checks
         Ddr::Events::FixityCheckEvent.for_object(self)

data/lib/ddr/models/indexing.rb

@@ -20,6 +20,7 @@ module Ddr
           ACCESS_ROLE           => roles.to_json,
           ADMIN_SET             => admin_set,
           ASPACE_ID             => aspace_id,
+          ATTACHED_FILES_HAVING_CONTENT => attached_files_having_content.keys,
           BOX_NUMBER_FACET      => desc_metadata_values('box_number'),
           CONTRIBUTOR_FACET     => desc_metadata_values('contributor'),
           CREATOR_FACET         => creator,
@@ -90,6 +91,7 @@ module Ddr
         if is_a? Item
           fields[ADMIN_SET_FACET] = admin_set_facet
           fields[COLLECTION_FACET] = collection_facet
+          fields[ALL_TEXT] = all_text
         end
         fields
       end

data/lib/ddr/models/solr_document.rb

@@ -131,6 +131,10 @@ module Ddr::Models
       has_datastream?(Ddr::Datastreams::CONTENT)
     end
 
+    def has_extracted_text?
+      has_datastream?(Ddr::Datastreams::EXTRACTED_TEXT)
+    end
+
     def content_ds
       datastreams[Ddr::Datastreams::CONTENT]
     end

data/lib/ddr/models/structure.rb

@@ -17,6 +17,10 @@ module Ddr
         xpath("//xmlns:structMap[@TYPE='#{type}']").first
       end
 
+      def fptr_nodes(type='default')
+        structMap_node(type).xpath('//xmlns:fptr')
+      end
+
       def as_xml_document
         __getobj__
       end
@@ -49,4 +53,4 @@ module Ddr
 
     end
   end
-end
+end

data/lib/ddr/models/version.rb

@@ -1,5 +1,5 @@
 module Ddr
   module Models
-    VERSION = "2.4.7"
+    VERSION = "2.4.8"
   end
 end

data/spec/fixtures/extractedText1.txt

@@ -0,0 +1 @@
+Lorem ipsum dolor sit amet, consectetur adipiscing elit. Duis scelerisque diam in lacus sollicitudin maximus. Duis molestie rutrum hendrerit. Vestibulum lobortis augue a ultrices mollis. Mauris semper eu odio posuere scelerisque. Donec vel condimentum nunc. Cum sociis natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Morbi fermentum, sem vel interdum suscipit, lacus risus ultricies sem, in lacinia ligula risus ut est. Integer gravida, orci ac lobortis blandit, eros tortor bibendum turpis, quis semper velit erat in augue.

data/spec/fixtures/extractedText2.txt

@@ -0,0 +1 @@
+Proin convallis nibh quis ex iaculis, eget porttitor elit consequat. Duis in tincidunt ipsum. Aenean imperdiet lacinia libero nec egestas. Fusce a metus volutpat, vehicula augue accumsan, mattis lorem. Donec ex tellus, tempus vitae orci sit amet, sollicitudin pharetra velit. Fusce egestas nunc vitae placerat elementum. Nullam pharetra porta eros, quis tempor sem pulvinar a. Ut in neque cursus, hendrerit neque in, congue est. Phasellus neque mauris, eleifend in sem ac, posuere tincidunt odio. Duis lectus nibh, ultricies a elementum a, consectetur vitae turpis.

data/spec/fixtures/extractedText3.txt

@@ -0,0 +1,3 @@
+Ut eu consequat turpis. Proin commodo facilisis pretium. Vestibulum tristique volutpat lobortis. Donec efficitur tincidunt porta. Sed dui massa, rutrum iaculis interdum at, scelerisque nec velit. Suspendisse potenti. Aliquam quis mauris sed magna convallis ullamcorper non ut ipsum.
+
+Donec dolor urna, malesuada id condimentum ut, feugiat consectetur lectus. Morbi dignissim nibh neque, a bibendum ex semper vel. Nunc pellentesque varius bibendum. Etiam posuere lorem nibh, sit amet iaculis dolor lobortis nec. Nam congue dui non lectus luctus scelerisque. Aliquam turpis orci, porta non dapibus eu, accumsan sed tortor. Phasellus placerat ac mauris sed vehicula. Sed dolor massa, fringilla vitae urna vel, efficitur porta eros. Ut massa nisi, bibendum id mollis quis, aliquet eu purus. Aliquam erat volutpat. Phasellus accumsan felis eget dui vulputate, eget feugiat diam eleifend. Donec facilisis massa odio, ac auctor eros ullamcorper quis. Donec sit amet malesuada lectus. Aenean egestas, leo id lacinia mattis, dui ex tempor tellus, vel mollis velit leo a ante.

