cul_hydra 1.0.0

data/app/models/dc_document.rb

@@ -0,0 +1,39 @@
+require "active-fedora"
+require "active_fedora_finders"
+class DcDocument < ActiveFedora::Base
+  include ::ActiveFedora::Associations
+  include ::ActiveFedora::FinderMethods::RepositoryMethods
+  include ::ActiveFedora::DatastreamCollections
+  include Hydra::ModelMethods
+  include Cul::Hydra::Models::Common
+  #alias :file_objects :resources
+
+  has_and_belongs_to_many :parts, :property => :cul_member_of, :class_name=>'ActiveFedora::Base'
+
+  def self.load_instance_from_solr(pid,solr_doc=nil)
+    if solr_doc.nil?
+      result = find_by_solr(pid)
+      raise "Object #{pid} not found in solr" if result.nil?
+      solr_doc = result.hits.first
+      #double check pid and id in record match
+      raise "Object #{pid} not found in Solr" unless !result.nil? && !solr_doc.nil? && pid == solr_doc[SOLR_DOCUMENT_ID]
+    else
+      raise "Solr document record id and pid do not match" unless pid == solr_doc[SOLR_DOCUMENT_ID]
+    end
+
+    create_date = solr_doc[::ActiveFedora::SolrService.solr_name(:system_create, :date)].nil? ? solr_doc[::ActiveFedora::SolrService.solr_name(:system_create, :date).to_s] : solr_doc[::ActiveFedora::SolrService.solr_name(:system_create, :date)]
+    modified_date = solr_doc[::ActiveFedora::SolrService.solr_name(:system_create, :date)].nil? ? solr_doc[::ActiveFedora::SolrService.solr_name(:system_modified, :date).to_s] : solr_doc[::ActiveFedora::SolrService.solr_name(:system_modified, :date)]
+    obj = DcDocument.new({:pid=>solr_doc[SOLR_DOCUMENT_ID],:create_date=>create_date,:modified_date=>modified_date})
+    obj.new_object = false
+      #set by default to load any dependent relationship objects from solr as well
+    obj.load_from_solr = true
+      #need to call rels_ext once so it exists when iterating over datastreams
+    obj.rels_ext
+    obj.datastreams.each_value do |ds|
+      if ds.respond_to?(:from_solr)
+        ds.from_solr(solr_doc) if ds.kind_of?(::ActiveFedora::MetadataDatastream) || ds.kind_of?(::ActiveFedora::NokogiriDatastream) || ( ds.kind_of?(::ActiveFedora::RelsExtDatastream))
+      end
+    end
+    obj
+  end
+end

data/app/models/generic_aggregator.rb

@@ -0,0 +1,68 @@
+require "active-fedora"
+require "active_fedora_finders"
+class GenericAggregator < ::ActiveFedora::Base
+  include ::ActiveFedora::FinderMethods::RepositoryMethods
+  include ::ActiveFedora::DatastreamCollections
+  include Hydra::ModelMethods
+  include Cul::Hydra::Models::Common
+  include Cul::Hydra::Models::Aggregator
+
+  has_many :parts, :property => :cul_member_of, :class_name=>'ActiveFedora::Base'
+
+  def route_as
+    "multipartitem"
+  end
+
+  def index_type_label
+    riquery = Cul::Hydra::Models::MEMBER_ITQL.gsub(/%PID%/, self.pid)
+    begin
+      docs = Cul::Hydra::Fedora.repository.find_by_itql riquery, limit: 2, format: 'count/json'
+      docs = JSON.parse(docs)['results']
+      size = docs.first && docs.first['count'] && docs.first['count'].to_i
+    rescue Exception=>e
+      Rails.logger.warn("#{self.class.name} failed to find children with TQL: #{e.message}")
+    end
+    size ||= self.parts(response_format: :solr, limit: 2).size
+    type_label_for(size)
+  end
+
+  def type_label_for(size=nil)
+    if size == 0
+      return "EMPTY"
+    elsif size == 1
+      return "SINGLE PART"
+    else
+      return "MULTIPART"
+    end
+  end
+
+  # set the index type label and any RI-based fields
+  # overridde
+  def set_size_labels(solr_doc={})
+    count = Cul::Hydra::RisearchMembers.get_direct_member_count(pid)
+    solr_doc["index_type_label_ssi"] = [type_label_for(count)]
+    solr_doc["cul_number_of_members_isi"] = count
+  end
+
+  def to_solr(solr_doc = Hash.new, opts={})
+    solr_doc = super(solr_doc, opts)
+    solr_doc
+  end
+
+  def proxies
+    datastreams['structMetadata'] ? datastreams['structMetadata'].proxies : []
+  end
+  def update_index
+    super
+    if has_struct_metadata?()
+      conn = ActiveFedora::SolrService.instance.conn
+      # delete by query proxyIn_ssi: internal_uri
+      conn.delete_by_query("proxyIn_ssi:#{RSolr.escape(internal_uri())}")
+
+      # reindex proxies
+      proxy_docs = proxies().collect {|p| p.to_solr}
+      conn.add(proxy_docs, params: {softCommit: true})
+      conn.commit
+    end
+  end
+end

