hydra-pbcore 2.4.0 → 3.0.0.rc1

checksums.yaml

@@ -0,0 +1,15 @@
+---
+!binary "U0hBMQ==":
+  metadata.gz: !binary |-
+    OTNhOTQ2MTY5MTMxMjc1YTJhYmEzMDc1M2RkMmU0NzlkOTVmODRlMg==
+  data.tar.gz: !binary |-
+    MzMwNjE5NjliMjc3YWQwZjViODIwNzk3MTQzMmJlYWZkNDRjMjM1Mg==
+!binary "U0hBNTEy":
+  metadata.gz: !binary |-
+    Nzc5Y2I1ZGQ5OTdmMGI0MjA3ZDQzYjYzNzlhODU2Zjc1YmFlZWY1ZTJmNzYz
+    NmYyM2UzMmM0YmJjYTQ0ZGJkNmM0YTVjYzMwNmQxNzFjNzg2OWZjYjVlNTJl
+    ZjU0NzhiYmEyNzU4NDNlNzY5YzBjYWMwODA3ZjkzNWQ4NTY1Mjc=
+  data.tar.gz: !binary |-
+    YTNhZDI5YmVhMWUyMmE3N2U2ZmRjY2JmZDM5MzUxNGMwYzg2ZjRlNGI5YzQx
+    OWIwNzI1ZjAyNzU2OGZjOTdlMWJhZjU2YWY0YWY1MWY1YmI0NDVmY2M2NTZm
+    MTk1NDUzY2Y4YTdmYjE1YmIzMjg3ZTlmNTI4MzUzZDMyNjEyMzc=

data/.gitignore

@@ -6,7 +6,6 @@
 .yardoc
 .rvmrc
 .ruby-version
-Gemfile.lock
 InstalledFiles
 _yardoc
 bin
@@ -21,3 +20,4 @@ test/version_tmp
 tmp/*
 !tmp/.gitkeep
 *.sw[opn]
+config/*

data/Gemfile.lock

@@ -0,0 +1,135 @@
+PATH
+  remote: .
+  specs:
+    hydra-pbcore (3.0.0.rc1)
+      active-fedora
+      solrizer
+
+GEM
+  remote: https://rubygems.org/
+  specs:
+    active-fedora (6.4.3)
+      activesupport (>= 3.0.0)
+      deprecation
+      mediashelf-loggable
+      nom-xml (>= 0.5.1)
+      om (~> 3.0.0)
+      rdf
+      rdf-rdfxml (= 1.0.1)
+      rsolr
+      rubydora (~> 1.6, >= 1.6.5)
+    activemodel (4.0.0)
+      activesupport (= 4.0.0)
+      builder (~> 3.1.0)
+    activesupport (4.0.0)
+      i18n (~> 0.6, >= 0.6.4)
+      minitest (~> 4.2)
+      multi_json (~> 1.3)
+      thread_safe (~> 0.1)
+      tzinfo (~> 0.3.37)
+    addressable (2.3.5)
+    atomic (1.1.10)
+    builder (3.1.4)
+    coderay (1.0.9)
+    columnize (0.3.6)
+    daemons (1.1.9)
+    debugger (1.6.1)
+      columnize (>= 0.3.1)
+      debugger-linecache (~> 1.2.0)
+      debugger-ruby_core_source (~> 1.2.3)
+    debugger-linecache (1.2.0)
+    debugger-ruby_core_source (1.2.3)
+    deprecation (0.0.5)
+      activesupport
+    diff-lcs (1.2.4)
+    equivalent-xml (0.3.0)
+      nokogiri (>= 1.4.3)
+    fastercsv (1.5.5)
+    hooks (0.3.1)
+    i18n (0.6.4)
+    json (1.8.0)
+    mediashelf-loggable (0.4.9)
+    method_source (0.8.2)
+    mime-types (1.23)
+    mini_portile (0.5.1)
+    minitest (4.7.5)
+    multi_json (1.7.7)
+    nokogiri (1.6.0)
+      mini_portile (~> 0.5.0)
+    nom-xml (0.5.1)
+      activesupport
+      i18n
+      nokogiri
+    om (3.0.2)
+      activemodel
+      activesupport
+      deprecation
+      mediashelf-loggable
+      nokogiri (>= 1.4.2)
+      solrizer (~> 3.1.0)
+    pry (0.9.12.2)
+      coderay (~> 1.0.5)
+      method_source (~> 0.8)
+      slop (~> 3.4)
+    rake (10.1.0)
+    rdf (1.0.7)
+      addressable (>= 2.2)
+    rdf-rdfxml (1.0.1)
+      rdf (>= 1.0)
+      rdf-xsd (>= 1.0)
+    rdf-xsd (1.0.2.1)
+      nokogiri (>= 1.5.0)
+      rdf (>= 0.3.4)
+    rdoc (4.0.1)
+      json (~> 1.4)
+    redcarpet (3.0.0)
+    rest-client (1.6.7)
+      mime-types (>= 1.16)
+    rsolr (1.0.9)
+      builder (>= 2.1.2)
+    rspec (2.14.1)
+      rspec-core (~> 2.14.0)
+      rspec-expectations (~> 2.14.0)
+      rspec-mocks (~> 2.14.0)
+    rspec-core (2.14.3)
+    rspec-expectations (2.14.0)
+      diff-lcs (>= 1.1.3, < 2.0)
+    rspec-mocks (2.14.1)
+    rubydora (1.6.5)
+      activemodel
+      activesupport
+      deprecation
+      equivalent-xml
+      fastercsv
+      hooks (~> 0.3.0)
+      mime-types
+      nokogiri
+      rest-client
+    slop (3.4.6)
+    solrizer (3.1.1)
+      activesupport
+      daemons
+      mediashelf-loggable (~> 0.4.7)
+      nokogiri
+      stomp
+      xml-simple
+    stomp (1.2.10)
+    thread_safe (0.1.0)
+      atomic
+    tzinfo (0.3.37)
+    xml-simple (1.1.2)
+    yard (0.8.6.2)
+
+PLATFORMS
+  ruby
+
+DEPENDENCIES
+  debugger
+  equivalent-xml
+  hydra-pbcore!
+  pry
+  rake
+  rdoc
+  redcarpet
+  rspec
+  yard

