Class: OpenTox::Parser::Owl::Dataset

Inherits:
Object
  • Object
show all
Includes:
OpenTox::Parser::Owl
Defined in:
lib/parser.rb

Overview

OWL-DL parser for datasets

Instance Attribute Summary collapse

Instance Method Summary collapse

Methods included from OpenTox::Parser::Owl

from_rdf, #load_metadata

Constructor Details

#initialize(uri, subjectid = nil) ⇒ OpenTox::Parser::Owl::Dataset

Create a new OWL-DL dataset parser

Parameters:

  • uri —

    Dataset URI



146
147
148
149
 
# File 'lib/parser.rb', line 146

def initialize(uri, subjectid=nil)
  super uri
  @dataset = ::OpenTox::Dataset.new(@uri, subjectid)
end

Instance Attribute Details

#uri=(value) ⇒ Object (writeonly)

Sets the attribute uri

Parameters:

  • value —

    the value to set the attribute uri to.



141
142
143
 
# File 'lib/parser.rb', line 141

def uri=(value)
  @uri = value
end

Instance Method Details

#load_features(subjectid = nil) ⇒ Hash

Read only features from a dataset service.

Returns:

  • (Hash) —

    Internal features representation



239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
 
# File 'lib/parser.rb', line 239

def load_features(subjectid=nil)
  if File.exist?(@uri)
    file = File.new(@uri)
  else
    file = Tempfile.new("ot-rdfxml")
    # do not concat /features to uri string, this would not work for dataset/R401577?max=3 
    uri = URI::parse(@uri)
    # PENDING
    # ambit models return http://host/dataset/id?feature_uris[]=sth but 
    # amibt dataset services does not support http://host/dataset/id/features?feature_uris[]=sth
    # and features are not inlcuded in http://host/dataset/id/features
    # -> load features from complete dataset
    uri.path = File.join(uri.path,"features") unless @uri=~/\?(feature_uris|page|pagesize)/
    uri = uri.to_s
    file.puts OpenTox::RestClientWrapper.get uri,{:subjectid => subjectid,:accept => "application/rdf+xml"},nil,false
    file.close
    to_delete = file.path
  end
  statements = []
  features = Set.new
  `rapper -i rdfxml -o ntriples #{file.path} 2>/dev/null`.each_line do |line|
    triple = line.chomp.split('> ').collect{|i| i.sub(/\s+.$/,'').gsub(/[<>"]/,'')}[0..2]
    statements << triple
    features << triple[0] if triple[1] == RDF.type and (triple[2] =~ /Feature|Substructure/) 
  end
  File.delete(to_delete) if to_delete
  statements.each do |triple|
    if features.include? triple[0]
      @dataset.features[triple[0]] = {} unless @dataset.features[triple[0]]
      if triple[1] == RDF.type
         @dataset.features[triple[0]][triple[1]] = [] unless @dataset.features[triple[0]][triple[1]]
         @dataset.features[triple[0]][triple[1]] << triple[2].split('^^').first
      else
        @dataset.features[triple[0]][triple[1]] = triple[2].split('^^').first
      end
    end
  end
  @dataset.features
end

#load_uri(subjectid = nil) ⇒ Hash

Read data from dataset service. Files can be parsed by setting #uri to a filename (after initialization with a real URI)

Examples:

Read data from an external service

parser = OpenTox::Parser::Owl::Dataaset.new "http://wwbservices.in-silico.ch/dataset/1"
dataset = parser.load_uri

Create dataset from RDF/XML file

dataset = OpenTox::Dataset.create
parser = OpenTox::Parser::Owl::Dataaset.new dataset.uri
parser.uri = "dataset.rdfxml" # insert your input file
dataset = parser.load_uri
dataset.save

Returns:

  • (Hash) —

    Internal dataset representation



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
229
230
231
232
233
234
235
 
# File 'lib/parser.rb', line 162

def load_uri(subjectid=nil)
  
  # avoid using rapper directly because of 2 reasons:
  # * http errors wont be noticed
  # * subjectid cannot be sent as header
  ##uri += "?subjectid=#{CGI.escape(subjectid)}" if subjectid
  ##`rapper -i rdfxml -o ntriples #{file} 2>/dev/null`.each_line do |line| 
  if File.exist?(@uri)
    file = File.new(@uri)
  else
    file = Tempfile.new("ot-rdfxml")
    file.puts OpenTox::RestClientWrapper.get @uri,{:subjectid => subjectid,:accept => "application/rdf+xml"},nil,false
    file.close
    to_delete = file.path
  end
  
  data = {}
  feature_values = {}
  feature = {}
  feature_accept_values = {}
  other_statements = {}
  `rapper -i rdfxml -o ntriples #{file.path} 2>/dev/null`.each_line do |line|
    triple = line.chomp.split(' ',3)
    triple = triple[0..2].collect{|i| i.sub(/\s+.$/,'').gsub(/[<>"]/,'')}
    case triple[1] 
    when /#{OT.values}/i
      data[triple[0]] = {:compound => "", :values => []} unless data[triple[0]]
      data[triple[0]][:values] << triple[2]  
    when /#{OT.value}/i
      feature_values[triple[0]] = triple[2] 
    when /#{OT.compound}/i
      data[triple[0]] = {:compound => "", :values => []} unless data[triple[0]]
      data[triple[0]][:compound] = triple[2]  
    when /#{OT.feature}/i
      feature[triple[0]] = triple[2]
    when /#{RDF.type}/i
      if triple[2]=~/#{OT.Compound}/i and !data[triple[0]]
        data[triple[0]] = {:compound => triple[0], :values => []} 
      end
    when /#{OT.acceptValue}/i # acceptValue in ambit datasets is only provided in dataset/<id> no in dataset/<id>/features  
      feature_accept_values[triple[0]] = [] unless feature_accept_values[triple[0]]
      feature_accept_values[triple[0]] << triple[2]
    else 
    end
  end
  File.delete(to_delete) if to_delete
  data.each do |id,entry|
    if entry[:values].size==0
      # no feature values add plain compounds
      @dataset.add_compound(entry[:compound])
    else
      entry[:values].each do |value_id|
        if feature_values[value_id]
          split = feature_values[value_id].split(/\^\^/)
          case split[-1]
          when XSD.double, XSD.float 
            value = split.first.to_f
          when XSD.boolean
            value = split.first=~/(?i)true/ ? true : false                
          else
            value = split.first
          end
        end
        @dataset.add entry[:compound],feature[value_id],value
      end
    end
  end
  load_features subjectid
  feature_accept_values.each do |feature, values|
    @dataset.features[feature][OT.acceptValue] = values
  end
  @dataset.metadata = load_metadata(subjectid)
  @dataset
end