cul_hydra 1.9.1 → 1.9.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 196168f0f3d0ca0289bec9980f0eaad07fad8d3ae1c594b35cc5255bddc72fb7
-  data.tar.gz: 5d323a16c37a06ef64b286d647f666c4013c9a1a75c56c71f4ce434f618739cc
+  metadata.gz: a7c942afbb5baf9638461f80723af40b30633892a7774d426ad4edabe6dbdab4
+  data.tar.gz: 549c735df93fed430ebb33708a59829f7f7228e664309d1f19f323f6de7ef4a6
 SHA512:
-  metadata.gz: 26c50ebe9f4132b295a7be90a865825b12b682428d2127f7f5210769f89fac78d4737f8a3f3fc4cea27156797cea743ab959fd52c32f1e52e7db2741f22b1f44
-  data.tar.gz: 03d3fb13ec720b1deb378c7be29b7a77f8fab281f3592a49980bd31d0fb605c2821c708f0636142dff067e520ad8deea6e4bd1252568356b6df75fd71bd0beea
+  metadata.gz: 0d5eba1acff06eb1789b529d0f4871a4d40ce48f76d6e1cbd5acff3162495b156fce85ed2910c581fc7f71548f0c42de3551343b5e85aeddbfbe63f6ea8062d1
+  data.tar.gz: 57d089ae06742089dd3e140307cf3cf787615196e69e8fbbc9a271fab3ff38f0ab4bf97554cab459d9cf1fa7ea66dedb2a17dbb49cb2480706e5712b5e61749b

data/app/models/concept.rb

@@ -9,7 +9,7 @@ class Concept < GenericAggregator
   rdf_types(RDF::CUL.Aggregator)
   rdf_types(RDF::PCDM.Object)
 
-  has_file_datastream :name => "descriptionText", :type=>::ActiveFedora::Datastream,
+  has_file_datastream :name => "descriptionText", :type=>Cul::Hydra::Datastreams::EncodedTextDatastream,
                  :versionable => false, :label => 'Textual Description of Concept',
                  :mimeType => 'text/markdown'
 
@@ -107,7 +107,15 @@ class Concept < GenericAggregator
 
   def to_solr(solr_doc = Hash.new, opts={})
     solr_doc = super(solr_doc, opts)
-    solr_doc[::ActiveFedora::SolrService.solr_name(:description_text, :displayable)] = description
+    description.tap do |description_value|
+      if description_value
+        unless description_ds.is_a? Cul::Hydra::Datastreams::EncodedTextDatastream
+          description_value = Cul::Hydra::Datastreams::EncodedTextDatastream.utf8able!(description_value).encode(Encoding::UTF_8)
+        end
+        description_field_name = ::ActiveFedora::SolrService.solr_name(:description_text, :displayable)
+        solr_doc[description_field_name] = description_value
+      end
+    end
     solr_doc
   end
 

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

@@ -0,0 +1,38 @@
+module Cul
+module Hydra
+module Datastreams
+class EncodedTextDatastream < ::ActiveFedora::Datastream
+  DEFAULT_PRIORITIES = [ Encoding::UTF_8, Encoding::ISO_8859_1, Encoding::WINDOWS_1252 ]
+  def initialize(digital_object=nil, dsid=nil, options={})
+    @encoding_priorities = options.delete(:encodings) || DEFAULT_PRIORITIES
+    super
+  end
+  def content=(value)
+    super(utf8able!(value).encode!(Encoding::UTF_8))
+  end
+
+  def content
+    utf8able!(super.to_s).encode!(Encoding::UTF_8)
+  end
+
+  def utf8able!(data)
+    EncodedTextDatastream.utf8able!(data, @encoding_priorities)
+  end
+
+  def self.utf8able!(data, encoding_priorities = DEFAULT_PRIORITIES)
+    return unless data
+    content_encoding = encoding_priorities.detect do |enc|
+      begin
+        data.force_encoding(enc).valid_encoding?
+      rescue Encoding::UndefinedConversionError, Encoding::InvalidByteSequenceError
+        false
+      end
+    end
+    raise "could not encode text datastream content" unless content_encoding
+    puts "using encoding #{content_encoding}"
+    data.force_encoding(content_encoding)
+  end
+end
+end
+end
+end

