confidential_info_redactor 0.0.12 → 0.0.13

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 1233d3ddc47135630231617aaebbabe48bdcaef5
-  data.tar.gz: 009da140352b1929a471a6a214f110369a6e7d22
+  metadata.gz: d698521a92a28fa94c1cd8f2c6317b4249a0a19d
+  data.tar.gz: ab825e72bb74de93b72efaac0364f4b01ba7b3fb
 SHA512:
-  metadata.gz: 2f09ed5dbe85f303577d3e1d11d61e83157e06fbb8764ef7b6b86960ce1265b9e560b8a880494264214399386156b405e24d1cb12cad90ec02ab8cbd13089a90
-  data.tar.gz: 5048ecb8b50262c157e2b6cd02b7d14d1f5f5c446ef5299df7a3efdf7efe1660301a22827b06208f96cab90a3c7d6d85d79c5c7d8f1e55e85799f4feaa3be9e4
+  metadata.gz: b91adab393e7137f24f9255a20b9d35cb812fbbb0513b0a3daf01944065d3659ef40864bc8013c7291d210593a60341e1dae898d6c5c97262029d7f0fefc8a5f
+  data.tar.gz: c554f97b0ce9fe0341ab983641bc347c415d770561acb150ce810236271c485cd0b395d8ebf8eef333372ebbb6719a703adca82cef36855022b8d565c1d196e1

data/lib/confidential_info_redactor/extractor.rb

@@ -34,7 +34,7 @@ module ConfidentialInfoRedactor
                   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/version.rb

@@ -1,3 +1,3 @@
 module ConfidentialInfoRedactor
-  VERSION = "0.0.12"
+  VERSION = "0.0.13"
 end

data/spec/confidential_info_redactor/extractor_spec.rb

@@ -141,10 +141,15 @@ RSpec.describe ConfidentialInfoRedactor::Extractor do
         expect(described_class.new(text: text, 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, language: 'de').extract).to eq(['Deutsche Bank'])
       end
+
+      it 'extracts the proper nouns from a text #003' do
+        text = 'Ich behielt diese Routine während und sogar während des Studiums an der Uni bei, und ich war damals froh, wenn ich pro Tag zwei ganze Mahlzeiten zu mir nahm.'
+        expect(described_class.new(text: text, language: 'de').extract).to eq([])
+      end
     end
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: confidential_info_redactor
 version: !ruby/object:Gem::Version
-  version: 0.0.12
+  version: 0.0.13
 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