Class: BulkOps::Parser

Inherits:
Object
  • Object
show all
Defined in:
lib/bulk_ops/parser.rb

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(prx, metadata_sheet = nil) ⇒ Parser

Returns a new instance of Parser.



25
26
27
28
29
30
31
 
# File 'lib/bulk_ops/parser.rb', line 25

def initialize prx, metadata_sheet=nil
  @proxy = prx
  @raw_data = (metadata_sheet || proxy.operation.metadata)
  @raw_row = @raw_data[@proxy.row_number]
  @metadata = {}
  @parsing_errors = []
end

Instance Attribute Details

#proxy ⇒ Object

Returns the value of attribute proxy.



4
5
6
 
# File 'lib/bulk_ops/parser.rb', line 4

def proxy
  @proxy
end

#raw_data ⇒ Object

Returns the value of attribute raw_data.



4
5
6
 
# File 'lib/bulk_ops/parser.rb', line 4

def raw_data
  @raw_data
end

#raw_row ⇒ Object

Returns the value of attribute raw_row.



4
5
6
 
# File 'lib/bulk_ops/parser.rb', line 4

def raw_row
  @raw_row
end

Class Method Details

.is_file_set?(metadata, row_number) ⇒ Boolean

Returns:

  • (Boolean) 


8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
 
# File 'lib/bulk_ops/parser.rb', line 8

def self.is_file_set? metadata, row_number
  return false unless metadata[row_number].present?
  # If the work type is explicitly specified, use that
  if (type_key = metadata[row_number].to_h.keys.find{|key| key.downcase.gsub(/[_\-\s]/,"").include?("worktype") })
    return true if metadata[row_number][type_key].downcase == "fileset" 
    return false if metadata[row_number][type_key].present?
  end
#    Otherwise, if there are any valid fields other than relationship or file fields, call it a work
  metadata[row_number].each do |field, value|
    next if BulkOps::Verification.is_file_field?(field)
    next if ["parent", "order"].include?(normalize_relationship_field_name(field))
    next if ["title","label"].include?(field.downcase.strip)
    return false
  end
  return true
end

.normalize_relationship_field_name(field) ⇒ Object 



320
321
322
323
 
# File 'lib/bulk_ops/parser.rb', line 320

def self.normalize_relationship_field_name field
  normfield = field.downcase.parameterize.gsub(/[_\s-]/,'')
  BulkOps::RELATIONSHIP_FIELDS.find{|rel_field| normfield == rel_field }
end

Instance Method Details

#connect_existing_work ⇒ Object 



66
67
68
69
70
71
72
 
# File 'lib/bulk_ops/parser.rb', line 66

def connect_existing_work
  return unless (column_name = operation.options["update_identifier"])
  return unless (key = @raw_row.keys.find{|key| key.to_s.parameterize.downcase.gsub("_","") == column_name.to_s.parameterize.downcase.gsub("_","")})
  return unless (value = @raw_row[key])
  return unless (work_id = find_work_id_from_unique_metadata(key, value)) 
  proxy.update(work_id: work_id)
end

#delete_file_set(fileset_id) ⇒ Object 



463
464
465
 
# File 'lib/bulk_ops/parser.rb', line 463

def delete_file_set fileset_id
  BulkOps::DeleteFileSetJob.perform_later(fileset_id, operation.user.email )
end

#disambiguate_columns ⇒ Object 



53
54
55
56
57
58
59
60
61
62
63
64
 
# File 'lib/bulk_ops/parser.rb', line 53

def disambiguate_columns
  #do nothing unless there are columns with the same header
  return unless (@raw_row.respond_to?(:headers) && (@raw_row.headers.uniq.length < @raw_row.length) )
  row = {}
  (0...@raw_row.length).each do |i|
    header = @raw_row.headers[i]
    value = @raw_row[i]
    # separate values in identical columns using the separator
    row[header] = (Array(row[header]) << value).join(BulkOps::SEPARATOR)
  end
  @raw_row = row
end

