Class: Document

Inherits:
Kithe::Work
  • Object
show all
Includes:
ActiveModel::Validations, AttrJson::Record::QueryScopes
Defined in:
app/models/document.rb,
app/models/document/reference.rb,
app/models/document/bbox_validator.rb,
app/models/document/date_validator.rb,
app/models/document/geom_validator.rb,
app/models/document/controlled_lists.rb,
app/models/document/date_range_validator.rb

Overview

Date Range Validation

Allow: YYYY-YYYY, *-YYYY, YYYY-*, Start YYYY == End YYYY Disallow: YYYX-YYYY, YYYY-, 2000-1999, YYYY-YYYY?

Defined Under Namespace

Classes: BboxValidator, ControlledLists, DateRangeValidator, DateValidator, GeomValidator, Reference

Instance Attribute Summary collapse

Instance Method Summary collapse

Instance Attribute Details

#skip_callbacks ⇒ Object

Returns the value of attribute skip_callbacks.



15
16
17
 
# File 'app/models/document.rb', line 15

def skip_callbacks
  @skip_callbacks
end

Instance Method Details

#a_downloadable_resource? ⇒ Boolean

Downloadable Resouce

Returns:

  • (Boolean) 


96
97
98
 
# File 'app/models/document.rb', line 96

def a_downloadable_resource?
  distributions_json.include?("downloadUrl")
end

#access_json ⇒ Object 



421
422
423
424
425
 
# File 'app/models/document.rb', line 421

def access_json
  access = {}
  access_urls.each { |au| access[au.institution_code] = au.access_url }
  access.to_json
end

#access_urls ⇒ Object

Institutional Access URLs



499
500
501
 
# File 'app/models/document.rb', line 499

def access_urls
  DocumentAccess.where(friendlier_id: friendlier_id).order(institution_code: :asc)
end

#apply_assets(distributions) ⇒ Object 



233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
 
# File 'app/models/document.rb', line 233

def apply_assets(distributions)
  # Distributable Document Assets
  # - Via DocumentAssets (Assets)
  # - With Downloadable URI
  if distributable_assets.present?
    distributable_assets.each do |asset|
      if asset.dct_references_uri_key == "download"
        distributions["http://schema.org/downloadUrl"] ||= []
        distributions["http://schema.org/downloadUrl"] << asset.to_aardvark_reference["http://schema.org/downloadUrl"]
      else
        distributions.merge!(asset.to_aardvark_reference)
      end
    end
  end

  distributions
end

#apply_downloads(distributions) ⇒ Object

BEFORE - Apply Downloads @TODO: Remove this once we’ve migrated to DocumentDistributions

  1. Native Aardvark Downloads

  2. Multiple Document Download Links

  3. Downloadable Document Assets



256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
 
# File 'app/models/document.rb', line 256

def apply_downloads(distributions)
  multiple_downloads = []

  dct_downloads = distributions["http://schema.org/downloadUrl"]

  logger.debug("Document#dct_downloads > init: #{dct_downloads}\n\n")

  # Native Aardvark Downloads
  # - Via CSV Import or via the webform
  if dct_downloads.present?
    dct_downloads.each do |download|
      multiple_downloads << {label: download_text(send(GeoblacklightAdmin::Schema.instance.solr_fields[:format])),
                            url: download}
    end
  end

  logger.debug("Document#multiple_downloads > aardvark: #{multiple_downloads.inspect}\n\n")

  # Multiple Document Download Links
  # - Via DocumentDownloads
  if document_downloads.present?
    multiple_downloads_array.each do |download|
      multiple_downloads << download
    end
  end

  logger.debug("Document#dct_downloads > document_downloads: #{multiple_downloads.inspect}\n\n")

  multiple_downloads = multiple_downloads.uniq { |d| [d[:label], d[:url]] } unless multiple_downloads.empty?

  distributions["http://schema.org/downloadUrl"] = multiple_downloads.flatten unless multiple_downloads.empty?
  distributions
end

#asset_label(asset) ⇒ Object 



225
226
227
228
229
230
231
 
# File 'app/models/document.rb', line 225

def asset_label(asset)
  if asset.label.present?
    asset.label
  else
    asset.title
  end
end

#available? ⇒ Boolean

GBL SolrDocument convience methods

Returns:

  • (Boolean) 


321
322
323
 
# File 'app/models/document.rb', line 321

def available?
  public? || same_institution?
end

#checked_endpoint(type) ⇒ Object

Provides a convenience method to access a SolrDocument’s References endpoint url without having to check and see if it is available :type => a string which if its a Geoblacklight::Constants::URI key

will return a coresponding Geoblacklight::Reference



409
410
411
412
 
# File 'app/models/document.rb', line 409

def checked_endpoint(type)
  type = distributions.send(type)
  type.endpoint if type.present?
end

#created_at_dt ⇒ Object 



427
428
429
 
# File 'app/models/document.rb', line 427

def created_at_dt
  created_at&.utc&.iso8601
end

#current_version ⇒ Object 



493
494
495
496
 
# File 'app/models/document.rb', line 493

def current_version
  # Will return 0 if no PaperTrail version exists yet
  versions&.last&.index || 0
end

#data_dictionary_download ⇒ Object 



375
376
377
 
# File 'app/models/document.rb', line 375

def data_dictionary_download
  distributions.data_dictionary.to_hash if distributions.data_dictionary.present?
end

#date_range_json ⇒ Object 



440
441
442
443
444
445
446
447
448
449
450
451
 
# File 'app/models/document.rb', line 440

def date_range_json
  date_ranges = []
  unless send(GeoblacklightAdmin::Schema.instance.solr_fields[:date_range]).all?(&:blank?)
    send(GeoblacklightAdmin::Schema.instance.solr_fields[:date_range]).each do |date_range|
      start_d, end_d = date_range.split("-")
      start_d = "*" if start_d == "YYYY" || start_d.nil?
      end_d = "*" if end_d == "YYYY" || end_d.nil?
      date_ranges << "[#{start_d} TO #{end_d}]" if start_d.present?
    end
  end
  date_ranges
end

#dct_references_s_to_csv(key, destination) ⇒ Object 



485
486
487
488
489
490
491
 
# File 'app/models/document.rb', line 485

def dct_references_s_to_csv(key, destination)
  send(destination).detect do |ref|
    ref.category == GeoblacklightAdmin::Schema.instance.dct_references_mappings[key]
  end.value
rescue NoMethodError
  nil
end

#dct_title_s ⇒ Object

Ensures a manually created “title” makes it into the attr_json “title”



436
437
438
 
# File 'app/models/document.rb', line 436

def dct_title_s
  title
end

#derive_dcat_bbox ⇒ Object

Convert GEOM for Solr Indexing



536
537
538
539
540
541
542
543
544
 
# File 'app/models/document.rb', line 536

def derive_dcat_bbox
  if send(GeoblacklightAdmin::Schema.instance.solr_fields[:bounding_box]).present?
    # "W,S,E,N" convert to "ENVELOPE(W,E,N,S)"
    w, s, e, n = send(GeoblacklightAdmin::Schema.instance.solr_fields[:bounding_box]).split(",")
    "ENVELOPE(#{w},#{e},#{n},#{s})"
  else
    ""
  end
end

#derive_dcat_centroid ⇒ Object 



546
547
548
549
550
551
552
553
 
# File 'app/models/document.rb', line 546

def derive_dcat_centroid
  if send(GeoblacklightAdmin::Schema.instance.solr_fields[:bounding_box]).present?
    w, s, e, n = send(GeoblacklightAdmin::Schema.instance.solr_fields[:bounding_box]).split(",")
    "#{(n.to_f + s.to_f) / 2},#{(e.to_f + w.to_f) / 2}"
  else
    ""
  end
end

#derive_locn_geometry ⇒ Object 



503
504
505
506
507
508
509
510
511
 
# File 'app/models/document.rb', line 503

def derive_locn_geometry
  if send(GeoblacklightAdmin::Schema.instance.solr_fields[:geometry]).present?
    send(GeoblacklightAdmin::Schema.instance.solr_fields[:geometry])
  elsif send(GeoblacklightAdmin::Schema.instance.solr_fields[:bounding_box]).present?
    derive_polygon
  else
    ""
  end
end

#derive_polygon ⇒ Object

Convert BBOX to GEOM Polygon



514
515
516
517
518
519
520
521
522
523
524
525
526
527
 
# File 'app/models/document.rb', line 514

def derive_polygon
  if send(GeoblacklightAdmin::Schema.instance.solr_fields[:bounding_box]).present?
    # Guard against a whole world polygons
    if send(GeoblacklightAdmin::Schema.instance.solr_fields[:bounding_box]) == "-180,-90,180,90"
      "ENVELOPE(-180,180,90,-90)"
    else
      # "W,S,E,N" convert to "POLYGON((W N, E N, E S, W S, W N))"
      w, s, e, n = send(GeoblacklightAdmin::Schema.instance.solr_fields[:bounding_box]).split(",")
      "POLYGON((#{w} #{n}, #{e} #{n}, #{e} #{s}, #{w} #{s}, #{w} #{n}))"
    end
  else
    ""
  end
end