data/spec/models/component_spec.rb

@@ -16,4 +16,68 @@ RSpec.describe Component, type: :model, components: true do
     its(:index_fields) { is_expected.to include(Ddr::Index::Fields::COLLECTION_URI => "info:fedora/test:1") }
   end
 
+  describe "extracted text" do
+    let(:parent) { FactoryGirl.create(:item) }
+    describe "when the child is not already associated with the parent" do
+      before {
+        subject.extractedText.content = fixture_file_upload('extractedText1.txt')
+        subject.save
+        subject.reload
+      }
+      it "updates the parent index when associated" do
+        expect(subject).to receive(:index_parent)
+        subject.parent = parent
+        subject.save
+      end
+    end
+    describe "when the child is already associated with the parent" do
+      before {
+        subject.parent = parent
+        subject.save
+      }
+      describe "when there is no extracted text" do
+        describe "and none is added" do
+          it "does not trigger an index update on the parent" do
+            expect(subject).not_to receive(:index_parent)
+            subject.save
+          end
+        end
+        describe "and extracted text is added" do
+          it "triggers an index update on the parent" do
+            expect(subject).to receive(:index_parent)
+            subject.extractedText.content = fixture_file_upload('extractedText1.txt')
+            subject.save
+          end
+        end
+      end
+      describe "when extracted text exists" do
+        before {
+          subject.extractedText.content = fixture_file_upload('extractedText1.txt')
+          subject.save
+          subject.reload
+        }
+        describe "and is removed" do
+          it "triggers an index update on the parent" do
+            pending "Deleting a datastream does not mark the content as changed"
+            expect(subject).to receive(:index_parent)
+            subject.extractedText.delete
+            subject.save
+          end
+        end
+        describe "and doesn't change" do
+          it "triggers an index update on the parent" do
+            expect(subject).to receive(:index_parent)
+            subject.save
+          end
+        end
+        describe "and changes" do
+          it "triggers an index update on the parent" do
+            expect(subject).to receive(:index_parent)
+            subject.extractedText.content = fixture_file_upload('extractedText2.txt')
+            subject.save
+          end
+        end
+      end
+    end
+  end
 end

data/spec/models/indexing_spec.rb

@@ -48,6 +48,9 @@ module Ddr::Models
       }
 
       its([Indexing::CONTENT_CREATE_DATE]) { is_expected.to eq "2016-01-22T21:50:33Z" }
+      its([Indexing::ATTACHED_FILES_HAVING_CONTENT]) {
+        is_expected.to contain_exactly("content", "RELS-EXT", "descMetadata", "adminMetadata")
+      }
     end
 
   end

data/spec/models/item_spec.rb

@@ -7,4 +7,27 @@ RSpec.describe Item, type: :model do
   it_behaves_like "it has an association", :has_many, :children, :is_part_of, "Component"
   it_behaves_like "a non-collection model"
 
+  describe "indexing text" do
+    let(:children) { FactoryGirl.build_list(:component, 5) }
+
+    let(:text1) { fixture_file_upload('extractedText1.txt', 'text/plain') }
+    let(:text2) { fixture_file_upload('extractedText2.txt', 'text/plain') }
+    let(:text3) { fixture_file_upload('extractedText3.txt', 'text/plain') }
+
+    before {
+      children[0].extractedText.content = text1
+      children[0].save
+      children[1].extractedText.content = text2
+      children[1].save
+      children[2].extractedText.content = text3
+      children[2].save
+      subject.children = children
+      subject.save
+    }
+
+    it "indexes the combined text of its children" do
+      expect(subject.index_fields[Ddr::Index::Fields::ALL_TEXT]).to contain_exactly(text1.read, text2.read, text3.read)
+    end
+  end
+
 end

data/spec/models/structure_spec.rb

@@ -5,9 +5,8 @@ module Ddr
   module Models
     RSpec.describe Structure, type: :model, structural_metadata: true do
 