#downcase_first_letter(str) ⇒ Object 



508
509
510
511
 
# File 'lib/bulk_ops/parser.rb', line 508

def downcase_first_letter(str)
  return "" unless str
  str[0].downcase + str[1..-1]
end

#find_collection(collection) ⇒ Object 



477
478
479
480
481
482
 
# File 'lib/bulk_ops/parser.rb', line 477

def find_collection(collection)
  cols = Collection.where(id: collection)
  cols += Collection.where(title: collection).select{|col| col.title.first == collection}
  return cols.last unless cols.empty?
  return false
end

#find_field_name(field) ⇒ Object 



504
505
506
 
# File 'lib/bulk_ops/parser.rb', line 504

def find_field_name(field)
  operation.find_field_name(field)
end

#find_or_create_collection(collection) ⇒ Object 



484
485
486
487
488
489
 
# File 'lib/bulk_ops/parser.rb', line 484

def find_or_create_collection(collection)
  col = find_collection(collection)
  return col if col
  return false if collection.to_i > 0
  col = Collection.create(title: [collection.to_s], depositor: operation.user.email, collection_type: Hyrax::CollectionType.find_by(title:"User Collection"))
end

#find_previous_parent(field = "parent") ⇒ Object 



325
326
327
328
329
330
331
332
333
 
# File 'lib/bulk_ops/parser.rb', line 325

def find_previous_parent field="parent"
  #Return the row number of the most recent preceding row that does
  # not itself have a parent defined
  i = 1;
  while (prev_row = raw_data[row_number - i])
    return (row_number - i) if prev_row[field].blank?
    i += 1
  end
end

#find_work_id_from_unique_metadata(field_name, value) ⇒ Object 



74
75
76
77
78
79
80
81
82
 
# File 'lib/bulk_ops/parser.rb', line 74

def find_work_id_from_unique_metadata field_name, value
  field_solr_name = schema.get_field(field_name).solr_name
  query = "_query_:\"{!dismax qf=#{field_solr_name}}#{value}\""
  response = ActiveFedora::SolrService.instance.conn.get(ActiveFedora::SolrService.select_path, params: { fq: query, rows: 1, start: 0})["response"]
  if response["numFound"] > 1
    report_error( :id_not_unique , "",  row_number: row_number, object_id: @proxy.id, options_name: field_name, option_values: value ) unless label
  end
  return response["docs"][0]["id"]
end

#findAuthUrl(auth, value) ⇒ Object 



396
397
398
399
400
401
402
403
404
405
406
407
408
 
# File 'lib/bulk_ops/parser.rb', line 396

def findAuthUrl(auth, value)
  value.strip!
  return nil if auth.nil?
  return nil unless (entries = Qa::Authorities::Local.subauthority_for(auth).search(value))
  entries.each do |entry|
    #require exact match
    next unless entry["label"].force_encoding('UTF-8') == value.force_encoding('UTF-8')
    url = entry["url"] || entry["id"]
#      url = localIdToUrl(url,auth) unless url =~ URI::regexp
    return url
  end
  return nil
end

#format_param_name(name) ⇒ Object 



496
497
498
 
# File 'lib/bulk_ops/parser.rb', line 496

def format_param_name(name)
  name.titleize.gsub(/\s+/, "").camelcase(:lower)
end

#format_visibility(value) ⇒ Object 



374
375
376
377
378
379
380
381
382
383
 
# File 'lib/bulk_ops/parser.rb', line 374

def format_visibility(value)
  case value.downcase
  when "public", "open", "true"
    return "open"
  when "campus", "ucsc", "institution"
    return "ucsc"
  when "restricted", "private", "closed", "false"
    return "restricted"
  end
end

#format_worktype(value) ⇒ Object 



365
366
367
368
369
370
371
372
 
# File 'lib/bulk_ops/parser.rb', line 365

def format_worktype(value)
  # format the value like a class name
  type = value.titleize.gsub(/[-_\s]/,'')
  # reject it if it isn't a defined class
  type = false unless Object.const_defined? type
  # fall back to the work type defined by the operation, or a standard "Work"
  return type ||= work_type || operation.work_type || "Work"
end

