Module: Cul::Hydra::Solrizer::ModsFieldable

Extended by:
ActiveSupport::Concern, ClassMethods
Includes:
Solrizer::DefaultDescriptors::Normal
Included in:
Datastreams::ModsDocument
Defined in:
lib/cul_hydra/solrizer/mods_fieldable.rb

Defined Under Namespace

Modules: ClassMethods

Constant Summary collapse

MODS_NS = 
{'mods'=>'http://www.loc.gov/mods/v3'}
ORIGIN_INFO_DATES = 
["dateCreated", "dateIssued", "dateOther"]

Instance Method Summary collapse

Methods included from ClassMethods

map_field, map_value, maps_field?, normalize, value_mapper

Instance Method Details

#all_subjects(node = mods) ⇒ Object 



301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 301

def all_subjects(node=mods)
  list_of_subjects = []

  node.xpath("./mods:subject[not(@authority) or @authority != 'Durst']/mods:topic", MODS_NS).collect do |n|
    list_of_subjects << ModsFieldable.normalize(n.text, true)
  end
  node.xpath("./mods:subject/mods:geographic", MODS_NS).collect do |n|
    list_of_subjects << ModsFieldable.normalize(n.text, true)
  end
  node.xpath("./mods:subject/mods:name", MODS_NS).collect do |n|
    list_of_subjects << ModsFieldable.normalize(n.text, true)
  end
  node.xpath("./mods:subject/mods:temporal", MODS_NS).collect do |n|
    list_of_subjects << ModsFieldable.normalize(n.text, true)
  end
  node.xpath("./mods:subject/mods:titleInfo", MODS_NS).collect do |n|
    list_of_subjects << ModsFieldable.normalize(n.text, true)
  end
  node.xpath("./mods:subject/mods:genre", MODS_NS).collect do |n|
    list_of_subjects << ModsFieldable.normalize(n.text, true)
  end

  return list_of_subjects
end

#alternative_titles(node = mods) ⇒ Object 



100
101
102
103
104
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 100

def alternative_titles(node=mods)
  node.xpath('./mods:titleInfo[@type and (@type="alternative" or @type="abbreviated" or @type="translated" or @type="uniform")]', MODS_NS).collect do |t|
    ModsFieldable.normalize(t.text)
  end
end

#collections ⇒ Object 



62
63
64
65
66
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 62

def collections
  mods.xpath("./mods:relatedItem[@type='host' and @displayLabel='Collection']", MODS_NS).collect do |p_node|
    ModsFieldable.normalize(main_title(p_node), true)
  end
end

#coordinates(node = mods) ⇒ Object 



356
357
358
359
360
361
362
363
364
365
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 356

def coordinates(node=mods)
  coordinate_values = []
  node.xpath("./mods:subject/mods:cartographics/mods:coordinates", MODS_NS).collect do |n|
    n = ModsFieldable.normalize(n.text, true)
    if n.match(/-*\d+\.\d+\s*,\s*-*\d+\.\d+\s*/) # Expected coordinate format: 40.123456,-73.5678
      coordinate_values << n
    end
  end
  coordinate_values
end

#date_notes(node = mods) ⇒ Object 



257
258
259
260
261
262
263
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 257

def date_notes(node=mods)
  date_notes = []
  node.xpath("./mods:note[@type = 'date' or @type = 'date source']", MODS_NS).collect do |n|
    date_notes << ModsFieldable.normalize(n.text, true)
  end
  return date_notes
end

#date_range_to_textual_date(start_year, end_year) ⇒ Object 



242
243
244
245
246
247
248
249
250
251
252
253
254
255
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 242

def date_range_to_textual_date(start_year, end_year)
  start_year = start_year.to_i.to_s # Remove zero-padding if present
  end_year = end_year.to_i.to_s # Remove zero-padding if present

  if start_year == end_year
    return [start_year]
  else
    return [('Between ' +
      (start_year.to_i > 0 ? start_year : start_year[1,start_year.length] + ' BCE') +
      ' and ' +
      (end_year.to_i > 0 ? (start_year.to_i > 0 ? end_year : end_year + ' CE') : end_year[1,end_year.length] + ' BCE')
    )]
  end
end