data/README.md

@@ -1,6 +1,6 @@
 # HydraPbcore
 
-[![Build Status](https://travis-ci.org/curationexperts/hydra-pbcore.png)](https://travis-ci.org/curationexperts/hydra-pbcore)
+[![Build Status](https://travis-ci.org/awead/hydra-pbcore.png)](https://travis-ci.org/awead/hydra-pbcore)
 
 A Hydra gem that offers PBCore datastream definitions using OM, as well as some other convenience
 methods such as inserting xml templates into existing documents and reordering your PBCore xml 
@@ -31,39 +31,46 @@ Use this with your hydra head to define a datastream in Fedora that will contain
     end
 
 Your descMetadata datastream will now have all the pbcore terms defined in HydraPbcore::Datastream::Document.
-There are also two additional datastream definitions:
 
-* HydraPbcore::Datastream::DigitalDocument
-* HydraPbcore::Datastream::Instantiation
+The HydraPbcore::Datastream::Instantiation datastream contains additional PBcore terms pertaining to 
+pbcoreInstantiation nodes.  HydraPbcore supports an atomistic Fedora model where there are multiple
+Fedora objects representing a single Pbcore document.  This supports a use case where multiple video
+files are represented as individual Fedora objects with the Instantiation datastream.
 
-DigitalDocument datastreams have the exact same terms as Document, except that there is no default physical 
-pbcoreInstantiation.  The Document datastreams comes with a single instantiation that represents a tape
-or other physical entity.  DigitalDocument datastreams assume born-digital content and must have added
-instantiations.  These instantiations are defined in the Instantiation datastream.
+### Atomistic Fedora Model
 
-### Additional Methods
+A parent pbcoreDocument object, with multiple instantiations, can be defined:
 
-HydraPbcore comes with a couple of additional features such as ordering your xml nodes so that your xml will 
-validate against the PBCore XML v.2 schema.  Additionally, there are several template methods that can be used
-to insert additional terms into your xml documents, such as contributors, publishers, as well as next and previous
-fields that specify which files come before and after one another in a multi-part born-digital video.
+    MyPbcoreDocument < ActiveFedora::Base
+      has_many :instantiations, :property => :is_part_of
+      has_metadata :name => "descMetadata", :type => HydraPbcore::Datastream::Document
+    end
 
-### Dates
+Then, a child pbcoreInstantiation object with one parent pbcoreDocument is:
 
