Module: Dor::Describable

Extended by:
ActiveSupport::Concern
Included in:
Abstract
Defined in:
lib/dor/models/describable.rb

Defined Under Namespace

Classes: CrosswalkError

Constant Summary collapse

MODS_TO_DC_XSLT = 
Nokogiri::XSLT(File.new(File.expand_path(File.dirname(__FILE__) + "/mods2dc.xslt")))
XMLNS_OAI_DC = 
'http://www.openarchives.org/OAI/2.0/oai_dc/'.freeze
XMLNS_DC = 
'http://purl.org/dc/elements/1.1/'.freeze

Class Method Summary collapse

Instance Method Summary collapse

Class Method Details

.get_collection_title(obj) ⇒ Object 



285
286
287
 
# File 'lib/dor/models/describable.rb', line 285

def self.get_collection_title(obj)
  obj.full_title
end

Instance Method Details

#add_access_conditions(doc) ⇒ Object

Note:

this method modifies the passed in doc

Create MODS accessCondition statements from rightsMetadata

Parameters:



71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
 
# File 'lib/dor/models/describable.rb', line 71

def add_access_conditions(doc)
  # clear out any existing accessConditions
  doc.xpath('//mods:accessCondition', 'mods' => 'http://www.loc.gov/mods/v3').each {|n| n.remove}
  rights = datastreams['rightsMetadata'].ng_xml

  rights.xpath('//use/human[@type="useAndReproduction"]').each do |use|
    txt = use.text.strip
    next if txt.empty?
    doc.root.element_children.last.add_next_sibling doc.create_element('accessCondition', txt, :type => 'useAndReproduction')
  end
  rights.xpath('//copyright/human[@type="copyright"]').each do |cr|
    txt = cr.text.strip
    next if txt.empty?
    doc.root.element_children.last.add_next_sibling doc.create_element('accessCondition', txt, :type => 'copyright')
  end
  rights.xpath("//use/machine[#{ci_compare('type', 'creativecommons')}]").each do |lic_type|
    next if lic_type.text =~ /none/i
    lic_text = rights.at_xpath("//use/human[#{ci_compare('type', 'creativecommons')}]").text.strip
    next if lic_text.empty?
    new_text = "CC #{lic_type.text}: #{lic_text}"
    doc.root.element_children.last.add_next_sibling doc.create_element('accessCondition', new_text, :type => 'license')
  end
  rights.xpath("//use/machine[#{ci_compare('type', 'opendatacommons')}]").each do |lic_type|
    next if lic_type.text =~ /none/i
    lic_text = rights.at_xpath("//use/human[#{ci_compare('type', 'opendatacommons')}]").text.strip
    next if lic_text.empty?
    new_text = "ODC #{lic_type.text}: #{lic_text}"
    doc.root.element_children.last.add_next_sibling doc.create_element('accessCondition', new_text, :type => 'license')
  end
end

#add_collection_reference(doc) ⇒ Void

Note:

this method modifies the passed in doc

Adds to desc metadata a relatedItem with information about the collection this object belongs to. For use in published mods and mods-to-DC conversion.

Parameters:

Returns:

  • (Void) 


149
150
151
152
153
154
155
156
157
158
159
160
161
162
 
# File 'lib/dor/models/describable.rb', line 149

def add_collection_reference(doc)
  return unless methods.include? :public_relationships
  collections = public_relationships.search('//rdf:RDF/rdf:Description/fedora:isMemberOfCollection',
                                   'fedora' => 'info:fedora/fedora-system:def/relations-external#',
                                   'rdf' => 'http://www.w3.org/1999/02/22-rdf-syntax-ns#')
  return if collections.empty?

  remove_related_item_nodes_for_collections(doc)

  collections.each do |collection_node|
    collection_druid = collection_node['rdf:resource'].gsub('info:fedora/', '')
    add_related_item_node_for_collection(doc, collection_druid)
  end
end

#add_constituent_relations(doc) ⇒ Void

expand constituent relations into relatedItem references – see JUMBO-18

Parameters:

  • doc (Nokogiri::XML) —

    public MODS XML being built

Returns:

  • (Void) 


167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
 
# File 'lib/dor/models/describable.rb', line 167

def add_constituent_relations(doc)
  public_relationships.search('//rdf:RDF/rdf:Description/fedora:isConstituentOf',
                                   'fedora' => 'info:fedora/fedora-system:def/relations-external#',
                                   'rdf' => 'http://www.w3.org/1999/02/22-rdf-syntax-ns#' ).each do |parent|
    # fetch the parent object to get title
    druid = parent['rdf:resource'].gsub(/^info:fedora\//, '')
    parent_item = Dor.find(druid)

    # create the MODS relation
    relatedItem = doc.create_element 'relatedItem'
    relatedItem['type'] = 'host'
    relatedItem['displayLabel'] = 'Appears in'

    # load the title from the parent's DC.title
    titleInfo = doc.create_element 'titleInfo'
    title = doc.create_element 'title'
    title.content = Dor::Describable.get_collection_title(parent_item)
    titleInfo << title
    relatedItem << titleInfo

    # point to the PURL for the parent
    location = doc.create_element 'location'
    url = doc.create_element 'url'
    url.content = "http://#{Dor::Config.stacks.document_cache_host}/#{druid.split(':').last}"
    location << url
    relatedItem << location

    # finish up by adding relation to public MODS
    doc.root << relatedItem
  end
end

#add_related_item_node_for_collection(doc, collection_druid) ⇒ Object 



109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
 
# File 'lib/dor/models/describable.rb', line 109

def add_related_item_node_for_collection(doc, collection_druid)
  begin
    collection_obj = Dor.find(collection_druid)
  rescue ActiveFedora::ObjectNotFoundError
    return nil
  end

  title_node         = Nokogiri::XML::Node.new('title', doc)
  title_node.content = Dor::Describable.get_collection_title(collection_obj)

  title_info_node = Nokogiri::XML::Node.new('titleInfo', doc)
  title_info_node.add_child(title_node)

  # e.g.:
  #   <location>
  #     <url>http://purl.stanford.edu/rh056sr3313</url>
  #   </location>
  loc_node = doc.create_element('location')
  url_node = doc.create_element('url')
  url_node.content = "https://#{Dor::Config.stacks.document_cache_host}/#{collection_druid.split(':').last}"
  loc_node << url_node

  type_node = Nokogiri::XML::Node.new('typeOfResource', doc)
  type_node['collection'] = 'yes'

  related_item_node = Nokogiri::XML::Node.new('relatedItem', doc)
  related_item_node['type'] = 'host'

  related_item_node.add_child(title_info_node)
  related_item_node.add_child(loc_node)
  related_item_node.add_child(type_node)

  doc.root.add_child(related_item_node)
end

#build_descMetadata_datastream(ds) ⇒ Object 



33
34
35
36
37
38
39
40
 
# File 'lib/dor/models/describable.rb', line 33

def build_descMetadata_datastream(ds)
  content = fetch_descMetadata_datastream
  return nil if content.nil?
  ds.dsLabel = 'Descriptive Metadata'
  ds.ng_xml = Nokogiri::XML(content)
  ds.ng_xml.normalize_text!
  ds.content = ds.ng_xml.to_xml
end

#fetch_descMetadata_datastream ⇒ Object 



27
28
29
30
31
 
# File 'lib/dor/models/describable.rb', line 27

def fetch_descMetadata_datastream
  candidates = datastreams['identityMetadata'].otherId.collect { |oid| oid.to_s }
  metadata_id = Dor::MetadataService.resolvable(candidates).first
  metadata_id.nil? ? nil : Dor::MetadataService.fetch(metadata_id.to_s)
end

#full_title ⇒ Object 



289
290
291
 
# File 'lib/dor/models/describable.rb', line 289

def full_title
  stanford_mods.sw_title_display
end

#generate_dublin_core(include_collection_as_related_item: true) ⇒ Nokogiri::Doc

Generates Dublin Core from the MODS in the descMetadata datastream using the LoC mods2dc stylesheet