#get_remote_id(value, authority: nil, property: nil) ⇒ Object 



491
492
493
494
 
# File 'lib/bulk_ops/parser.rb', line 491

def get_remote_id(value, authority: nil, property: nil)
  return false
  #TODO retrieve URL for this value from the specified remote authr
end

#get_removed_filesets(filestring) ⇒ Object 



448
449
450
451
452
453
454
455
456
457
458
459
460
461
 
# File 'lib/bulk_ops/parser.rb', line 448

def get_removed_filesets(filestring)
  file_ids = split_values(filestring)
  file_ids.select{|file_id| record_exists?(file_id)}

# This part handles filenames in addition to file ids. It doesn't work yet!
#    file_ids.map do |file_id| 
    # If the filename is the id of an existing record, keep that
#      next(file_id) if (record_exists?(file_id))
    # If this is the label (i.e.filename) of an existing fileset, use that fileset id
    # TODO MAKE THIS WORK!!
#      next(filename) if (filename_exists?(filename))
#      File.join(BulkOps::INGEST_MEDIA_PATH, filename_prefix, filename)
#    end
end

#getLocalAuth(field_name) ⇒ Object 



418
419
420
421
422
423
424
425
426
427
 
# File 'lib/bulk_ops/parser.rb', line 418

def getLocalAuth(field_name)
  field =  schema.get_property(field_name)
  # There is only ever one local authority per field, so just pick the first you find
  if vocs = field.vocabularies
    vocs.each do |voc|
      return voc["subauthority"] if voc["authority"].downcase == "local"
    end
  end
  return nil
end

#interpret_controlled_fields ⇒ Object 



84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
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
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
 
# File 'lib/bulk_ops/parser.rb', line 84

def interpret_controlled_fields 
  
  # The labels array tracks the contents of columns marked as labels,
  # which may require special validation
  labels = {}

  # This hash is populated with relevant data as we loop through the fields
  controlled_data = {}

  @raw_row.each do |field_name, value| 
    next if value.blank?  or field_name.blank?
    field_name = field_name.to_s

    #If our CSV interpreter is feeding us the headers as a line, ignore it.
    next if field_name == value

    #check if they are using the 'field_name.authority' syntax
    authority = nil
    if ((split=field_name.split('.')).count == 2)
      authority = split.last 
      field_name = split.first
    end

    # get the field name, if this column is a metadata field
    field_name_norm = find_field_name(field_name)
    field = schema.get_field(field_name_norm)

    # Ignore anything that isn't a controlled field
    next unless field.present? && field.controlled?

    # Keep track of label fields
    if field_name.downcase.ends_with?("label")
      next if operation.options["ignore_labels"]  
      labels[field_name_norm] ||= []
      labels[field_name_norm] += split_values value
      next unless operation.options["import_labels"]
    end

    remove = field_name.downcase.starts_with?("remove") || field_name.downcase.starts_with?("delete")
    
    # handle multiple values
    value_array = split_values(value)
    controlled_data[field_name_norm] ||= [] unless value_array.blank?
    value_array.each do |value|
      # Decide of we're dealing with a label or url
      # It's an ID if it's a URL and the name doesn't end in 'label'
      value.strip!
      if value =~ /^#{URI::regexp}$/ and !field_name.downcase.ends_with?("label")
        value_id = value
      #          label = WorkIndexer.fetch_remote_label(value)
      #          error_message =  "cannot fetch remote label for url: #{value}"
      #          report_error( :cannot_retrieve_label , error_message, url: value, row_number: row_number) unless label
      else
        # It's a label, so unescape it and get the id
        value = unescape_csv(value)
        value_id = get_remote_id(value, property: field_name_norm, authority: authority) || localAuthUrl(field_name_norm, value)
        #          label = value
        report_error(:cannot_retrieve_url, 
                     message: "cannot find or create url for controlled vocabulary label: #{value}", 
                     url: value, 
                     row_number: row_number) unless value_id
      end
      controlled_data[field_name_norm] << {id: value_id, remove: field_name.downcase.starts_with?("remove")}
    end
  end