data/app/models/generic_object.rb

@@ -0,0 +1,18 @@
+require "active-fedora"
+require "active_fedora_finders"
+class GenericObject < ::ActiveFedora::Base
+  extend ActiveModel::Callbacks
+  include ::ActiveFedora::FinderMethods::RepositoryMethods
+  include ::ActiveFedora::DatastreamCollections
+  include ::Hydra::ModelMethods
+  include Cul::Hydra::Models::Common
+  include Cul::Hydra::Models::Aggregator
+
+  has_many :parts, :property => :cul_member_of, :class_name=>'ActiveFedora::Base'
+
+  #alias :file_objects :resources
+
+  def route_as
+    "multipartitem"
+  end
+end

data/app/models/generic_resource.rb

@@ -0,0 +1,210 @@
+require "cul_image_props"
+require "mime/types"
+require "uri"
+require "open-uri"
+require "tempfile"
+require "active_fedora_finders"
+class GenericResource < ::ActiveFedora::Base
+  extend ActiveModel::Callbacks
+  include ::ActiveFedora::FinderMethods::RepositoryMethods
+  include ::ActiveFedora::DatastreamCollections
+  include ::Hydra::ModelMethods
+  include Cul::Hydra::Models::Common
+  include Cul::Hydra::Models::ImageResource
+  include Cul::Hydra::Fedora::UrlHelperBehavior
+  include ::ActiveFedora::RelsInt
+
+  has_and_belongs_to_many :containers, :property=>:cul_member_of, :class_name=>'ActiveFedora::Base'
+
+  IMAGE_EXT = {"image/bmp" => 'bmp', "image/gif" => 'gif', "image/jpeg" => 'jpg', "image/png" => 'png', "image/tiff" => 'tif', "image/x-windows-bmp" => 'bmp'}
+  WIDTH = RDF::URI(ActiveFedora::Predicates.find_graph_predicate(:image_width))
+  LENGTH = RDF::URI(ActiveFedora::Predicates.find_graph_predicate(:image_length))
+
+  has_datastream :name => "content", :type=>::ActiveFedora::Datastream, :versionable => true
+  has_metadata :name=>"RELS-INT", :type=>ActiveFedora::RelsInt::Datastream
+
+  def assert_content_model
+    super
+    add_relationship(:rdf_type, RDF::CUL.Resource.to_s)
+  end
+
+  def route_as
+    "resource"
+  end
+
+  def index_type_label
+    "FILE ASSET"
+  end
+
+  def thumbnail_info
+    thumb = relsint.relationships(datastreams['content'],:foaf_thumb).first
+    if thumb
+      t_dsid = thumb.object.to_s.split('/')[-1]
+      return {:url=>"#{ActiveFedora.fedora_config[:url]}/objects/#{pid}/datastreams/#{t_dsid}/content",:mime=>datastreams[t_dsid].mimeType}
+    else
+      return {:url=>image_url("cul_hydra/crystal/file.png"),:mime=>'image/png'}
+    end
+  end
+
+  def to_solr(solr_doc = Hash.new, opts={})
+    solr_doc = super
+
+    unless solr_doc["extent_ssim"] || self.datastreams["content"].nil?
+      if self.datastreams["content"].dsSize.to_i > 0
+        solr_doc["extent_ssim"] = [self.datastreams["content"].dsSize]
+      else
+        repo = ActiveFedora::Base.connection_for_pid(pid)
+        ds_parms = {pid: pid, dsid: "content", method: :head}
+        repo.datastream_dissemination(ds_parms) do |res|
+          solr_doc["extent_ssim"] = res['Content-Length']
+        end
+      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
+
+    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
+    end
+    relationships(:original_name).each do |original_name|
+      solr_doc["original_name_tesim"] ||= []
+      original_name = original_name.object.to_s.split('/').join(' ')
+      solr_doc["original_name_tesim"] << original_name.strip
+    end
+
+    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 = []
+    if (rels = rels_int.instance_variable_get :@solr_hash)
+      # this was loaded from solr
+      rels.each do |dsuri, props|
+        if dsuri =~ /\/content$/ or not props[FORMAT_OF_PREDICATE].blank?
+          dsid =  dsuri.split('/')[-1]
+          puts props.inspect
+          res = datastream_as_resource(dsid, props.with_indifferent_access)