#direct_download ⇒ Object 



350
351
352
 
# File 'app/models/document.rb', line 350

def direct_download
  distributions.download.to_hash if distributions.download.present?
end

#display_note ⇒ Object 



354
355
356
 
# File 'app/models/document.rb', line 354

def display_note
  send(Settings.FIELDS.DISPLAY_NOTE) || ""
end

#distributable_assets ⇒ Object 



54
55
56
 
# File 'app/models/document.rb', line 54

def distributable_assets
  document_assets.select { |a| a.dct_references_uri_key.present? }
end

#distributions ⇒ Object

Distributions

  • BEFORE DocumentDistributions

    • Use dct_references_s

    • Use multiple downloads

    • Use distributable assets

  • AFTER DocumentDistributions

    • Use document_distributions

    • Use distributable assets

@TODO: Remove BEFORE path once we’ve migrated to DocumentDistributions



129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
 
# File 'app/models/document.rb', line 129

def distributions
  distributions = {}

  logger.debug("Document#distributions > ENV['GBL_ADMIN_REFERENCES_MIGRATED']: #{ENV["GBL_ADMIN_REFERENCES_MIGRATED"]}")

  # AFTER - Add DocumentDistributions to distributions
  if ENV["GBL_ADMIN_REFERENCES_MIGRATED"] == "true"
    distributions = document_distributions.to_aardvark_distributions

    logger.debug("Document#distributions > document_distributions: #{distributions}")
  else
    # BEFORE - Prep value arrays
    # @TODO: Remove this once we've migrated to DocumentDistributions
    send(GeoblacklightAdmin::Schema.instance.solr_fields[:reference]).each do |ref|
      if ref.category.present?
        distributions[Document::Reference::REFERENCE_VALUES[ref.category.to_sym][:uri]] = []
      end
    end

    # BEFORE - Seed value arrays
    # @TODO: Remove this once we've migrated to DocumentDistributions
    send(GeoblacklightAdmin::Schema.instance.solr_fields[:reference]).each do |ref|
      if ref.category.present?
        distributions[Document::Reference::REFERENCE_VALUES[ref.category.to_sym][:uri]] << ref.value
      end
    end
    logger.debug("\n\nDocument#distributions > seeded: #{distributions}")

    # BEFORE - Apply Multiple Downloads
    # @TODO: Remove this once we've migrated to DocumentDistributions
    distributions = apply_downloads(distributions)
    logger.debug("Document#distributions > downloads: #{distributions}")
  end

  # BEFORE & AFTER - Apply Distributable Assets
  distributions = apply_assets(distributions)
  logger.debug("Document#distributions > assets: #{distributions}")

  # Need to flatten the arrays here to avoid the following potential error:
  # - ArgumentError: Please use symbols for polymorphic route arguments.
  # - Via: app/helpers/geoblacklight_helper.rb:224:in `render_references_url'
  distributions.each do |key, value|
    next if key == "http://schema.org/downloadUrl"
    if value.is_a?(Array) && value.length == 1
      distributions[key] = value.first
    end
  end

  distributions
end

#distributions_csv ⇒ Object 



196
197
198
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
 
# File 'app/models/document.rb', line 196

def distributions_csv
  # Initialize CSV
  # - [document_id, category, value, label]
  csv = []

  distributions.each do |key, value|
    if key == "http://schema.org/downloadUrl" || key == :"http://schema.org/downloadUrl"
      value.each do |download|
        logger.debug("Document#distributions_csv > download: #{download.inspect}")

        csv << [
          friendlier_id,
          ReferenceType.find_by(reference_uri: key).name,
          download["url"],
          download["label"]
        ]
      end
    else
      csv << [
        friendlier_id,
        ReferenceType.find_by(reference_uri: key)&.name,
        value,
        nil
      ]
    end
  end
  csv
end