-Date fields in solr have to be in ISO8601 format, as opposed to text fields which may contain dates, but are treated differently.
-The difference is with the former, solr is treating the value of the field as an actual date, and as such, it can do time-based
-querying.  The latter, is just a string and solr would search it accordingly.  Currently, two types of date fields are provided:
-"_dt" for single-valued date fields, and "_dts" for multi-valued date fields.  Solr cannot sort on multi-valued date fields;
-otherwise, the two functionally identical.  By indexing your date fields as:
+    MyPbcoreInstantiation < ActiveFedora::Base
+      belongs_to :parent, :property => :is_part_of, :class_name => "MyPbcoreDocument"
+      has_metadata :name => "descMetadata", :type => HydraPbcore::Datastream::Instantiation
+    end
 
-    :index_as => [:displayable, :converted_date]
-    :index_as => [:displayable, :converted_multi_date]
+The objects can then be linked together using Fedora's RDF datastream via ActiveFedora's methods:
+
+    > doc = MyPbcoreDocument.new
+    > instantiation = MyPbcoreInstantiation.new
+    > doc.instantiations << instantiation
+
+### Additional Methods
 
-the content of your date fields will be converted to ISO8601 dates, and the original content will be preserved for display.
-So you could enter a date such as "2001" or "2004-10", and each will be stored for display as such, but will also be 
-indexed as a date in Solr with the value "2001-01-01T00:00:00Z" and "2004-10-01T00:00:00Z"
+Complete PBcore documents can be assembled with their related instantiations using some of HydraPbcore's 
+convenience methods, which will reorder all nodes so the resulting document can be validated
+according to the PBCore XML v.2 schema:
 
-Note: these functions will likely change with the upgrade to Solrizer 3.0.
+    > doc = HydraPbcore::Datastream::Document.new
+    > instantiation1 = HydraPbcore::Datastream::Instantiation.new
+    > instantiation2 = HydraPbcore::Datastream::Instantiation.new
+    > pbcore = doc.to_pbcore_xml(instantiation1, instantiation2)
+    > pbcore.valid_pbcore?
+    => true
 
 ## Testing
 

data/Rakefile

@@ -5,7 +5,7 @@ require 'rspec/core/rake_task'
 # Rake task for getting an irb session
 desc "Open an irb session preloaded with this library"
 task :console do
-  sh "irb -rubygems -I lib -r hydra-pbcore.rb"
+  sh "bin/pry -rubygems -I lib -e \"require 'hydra-pbcore'\""
 end
 
 RSpec::Core::RakeTask.new(:spec)

data/hydra-pbcore.gemspec

@@ -18,7 +18,7 @@ Gem::Specification.new do |gem|
 
   # Dependencies
   gem.add_dependency('active-fedora')
-  gem.add_dependency('solrizer', '< 3.0.0')
+  gem.add_dependency('solrizer')
   gem.add_development_dependency('yard')
   gem.add_development_dependency('redcarpet')
   # For Development
@@ -26,4 +26,6 @@ Gem::Specification.new do |gem|
   gem.add_development_dependency 'rspec'
   gem.add_development_dependency 'rdoc'
   gem.add_development_dependency 'equivalent-xml'
+  gem.add_development_dependency 'debugger'
+  gem.add_development_dependency 'pry'
 end

data/lib/hydra-pbcore.rb

@@ -3,6 +3,7 @@ require "nokogiri"
 require "solrizer"
 require "om"
 require "active-fedora"
+require "yaml"
 
 module HydraPbcore
   extend ActiveSupport::Autoload
@@ -56,6 +57,24 @@ module HydraPbcore
     "instantiationExtension",
   ]
 
+  Config = {
+    "institution" => "Rock and Roll Hall of Fame and Museum",
+    "relator"     => "MARC relator terms",
+    "address"     => "Rock and Roll Hall of Fame and Museum,\n2809 Woodland Ave.,\nCleveland, OH, 44115\n216-515-1956\nlibrary@rockhall.org"
+  }
+
+  # load our config file
+  def self.config opts = {}
+    if File.exists? "config/pbcore.yml"
+      config = ::YAML::load(IO.read("config/pbcore.yml"))
+      Config.keys.collect { |k| config[k] = Config[k] if config[k].nil? }
+    else
+      logger.info "pbcore.yml file not found, using default values"
+      config = Config
+    end
+    return config
+  end
+
   def self.version
     HydraPbcore::VERSION
   end
@@ -74,9 +93,7 @@ module HydraPbcore
     xsd.validate(xml)
   end
 
