Class: Bulkrax::Importer
Constant Summary
collapse
- DEFAULT_OBJECT_TYPES =
%w[collection work file_set relationship].freeze
Instance Attribute Summary collapse
Class Method Summary
collapse
Instance Method Summary
collapse
Methods included from StatusInfo
#current_status, #failed?, #last_error, #set_status_info, #status_at, #succeeded?
#file?, #increment_counters, #key_without_numbers, #keys_without_numbers, #last_imported_at, #next_import_at, #parser, #parser_class, #zip?
Instance Attribute Details
#current_run ⇒ Object
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
|
# File 'app/models/bulkrax/importer.rb', line 106
def current_run
return @current_run if @current_run.present?
@current_run = self.importer_runs.create!
return @current_run if file? && zip?
entry_counts = {
total_work_entries: self.limit || parser.works_total,
total_collection_entries: parser.collections_total,
total_file_set_entries: parser.file_sets_total
}
@current_run.update!(entry_counts)
@current_run
end
|
#file ⇒ Object
Returns the value of attribute file.
23
24
25
|
# File 'app/models/bulkrax/importer.rb', line 23
def file
@file
end
|
#file_style ⇒ Object
Returns the value of attribute file_style.
23
24
25
|
# File 'app/models/bulkrax/importer.rb', line 23
def file_style
@file_style
end
|
#only_updates ⇒ Object
Returns the value of attribute only_updates.
23
24
25
|
# File 'app/models/bulkrax/importer.rb', line 23
def only_updates
@only_updates
end
|
Class Method Details
.frequency_enums ⇒ Object
86
87
88
89
90
91
|
# File 'app/models/bulkrax/importer.rb', line 86
def self.frequency_enums
[['Daily', 'P1D'], ['Monthly', 'P1M'], ['Yearly', 'P1Y'], ['Once (on save)', 'PT0S']]
end
|
.safe_uri_filename(uri) ⇒ Object
26
27
28
29
30
31
32
33
|
# File 'app/models/bulkrax/importer.rb', line 26
def self.safe_uri_filename(uri)
r = Faraday.head(uri.to_s)
return CGI.parse(r.['content-disposition'])["filename"][0].delete("\"")
rescue
filename = File.basename(uri.to_s)
filename.delete!('/')
filename.presence || SecureRandom.uuid
end
|
Instance Method Details
#default_field_mapping ⇒ Object
72
73
74
75
76
77
78
79
80
|
# File 'app/models/bulkrax/importer.rb', line 72
def default_field_mapping
return self.field_mapping if parser.import_fields.nil?
ActiveSupport::HashWithIndifferentAccess.new(
parser.import_fields.reject(&:nil?).map do |m|
Bulkrax.default_field_mapping.call(m)
end.inject(:merge)
)
end
|
#errored_entries_csv_path ⇒ Object
186
187
188
|
# File 'app/models/bulkrax/importer.rb', line 186
def errored_entries_csv_path
@errored_entries_csv_path ||= File.join(parser.base_path, "import_#{path_string}_errored_entries.csv")
end
|
#frequency ⇒ Object
97
98
99
100
|
# File 'app/models/bulkrax/importer.rb', line 97
def frequency
f = self[:frequency] || "PT0S"
ISO8601::Duration.new(f)
end
|
#frequency=(frequency) ⇒ Object
93
94
95
|
# File 'app/models/bulkrax/importer.rb', line 93
def frequency=(frequency)
self[:frequency] = ISO8601::Duration.new(frequency).to_s
end
|
#import_collections ⇒ Object
142
143
144
|
# File 'app/models/bulkrax/importer.rb', line 142
def import_collections
import_objects(['collection'])
end
|
#import_file_sets ⇒ Object
146
147
148
|
# File 'app/models/bulkrax/importer.rb', line 146
def import_file_sets
import_objects(['file_set'])
end
|
The format for metadata for the incoming import; corresponds to an Entry class
172
173
174
|
# File 'app/models/bulkrax/importer.rb', line 172
def import_metadata_format
[['CSV', 'Bulkrax::CsvEntry'], ['RDF (N-Triples)', 'Bulkrax::RdfEntry']]
end
|
#import_objects(types_array = nil) ⇒ Object
156
157
158
159
160
161
162
163
|
# File 'app/models/bulkrax/importer.rb', line 156
def import_objects(types_array = nil)
self.only_updates ||= false
self.save if self.new_record? types = types_array || DEFAULT_OBJECT_TYPES
parser.create_objects(types)
rescue StandardError => e
set_status_info(e)
end
|
#import_relationships ⇒ Object
150
151
152
|
# File 'app/models/bulkrax/importer.rb', line 150
def import_relationships
import_objects(['relationship'])
end
|
#import_works ⇒ Object
138
139
140
|
# File 'app/models/bulkrax/importer.rb', line 138
def import_works
import_objects(['work'])
end
|
#importer_unzip_path ⇒ Object
If the import data is zipped, unzip it to this path
182
183
184
|
# File 'app/models/bulkrax/importer.rb', line 182
def importer_unzip_path
@importer_unzip_path ||= File.join(parser.base_path, "import_#{path_string}")
end
|
#last_run ⇒ Object
122
123
124
|
# File 'app/models/bulkrax/importer.rb', line 122
def last_run
@last_run ||= self.importer_runs.last
end
|
#mapping ⇒ Object
If field_mapping is empty, setup a default based on the export_properties
59
60
61
62
63
64
65
66
67
68
69
70
|
# File 'app/models/bulkrax/importer.rb', line 59
def mapping
@mapping ||= if self.field_mapping.blank? || self.field_mapping == [{}]
if parser.import_fields.present? || self.field_mapping == [{}]
default_field_mapping
end
else
default_field_mapping.merge(self.field_mapping)
end
end
|
196
197
198
|
# File 'app/models/bulkrax/importer.rb', line 196
def metadata_only?
parser.parser_fields['metadata_only'] == true
end
|
#parser_fields ⇒ Object
82
83
84
|
# File 'app/models/bulkrax/importer.rb', line 82
def parser_fields
self[:parser_fields] || {}
end
|
#path_string ⇒ Object
190
191
192
193
194
|
# File 'app/models/bulkrax/importer.rb', line 190
def path_string
"#{self.id}_#{self.created_at.strftime('%Y%m%d%H%M%S')}_#{self.importer_runs.last.id}"
rescue
"#{self.id}_#{self.created_at.strftime('%Y%m%d%H%M%S')}"
end
|
#record_status ⇒ Object
43
44
45
46
47
48
49
50
51
52
53
54
55
56
|
# File 'app/models/bulkrax/importer.rb', line 43
def record_status
importer_run = ImporterRun.find(current_run.id) return if importer_run.enqueued_records.positive? if importer_run.failed_records.positive?
if importer_run.invalid_records.present?
e = Bulkrax::ImportFailed.new('Failed with Invalid Records', importer_run.invalid_records.split("\n"))
importer_run.importer.set_status_info(e)
else
importer_run.importer.set_status_info('Complete (with failures)')
end
else
importer_run.importer.set_status_info('Complete')
end
end
|
#replace_files ⇒ Object
130
131
132
|
# File 'app/models/bulkrax/importer.rb', line 130
def replace_files
self.parser_fields['replace_files']
end
|
#schedulable? ⇒ Boolean
102
103
104
|
# File 'app/models/bulkrax/importer.rb', line 102
def schedulable?
frequency.to_seconds != 0
end
|
#seen ⇒ Object
126
127
128
|
# File 'app/models/bulkrax/importer.rb', line 126
def seen
@seen ||= {}
end
|
#status ⇒ Object
35
36
37
38
39
40
41
|
# File 'app/models/bulkrax/importer.rb', line 35
def status
if self.validate_only
'Validated'
else
super
end
end
|
#unique_collection_identifier(id) ⇒ Object
Prepend the base_url to ensure unique set identifiers
167
168
169
|
# File 'app/models/bulkrax/importer.rb', line 167
def unique_collection_identifier(id)
"#{self.parser_fields['base_url'].split('/')[2]}_#{id}"
end
|
#update_files ⇒ Object
134
135
136
|
# File 'app/models/bulkrax/importer.rb', line 134
def update_files
self.parser_fields['update_files']
end
|