#dates(node = mods) ⇒ Object 



133
134
135
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 133

def dates(node=mods)
  # get all the dateIssued with keyDate = 'yes', but not point = 'end'
end

#durst_subjects(node = mods) ⇒ Object 



326
327
328
329
330
331
332
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 326

def durst_subjects(node=mods)
  list_of_subjects = []
  node.xpath("./mods:subject[@authority='Durst']/mods:topic", MODS_NS).collect do |n|
    list_of_subjects << ModsFieldable.normalize(n.text, true)
  end
  return list_of_subjects
end

#formats(node = mods) ⇒ Object 



137
138
139
140
141
142
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 137

def formats(node=mods)
  # get all the form values with authority != 'marcform'
  node.xpath("./mods:physicalDescription/mods:form[@authority != 'marcform']", MODS_NS).collect do |n|
    ModsFieldable.normalize(n.text)
  end
end

#item_in_context_url(node = mods) ⇒ Object 



277
278
279
280
281
282
283
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 277

def item_in_context_url(node=mods)
  item_in_context_url_val = []
  node.xpath("./mods:location/mods:url[@access='object in context' and @usage='primary display']", MODS_NS).collect do |n|
    item_in_context_url_val << ModsFieldable.normalize(n.text, true)
  end
  item_in_context_url_val
end

#key_date_range(node = mods) ⇒ Object 



222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 222

def key_date_range(node=mods)
  dates = []
  encodings = ['w3cdtf','iso8601']
  ORIGIN_INFO_DATES.each do |element|
    encodings.each do |encoding|
      xpath = "./mods:originInfo/mods:#{element}[(@keyDate) and (@encoding = '#{encoding}')]"
      node.xpath(xpath, MODS_NS).collect do |n|
        range = [ModsFieldable.normalize(n.text, true)]
        if n['point'] != 'end'
          n.xpath("../mods:#{element}[(@encoding = '#{encoding}' and @point = 'end')]", MODS_NS).each do |ep|
            range << ModsFieldable.normalize(ep.text, true)
          end
        end
        dates << range
      end
    end
  end
  return dates.first || dates
end

#main_title(node = mods) ⇒ Object 



82
83
84
85
86
87
88
89
90
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 82

def main_title(node=mods)
  # include only the untyped [!@type] titleInfo
  t = node.xpath('./mods:titleInfo[not(@type)]', MODS_NS).first
  if t
    ModsFieldable.normalize(t.text)
  else
    nil
  end
end

#mods ⇒ Object 



52
53
54
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 52

def mods
  ng_xml.xpath('/mods:mods', MODS_NS).first
end

#names(role_authority = nil, role = nil) ⇒ Object 



106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 106

def names(role_authority=nil, role=nil)
  # get all the name nodes
  # keep all child text except the role terms
  xpath = "./mods:name"
  unless role_authority.nil?
    xpath << "/mods:role/mods:roleTerm[@authority='#{role_authority.to_s}'"
    unless role.nil?
      xpath << " and normalize-space(text()) = '#{role.to_s.strip}'"
    end
    xpath << "]/ancestor::mods:name"
  end
  names = mods.xpath(xpath, MODS_NS).collect do |node|
    base_text = node.xpath('./mods:namePart', MODS_NS).collect { |c| c.text }.join(' ')
    ModsFieldable.normalize(base_text, true)
  end

  # Note: Removing subject names from name field extraction.
  # See: https://issues.cul.columbia.edu/browse/DCV-231 and https://issues.cul.columbia.edu/browse/SCV-102
  #xpath = "./mods:subject" + xpath[1,xpath.length]
  #mods.xpath(xpath, MODS_NS).each do |node|
  #  base_text = node.xpath('./mods:namePart', MODS_NS).collect { |c| c.text }.join(' ')
  #  names << ModsFieldable.normalize(base_text, true)
  #end

  names
end

#non_date_notes(node = mods) ⇒ Object 



265
266
267
268
269
270
271
272
273
274
275
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 265

def non_date_notes(node=mods)
  non_date_notes = []
  node.xpath("./mods:note[not(@type) or (@type != 'date' and @type != 'date source' and @type != 'filename')]", MODS_NS).collect do |n|
if n.attr('type') == 'view direction'
	non_date_notes << 'View Direction: ' + ModsFieldable.normalize(n.text, true)
else
	non_date_notes << ModsFieldable.normalize(n.text, true)
end
  end
  return non_date_notes
end

