confidential_info_redactor_lite 0.0.17 → 0.0.18

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 773956f15f755ab2aaf70ad75acfdd4c0ed2d9c1
-  data.tar.gz: 4ad0c3f4555050f33608039ff40627b31e76a848
+  metadata.gz: 52f6dc75062fd5b96bf6ebddda6091ece0984a5f
+  data.tar.gz: f1e59cf4d26237d79076474ca22ee5eaeb27ae41
 SHA512:
-  metadata.gz: 278a320f5d5e8e835abab2e62a6924629bec0b9835a18c82cbf5fa41dc7bdf5996d70411f4a478abe6b322cbc89aafd67f699ea9e7cc58c86127f7995d149c8e
-  data.tar.gz: 954e633068ee089384440851966fc7e6fee8e4a2fbc47ec7cf3f68b6789610d1e1a123589ed7d0a33fd82a703114a1a64b5d7555e0fcd7520f85f931a94c33c7
+  metadata.gz: 9295558573274c81494ad53a53f61da538cba25a8bf5b32035a0af7baa78f8e2b83b63ecd3516902fc2457acfda037e5a45bf9106a4976d8c7489554182ebb34
+  data.tar.gz: 3b881aef75c2df2c7507034feb57ab02deba5ba4735bf50ca00c63d83eee034fd0c92a53b80a0fdf77203a8a19434c70850847cf08b9b3adc1b9796acd0e21af

data/lib/confidential_info_redactor_lite/extractor.rb

@@ -25,7 +25,7 @@ module ConfidentialInfoRedactorLite
                   tracker = false if corpus.include?(token.downcase)
                 end
               end
-              extracted_terms << t.gsub(/[\?\)\(\!\\\/\"\:\;\,]/, '').gsub(/\'$/, '').gsub(/”/,'').gsub(/\.\z/, '').strip unless corpus.include?(t.downcase.gsub(/[\?\.\)\(\!\\\/\"\:\;]/, '').gsub(/”/,'').gsub(/\'$/, '').strip) || !tracker
+              extracted_terms << t.gsub(/[\?\)\(\!\\\/\"\:\;\,]/, '').gsub(/\'$/, '').gsub(/”/,'').gsub(/\.\z/, '').strip unless corpus.include?(t.downcase.gsub(/[\?\.\)\(\!\\\/\"\:\;]/, '').gsub(/”/,'').gsub(/\'$/, '').strip) || !tracker || (corpus.include?(t.downcase.gsub(/[\?\.\)\(\!\\\/\"\:\;]/, '').gsub(/”/,'').gsub(/\'$/, '').strip[0...-2]) && t.downcase.gsub(/[\?\.\)\(\!\\\/\"\:\;]/, '').gsub(/”/,'').gsub(/\'$/, '').strip[-2..-1].eql?('en')) || (corpus.include?(t.downcase.gsub(/[\?\.\)\(\!\\\/\"\:\;]/, '').gsub(/”/,'').gsub(/\'$/, '').strip[0...-2]) && t.downcase.gsub(/[\?\.\)\(\!\\\/\"\:\;]/, '').gsub(/”/,'').gsub(/\'$/, '').strip[-2..-1].eql?('es')) || (corpus.include?(t.downcase.gsub(/[\?\.\)\(\!\\\/\"\:\;]/, '').gsub(/”/,'').gsub(/\'$/, '').strip[0...-2]) && t.downcase.gsub(/[\?\.\)\(\!\\\/\"\:\;]/, '').gsub(/”/,'').gsub(/\'$/, '').strip[-2..-1].eql?('er')) || (corpus.include?(t.downcase.gsub(/[\?\.\)\(\!\\\/\"\:\;]/, '').gsub(/”/,'').gsub(/\'$/, '').strip[0...-1]) && t.downcase.gsub(/[\?\.\)\(\!\\\/\"\:\;]/, '').gsub(/”/,'').gsub(/\'$/, '').strip[-1].eql?('s'))
             end
           end
         end

data/lib/confidential_info_redactor_lite/version.rb

@@ -1,3 +1,3 @@
 module ConfidentialInfoRedactorLite
-  VERSION = "0.0.17"
+  VERSION = "0.0.18"
 end

data/spec/confidential_info_redactor_lite/extractor_spec.rb

@@ -142,10 +142,15 @@ RSpec.describe ConfidentialInfoRedactorLite::Extractor do
         expect(described_class.new(text: text, corpus: corpus, language: 'de').extract).to eq(['Deutschen Bank'])
       end
 
-      it 'extracts the proper nouns from a text #001' do
+      it 'extracts the proper nouns from a text #002' do
         text = 'Viele Mitarbeiter der Deutsche Bank suchen eine andere Arbeitsstelle.'
         expect(described_class.new(text: text, corpus: corpus, language: 'de').extract).to eq(['Deutsche Bank'])
       end
+
+      it 'extracts the proper nouns from a text #003' do
+        text = 'Viele de Mitarbeiters der Deutsche Bank suchen eine andere Arbeitsstelle.'
+        expect(described_class.new(text: text, corpus: corpus, language: 'de').extract).to eq(['Deutsche Bank'])
+      end
     end
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: confidential_info_redactor_lite
 version: !ruby/object:Gem::Version
-  version: 0.0.17
+  version: 0.0.18
 platform: ruby
 authors:
 - Kevin S. Dias
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-04-23 00:00:00.000000000 Z
+date: 2015-04-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler