Class: PUIIndexer

Inherits:
PeriodicIndexer show all
Defined in:
indexer/app/lib/pui_indexer.rb

Constant Summary collapse

PUI_RESOLVES = 
AppConfig[:record_inheritance_resolves]

Constants inherited from PeriodicIndexer

PeriodicIndexer::WORKER_STATUS_INDEX_ERROR, PeriodicIndexer::WORKER_STATUS_INDEX_SUCCESS, PeriodicIndexer::WORKER_STATUS_NOTHING_INDEXED

Constants inherited from IndexerCommon

IndexerCommon::MAX_PENDING_DELETES

Class Method Summary collapse

Instance Method Summary collapse

Methods inherited from PeriodicIndexer

#handle_deletes, #log, #run, #run_index_round, #start_worker_thread

Methods inherited from IndexerCommon

#add_agents, #add_arks, add_attribute_to_resolve, #add_audit_info, #add_batch_hook, #add_delete_hook, #add_document_prepare_hook, #add_extents, #add_extra_documents_hook, add_indexer_initialize_hook, #add_level, #add_subjects, #add_subjects_subrecord, #add_summary, #add_years, #apply_pui_fields, #clean_for_sort, #clean_whitespace, #dedupe_by_uri, #delete_records, #do_http_request, #enum_fields, extract_string_values, generate_permutations_for_identifier, generate_sort_string_for_identifier, generate_years_for_date_range, #get_record_scope, #index_batch, #index_records, #is_repository_unpublished?, #login, pause, paused?, #paused?, #record_has_children, #records_with_children, #reset_session, #sanitize_json, #send_commit, #solr_url, #trim_ark_value

Constructor Details

#initialize(backend = nil, state = nil, name) ⇒ PUIIndexer

Returns a new instance of PUIIndexer.



12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
 
# File 'indexer/app/lib/pui_indexer.rb', line 12

def initialize(backend = nil, state = nil, name)
  state_class = AppConfig[:index_state_class].constantize
  index_state = state || state_class.new("indexer_pui_state")

  super(backend, index_state, name)

  # Set up our JSON schemas now that we know the JSONModels have been loaded
  RecordInheritance.prepare_schemas

  @time_to_sleep = AppConfig[:pui_indexing_frequency_seconds].to_i
  @thread_count = AppConfig[:pui_indexer_thread_count].to_i
  @records_per_thread = AppConfig[:pui_indexer_records_per_thread].to_i

  @unpublished_records = java.util.Collections.synchronizedList(java.util.ArrayList.new)
end

Class Method Details

.get_indexer(state = nil, name = "PUI Indexer") ⇒ Object 



37
38
39
 
# File 'indexer/app/lib/pui_indexer.rb', line 37

def self.get_indexer(state = nil, name = "PUI Indexer")
  indexer = self.new(state, name)
end

Instance Method Details

#add_infscroll_docs(resource_uris, batch) ⇒ Object 



126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
 
# File 'indexer/app/lib/pui_indexer.rb', line 126

def add_infscroll_docs(resource_uris, batch)
  resource_uris.each do |resource_uri|
    json = JSONModel::HTTP.get_json(resource_uri + '/ordered_records')

    batch << {
      'id' => "#{resource_uri}/ordered_records",
      'uri' => "#{resource_uri}/ordered_records",
      'pui_parent_id' => resource_uri,
      'publish' => "true",
      'primary_type' => "resource_ordered_records",
      'types' => ['pui'],
      'json' => ASUtils.to_json(json)
    }
  end
end

#add_notes(doc, record) ⇒ Object 



120
121
122
123
124
 
# File 'indexer/app/lib/pui_indexer.rb', line 120

def add_notes(doc, record)
  if record['record']['notes']
    doc['notes_published'] = IndexerCommon.extract_string_values(record['record']['notes'], :published_only)
  end
end

#build_fullrecord(doc, record) ⇒ Object 



116
117
118
 
# File 'indexer/app/lib/pui_indexer.rb', line 116

def build_fullrecord(doc, record)
  doc['fullrecord_published'] = IndexerCommon.extract_string_values(record['record'], :published_only)
end

#configure_doc_rules ⇒ Object 



51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
 
# File 'indexer/app/lib/pui_indexer.rb', line 51

def configure_doc_rules
  super

  record_has_children('resource')
  record_has_children('archival_object')
  record_has_children('digital_object')
  record_has_children('digital_object_component')
  record_has_children('classification')
  record_has_children('classification_term')


  add_document_prepare_hook {|doc, record|

    if RecordInheritance.has_type?(doc['primary_type'])
      parent_id = doc['id']
      doc['id'] = "#{parent_id}#pui"
      doc['pui_parent_id'] = parent_id
      doc['types'] ||= []
      doc['types'] << 'pui'
      doc['types'] << "pui_#{doc['primary_type']}"
      doc['types'] << 'pui_record'
      doc['types'] << 'pui_only'
    end
  }
end

#fetch_records(type, ids, resolve) ⇒ Object 



28
29
30
31
32
33
34
35
 
# File 'indexer/app/lib/pui_indexer.rb', line 28

def fetch_records(type, ids, resolve)
  records = JSONModel(type).all(:id_set => ids.join(","), 'resolve[]' => resolve)
  if RecordInheritance.has_type?(type)
    RecordInheritance.merge(records, :direct_only => true)
  else
    records
  end
end

#final_doc_rules ⇒ Object

Run the final doc rules after all the hooks have been added This allows plugins to access ancestor data in PUI records before it is removed



79
80
81
82
83
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
 
# File 'indexer/app/lib/pui_indexer.rb', line 79

def final_doc_rules
  # this runs after the hooks in indexer_common, so we can overwrite with confidence
  add_document_prepare_hook {|doc, record|
    if RecordInheritance.has_type?(doc['primary_type'])
      # special handling for json because we need to include indirectly inherited
      # fields too - the json sent to indexer_common only has directly inherited
      # fields because only they should be indexed.
      # so we remerge without the :direct_only flag, and we remove the ancestors
      doc['json'] = ASUtils.to_json(RecordInheritance.merge(record['record'],
                                                            :remove_ancestors => true))

      # special handling for title because it is populated from display_string
      # in indexer_common and display_string is not changed in the merge process
      doc['title'] = record['record']['title'] if record['record']['title']

      # special handling for fullrecord because we don't want the ancestors indexed.
      # we're now done with the ancestors, so we can just delete them from the record
      record['record'].delete('ancestors')
      # we don't want container_profile or top_container notes indexed for the public either
      if record['record']['instances']
        record['record']['instances'].each do |instance|
          if instance['sub_container'] && instance['sub_container']['top_container']
            top_container = instance['sub_container']['top_container']
            if top_container['_resolved']
              top_container['_resolved'].delete('internal_note')
              if top_container['_resolved']['container_profile'] && top_container['_resolved']['container_profile']['_resolved']
                top_container['_resolved']['container_profile']['_resolved'].delete('notes')
              end
            end
          end
        end
      end
      build_fullrecord(doc, record)
    end
  }
end

#index_round_complete(repository) ⇒ Object 



159
160
161
162
163
164
165
166
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
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
 
# File 'indexer/app/lib/pui_indexer.rb', line 159

def index_round_complete(repository)
  # Index any trees in `repository`
  tree_types = [[:resource, :archival_object],
                [:digital_object, :digital_object_component],
                [:classification, :classification_term]]

  start = Time.now
  checkpoints = []
  update_mtimes = false
  state_type = 'tree'

  tree_uris = []

  tree_types.each do |pair|
    root_type = pair.first
    node_type = pair.last

    checkpoints << [repository, root_type, start]
    checkpoints << [repository, node_type, start]

    last_root_node_mtime = [@state.get_last_mtime(repository.id, root_type, state_type) - @window_seconds, 0].max
    last_node_mtime = [@state.get_last_mtime(repository.id, node_type, state_type) - @window_seconds, 0].max

    root_node_ids = Set.new(JSONModel::HTTP.get_json(JSONModel(root_type).uri_for, :all_ids => true, :modified_since => last_root_node_mtime))
    node_ids = JSONModel::HTTP.get_json(JSONModel(node_type).uri_for, :all_ids => true, :modified_since => last_node_mtime)