  # Actually add all the data
  controlled_data.each do |property_name, data|
    @metadata["#{property_name}_attributes"] ||= [] unless data.blank?
    data.uniq.each do |datum| 
      atts = {"id" => datum[:id]}
      atts["_delete"] = true if datum[:remove]
      @metadata["#{property_name}_attributes"] << atts
    end
  end
end

#interpret_data(raw_row: nil, raw_data: nil, proxy: nil) ⇒ Object 



33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
 
# File 'lib/bulk_ops/parser.rb', line 33

def interpret_data raw_row: nil, raw_data: nil, proxy: nil
  @raw_row = raw_row if raw_row.present?
  @proxy = proxy if proxy.present?
  @raw_data = raw_data if raw_data.present?
  disambiguate_columns
  setAdminSet
  #The order here matters a little: interpreting the relationship fields specifies containing collections,
  # which may have opinions about whether we should inherit metadata from parent works
  interpret_relationship_fields
  setMetadataInheritance
  interpret_option_fields
  interpret_file_fields
  interpret_controlled_fields
  interpret_scalar_fields
  connect_existing_work 
  @proxy.update(status: "ERROR", message: "error parsing spreadsheet line") if @parsing_errors.present?
  @proxy.proxy_errors = (@proxy.proxy_errors || []) + @parsing_errors
  return @metadata
end

#interpret_file_fields ⇒ Object 



178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
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
224
225
226
227
228
229
230
231
 
# File 'lib/bulk_ops/parser.rb', line 178

def interpret_file_fields
  # This method handles file additions and deletions from the spreadsheet
  # if additional files need to be deleted because the update is set to replace
  # some or all existing files, those replacement-related deletions are handled
  # by the BulkOps::Operation.
  #

  @raw_row.each do |field, value|
    next if value.blank?  or field.blank?
    field = field.to_s
    #If our CSV interpreter is feeding us the headers as a line, ignore it.
    next if field == value

    # Check if this is a file field, and whether we are removing or adding a file
    next unless (action = BulkOps::Verification.is_file_field?(field))
    
    # Move on if this field is the name of another property (e.g. masterFilename)
    next if find_field_name(field)
    
    # Check if we are removing a file
    if action == "remove"
      get_removed_filesets(value).each { |fileset_id| delete_file_set(file_set_id) } 
    else
      # Add a file
      operation.get_file_paths(value).each do |filepath|
        begin
          uploaded_file = Hyrax::UploadedFile.create(file:  File.open(filepath), user: operation.user)
          (@metadata[:uploaded_files] ||= []) << uploaded_file.id unless uploaded_file.id.nil?
        rescue Exception => e  
          report_error(:upload_error,
                       message: "Error opening file: #{ filepath } -- #{e}",
                       file: File.join(BulkOps::INGEST_MEDIA_PATH,filename),
                       row_number: row_number)
        end
      end
    end

    # Check if any of the upcoming rows are child filesets
    i = 1
    while self.class.is_file_set?(@metadata,row_number+i)
      child_row.each do |field,value|
        next if value.blank?
        title = value if ["title","label"].include?(field.downcase.strip)
        if BulkOps::Verification.is_file_field?(field)
          operation.get_file_paths(value).each do |filepath|
            uploaded_file = Hyrax::UploadedFile.create(file:  File.open(filepath), user: operation.user)
          end
        end
      end
      i+=1
    end

  end
end

#interpret_option_fields ⇒ Object 



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
 
# File 'lib/bulk_ops/parser.rb', line 233

def interpret_option_fields
  @raw_row.each do |field,value|
    next if value.blank? or field.blank?
    field = field.to_s
    next if value == field

    normfield = field.downcase.parameterize.gsub(/[_\s-]/,'')
    if ["visibility", "public"].include?(normfield)
      @proxy.update(visibility: format_visibility(value))

    end
    if ["worktype","model","type"].include?(normfield)
      @proxy.update(work_type: format_worktype(value) )
    end
    if ["referenceidentifier", 
        "referenceid", 
        "refid",
        "referenceidentifiertype", 
        "referenceidtype", 
        "refidtype", 
        "relationshipidentifier",
        "relationshipid",
        "relationshipidentifiertype",
        "relationshipidtype",
        "relid",
        "relidtype"].include?(normfield)
      @proxy.update(reference_identifier: format_reference_id(value))
    end
  end
end