#distributions_json ⇒ Object

Distributions JSON

  • Indexes to Solr as dct_distributions_s



182
183
184
185
186
187
188
189
190
191
192
193
194
 
# File 'app/models/document.rb', line 182

def distributions_json
  if ENV["GBL_ADMIN_REFERENCES_MIGRATED"] == "true"
    logger.debug("Document#distributions_json > using document_distributions")
    distributions = document_distributions.to_aardvark_distributions
    distributions = apply_assets(distributions)
    distributions.to_json
  else
    logger.debug("Document#distributions > #{distributions.inspect}")
    logger.debug("Document#distributions_json > using distributions")
    logger.warn("Deprecation warning: AttrJSON-based dct_references_s will not be supported soon.")
    self.distributions.to_json
  end
end

#document_assets ⇒ Object

DocumentAssets - Thumbnails, Attachments, etc @TODO: Redundant? Kithe also includes a members association



48
49
50
51
52
 
# File 'app/models/document.rb', line 48

def document_assets
  scope = Kithe::Asset
  scope = scope.where(parent_id: id).order(position: :asc)
  scope.includes(:parent)
end

#download_text(format) ⇒ Object

Wraps download text with proper_case_format



311
312
313
314
315
316
 
# File 'app/models/document.rb', line 311

def download_text(format)
  download_format = proper_case_format(format)
  download_format.to_s.html_safe
rescue
  format.to_s.html_safe
end

#downloadable? ⇒ Boolean

Returns:

  • (Boolean) 


346
347
348
 
# File 'app/models/document.rb', line 346

def downloadable?
  (direct_download || download_types.present? || iiif_download) && available?
end

#external_url ⇒ Object 



379
380
381
 
# File 'app/models/document.rb', line 379

def external_url
  distributions.url&.endpoint
end

#file_format ⇒ Object 



400
401
402
 
# File 'app/models/document.rb', line 400

def file_format
  send(Settings.FIELDS.FORMAT) || ""
end

#gbl_mdModified_dt ⇒ Object 



431
432
433
 
# File 'app/models/document.rb', line 431

def gbl_mdModified_dt
  updated_at&.utc&.iso8601
end

#geom_field ⇒ Object 



387
388
389
 
# File 'app/models/document.rb', line 387

def geom_field
  send(Settings.FIELDS.GEOMETRY) || ""
end

#geometry ⇒ Object 



391
392
393
394
 
# File 'app/models/document.rb', line 391

def geometry
  # @TODO
  # @geometry ||= Geoblacklight::Geometry.new(geom_field)
end

#hgl_download ⇒ Object 



358
359
360
 
# File 'app/models/document.rb', line 358

def hgl_download
  distributions.hgl.to_hash if distributions.hgl.present?
end

#iiif_download ⇒ Object 



371
372
373
 
# File 'app/models/document.rb', line 371

def iiif_download
  distributions.iiif.to_hash if distributions.iiif.present?
end

#iso_language_mapping ⇒ Object

Convert three char language code to proper string



556
557
558
559
560
561
562
563
564
565
 
# File 'app/models/document.rb', line 556

def iso_language_mapping
  mapping = []

  if send(GeoblacklightAdmin::Schema.instance.solr_fields[:language]).present?
    send(GeoblacklightAdmin::Schema.instance.solr_fields[:language]).each do |lang|
      mapping << GeoblacklightAdmin::IsoLanguageCodes.call[lang]
    end
  end
  mapping
end

#item_viewer ⇒ Object 



11
12
13
 
# File 'app/models/document.rb', line 11

def item_viewer
  GeoblacklightAdmin::ItemViewer.new(distributions)
end

#itemtype ⇒ Object 



383
384
385
 
# File 'app/models/document.rb', line 383

def itemtype
  "http://schema.org/Dataset"
end

#local? ⇒ Boolean

Returns:

  • (Boolean) 


333
334
335
336
 
# File 'app/models/document.rb', line 333

def local?
  local = send(Settings.FIELDS.PROVIDER) || ""
  local.casecmp(Settings.INSTITUTION)&.zero?
end

#local_restricted? ⇒ Boolean

Returns:

  • (Boolean) 


329
330
331
 
# File 'app/models/document.rb', line 329

def local_restricted?
  local? && restricted?
end

#multiple_downloads_array ⇒ Object

BEFORE - Multiple Downloads Array @TODO: Remove this once we’ve migrated to DocumentDistributions



292
293
294
 
# File 'app/models/document.rb', line 292

def multiple_downloads_array
  document_downloads.collect { |d| {label: d.label, url: d.value} }
end