    node_ids.each_slice(@records_per_thread) do |ids|
      node_records = JSONModel(node_type).all(:id_set => ids.join(","), 'resolve[]' => [])

      node_records.each do |record|
        root_node_ids << JSONModel.parse_reference(record[root_type.to_s]['ref']).fetch(:id)
      end
    end

    tree_uris.concat(root_node_ids.map {|id| JSONModel(root_type).uri_for(id) })
  end

  batch = IndexBatch.new

  add_infscroll_docs(tree_uris.select {|uri| JSONModel.parse_reference(uri).fetch(:type) == 'resource'},
                     batch)

  tree_indexer = LargeTreeDocIndexer.new(batch)
  tree_indexer.add_largetree_docs(tree_uris)

  if batch.length > 0
    log "Indexed #{batch.length} additional PUI records in repository #{repository.repo_code}"

    index_batch(batch, nil, :parent_id_field => 'pui_parent_id')
    send_commit
    update_mtimes = true
  end

  if tree_indexer.deletes.length > 0
    tree_indexer.deletes.each_slice(100) do |deletes|
      delete_records(deletes, :parent_id_field => 'pui_parent_id')
    end
  end

  handle_deletes(:parent_id_field => 'pui_parent_id')

  # Delete any unpublished records and decendents
  delete_records(@unpublished_records, :parent_id_field => 'pui_parent_id')
  @unpublished_records.clear()

  checkpoints.each do |repository, type, start|
    @state.set_last_mtime(repository.id, type, start, state_type) if update_mtimes
  end

end

#record_types ⇒ Object 



45
46
47
48
49
 
# File 'indexer/app/lib/pui_indexer.rb', line 45

def record_types
  # We only want to index the record types we're going to make separate
  # PUI-specific versions of...
  (super.select {|type| RecordInheritance.has_type?(type)} + [:archival_object]).uniq
end

#repositories_updated_action(updated_repositories) ⇒ Object 



234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
 
# File 'indexer/app/lib/pui_indexer.rb', line 234

def repositories_updated_action(updated_repositories)

  updated_repositories.each do |repository|

    if !repository['record']['publish']

      # Delete PUI-only Solr documents in case this is the first index run after the repository has been unpublished
      req = Net::HTTP::Post.new("#{solr_url.path}/update")
      req['Content-Type'] = 'application/json'
      delete_request = {:delete => {'query' => "repository:\"#{repository['uri']}\" AND types:pui_only"}}
      req.body = delete_request.to_json
      response = do_http_request(solr_url, req)
      if response.code == '200'
        Log.info "Deleted PUI-only documents in private repository #{repository['record']['repo_code']}: #{response}"
      else
        Log.error "SolrIndexerError when deleting PUI-only records in private repository #{repository['record']['repo_code']}: #{response.body}"
      end

    end

  end

end

#resolved_attributes ⇒ Object 



41
42
43
 
# File 'indexer/app/lib/pui_indexer.rb', line 41

def resolved_attributes
  super + PUI_RESOLVES
end

#skip_index_doc?(doc) ⇒ Boolean

Returns:

  • (Boolean) 


151
152
153
154
155
156
157
 
# File 'indexer/app/lib/pui_indexer.rb', line 151

def skip_index_doc?(doc)
  published = doc['publish']

  stage_unpublished_for_deletion(doc['id']) unless published

  !published
end

#skip_index_record?(record) ⇒ Boolean

Returns:

  • (Boolean) 


142
143
144
145
146
147
148
 
# File 'indexer/app/lib/pui_indexer.rb', line 142

def skip_index_record?(record)
  published = record['record']['publish']

  stage_unpublished_for_deletion("#{record['record']['uri']}#pui") unless published

  !published
end

#stage_unpublished_for_deletion(doc_id) ⇒ Object 



230
231
232
 
# File 'indexer/app/lib/pui_indexer.rb', line 230

def stage_unpublished_for_deletion(doc_id)
  @unpublished_records.add(doc_id) if doc_id =~ /#pui$/
end