-  autoload :Mapper
   autoload :Methods
-  autoload :Conversions
   autoload :Behaviors
   autoload :Templates
   autoload :Datastream

data/lib/hydra_pbcore/datastream.rb

@@ -2,5 +2,4 @@ module HydraPbcore::Datastream
   extend ActiveSupport::Autoload
   autoload :Document
   autoload :Instantiation
-  autoload :Deprecated
 end

data/lib/hydra_pbcore/datastream/document.rb

@@ -1,22 +1,17 @@
 module HydraPbcore::Datastream
 class Document < ActiveFedora::OmDatastream
 
-  class_attribute :institution, :relator
-  self.institution = "Rock and Roll Hall of Fame and Museum"
-  self.relator     = "MARC relator terms"
-
   include HydraPbcore::Methods
   include HydraPbcore::Templates
-  include HydraPbcore::Conversions
 
   set_terminology do |t|
     t.root(:path=>"pbcoreDescriptionDocument")
 
     t.pbc_id(:path=>"pbcoreIdentifier", 
-      :attributes=>{ :source=>self.institution, :annotation=>"PID" }
+      :attributes=>{ :source=>HydraPbcore.config["institution"], :annotation=>"PID" }
     )
 
-    t.title(:path=>"pbcoreTitle", :attributes=>{ :titleType=>"Main" }, :index_as => [:searchable, :displayable])
+    t.title(:path=>"pbcoreTitle", :attributes=>{ :titleType=>"Main" }, :index_as => [:searchable, :displayable, :sortable])
     t.alternative_title(:path=>"pbcoreTitle", :attributes=>{ :titleType=>"Alternative" },
       :index_as => [:searchable, :displayable]
     )
@@ -46,7 +41,7 @@ class Document < ActiveFedora::OmDatastream
       :index_as => [:displayable]
     )
     t.rh_subject(:path=>"pbcoreSubject", 
-      :attributes=>{ :source=>self.institution },
+      :attributes=>{ :source=>HydraPbcore.config["institution"] },
       :index_as => [:displayable]
     )
 
@@ -120,7 +115,7 @@ class Document < ActiveFedora::OmDatastream
     )
     t.event_date(:path=>"pbcoreCoverage/coverage", 
       :attributes => {:annotation=>"Event Date"},
-      :index_as => [:not_searchable, :converted_multi_date, :displayable]
+      :index_as => [:dateable, :displayable]
     )
 
     # Creator names and roles
@@ -134,7 +129,7 @@ class Document < ActiveFedora::OmDatastream
     # Contributor names and roles
     t.contributor(:path=>"pbcoreContributor") do
       t.name_(:path=>"contributor")
-      t.role_(:path=>"contributorRole", :attributes=>{ :source=>self.relator })
+      t.role_(:path=>"contributorRole", :attributes=>{ :source=>HydraPbcore.config["relator"] })
     end
     t.contributor_name(:ref=>[:contributor, :name], :index_as => [:searchable, :facetable])
     t.contributor_role(:ref=>[:contributor, :role], :index_as => [:searchable, :displayable])
@@ -162,7 +157,7 @@ class Document < ActiveFedora::OmDatastream
       xml.pbcoreDescriptionDocument("xmlns:xsi"=>"http://www.w3.org/2001/XMLSchema-instance",
         "xsi:schemaLocation"=>"http://www.pbcore.org/PBCore/PBCoreNamespace.html") {
 
-        xml.pbcoreIdentifier(:source=>self.institution, :annotation=>"PID")
+        xml.pbcoreIdentifier(:source=>HydraPbcore.config["institution"], :annotation=>"PID")
         xml.pbcoreTitle(:titleType=>"Main")
         xml.pbcoreDescription(:descriptionType=>"Description",
           :descriptionTypeSource=>"pbcoreDescription/descriptionType",

data/lib/hydra_pbcore/datastream/instantiation.rb

@@ -1,13 +1,8 @@
 module HydraPbcore::Datastream
 class Instantiation < ActiveFedora::OmDatastream
 
-  class_attribute :institution, :address
-  self.institution = "Rock and Roll Hall of Fame and Museum"
-  self.address     = "Rock and Roll Hall of Fame and Museum,\n2809 Woodland Ave.,\nCleveland, OH, 44115\n216-515-1956\nlibrary@rockhall.org"
-
   include HydraPbcore::Methods
   include HydraPbcore::Templates
-  include HydraPbcore::Conversions
 