#non_item_in_context_url(node = mods) ⇒ Object 



285
286
287
288
289
290
291
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 285

def non_item_in_context_url(node=mods)
	non_item_in_context_url_val = []
    node.xpath("./mods:location/mods:url[not(@access='object in context')]", MODS_NS).collect do |n|
      non_item_in_context_url_val << ModsFieldable.normalize(n.text, true)
    end
    non_item_in_context_url_val
end

#origin_info_place(node = mods) ⇒ Object 



334
335
336
337
338
339
340
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 334

def origin_info_place(node=mods)
  places = []
  node.xpath("./mods:originInfo/mods:place/mods:placeTerm", MODS_NS).collect do |n|
    places << ModsFieldable.normalize(n.text, true)
  end
  return places
end

#origin_info_place_for_display(node = mods) ⇒ Object 



342
343
344
345
346
347
348
349
350
351
352
353
354
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 342

def origin_info_place_for_display(node=mods)
  # If there are multiple origin_info place elements, choose only the ones without valueURI attributes.  Otherwise show the others.
  places_with_uri = []
  places_without_uri = []
  node.xpath("./mods:originInfo/mods:place/mods:placeTerm[@valueURI]", MODS_NS).collect do |n|
    places_with_uri << ModsFieldable.normalize(n.text, true)
  end
  node.xpath("./mods:originInfo/mods:place/mods:placeTerm[not(@valueURI)]", MODS_NS).collect do |n|
    places_without_uri << ModsFieldable.normalize(n.text, true)
  end

  return (places_without_uri.length > 0 ? places_without_uri : places_with_uri)
end

#project_url(node = mods) ⇒ Object 



293
294
295
296
297
298
299
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 293

def project_url(node=mods)
  project_url_val = []
  node.xpath("./mods:relatedItem[@type='host' and @displayLabel='Project']/mods:location/mods:url", MODS_NS).collect do |n|
    project_url_val << ModsFieldable.normalize(n.text, true)
  end
  project_url_val
end

#projects ⇒ Object 



56
57
58
59
60
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 56

def projects
  mods.xpath("./mods:relatedItem[@type='host' and @displayLabel='Project']", MODS_NS).collect do |p_node|
    ModsFieldable.normalize(main_title(p_node), true)
  end
end

#repository_code(node = mods) ⇒ Object 



144
145
146
147
148
149
150
151
152
153
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 144

def repository_code(node=mods)
  # get the location/physicalLocation[@authority = 'marcorg']
  repo_code_node = node.xpath("./mods:location/mods:physicalLocation[@authority = 'marcorg']", MODS_NS).first

  if repo_code_node
    ModsFieldable.normalize(repo_code_node.text)
  else
    return nil
  end
end

#repository_text(node = mods) ⇒ Object 



155
156
157
158
159
160
161
162
163
164
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 155

def repository_text(node=mods)
  # get the location/physicalLocation[not(@authority)]
  repo_text_node = node.xpath("./mods:location/mods:physicalLocation[not(@authority)]", MODS_NS).first

  if repo_text_node
    ModsFieldable.normalize(repo_text_node.text)
  else
    return nil
  end
end

#shelf_locators(node = mods) ⇒ Object 



192
193
194
195
196
197
198
199
200
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 192

def shelf_locators(node=mods)
  values = node.xpath("./mods:location/mods:shelfLocator", MODS_NS).collect do |n|
    ModsFieldable.normalize(n.text, true)
  end
  values += node.xpath("./mods:location/mods:holdingSimple/mods:copyInformation/mods:shelfLocator", MODS_NS).collect do |n|
    ModsFieldable.normalize(n.text, true)
  end
  values
end

#sort_title(node = mods) ⇒ Object 



68
69
70
71
72
73
74
75
76
77
78
79
80
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 68

def sort_title(node=mods)
  # include only the untyped [!@type] titleInfo, exclude noSort
  base_text = ''
  t = node.xpath('./mods:titleInfo[not(@type)]', MODS_NS).first
  if t
    t.children.each do |child|
      base_text << child.text unless child.name == 'nonSort'
    end
  end
  base_text = ModsFieldable.normalize(base_text, true)
  base_text = nil if base_text.empty?
  base_text
end

#sublocation(node = mods) ⇒ Object 



202
203
204
205
206
207
208
209
210
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 202

def sublocation(node=mods)
  values = node.xpath("./mods:location/mods:sublocation", MODS_NS).collect do |n|
    ModsFieldable.normalize(n.text, true)
  end
  values += node.xpath("./mods:location/mods:holdingSimple/mods:copyInformation/mods:sublocation", MODS_NS).collect do |n|
    ModsFieldable.normalize(n.text, true)
  end
  values
end

#textual_dates(node = mods) ⇒ Object 



212
213
214
215
216
217
218
219
220
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 212

def textual_dates(node=mods)
  dates = []
  ORIGIN_INFO_DATES.each do |element|
    node.xpath("./mods:originInfo/mods:#{element}[not(@keyDate) and not(@point) and not(@encoding)]", MODS_NS).collect do |n|
      dates << ModsFieldable.normalize(n.text, true)
    end
  end
  return dates
end

#titles(node = mods) ⇒ Object 



92
93
94
95
96
97
98
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 92

def titles(node=mods)
  # all titles without descending into relatedItems
  # For now, this only includes the main title and selected alternate_titles
  all_titles = []
  all_titles << main_title unless main_title.nil?
  all_titles += alternative_titles unless alternative_titles.nil?
end

#to_solr(solr_doc = {}) ⇒ Object 



367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 367

def to_solr(solr_doc={})
  solr_doc = (defined? super) ? super : solr_doc

  return solr_doc if mods.nil? 	# There is no mods.  Return because there is nothing to process, otherwise NoMethodError will be raised by subsequent lines.

  solr_doc["all_text_teim"] ||= []

  solr_doc["title_si"] = sort_title
  solr_doc["title_ssm"] = titles
  solr_doc["alternative_title_ssm"] = alternative_titles
  solr_doc["all_text_teim"] += solr_doc["alternative_title_ssm"]
  solr_doc["lib_collection_sim"] = collections
  solr_doc["lib_name_sim"] = names
  solr_doc["lib_name_teim"] = solr_doc["lib_name_sim"]
  solr_doc["all_text_teim"] += solr_doc["lib_name_teim"]
  solr_doc["lib_all_subjects_ssm"] = all_subjects
  solr_doc["durst_subjects_ssim"] = durst_subjects
  solr_doc["lib_all_subjects_teim"] = solr_doc["lib_all_subjects_ssm"]
  solr_doc["all_text_teim"] += solr_doc["lib_all_subjects_teim"]
  solr_doc["lib_name_ssm"] = solr_doc["lib_name_sim"]
  solr_doc["lib_author_sim"] = names(:marcrelator, 'aut')
  solr_doc["lib_recipient_sim"] = names(:marcrelator, 'rcp')
  solr_doc["lib_format_sim"] = formats
  solr_doc["lib_shelf_sim"] = shelf_locators
  solr_doc['location_shelf_locator_ssm'] = solr_doc["lib_shelf_sim"]
  solr_doc["all_text_teim"] += solr_doc["lib_shelf_sim"]
  solr_doc['lib_sublocation_sim'] = sublocation
  solr_doc['lib_sublocation_ssm'] = solr_doc['lib_sublocation_sim']
  solr_doc["all_text_teim"] += solr_doc['lib_sublocation_sim']
  solr_doc["lib_date_textual_ssm"] = textual_dates
  solr_doc["lib_date_notes_ssm"] = date_notes
  solr_doc["lib_non_date_notes_ssm"] = non_date_notes
  solr_doc["lib_item_in_context_url_ssm"] = item_in_context_url
  solr_doc["lib_non_item_in_context_url_ssm"] = non_item_in_context_url
  solr_doc["lib_project_url_ssm"] = project_url
  solr_doc["origin_info_place_ssm"] = origin_info_place
  solr_doc["origin_info_place_for_display_ssm"] = origin_info_place_for_display

  repo_marc_code = repository_code
  unless repo_marc_code.nil?
    solr_doc["lib_repo_short_ssim"] = [translate_repo_marc_code(repo_marc_code, 'short')]
    solr_doc["lib_repo_long_sim"] = [translate_repo_marc_code(repo_marc_code, 'long')]
    solr_doc["lib_repo_full_ssim"] = [translate_repo_marc_code(repo_marc_code, 'full')]
  end
  solr_doc["lib_repo_text_ssm"] = repository_text