#oembed ⇒ Object 



362
363
364
 
# File 'app/models/document.rb', line 362

def oembed
  distributions.oembed.endpoint if distributions.oembed.present?
end

#proper_case_format(format) ⇒ Object

From GBL

Looks up properly formatted names for formats



300
301
302
303
304
305
306
 
# File 'app/models/document.rb', line 300

def proper_case_format(format)
  if I18n.exists?("geoblacklight.formats.#{format.to_s.parameterize(separator: "_")}")
    I18n.t("geoblacklight.formats.#{format.to_s.parameterize(separator: "_")}")
  else
    format
  end
end

#public? ⇒ Boolean

Returns:

  • (Boolean) 


325
326
327
 
# File 'app/models/document.rb', line 325

def public?
  rights_field_data.present? && rights_field_data.casecmp("public").zero?
end

#raw_solr_document ⇒ Object 



77
78
79
 
# File 'app/models/document.rb', line 77

def raw_solr_document
  Blacklight.default_index.connection.get("select", {params: {q: "id:\"#{geomg_id_s}\""}})["response"]["docs"][0]
end

#restricted? ⇒ Boolean

Returns:

  • (Boolean) 


338
339
340
 
# File 'app/models/document.rb', line 338

def restricted?
  rights_field_data.blank? || rights_field_data.casecmp("restricted").zero?
end

#rights_field_data ⇒ Object 



342
343
344
 
# File 'app/models/document.rb', line 342

def rights_field_data
  send(Settings.FIELDS.ACCESS_RIGHTS) || ""
end

#same_institution? ⇒ Boolean

Returns:

  • (Boolean) 


366
367
368
369
 
# File 'app/models/document.rb', line 366

def same_institution?
  institution = send(Settings.FIELDS.PROVIDER) || ""
  institution.casecmp(Settings.INSTITUTION.downcase).zero?
end

#set_geometry ⇒ Object 



529
530
531
532
533
 
# File 'app/models/document.rb', line 529

def set_geometry
  return unless locn_geometry.blank? && self&.dcat_bbox&.present?

  self.locn_geometry = derive_polygon
end

#solr_year_json ⇒ Object Also known as: gbl_indexYear_im 



453
454
455
456
457
458
 
# File 'app/models/document.rb', line 453

def solr_year_json
  return [] if send(GeoblacklightAdmin::Schema.instance.solr_fields[:date_range]).blank?

  start_d, _end_d = send(GeoblacklightAdmin::Schema.instance.solr_fields[:date_range]).first.split("-")
  [start_d] if start_d.presence
end

#state_machine ⇒ Object

@TODO: Rename this to publication_state_machine



64
65
66
 
# File 'app/models/document.rb', line 64

def state_machine
  @state_machine ||= DocumentStateMachine.new(self, transition_class: DocumentTransition)
end

#thumbnail ⇒ Object

Thumbnail is a special case of document_assets



417
418
419
 
# File 'app/models/document.rb', line 417

def thumbnail
  members.find { |m| m.respond_to?(:thumbnail) && m.thumbnail? }
end

#thumbnail_state_machine ⇒ Object 



73
74
75
 
# File 'app/models/document.rb', line 73

def thumbnail_state_machine
  @thumbnail_state_machine ||= DocumentThumbnailStateMachine.new(self, transition_class: DocumentThumbnailTransition)
end

#to_csv ⇒ Object

Export Transformations - to_*



462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
 
# File 'app/models/document.rb', line 462

def to_csv
  attributes = GeoblacklightAdmin::Schema.instance.exportable_fields
  attributes.map do |key, value|
    if value[:delimited]
      send(value[:destination])&.join("|")
    elsif value[:destination] == "dct_references_s"
      # @TODO: Downloads need to be handled differently
      # - Need to support multiple entries per key here
      # - Need to respect label and url
      dct_references_s_to_csv(key, value[:destination])
    elsif value[:destination] == "b1g_publication_state_s"
      send(:current_state)
    else
      next if send(value[:destination]).blank?
      send(value[:destination])
    end
  end
end

#to_traject ⇒ Object 



481
482
483
 
# File 'app/models/document.rb', line 481

def to_traject
  Kithe::Model.find_by_friendlier_id(friendlier_id).update_index(writer: Traject::DebugWriter.new({}))
end

#wxs_identifier ⇒ Object 



396
397
398
 
# File 'app/models/document.rb', line 396

def wxs_identifier
  send(Settings.FIELDS.WXS_IDENTIFIER) || ""
end