confidential_info_redactor 1.0.1 → 1.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
-SHA1:
-  metadata.gz: ea6f9a1d3391a351050aaaca8cc00fff7d54b794
-  data.tar.gz: b6892d0b8e61b4d3dd92f2aba3387396e6cf5f65
+SHA256:
+  metadata.gz: 64cb868a941bb737cc49f0ea98334c89cb719a9fc01e3ada9997e29df07347d1
+  data.tar.gz: 158de52cd0bfb45766e66e45a200590bb0b95282bb49688161393691f75bb92b
 SHA512:
-  metadata.gz: 63937326d768ceeb5b018bee2fb92831dd87ec8b1f1fe9533b999f97383540cef93282eff2a02fa469a77160cf6df717882b188d6b573f6e2162636fcea9e0dd
-  data.tar.gz: 93fd0e9d80d33927233b1232e1d283d347c547a9ab559be307dd1bfb111941bc19083357adabbbfa590486a09177b13add4a77cbe050643da1e3e1f402b22e58
+  metadata.gz: 6c455ad2194369013dc3a451196d9df40b9df359eeea0e62de3a5650398a3d139beed9b3db2638728f498fbfedb281b35ed9e8e3d64d311de9709370366358fa
+  data.tar.gz: 5174e8f4ef98a9194b0554bee6e326072bea9e3b4b03187db55b50f4cf4df39c38940bdc6283badb1c62ea63d4e3e82463b135ebff352d38e7c4b38446b44f14

data/README.md

@@ -62,6 +62,13 @@ ConfidentialInfoRedactor::Redactor.new.hyperlinks(text)
 ConfidentialInfoRedactor::Redactor.new(tokens: tokens).proper_nouns(text)
 # => '<redacted> announced a merger with <redacted> that will happen on December 15th, 2020 for $200,000,000,000. Please contact <redacted> at j.smith@example.com or visit http://www.super-fake-merger.com.'
 
+# Possessive proper nouns are handled automatically
+text = "The teacher took John's book from the shelf."
+tokens = ConfidentialInfoRedactor::Extractor.new.extract(text)
+# => ["John"]
+ConfidentialInfoRedactor::Redactor.new(tokens: tokens).redact(text)
+# => "The teacher took <redacted>'s book from the shelf."
+
 # It is possible to 'turn off' any of the specific redactors
 ConfidentialInfoRedactor::Redactor.new(tokens: tokens, ignore_numbers: true).redact(text)
 # => '<redacted> announced a merger with <redacted> that will happen on <redacted date> for $200,000,000,000. Please contact <redacted> at <redacted> or visit <redacted>.'

data/lib/confidential_info_redactor/extractor.rb

@@ -33,11 +33,11 @@ module ConfidentialInfoRedactor
     private
 
     def extract_preliminary_terms(segment)
-      segment.to_s.gsub(EXTRACT_REGEX).map { |match| match unless corpus.include?(match.downcase.gsub(PUNCTUATION_REGEX, '').gsub(/\'$/, '')) }.compact
+      segment.to_s.gsub(EXTRACT_REGEX).map { |match| match unless corpus.include?(match.downcase.gsub(PUNCTUATION_REGEX, '').gsub(/\'s\z/, '').gsub(/\'$/, '')) }.compact
     end
 
     def clean_token(token)
-      token.gsub(PUNCTUATION_REGEX, '').gsub(/\'$/, '').gsub(/\.\z/, '').strip
+      token.gsub(PUNCTUATION_REGEX, '').gsub(/\'s\z/, '').gsub(/\'$/, '').gsub(/\.\z/, '').strip
     end
 
     def non_confidential_token?(token, includes_confidential)

data/lib/confidential_info_redactor/redactor.rb

@@ -75,6 +75,7 @@ module ConfidentialInfoRedactor
 
     def redact_tokens(txt)
       tokens.sort_by{ |x| x.split.count }.reverse.each do |token|
+        txt.gsub!(/(?<=\s|^|\")#{Regexp.escape(token)}(?=['\u2019]s\b)/, "#{token_text}")
         txt.gsub!(/(?<=\s|^|\")#{Regexp.escape(token)}(?=\W|$)/, "#{token_text}")
       end
       txt.strip

data/lib/confidential_info_redactor/version.rb

@@ -1,3 +1,3 @@
 module ConfidentialInfoRedactor
-  VERSION = "1.0.1"
+  VERSION = "1.0.2"
 end

data/spec/confidential_info_redactor/extractor_spec.rb

@@ -143,6 +143,11 @@ RSpec.describe ConfidentialInfoRedactor::Extractor do
         text = 'John'
         expect(described_class.new(language: 'en').extract(text)).to eq(['John'])
       end
+
+      it 'extracts the proper nouns from a text #018' do
+        text = "The teacher took John's book from the shelf."
+        expect(described_class.new(language: 'en').extract(text)).to eq(['John'])
+      end
     end
 
     context 'German (de)' do

data/spec/confidential_info_redactor/redactor_spec.rb

@@ -181,5 +181,19 @@ RSpec.describe ConfidentialInfoRedactor::Redactor do
       text = 'My Transformation - avoid Trans.'
       expect(described_class.new(language: 'en', tokens: tokens, number_text: '**redacted number**', date_text: '^^redacted date^^', token_text: '*****', hyperlink_text: '*****', email_text: '*****').redact(text)).to eq('My Transformation - avoid *****.')
     end
+
+    it 'redacts all confidential information from a text #008' do
+      text = "The teacher took John's book from the shelf."
+      tokens = ConfidentialInfoRedactor::Extractor.new.extract(text)
+      expect(tokens).to eq(['John'])
+      expect(described_class.new(language: 'en', tokens: tokens).redact(text)).to eq("The teacher took <redacted>'s book from the shelf.")
+    end
+
+    it 'redacts all confidential information from a text #009' do
+      text = "The cat's toy rolled under the couch and disappeared."
+      tokens = ConfidentialInfoRedactor::Extractor.new.extract(text)
+      expect(tokens).to eq([])
+      expect(described_class.new(language: 'en', tokens: tokens).redact(text)).to eq("The cat's toy rolled under the couch and disappeared.")
+    end
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: confidential_info_redactor
 version: !ruby/object:Gem::Version
-  version: 1.0.1
+  version: 1.0.2
 platform: ruby
 authors:
 - Kevin S. Dias
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2016-08-11 00:00:00.000000000 Z
+date: 2026-03-11 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -98,7 +98,7 @@ homepage: https://github.com/diasks2/confidential_info_redactor
 licenses:
 - MIT
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -113,9 +113,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.4.1
-signing_key: 
+rubygems_version: 3.3.26
+signing_key:
 specification_version: 4
 summary: Semi-automatically redact confidential information from a text
 test_files: