cul_hydra 1.9.1 → 1.9.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 196168f0f3d0ca0289bec9980f0eaad07fad8d3ae1c594b35cc5255bddc72fb7
-  data.tar.gz: 5d323a16c37a06ef64b286d647f666c4013c9a1a75c56c71f4ce434f618739cc
+  metadata.gz: fee2757a484cdb805382ca83fcc8c21354353c85a0d05e55709f090f37d049e3
+  data.tar.gz: 2a642263bb0225de2cb0f6bc2b1b566de3cd8d7e34e78d4dc4be0b76b3d002ab
 SHA512:
-  metadata.gz: 26c50ebe9f4132b295a7be90a865825b12b682428d2127f7f5210769f89fac78d4737f8a3f3fc4cea27156797cea743ab959fd52c32f1e52e7db2741f22b1f44
-  data.tar.gz: 03d3fb13ec720b1deb378c7be29b7a77f8fab281f3592a49980bd31d0fb605c2821c708f0636142dff067e520ad8deea6e4bd1252568356b6df75fd71bd0beea
+  metadata.gz: 7ac25bc59536aa135ae8a5eda1a3d26faf2dfdbc540e8f272ec73a208a956af3043a85ec7ba2741b466461c431ebfc7cc2abb65398e1ad3c8047bc84ba4cb17a
+  data.tar.gz: c8175c178d6deab4403fa453bb7e441563aff35f1b67e17ee29d215658020aa4d2645d59808d6019766f17a8a600451787911ba62e1b9fe1cf67b80769523274

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 +1,40 @@
+module Cul
+module Hydra
+module Datastreams
+class EncodedTextDatastream < ::ActiveFedora::Datastream
+  DEFAULT_PRIORITIES = [ Encoding::UTF_8, Encoding::WINDOWS_1252, Encoding::ISO_8859_1 ]
+
+  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
+    data = data.read if data.is_a? IO
+    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
+    data.force_encoding(content_encoding)
+  end
+end
+end
+end
+end

data/app/models/generic_resource.rb

@@ -68,7 +68,8 @@ class GenericResource < ::ActiveFedora::Base
     solr_doc["fulltext_tesim"] = []
     unless self.datastreams["fulltext"].nil?
       solr_doc["fulltext_tesim"].concat(solr_doc["title_display_ssm"]) unless solr_doc["title_display_ssm"].nil? or solr_doc["title_display_ssm"].length == 0
-      solr_doc["fulltext_tesim"] << self.datastreams["fulltext"].content
+      utf8able = Cul::Hydra::Datastreams::EncodedTextDatastream.utf8able!(self.datastreams["fulltext"].content)
+      solr_doc["fulltext_tesim"] << utf8able.encode(Encoding::UTF_8)
     end
     relationships(:original_name).each do |original_name|
       solr_doc["original_name_tesim"] ||= []

data/fixtures/spec/BLOB/description-cp1252.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/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/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

@@ -1,8 +1,12 @@
+require "active-fedora"
+
 module Cul::Hydra::Indexer
 
   NUM_FEDORA_RETRY_ATTEMPTS = 3
   DELAY_BETWEEN_FEDORA_RETRY_ATTEMPTS = 5.seconds
-
+  DEFAULT_INDEX_OPTS = {
+    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?
       raise 'Please supply a pid (e.g. rake recursively_index_fedora_objects pid=ldpd:123)'
@@ -60,36 +64,59 @@ module Cul::Hydra::Indexer
   end
   def self.recursively_index_fedora_objects(top_pid, pids_to_omit=nil, skip_generic_resources=false, verbose_output=false)
 
+    index_opts = { skip_generic_resources: skip_generic_resources, verbose_output: verbose_output }
     descend_from(top_pid, pids_to_omit, verbose_output) do |pid|
-      self.index_pid(pid, skip_generic_resources, verbose_output)
+      self.index_pid(pid, index_opts)
     end
+  end
 
+  # this is a compatibility method for bridging the previously used postional arguments to
+  # keyword arguments by extracting an opts hash from varargs
+  # legacy positional opts signature: skip_resources = false, verbose_output = false, softcommit = true
+  # keyword defaults are in DEFAULT_INDEX_OPTS
+  # @param args [Array] a list of arguments ending with an options hash
+  # @return options hash
+  def self.extract_index_opts(args)
+    args = args.dup # do not modify the original list
+    # extract opts hash
+    index_opts = (args.last.is_a? Hash) ? args.pop : {}
+    # symbolize keys and reverse merge defaults
+    index_opts = index_opts.map {|k,v| [k.to_sym, v] }.to_h
+    index_opts = DEFAULT_INDEX_OPTS.merge(index_opts)
+    # assign any legacy positional arguments, permitting explicit nils
+    unless args.empty?
+      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
+    index_opts
   end
 
