Class: Puree::Dataset

Inherits:
Resource show all
Defined in:
lib/puree/dataset.rb

Overview

Dataset resource

Instance Method Summary collapse

Methods inherited from Resource

#content, #created, #get, #modified, #response, #set_content, #uuid

Constructor Details

#initialize(endpoint: nil, username: nil, password: nil) ⇒ Dataset

Returns a new instance of Dataset.

Parameters:

  • endpoint (String) (defaults to: nil)
  • optional —

    username [String]

  • optional —

    password [String]



10
11
12
13
14
15
 
# File 'lib/puree/dataset.rb', line 10

def initialize(endpoint: nil, username: nil, password: nil)
  super(api: :dataset,
        endpoint: endpoint,
        username: username,
        password: password)
end

Instance Method Details

#access ⇒ String

Open access permission

Returns:

  • (String) 


214
215
216
217
 
# File 'lib/puree/dataset.rb', line 214

def access
  data = node 'openAccessPermission'
  !data.nil? && !data.empty? ? data['term']['localizedString']["__content__"].strip : ''
end

#associated ⇒ Hash

Combines project and publication

Returns:

  • (Hash) 


58
59
60
61
62
63
64
65
66
67
68
69
70
 
# File 'lib/puree/dataset.rb', line 58

def associated
  path = '//associatedContent//relatedContent'
  xpath_result =  xpath_query path
  data_arr = []
  xpath_result.each { |i|
    data = {}
    data['type'] = i.xpath('typeClassification').text.strip
    data['title'] = i.xpath('title').text.strip
    data['uuid'] = i.attr('uuid').strip
    data_arr << data
  }
  data_arr.uniq
end

#available ⇒ Hash

Date made available

Returns:

  • (Hash) 


178
179
180
181
 
# File 'lib/puree/dataset.rb', line 178

def available
  data = node('dateMadeAvailable')
  Puree::Date.normalise(data)
end

#description ⇒ String

Description

Returns:

  • (String) 


115
116
117
118
119
 
# File 'lib/puree/dataset.rb', line 115

def description
  path = '//descriptions/classificationDefinedField/value/localizedString'
  xpath_result = xpath_query path
  xpath_result ? xpath_result.text.strip : ''
end

#doi ⇒ String

Digital Object Identifier

Returns:

  • (String) 


256
257
258
259
260
 
# File 'lib/puree/dataset.rb', line 256

def doi
  path = '//content/doi'
  xpath_result = xpath_query path
  xpath_result ? xpath_result.text.strip : ''
end

#file ⇒ Array<Hash>

Supporting file

Returns:

  • (Array<Hash>) 


223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
 
# File 'lib/puree/dataset.rb', line 223

def file
  path = '//documents/document'
  xpath_result = xpath_query path

  docs = []

  xpath_result.each do |d|
    doc = {}
    # doc['id'] = f.xpath('id').text.strip
    doc['name'] = d.xpath('fileName').text.strip
    doc['mime'] = d.xpath('mimeType').text.strip
    doc['size'] = d.xpath('size').text.strip
    doc['url'] = d.xpath('url').text.strip
    doc['title'] = d.xpath('title').text.strip
    # doc['createdDate'] = d.xpath('createdDate').text.strip
    # doc['visibleOnPortalDate'] = d.xpath('visibleOnPortalDate').text.strip
    # doc['limitedVisibility'] = d.xpath('limitedVisibility').text.strip

    license = {}
    license_name = d.xpath('documentLicense/term/localizedString').text.strip
    license['name'] = license_name
    license_url = d.xpath('documentLicense/description/localizedString').text.strip
    license['url'] = license_url
    doc['license'] = license
    docs << doc

  end
  docs.uniq
end

#geographical ⇒ Array<String>

Geographical coverage

Returns:

  • (Array<String>) 


186
187
188
189
190
191
192
193
194
 
# File 'lib/puree/dataset.rb', line 186

def geographical
  data = node 'geographicalCoverage'
  if !data.nil? && !data.empty?
    data = data['localizedString']["__content__"]
    data.is_a?(Array) ? data.uniq : data.split(',').map(&:strip).uniq
  else
    []
  end
end

#keyword ⇒ Array<String>

Keyword

Returns:

  • (Array<String>) 


105
106
107
108
109
110
 
# File 'lib/puree/dataset.rb', line 105

def keyword
  path = '//keywordGroups/keywordGroup/keyword/userDefinedKeyword/freeKeyword'
  xpath_result =  xpath_query path
  data_arr = xpath_result.map { |i| i.text.strip }
  data_arr.uniq
end

#link ⇒ Array<Hash>

Link

Returns:

  • (Array<Hash>) 


20
21
22
23
24
25
26
27
28
29
30
31
 
# File 'lib/puree/dataset.rb', line 20

def link
  path = '//links/link'
  xpath_result = xpath_query path
  data = []
  xpath_result.each { |i|
    o = {}
    o['url'] = i.xpath('url').text.strip
    o['description'] = i.xpath('description').text.strip
    data << o
  }
  data.uniq
end

#metadata ⇒ Hash

All metadata

Returns:

  • (Hash) 


271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
 
# File 'lib/puree/dataset.rb', line 271

def metadata
  o = super
  o['access'] = access
  o['associated'] = associated
  o['available'] = available
  o['description'] = description
  o['doi'] = doi
  o['file'] = file
  o['geographical'] = geographical
  o['keyword'] = keyword
  o['link'] = link
  o['organisation'] = organisation
  o['person'] = person
  o['project'] = project
  o['production'] = production
  o['publication'] = publication
  o['publisher'] = publisher
  o['temporal'] = temporal
  o['title'] = title
  o
end

#organisation ⇒ Hash

Organisation

Returns:

  • (Hash) 


36
37
38
39
40
41
42
43
44
 
# File 'lib/puree/dataset.rb', line 36

def organisation
  path = '//content/managedBy'
  xpath_result =  xpath_query path
  o = {}
  o['uuid'] = xpath_result.xpath('@uuid').text.strip
  o['name'] = xpath_result.xpath('name/localizedString').text.strip
  o['type'] = xpath_result.xpath('typeClassification/term/localizedString').text.strip
  o
end

#person ⇒ Hash

Person (internal, external, other)

Returns:

  • (Hash) 


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
160
161
162
163
164
165
166
167
168
169
170
171
 
# File 'lib/puree/dataset.rb', line 124

def person
  data = node('persons')
  persons = {}
  if !data.nil? && !data.empty?
    data = data['dataSetPersonAssociation']
  else
    return persons
  end
  internal_persons = []
  external_persons = []
  other_persons = []
  case data
    when Array
      data.each do |d|
        person = generic_person d
        if d.key? 'person'
          person['uuid'] = d['person']['uuid'].strip
          internal_persons << person
        end
        if d.key? 'externalPerson'
          person['uuid'] = d['externalPerson']['uuid'].strip
          external_persons << person
        end
        if !d.key?('person') && !d.key?('externalPerson')
          person['uuid'] = ''
          other_persons << person
        end
      end
    when Hash
      person = generic_person data
      if data.key? 'person'
        person['uuid'] = data['person']['uuid'].strip
        internal_persons << person
      end
      if data.key? 'externalPerson'
        person['uuid'] = data['externalPerson']['uuid'].strip
        external_persons << person
      end
      if !data.key?('person') && !data.key?('externalPerson')
        person['uuid'] = ''
        other_persons << person
      end
  end
  persons['internal'] = internal_persons.uniq
  persons['external'] = external_persons.uniq
  persons['other'] = other_persons.uniq
  persons
end

#production ⇒ Hash

Date of data production

Returns:

  • (Hash) 


199
200
201
 
# File 'lib/puree/dataset.rb', line 199

def production
  temporal_range 'dateOfDataProduction', 'endDateOfDataProduction'
end

#project ⇒ Array<Hash>

Project

Returns:

  • (Array<Hash>) 


75
76
77
 
# File 'lib/puree/dataset.rb', line 75

def project
  associated_type('Research').uniq
end

#publication ⇒ Array<Hash>

Publication

Returns:

  • (Array<Hash>) 


82
83
84
85
86
87
88
89
90
 
# File 'lib/puree/dataset.rb', line 82

def publication
  data_arr = []
  associated.each do |i|
    if i['type'] != 'Research'
      data_arr << i
    end
  end
  data_arr.uniq
end

#publisher ⇒ String

Publisher

Returns:

  • (String) 


49
50
51
52
53
 
# File 'lib/puree/dataset.rb', line 49

def publisher
  path = '//publisher/name'
  xpath_result = xpath_query path
  xpath_result ? xpath_result.text.strip : ''
end

#temporal ⇒ Hash

Temporal coverage

Returns:

  • (Hash) 


207
208
209
 
# File 'lib/puree/dataset.rb', line 207

def temporal
  temporal_range 'temporalCoverageStartDate', 'temporalCoverageEndDate'
end

#title ⇒ String

Title

Returns:

  • (String) 


96
97
98
99
100
 
# File 'lib/puree/dataset.rb', line 96

def title
  path = '//title/localizedString'
  xpath_result = xpath_query path
  xpath_result ? xpath_result.text.strip : ''
end