-      let(:structure) { FactoryGirl.build(:multiple_struct_maps_structure) }
-
       describe "#struct_maps" do
+        let(:structure) { FactoryGirl.build(:multiple_struct_maps_structure) }
         let(:struct_maps) { structure.struct_maps }
         it "should include struct divs for each struct map" do
           expect(struct_maps.keys).to match_array([ 'default', 'reverse' ])
@@ -16,6 +15,16 @@ module Ddr
         end
       end
 
+      describe "#fptr_nodes" do
+        let(:structure) { FactoryGirl.build(:nested_structure) }
+        it "should return all fptr nodes" do
+          results = structure.fptr_nodes
+          expect(results.size).to eq(3)
+          ids = results.map { |entry| entry['CONTENTIDS'] }
+          expect(ids).to match_array([ 'info:fedora/test:5', 'info:fedora/test:6', 'info:fedora/test:7' ])
+        end
+      end
+
     end
   end
-end
+end

data/spec/support/ezid_mock_identifier.rb

@@ -6,8 +6,8 @@ module Ezid
 
     self.defaults = {}
 
-    def reload; self; end
-    def reset; self; end
+    def load_metadata; self; end
+    def reset_metadata; self; end
 
     private
 

data/spec/support/shared_examples_for_has_content.rb

@@ -31,14 +31,12 @@ RSpec.shared_examples "an object that can have content" do
   describe "extracted text" do
     describe "when it is not present" do
       its(:has_extracted_text?) { should be false }
-      its(:to_solr) { should_not include(Ddr::Index::Fields::EXTRACTED_TEXT) }
+      its(:to_solr) { is_expected.not_to include(Ddr::Index::Fields::EXTRACTED_TEXT) }
     end
     describe "when it is present" do
       before { subject.extractedText.content = "This is my text. See Spot run." }
       its(:has_extracted_text?) { should be true }
-      it "should be indexed" do
-        expect(subject.to_solr[Ddr::Index::Fields::EXTRACTED_TEXT]).to eq("This is my text. See Spot run.")
-      end
+      its(:to_solr) { is_expected.to include(Ddr::Index::Fields::EXTRACTED_TEXT) }
     end
   end
 

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: ddr-models
 version: !ruby/object:Gem::Version
-  version: 2.4.7
+  version: 2.4.8
 platform: ruby
 authors:
 - Jim Coble
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-02-23 00:00:00.000000000 Z
+date: 2016-03-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rails
@@ -129,20 +129,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.1'
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 1.1.1
+        version: 1.4.2
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.1'
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 1.1.1
+        version: 1.4.2
 - !ruby/object:Gem::Dependency
   name: resque
   requirement: !ruby/object:Gem::Requirement
@@ -639,6 +633,9 @@ files:
 - spec/fixtures/8bit.tif
 - spec/fixtures/arrow1rightred_e0.gif
 - spec/fixtures/bird.jpg
+- spec/fixtures/extractedText1.txt
+- spec/fixtures/extractedText2.txt
+- spec/fixtures/extractedText3.txt
 - spec/fixtures/fits/document.xml
 - spec/fixtures/fits/image.xml
 - spec/fixtures/imageA.tif
@@ -811,6 +808,9 @@ test_files:
 - spec/fixtures/8bit.tif
 - spec/fixtures/arrow1rightred_e0.gif
 - spec/fixtures/bird.jpg
+- spec/fixtures/extractedText1.txt
+- spec/fixtures/extractedText2.txt
+- spec/fixtures/extractedText3.txt
 - spec/fixtures/fits/document.xml
 - spec/fixtures/fits/image.xml
 - spec/fixtures/imageA.tif