-  def self.index_pid(pid, skip_generic_resources=false, verbose_output=false, softcommit=true)
+  def self.index_pid(pid, *args)
     # We found an object with the desired PID. Let's reindex it
+    index_opts = extract_index_opts(args)
     begin
       active_fedora_object = nil
 
       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 softcommit
+            if index_opts[:softcommit]
               active_fedora_object.update_index
             else
               # Using direct solr query to update document without soft commiting
               ActiveFedora::SolrService.add(active_fedora_object.to_solr)
             end
-            puts 'done.' if verbose_output
+            puts 'done.' if index_opts[:verbose_output]
           end
           break
         rescue RestClient::RequestTimeout, Errno::EHOSTUNREACH => e
           remaining_attempts = (NUM_FEDORA_RETRY_ATTEMPTS-1) - i
           if remaining_attempts == 0
-            raise e
+            raise
           else
             Rails.logger.error "Error: Could not connect to fedora. (#{e.class.to_s + ': ' + e.message}).  Will retry #{remaining_attempts} more #{remaining_attempts == 1 ? 'time' : 'times'} (after a #{DELAY_BETWEEN_FEDORA_RETRY_ATTEMPTS} second delay)."
             sleep DELAY_BETWEEN_FEDORA_RETRY_ATTEMPTS
@@ -102,15 +129,18 @@ module Cul::Hydra::Indexer
             sleep 5
           else
             # Other RuntimeErrors should be passed on
-            raise e
+            raise
           end
         end
       end
     rescue SystemExit, Interrupt => e
       # Allow system interrupt (ctrl+c)
-      raise e
+      raise
     rescue Exception => e
       puts "Encountered problem with #{pid}.  Skipping record.  Exception class: #{e.class.name}.  Message: #{e.message}"
+      if index_opts[:reraise]
+        raise
+      end
     end
   end
 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 = '1.9.1'
+    VERSION = '1.9.6'
     def self.version
       VERSION
     end

data/lib/tasks/index.rake

@@ -91,10 +91,11 @@ namespace :cul_hydra do
       pool = Thread.pool(thread_pool_size)
       mutex = Mutex.new
 
+      index_opts = { skip_generic_resources: skip_generic_resources, verbose_output: false }
       pids.each do |pid|
         pool.process {
           
-          Cul::Hydra::Indexer.index_pid(pid, skip_generic_resources, false)
+          Cul::Hydra::Indexer.index_pid(pid, index_opts)
           
           mutex.synchronize do
             counter += 1
@@ -124,7 +125,8 @@ namespace :cul_hydra do
         next
       end
 
-      skip_generic_resources = (ENV['skip_generic_resources'] == 'true')
+      index_opts = { verbose_output: false }
+      index_opts[:skip_generic_resources] = (ENV['skip_generic_resources'] == 'true')
 
       start_time = Time.now
       pids = Cul::Hydra::RisearchMembers.get_publish_target_member_pids(publish_target_pid, true)
@@ -133,7 +135,7 @@ namespace :cul_hydra do
       counter = 0
 
       pids.each do |pid|
-        Cul::Hydra::Indexer.index_pid(pid, skip_generic_resources, false)
+        Cul::Hydra::Indexer.index_pid(pid, index_opts)
         counter += 1
         puts "Indexed #{counter} of #{total} | #{Time.now - start_time} seconds"
       end

metadata

@@ -1,15 +1,15 @@
 --- !ruby/object:Gem::Specification
 name: cul_hydra
 version: !ruby/object:Gem::Version
-  version: 1.9.1
+  version: 1.9.6
 platform: ruby
 authors:
 - Benjamin Armintor
 - Eric O'Hanlon
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2020-03-10 00:00:00.000000000 Z
+date: 2020-11-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rails
@@ -31,14 +31,14 @@ dependencies:
     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
@@ -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 +451,8 @@ files:
 - fixtures/cmodels/ldpd_sdep.StaticImageCore.xml
 - fixtures/cmodels/ore_Proxy.xml
 - fixtures/cmodels/pcdm_Collection.xml
+- fixtures/spec/BLOB/description-cp1252.txt
+- 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 +462,6 @@ 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
@@ -493,6 +492,7 @@ 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
@@ -548,7 +548,7 @@ files:
 homepage: https://github.com/cul/cul_hydra
 licenses: []
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - app
@@ -566,8 +566,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.3
-signing_key: 
+rubygems_version: 3.0.8
+signing_key:
 specification_version: 4
 summary: ActiveFedora, OM, and Solrizer implementations for CUL repository apps
 test_files: []

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>