  project_titles = projects
  unless project_titles.nil?
    solr_doc["lib_project_short_ssim"] = []
    solr_doc["lib_project_full_ssim"] = []
    project_titles.each {|project_title|
      solr_doc["lib_project_short_ssim"] << translate_project_title(project_title, 'short')
      solr_doc["lib_project_full_ssim"] << translate_project_title(project_title, 'full')
    }
    solr_doc["lib_project_short_ssim"].uniq!
    solr_doc["lib_project_full_ssim"].uniq!
  end

  # Create convenient start and end date values based on one of the many possible originInfo/dateX elements.
  start_date, end_date = key_date_range
  start_year = nil
  end_year = nil

  if start_date.present?

start_year = nil
end_year = nil

start_date = nil if start_date == 'uuuu'
end_date = nil if end_date == 'uuuu'
start_date = start_date.gsub('u', '0') unless start_date.nil?
end_date = end_date.gsub('u', '0') unless end_date.nil?

    end_date = start_date if end_date.blank?
    start_date = end_date if start_date.blank?

    year_regex = /^(-?\d{1,4}).*/

unless start_date.blank?
	start_year_match = start_date.match(year_regex)
	if start_year_match && start_year_match.captures.length > 0
		start_year = start_year_match.captures[0]
		start_year = zero_pad_year(start_year)
		solr_doc["lib_start_date_year_itsi"] = start_year.to_i # TrieInt version for searches
	end
end

unless end_date.blank?
	end_year_match = end_date.match(year_regex)
	if end_year_match && end_year_match.captures.length > 0
		end_year = end_year_match.captures[0]
		end_year = zero_pad_year(end_year)
		solr_doc["lib_end_date_year_itsi"] = end_year.to_i # TrieInt version for searches
	end
end

    solr_doc["lib_date_year_range_si"] = start_year + '-' + end_year if start_year && end_year
    solr_doc["lib_date_year_range_ss"] = solr_doc["lib_date_year_range_si"]

    # When no textual date is available, fall back to other date data (if available)
    if solr_doc["lib_date_textual_ssm"].blank?
      solr_doc["lib_date_textual_ssm"] = date_range_to_textual_date(start_year.to_i, end_year.to_i)
    end
  end

  # Geo data
  solr_doc["geo"] = coordinates

  solr_doc.each do |k, v|
    if self.class.maps_field? k
      solr_doc[k] = self.class.map_value(k, v)
    end
  end

  solr_doc
end

#translate_project_title(project_title, type) ⇒ Object 



180
181
182
183
184
185
186
187
188
189
190
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 180

def translate_project_title(project_title, type)
  normalized_project_title = ModsFieldable.normalize(project_title)

  if type == 'short'
    return translate_with_default(SHORT_PROJ, normalized_project_title, normalized_project_title)
  elsif type == 'full'
    return translate_with_default(FULL_PROJ, normalized_project_title, normalized_project_title)
  end

  return nil
end

#translate_repo_marc_code(code, type) ⇒ Object 



166
167
168
169
170
171
172
173
174
175
176
177
178
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 166

def translate_repo_marc_code(code, type)
  #code = ModsFieldable.normalize(code)

  if type == 'short'
    return translate_with_default(SHORT_REPO, code, 'Non-Columbia Location')
  elsif type == 'long'
    return translate_with_default(LONG_REPO, code, 'Non-Columbia Location')
  elsif type == 'full'
    return translate_with_default(FULL_REPO, code, 'Non-Columbia Location')
  end

  return nil
end

#zero_pad_year(year) ⇒ Object 



484
485
486
487
488
489
490
491
492
493
 
# File 'lib/cul_hydra/solrizer/mods_fieldable.rb', line 484

def zero_pad_year(year)
  year = year.to_s
  is_negative = year.start_with?('-')
  year_without_sign = (is_negative ? year[1, year.length]: year)
  if year_without_sign.length < 4
    year_without_sign = year_without_sign.rjust(4, '0')
  end

  return (is_negative ? '-' : '') + year_without_sign
end