data/fixtures/spec/BLOB/description-utf8.txt

@@ -0,0 +1 @@
+In 1917, Harrison founded the first organization (The Liberty League) and the first newspaper (The Voice) of the “New Negro Movement” and he published his first book, The Negro and the Nation. He opposed positions taken by Joel E. Spingarn and W.E.B. Du Bois of the NAACP during the First World War and, along with William Monroe Trotter and others he organized the 1918 Liberty Congress. The Congress, the major Black protest effort during the war, demanded enforcement of the Thirteenth, Fourteenth, and Fifteenth Amendments and federal anti-lynching legislation. Beginning in 1920, he became the principal editor of Marcus Garvey's Negro World, which he reshaped into a leading political and literary publication of the era. In its pages, he discussed history, politics, theater, international affairs, religion, and science. He also created a "Poetry for the People" feature, a “West Indian News Notes” column, and what he described as the first regular book review section by a Black author in “Negro newspaperdom.” In 1920 he also published his second book, When Africa Awakes: The “Inside Story” of the Stirrings and Strivings of the New Negro in the Western World. Later, he would criticize Garvey's methods and actions.

data/lib/cul_hydra/version.rb

@@ -1,6 +1,6 @@
 module Cul
   module Hydra
-    VERSION = '1.9.1'
+    VERSION = '1.9.2'
     def self.version
       VERSION
     end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: cul_hydra
 version: !ruby/object:Gem::Version
-  version: 1.9.1
+  version: 1.9.2
 platform: ruby
 authors:
 - Benjamin Armintor
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-03-10 00:00:00.000000000 Z
+date: 2020-03-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rails
@@ -394,6 +394,7 @@ files:
 - app/models/content_aggregator.rb
 - app/models/cul/hydra/datastreams/access_control_metadata.rb
 - app/models/cul/hydra/datastreams/dc_metadata.rb
+- app/models/cul/hydra/datastreams/encoded_text_datastream.rb
 - app/models/cul/hydra/datastreams/mods_document.rb
 - app/models/cul/hydra/datastreams/rels_int.rb
 - app/models/cul/hydra/datastreams/struct_metadata.rb
@@ -402,14 +403,11 @@ files:
 - app/models/generic_object.rb
 - app/models/generic_resource.rb
 - bin/rails
-- config/fedora.yml
 - config/jetty.yml
 - config/locales/ldpd_hydra.en.yml
 - config/predicate_mappings.yml
-- config/solr.yml
 - config/solr_mappings.yml
 - config/solr_value_maps.yml
-- config/subs.yml
 - fixtures/cmodels/ldpd_ADLMetadata.xml
 - fixtures/cmodels/ldpd_AESMetadata.xml
 - fixtures/cmodels/ldpd_BagAggregator.xml
@@ -453,6 +451,7 @@ files:
 - fixtures/cmodels/ldpd_sdep.StaticImageCore.xml
 - fixtures/cmodels/ore_Proxy.xml
 - fixtures/cmodels/pcdm_Collection.xml
+- fixtures/spec/BLOB/description-utf8.txt
 - fixtures/spec/BLOB/dlc.md
 - fixtures/spec/BLOB/test001.jpg
 - fixtures/spec/CUL_ACCESS/access-closed.xml
@@ -462,7 +461,6 @@ files:
 - fixtures/spec/CUL_MODS/mods-001.xml
 - fixtures/spec/CUL_MODS/mods-access-condition.xml
 - fixtures/spec/CUL_MODS/mods-all.xml
-- fixtures/spec/CUL_MODS/mods-archival-context-2.xml
 - fixtures/spec/CUL_MODS/mods-archival-context.xml
 - fixtures/spec/CUL_MODS/mods-bad-repo.xml
 - fixtures/spec/CUL_MODS/mods-date-created-range.xml
@@ -566,7 +564,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.3
+rubygems_version: 3.1.2
 signing_key: 
 specification_version: 4
 summary: ActiveFedora, OM, and Solrizer implementations for CUL repository apps