   # Note: this is not a complete PBCore document, just an instantiation node
   set_terminology do |t|
@@ -17,7 +12,7 @@ class Instantiation < ActiveFedora::OmDatastream
       :path => "instantiationIdentifier",
       :attributes=>{
         :annotation=>"Filename",
-        :source=>self.institution
+        :source=>HydraPbcore.config["institution"]
       },
       :index_as => [:displayable]
     )
@@ -25,12 +20,12 @@ class Instantiation < ActiveFedora::OmDatastream
       :path => "instantiationIdentifier",
       :attributes=>{
         :annotation=>"Barcode",
-        :source=>self.institution
+        :source=>HydraPbcore.config["institution"]
       },
-      :index_as => [:displayable]
+      :index_as => [:searchable, :displayable]
     )
     t.instantiationDate(:attributes=>{ :dateType=>"created" },
-      :index_as => [:not_searchable, :converted_date, :displayable]
+      :index_as => [:dateable, :displayable]
     )
     t.instantiationDigital(:attributes=>{ :source=>"EBU file formats" },
       :index_as => [:searchable, :facetable]
@@ -94,7 +89,7 @@ class Instantiation < ActiveFedora::OmDatastream
     )
 
     t.instantiationRelation do
-      t.instantiationRelationIdentifier(:attributes=>{ :source=>self.institution })
+      t.instantiationRelationIdentifier(:attributes=>{ :source=>HydraPbcore.config["institution"] })
     end
     # The file we're describing at the root of this document preceeds the file marked "next"
     t.next_inst(:ref => [:pbcoreInstantiation, :instantiationRelation],

data/lib/hydra_pbcore/methods.rb

@@ -73,10 +73,4 @@ module HydraPbcore::Methods
     end
   end
 
-  # Overrides Solrizer::XML::TerminologyBasedSolrizer.to_solr to use our own mapper
-  def to_solr(solr_doc = Hash.new)
-    Solrizer.default_field_mapper = HydraPbcore::Mapper.new
-    super(solr_doc)
-  end
-
 end

data/lib/hydra_pbcore/templates.rb

@@ -21,7 +21,7 @@ module HydraPbcore::Templates
       define_template :contributor do |xml, author, role|
         xml.pbcoreContributor {
           xml.contributor(author)
-          xml.contributorRole(role, :source=>self.relator) 
+          xml.contributorRole(role, :source=>HydraPbcore.config["relator"]) 
         }
       end
 
@@ -30,7 +30,7 @@ module HydraPbcore::Templates
           xml.instantiationRelationType(:annotation=>"One of a multi-part instantiation") {
             xml.text "Follows in Sequence"
           }
-          xml.instantiationRelationIdentifier(file, :source=>self.institution)
+          xml.instantiationRelationIdentifier(file, :source=>HydraPbcore.config["institution"])
         }
       end
 
@@ -39,7 +39,7 @@ module HydraPbcore::Templates
           xml.instantiationRelationType(:annotation=>"One of a multi-part instantiation") {
             xml.text "Precedes in Sequence"
           }
-          xml.instantiationRelationIdentifier(file, :source=>self.institution)
+          xml.instantiationRelationIdentifier(file, :source=>HydraPbcore.config["institution"])
         }
       end
 
@@ -108,7 +108,7 @@ module HydraPbcore::Templates
 
       xml.pbcoreInstantiation {
 
-        xml.instantiationIdentifier(:annotation=>"Filename", :source=>self.institution)
+        xml.instantiationIdentifier(:annotation=>"Filename", :source=>HydraPbcore.config["institution"])
         xml.instantiationDate(:dateType=>"created")
         xml.instantiationDigital(:source=>"EBU file formats")
         xml.instantiationLocation
@@ -163,12 +163,12 @@ module HydraPbcore::Templates
       xml.pbcoreInstantiation {
 
         # Item details
-        xml.instantiationIdentifier(:annotation=>"Barcode", :source=>self.institution)
+        xml.instantiationIdentifier(:annotation=>"Barcode", :source=>HydraPbcore.config["institution"])
         xml.instantiationDate(:dateType=>"created")
         xml.instantiationPhysical(:source=>"PBCore instantiationPhysical")
         xml.instantiationStandard
         xml.instantiationLocation {
-          xml.text self.address
+          xml.text HydraPbcore.config["address"]
         }
         xml.instantiationMediaType(:source=>"PBCore instantiationMediaType") {
           xml.text "Moving image"