#interpret_relationship_fields ⇒ Object 



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
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
 
# File 'lib/bulk_ops/parser.rb', line 264

def interpret_relationship_fields
  @raw_row.each do |field,value|
    next if value.blank?  or field.blank?
    field = field.to_s              
    value = unescape_csv(value)
    identifer_type = reference_identifier

    next if value == field

    # Correctly interpret the notation "parent:id", "parent id" etc in a column header
    if (split = field.split(/[:_\-\s]/)).count == 2
      identifier_type = split.last
      relationship_type = split.first.to_s
    else
      relationship_type = field
    end

    relationship_type = self.class.normalize_relationship_field_name(relationship_type)
    case relationship_type
    when "order"
      # If the field specifies the object's order among siblings 
      @proxy.update(order: value.to_f)
      next
    when "collection"
      # If the field specifies the name or ID of a collection,
      # find or create the collection and update the metadata to match
      col = find_or_create_collection(value)
      ( @metadata[:member_of_collection_ids] ||= [] ) << col.id if col
      next
    when "parent", "child"
      
      # correctly interpret the notation "id:a78C2d81"
      identifier_type, object_identifier = interpret_relationship_value(identifier_type, value)
      
      relationship_parameters =  { work_proxy_id: @proxy.id,
                                   identifier_type: identifier_type,
                                   relationship_type: relationship_type,
                                   object_identifier: object_identifier,
                                   status: "new"}
      
      #add previous sibling link if necessary
      previous_value = @raw_data[row_number-1][field]
      # Check if this is a parent relationship, and the previous row also has one
      if previous_value.present? && (relationship_type == "parent")
        # Check if the previous row has the same parent as this row
        if object_identifier == interpret_relationship_value(identifier_type, previous_value, field).last
          # If so, set the previous sibling parameter on the relationshp 
          #    to the id for the proxy associated with the previous row
          relationship_parameters[:previous_sibling] = operation.work_proxies.find_by(row_number: row_number-1).id 
        end
      end
      BulkOps::Relationship.create(relationship_parameters)
    end
  end
end

#interpret_relationship_value(id_type, value, field = "parent") ⇒ Object 



335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
 
# File 'lib/bulk_ops/parser.rb', line 335

def interpret_relationship_value id_type, value, field="parent"
  #Handle "id:20kj4259" syntax if it hasn't already been handled
  if (split = value.to_s.split(":")).count == 2
    id_type, value = split.first
    value = split.last
  end
  #Handle special shorthand syntax for refering to relative row numbers
  if id_type == "row"
    #if the value is an integer
    if value =~ /\A[-+]?[0-9]+\z/
      if value.to_i < 0
      # if given a negative integer, count backwards from the current row (remember that value.to_i is negative)
        return [id_type,row_number + value.to_i]
      elsif value.to_i > 0
        # if given a positive integer, remove the row offset
        value = (value.to_i - BulkOps::ROW_OFFSET).to_s
      end
    elsif value.to_s.downcase.include?("prev")
      # if given any variation of the word "previous", get the first preceding row with no parent of its own
      return [id_type,find_previous_parent(field)]
    end
  end
  return [id_type,value]
end

#interpret_scalar_fields ⇒ Object 



161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
 
# File 'lib/bulk_ops/parser.rb', line 161

def interpret_scalar_fields
  @raw_row.each do |field, values| 
   next if values.blank? or field.nil? or field == values
   # get the field name, if this column is a metadata field
   next unless field_name = find_field_name(field.to_s)
   field = schema.get_field(field_name)
   # Ignore controlled fields
   next if field.controlled?
   split_values(values).each do |value|
     next if value.blank?
     value = value.strip.encode('utf-8', :invalid => :replace, :undef => :replace, :replace => '_') unless value.blank?
     value = unescape_csv(value)
     (@metadata[field_name] ||= []) << value
    end
 end
end