+          results << res
+        end
+      end
+    else
+      content_uri = RDF::URI("info:fedora/#{self.pid}/content")
+      dsuris = [content_uri]
+      results = []
+      # read the graph
+      datastreams.each do |k, v|
+        rels = rels_int.relationships(v, :format_of, content_uri)
+        dsuris << rels[0].subject unless rels.blank?
+      end
+
+      dsuris.each do |dsuri|
+        dsid = dsuri.to_s.split('/')[-1]
+        width_rel = rels_int.relationships(dsuri, :image_width)[0]
+        length_rel = rels_int.relationships(dsuri, :image_length)[0]
+        extent_rel = rels_int.relationships(dsuri, :extent)[0]
+        props = {extent: [], image_width: [], image_length: []}
+        props[:extent] << extent_rel.object.to_s unless extent_rel.blank?
+        props[:image_width] << width_rel.object.to_s unless width_rel.blank?
+        props[:image_length] << length_rel.object.to_s unless length_rel.blank?
+        results << datastream_as_resource(dsid, props)
+      end
+    end
+    results
+  end
+
+  def zooming?
+    (zr = rels_int.relationships(datastreams['content'], :foaf_zooming) and not zr.first.blank?)
+  end
+
+  def with_ds_resource(ds_id, fedora_content_filesystem_mounted=false, &block)
+
+    ds = self.datastreams[ds_id]
+
+    puts 'dsLocation.start_with?(self.pid) : ' + ds.dsLocation.start_with?(self.pid).to_s
+
+    # If the dsLocation starts with the pid, that means that we're dealing with an internally-managed ds,
+    # so we can't reference the file directly even if we do have the fedora content filesystem mounted.
+    if ! ds.dsLocation.start_with?(self.pid) && fedora_content_filesystem_mounted
+      if ds.dsLocation =~ /^file:\//
+        dsLocation = ds.dsLocation.sub(/^file:\/+/,'/')
+        path = URI.unescape(dsLocation)
+      else
+        path = ds.dsLocation
+      end
+
+      yield(path)
+    else
+      internal_uri = "info:fedora/#{self.pid}/#{ds_id}"
+      # No local fedora mount, need to download content over http[s]
+
+      file_basename = File.basename(ds.dsLocation.gsub(/^file:/,''))
+      file_extension = File.extname(file_basename)
+
+      # In some cases, we actually do want to know the original extension of the file, so we'll preserve it in the temp file filename
+      temp_file = Tempfile.new([file_basename, file_extension])
+      begin
+        parts = internal_uri.split('/')
+        open(temp_file.path, 'wb') do |blob|
+          repo = ActiveFedora::Base.connection_for_pid(parts[1])
+          repo.datastream_dissemination({:dsid=>parts[2], :pid=>parts[1], :finished=>false}) do |res|
+            res.read_body do |seg|
+              blob << seg
+            end
+          end
+        end
+        yield(temp_file.path)
+      ensure
+        temp_file.unlink
+      end
+    end
+
+  end
+
+  private
+  def datastream_as_resource(dsid, props={})
+    ds = datastreams[dsid]
+    raise "No resource at info:fedora/#{pid}/#{dsid}" unless ds
+    res = {}
+    res[:uri] = self.pid
+    res[:block] = dsid
+    res[:mime_type] = ds.mimeType
+    res[:content_models] = ["Datastream"]
+    res[:file_size] = ds.dsSize.to_s
+    if res[:file_size] == "0" and props[:extent]
+      res[:file_size] = (props[:extent].first || "0")
+    end
+    res[:size] = (res[:file_size].to_i / 1024).to_s + " Kb"
+    res[:width] = (props[:image_width] and props[:image_width].first) || "0"
+    res[:height] = (props[:image_length] and props[:image_length].first) || "0"
+    res[:dimensions] = "#{res[:width]} x #{res[:height]}"
+    base_filename = pid.gsub(/\:/,"")
+    res[:filename] = base_filename + "." + dsid + "." + ds.mimeType.gsub(/^[^\/]+\//,"")
+   res
+  end
+end