data/config/fedora.yml

@@ -1,26 +0,0 @@
-development:
-  user: fedoraAdmin
-  password: f+BULUS*^
-  url: http://repository.cul.columbia.edu:8080/fedora
-  datastreams_root: /ifs/cul/repo/archive/repo-datastreams
-  time_zone: "America/New_York"
-test: &TEST  
-  user: fedoraAdmin
-  password: fedoraAdmin
-  url: <%= "http://127.0.0.1:#{ENV['TEST_JETTY_PORT'] || 8983}/fedora-test" %>
-  datastreams_root: /ifs/cul/repo/archive/repo-datastreams
-  time_zone: "America/New_York"
-production:
-  user: fedoraAdmin
-  password: f+BULUS*^
-  url: http://repository.cul.columbia.edu:8080/fedora
-  datastreams_root: /ifs/cul/repo/archive/repo-datastreams
-  time_zone: "America/New_York"
-dcv_test:
-  user: fedoraAdmin
-  password: f+BULUS*^
-  url: http://repository.cul.columbia.edu:8080/fedora
-  datastreams_root: /ifs/cul/repo/archive/repo-datastreams
-  time_zone: "America/New_York"
-cucumber:
-  <<: *TEST

data/config/solr.yml

@@ -1,17 +0,0 @@
-# This is a sample config file that does not have multiple solr instances. You will also need to be sure to
-# edit the fedora.yml file to match the solr URL for active-fedora. 
-#  url: http://katana.cul.columbia.edu:8080/solr-4.7/dcv_private_dev
-test: &TEST
-  url: <%= "http://127.0.0.1:#{ENV['TEST_JETTY_PORT'] || 8983}/solr/test" %>
-dcv_dev: &DEV
-  url: http://spatha.cul.columbia.edu:8080/solr-4.7/dcv_private_test
-  dcv_core: 'dcv_private_dev'
-dcv_test: &IFP
-  url: http://spatha.cul.columbia.edu:8080/solr-4.7/dcv_private_test
-  dcv_core: 'dcv_private_test'
-cucumber:
-  <<: *TEST
-production: &PROD
-  url: http://spatha.cul.columbia.edu:8080/solr-4.7/dcv_prod
-development:
-  <<: *PROD

data/config/subs.yml

@@ -1,10 +0,0 @@
-old:
-  fedora_server: http://sayers.cul.columbia.edu:8080
-  php_server: http://bach.cul.columbia.edu/dev
-  djatoka_server: http://iris.cul.columbia.edu:8080
-development:
-  fedora_server: http://repository.cul.columbia.edu:8080
-  php_server: http://fedora-svc.cul.columbia.edu/dev
-test:
-  fedora_server: http://repository.cul.columbia.edu:8080
-  php_server: http://fedora-svc.cul.columbia.edu/dev

data/fixtures/spec/CUL_MODS/mods-archival-context-2.xml

@@ -1,22 +0,0 @@
-<?xml version="1.0"?>
-<mods:mods xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:mods="http://www.loc.gov/mods/v3" version="3.6" xsi:schemaLocation="http://www.loc.gov/mods/v3 http://www.loc.gov/standards/mods/v3/mods-3-6.xsd">
-  <mods:titleInfo>
-    <mods:title>Agreement between Gunnar Myrdal and Harper &amp; Brothers for the publication of "An American Dilemma," January 8, 1943</mods:title>
-  </mods:titleInfo>
-  <mods:relatedItem type="host" displayLabel="Project">
-    <mods:titleInfo>
-      <mods:title>Carnegie Digital Past and Future</mods:title>
-    </mods:titleInfo>
-  </mods:relatedItem>
-  <mods:relatedItem type="host" displayLabel="Collection">
-    <mods:titleInfo valueURI="http://id.library.columbia.edu/term/72fcfd07-f7db-4a18-bdb2-beb0abce071c">
-      <mods:title>Carnegie Corporation of New York Records</mods:title>
-    </mods:titleInfo>
-    <mods:identifier type="CLIO">4079753</mods:identifier>
-    <mods:part>
-      <mods:detail type="Series" level="1">
-        <mods:title>Series II. Files on Microfilm</mods:title>
-      </mods:detail>
-    </mods:part>
-  </mods:relatedItem>
-</mods:mods>