Should not be used for the Fedora DC datastream

Returns:

  • (Nokogiri::Doc) —

    the DublinCore XML document object

Raises:

  • (CrosswalkError) —

    Raises an Exception if the generated DC is empty or has no children



46
47
48
49
50
51
52
53
54
 
# File 'lib/dor/models/describable.rb', line 46

def generate_dublin_core(include_collection_as_related_item: true)
  desc_md = descMetadata.ng_xml.dup(1)
  add_collection_reference(desc_md) if include_collection_as_related_item
  dc_doc = MODS_TO_DC_XSLT.transform(desc_md)
  dc_doc.xpath('/oai_dc:dc/*[count(text()) = 0]', oai_dc: XMLNS_OAI_DC).remove # Remove empty nodes
  raise CrosswalkError, "Dor::Item#generate_dublin_core produced incorrect xml (no root):\n#{dc_doc.to_xml}" if dc_doc.root.nil?
  raise CrosswalkError, "Dor::Item#generate_dublin_core produced incorrect xml (no children):\n#{dc_doc.to_xml}" if dc_doc.root.children.size == 0
  dc_doc
end

#generate_public_desc_md ⇒ String

Returns Public descriptive medatada XML.

Returns:

  • (String) —

    Public descriptive medatada XML



57
58
59
60
61
62
63
64
65
66
 
# File 'lib/dor/models/describable.rb', line 57

def generate_public_desc_md
  doc = descMetadata.ng_xml.dup(1)
  add_collection_reference(doc)
  add_access_conditions(doc)
  add_constituent_relations(doc)
  doc.xpath('//comment()').remove
  new_doc = Nokogiri::XML(doc.to_xml) { |x| x.noblanks }
  new_doc.encoding = 'UTF-8'
  new_doc.to_xml
end

#remove_related_item_nodes_for_collections(doc) ⇒ Object

Remove existing relatedItem entries for collections from descMetadata



103
104
105
106
107
 
# File 'lib/dor/models/describable.rb', line 103

def remove_related_item_nodes_for_collections(doc)
  doc.search('/mods:mods/mods:relatedItem[@type="host"]/mods:typeOfResource[@collection=\'yes\']', 'mods' => 'http://www.loc.gov/mods/v3').each do |node|
    node.parent.remove
  end
end

#set_desc_metadata_using_label(force = false) ⇒ String

Returns descMetadata.content XML.

Parameters:

  • force (Boolean) (defaults to: false) —

    Overwrite existing XML

Returns:

  • (String) —

    descMetadata.content XML



268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
 
# File 'lib/dor/models/describable.rb', line 268

def set_desc_metadata_using_label(force = false)
  unless force || descMetadata.new?
    raise 'Cannot proceed, there is already content in the descriptive metadata datastream: ' + descMetadata.content.to_s
  end
  label = self.label
  builder = Nokogiri::XML::Builder.new { |xml|
    xml.mods(
      'xmlns' => 'http://www.loc.gov/mods/v3', 'xmlns:xsi' => 'http://www.w3.org/2001/XMLSchema-instance', :version => '3.3',
      'xsi:schemaLocation' => 'http://www.loc.gov/mods/v3 http://www.loc.gov/standards/mods/v3/mods-3-3.xsd') {
      xml.titleInfo {
        xml.title label
      }
    }
  }
  descMetadata.content = builder.to_xml
end

#stanford_mods(content = nil, ns_aware = true) ⇒ Object

intended for read-access, “as SearchWorks would see it”, mostly for to_solr()

Parameters:

  • content (Nokogiri::XML::Document) (defaults to: nil) —

    Nokogiri descMetadata document (overriding internal data)

  • ns_aware (boolean) (defaults to: true) —

    namespace awareness toggle for from_nk_node()