data/app/models/jp2_image_aggregator.rb

@@ -0,0 +1,34 @@
+require "active-fedora"
+require "active_fedora_finders"
+class JP2ImageAggregator < ResourceAggregator
+
+  has_datastream :name => "SOURCE", :type=>::ActiveFedora::Datastream, :mimeType=>"image/jp2", :controlGroup=>'E'
+
+  def route_as
+    "zoomingimage"
+  end
+
+  def index_type_label
+    "PART"
+  end
+
+  def to_solr(solr_doc = Hash.new, opts={})
+    solr_doc = super
+    source = self.datastreams["SOURCE"]
+    source.profile
+    if source.controlGroup == 'E'
+      solr_doc["rft_id_ss"] = source.dsLocation
+    else
+      rc = ActiveFedora::RubydoraConnection.instance.connection
+      url = rc.config["url"]
+      uri = URI::parse(url)
+      url = "#{uri.scheme}://#{uri.host}:#{uri.port}/fedora/objects/#{pid}/datastreams/SOURCE/content"
+      solr_doc["rft_id_ss"] = url
+    end
+    solr_doc
+  end
+  
+  def thumbnail_info
+    {:url => "#{ActiveFedora.fedora_config[:url]}/objects/#{pid}/methods/ldpd:sdef.Image/getView?max=250", :mime => 'image/jpeg'}
+  end
+end

data/app/models/mets_structured_aggregator.rb

@@ -0,0 +1,18 @@
+require "active-fedora"
+require "active_fedora_finders"
+class METSStructuredAggregator < ::ActiveFedora::Base
+  extend ActiveModel::Callbacks
+  include ::ActiveFedora::FinderMethods::RepositoryMethods
+  include ::ActiveFedora::DatastreamCollections
+  include ::Hydra::ModelMethods
+  include Cul::Hydra::Models::Common
+  include Cul::Hydra::Models::Aggregator
+
+  has_many :parts, :property => :cul_member_of, :class_name=>'ActiveFedora::Base'
+
+  #alias :file_objects :resources
+
+  def route_as
+    "multipartitem"
+  end
+end

data/app/models/resource.rb

@@ -0,0 +1,78 @@
+require "active-fedora"
+require "active_fedora_finders"
+require "cul_image_props"
+require "mime/types"
+require "uri"
+class Resource < ActiveFedora::Base
+  include ::ActiveFedora::FinderMethods::RepositoryMethods
+  include ::ActiveFedora::DatastreamCollections
+#  include ::Hydra::ModelMethods
+  include Cul::Hydra::Models::Common
+  include Cul::Hydra::Models::Resource
+  include Cul::Hydra::Models::ImageResource
+
+  belongs_to :container, :property=>:cul_member_of, :class_name=>'ActiveFedora::Base'
+
+#  CUL_WIDTH = "http://purl.oclc.org/NET/CUL/RESOURCE/STILLIMAGE/BASIC/imageWidth"
+#  CUL_LENGTH = "http://purl.oclc.org/NET/CUL/RESOURCE/STILLIMAGE/BASIC/imageLength"
+#  FORMAT = "http://purl.org/dc/elements/1.1/format"
+#  MEMBER_OF = "http://purl.oclc.org/NET/CUL/memberOf"
+
+  def route_as
+    "resource"
+  end
+
+  def index_type_label
+    "FILE RESOURCE"
+  end
+
+  def to_solr(solr_doc = Hash.new, opts={})
+    super
+    unless solr_doc["extent_ssi"] || self.datastreams["CONTENT"].nil?
+      solr_doc["extent_ssi"] = [self.datastreams["CONTENT"].size]
+    end
+    solr_doc
+  end
+
+  def set_title_and_label(new_title, opts={})
+    if opts[:only_if_blank]
+      if self.label.nil? || self.label.empty?
+        self.label = new_title
+        self.set_title( new_title )
+      end
+    else
+      self.label = new_title
+      set_title( new_title )
+    end
+  end
+
+  # Set the title and label on the current object
+  #
+  # @param [String] new_title
+  # @param [Hash] opts (optional) hash of configuration options
+  def set_title(new_title, opts={})
+    if has_desc? 
+      desc_metadata_ds = self.datastreams["descMetadata"]
+      if desc_metadata_ds.respond_to?(:title_values)
+        desc_metadata_ds.title_values = new_title
+      else
+        desc_metadata_ds.title = new_title
+      end
+    end
+  end
+  
+  def thumbnail_info
+    # do the triples indicate this is a thumb? fetch
+    if long <= 251
+      mime = object_relations[FORMAT].first
+      url = {:url=>"#{ActiveFedora.fedora_config[:url]}/objects/#{self.pid}/datastreams/CONTENT/content", :mime=>mime}
+    else
+      if object_relations[MEMBER_OF].blank?
+        return {:url=>image_url("cul_hydra/crystal/file.png"),:mime=>'image/png'}
+      else
+        url = ActiveFedora::Base.find(parents.first, :cast => true).thumbnail_info
+      end
+    end
+    return url
+  end
+end