#localAuthUrl(property, value) ⇒ Object 



471
472
473
474
475
 
# File 'lib/bulk_ops/parser.rb', line 471

def localAuthUrl(property, value) 
  return value if (auth = getLocalAuth(property)).nil?
  url =   findAuthUrl(auth, value) ||  mintLocalAuthUrl(auth,value)
  return url
end

#localIdToUrl(id, auth_name) ⇒ Object 



410
411
412
413
414
415
416
 
# File 'lib/bulk_ops/parser.rb', line 410

def localIdToUrl(id,auth_name) 
  root_urls = {production: "https://digitalcollections.library.ucsc.edu",
               staging: "http://digitalcollections-staging.library.ucsc.edu",
               development: "http://#{Socket.gethostname}",
               test: "http://#{Socket.gethostname}"}
  return "#{root_urls[Rails.env.to_sym]}/authorities/show/local/#{auth_name}/#{id}"
end

#mintLocalAuthUrl(auth_name, value) ⇒ Object 



386
387
388
389
390
391
392
393
394
 
# File 'lib/bulk_ops/parser.rb', line 386

def mintLocalAuthUrl(auth_name, value)
  value.strip!
  id = value.parameterize
  auth = Qa::LocalAuthority.find_or_create_by(name: auth_name)
  entry = Qa::LocalAuthorityEntry.create(local_authority: auth,
                                         label: value,
                                         uri: id)
  return localIdToUrl(id,auth_name)
end

#record_exists?(id) ⇒ Boolean

Returns:

  • (Boolean) 


467
468
469
 
# File 'lib/bulk_ops/parser.rb', line 467

def record_exists? id
  operation.record_exists? id
end

#report_error(type, message, **args) ⇒ Object 



441
442
443
444
445
446
 
# File 'lib/bulk_ops/parser.rb', line 441

def report_error type, message, **args
  puts "ERROR MESSAGE: #{message}"
  @proxy.update(status: "error", message: message)
  args[:type]=type
  (@parsing_errors ||= []) <<  BulkOps::Error.new(**args)
end

#schema ⇒ Object 



500
501
502
 
# File 'lib/bulk_ops/parser.rb', line 500

def schema
  ScoobySnacks::METADATA_SCHEMA
end

#setAdminSet ⇒ Object 



429
430
431
432
433
434
 
# File 'lib/bulk_ops/parser.rb', line 429

def setAdminSet 
  return if @metadata[:admin_set_id]
  asets = AdminSet.where({title: "Bulk Ingest Set"})
  asets = AdminSet.find('admin_set/default') if asets.blank?
  @metadata[:admin_set_id] = Array(asets).first.id unless asets.blank?
end

#setMetadataInheritance ⇒ Object 



436
437
438
439
 
# File 'lib/bulk_ops/parser.rb', line 436

def setMetadataInheritance
  return if @metadata[:metadataInheritance].present?
  @metadata[:metadataInheritance] = operation.options["metadataInheritance"] unless operation.options["metadataInheritance"].blank?
end

#split_values(value_string) ⇒ Object 



513
514
515
516
517
 
# File 'lib/bulk_ops/parser.rb', line 513

def split_values value_string
  # Split values on all un-escaped separator character (escape character is '\')
  # Then replace all escaped separator charactors with un-escaped versions
  value_string.split(/(?<!\\)#{BulkOps::SEPARATOR}/).map{|val| val.gsub("\\#{BulkOps::SEPARATOR}",BulkOps::SEPARATOR).strip}
end

#unescape_csv(value) ⇒ Object 



360
361
362
 
# File 'lib/bulk_ops/parser.rb', line 360

def unescape_csv(value)
  value.gsub(/\\(['";,])/,'\1')
end