20
21
22
23
24
25
 
# File 'lib/dor/models/describable.rb', line 20

def stanford_mods(content = nil, ns_aware = true)
  m = Stanford::Mods::Record.new
  desc = content.nil? ? descMetadata.ng_xml : content
  m.from_nk_node(desc.root, ns_aware)
  m
end

#to_solr(solr_doc = {}, *args) ⇒ Object 



199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
 
# File 'lib/dor/models/describable.rb', line 199

def to_solr(solr_doc = {}, *args)
  solr_doc = super solr_doc, *args
  mods_sources = {
    'sw_display_title_tesim'      => :sw_title_display,
    'sw_author_ssim'              => :main_author_w_date,
    'sw_author_tesim'             => :main_author_w_date,
    'sw_author_sort_ssi'          => :sw_sort_author,
    'sw_language_ssim'            => :sw_language_facet,
    'sw_language_tesim'           => :sw_language_facet,
    'sw_genre_ssim'               => :sw_genre,
    'sw_genre_tesim'              => :sw_genre,
    'sw_format_ssim'              => :format_main,   # basically sw_typeOfResource_ssim
    'sw_format_tesim'             => :format_main,   # basically sw_typeOfResource_tesim
    'sw_topic_ssim'               => :topic_facet,
    'sw_topic_tesim'              => :topic_facet,
    'sw_subject_temporal_ssim'    => :era_facet,
    'sw_subject_temporal_tesim'   => :era_facet,
    'sw_subject_geographic_ssim'  => :geographic_facet,
    'sw_subject_geographic_tesim' => :geographic_facet,
    'mods_typeOfResource_ssim'    => [:term_values, :typeOfResource],
    'mods_typeOfResource_tesim'   => [:term_values, :typeOfResource]
  }
  keys = mods_sources.keys.concat(%w( metadata_format_ssim ))
  keys.each { |key|
    solr_doc[key] ||= []     # initialize multivalue targts if necessary
  }

  solr_doc['metadata_format_ssim'] << 'mods'
  begin
    dc_doc = generate_dublin_core(include_collection_as_related_item: false)
    # we excluding the generated collection relation here; we instead get the collection
    # title from Dor::Identifiable.
    dc_doc.xpath('/oai_dc:dc/*', oai_dc: XMLNS_OAI_DC).each do |node|
      add_solr_value(solr_doc, "public_dc_#{node.name}", node.text, :string, [:stored_searchable])
    end
    creator = ''
    dc_doc.xpath('//dc:creator', dc: XMLNS_DC).each do |node|
      creator = node.text
    end
    title = ''
    dc_doc.xpath('//dc:title', dc: XMLNS_DC).each do |node|
      title = node.text
    end
    creator_title = creator + title
    add_solr_value(solr_doc, 'creator_title', creator_title, :string, [:stored_sortable])
  rescue CrosswalkError => e
    Dor.logger.warn "Cannot index #{pid}.descMetadata: #{e.message}"
  end

  begin
    mods = stanford_mods
    mods_sources.each_pair do |solr_key, meth|
      vals = meth.is_a?(Array) ? mods.send(meth.shift, *meth) : mods.send(meth)
      solr_doc[solr_key].push *vals unless vals.nil? || vals.empty?
      # asterisk to avoid multi-dimensional array: push values, not the array
    end
    solr_doc['sw_pub_date_sort_ssi' ] = mods.pub_year_sort_str  # e.g. '0800'
    solr_doc['sw_pub_date_sort_isi' ] = mods.pub_year_int  # e.g. '0800'
    solr_doc['sw_pub_date_facet_ssi'] = mods.pub_year_display_str # e.g. '9th century'
  end
  # some fields get explicit "(none)" placeholder values, mostly for faceting
  %w(sw_language_tesim sw_genre_tesim sw_format_tesim).each { |key| solr_doc[key] = ['(none)'] if solr_doc[key].empty? }
  # otherwise remove empties
  keys.each { |key| solr_doc.delete(key) if solr_doc[key].nil? || solr_doc[key].empty?}
  solr_doc
end