data-anonymization 0.8.0 → 0.8.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 1eed8d87b9e7060ee7d3813c77b52b838f789d92
-  data.tar.gz: bc475f130cf331debd4cbc835ab5391aa38c3129
+  metadata.gz: 6455e294c22b99a0181092f93c506060c5c9b383
+  data.tar.gz: 4da392d7b4712d941b9cab8206d178bae89eb288
 SHA512:
-  metadata.gz: e8e6e315a8468c1208c3fd65cf3ea575bbce766290cbbfc9fe323a329a438a02abe63d3751dabda4dd7a1e39a147f0d75aaf88b95beb4c6f09cef7969730b09a
-  data.tar.gz: aef7a9d30b0d1868435445f91a1cbe6412c8628eabcaffa2e51ea74fa721d1a2f3a53aa54295aabcfee229ae7b9333065ab119056c40946dca39ffc78a520ed7
+  metadata.gz: 5b0da43a74bf21505f6e462ac73c17fff694f24082478f29fee83c4a9746e2703cc6497fb0e333b811992a806fdec5cb8b3bbcabefe667548577809445c86148
+  data.tar.gz: ba546b36a7681cdafb7cbcb0d98025ee357194e3b1ab5eedb4825fa484a47d8b50e32d0890ef7e2c0cb9f9327bd066fd76986280d8f5955b963827fbdd44c991

data/.gitignore

@@ -18,4 +18,5 @@ tmp
 .idea
 sample-data/chinook-empty.sqlite
 tmp
-examples/mongodb_whitelist_generated.rb
+examples/mongodb_whitelist_generated.rb
+data

data/.ruby-version

@@ -1 +1 @@
-ruby-2.3.1
+ruby-2.4.1

data/.travis.yml

@@ -1,8 +1,9 @@
 language: ruby
 services:
   - mongodb
-before_install: gem install bundler -v 1.12.5
+before_install: gem install bundler -v 1.15.3
 before_script: rake empty_dest
 rvm:
-  - 2.2.5
-  - 2.3.1
+  - 2.2.6
+  - 2.3.3
+  - 2.4.1

data/README.md

@@ -70,6 +70,10 @@ Postgresql database having **composite primary key**
 
 ## Changelog
 
+#### 0.8.1 (Aug 19, 2017)
+1. Multi-threading support added by [stanislav-tyutin](https://github.com/stanislav-tyutin) using Pull Request.  
+2. Fixed to work with Ruby 2.4.x, issue with Integer data type
+
 #### 0.8.0 (Oct 31, 2016)
 1. Upgraded to rails 5.x 
 

data/data-anonymization.gemspec

@@ -25,6 +25,6 @@ Gem::Specification.new do |gem|
   gem.add_dependency('rgeo', '~> 0.5')
   gem.add_dependency('rgeo-geojson', '~> 0.4')
   gem.add_dependency('powerbar', '~> 1.0')
-  gem.add_dependency('parallel', '~> 1.9')
-  gem.add_dependency('thor', '~> 0.19')
+  gem.add_dependency('parallel', '~> 1.12')
+  gem.add_dependency('thor', '~> 0.20')
 end

data/examples/blacklist_dsl.rb

@@ -1,4 +1,4 @@
-system "bundle exec ruby examples/whitelist_dsl.rb"
+system 'bundle exec ruby examples/whitelist_dsl.rb'
 
 require 'data-anonymization'
 
@@ -18,7 +18,7 @@ database 'Chinook' do
     anonymize('City').using FieldStrategy::RandomCity.region_US
     anonymize('State').using FieldStrategy::RandomProvince.region_US
     anonymize('PostalCode').using FieldStrategy::RandomZipcode.region_US
-    anonymize('Country') {|field| "USA" }
+    anonymize('Country') {|field| 'USA'}
     anonymize('Phone').using FieldStrategy::RandomPhoneNumber.new
     anonymize('Fax').using FieldStrategy::RandomPhoneNumber.new
     anonymize('Email').using FieldStrategy::StringTemplate.new('test+#{row_number}@gmail.com')
@@ -33,7 +33,7 @@ database 'Chinook' do
     anonymize('City').using FieldStrategy::RandomCity.region_US
     anonymize('State').using FieldStrategy::RandomProvince.region_US
     anonymize('PostalCode').using FieldStrategy::RandomZipcode.region_US
-    anonymize('Country') {|field| "USA" }
+    anonymize('Country') {|field| 'USA'}
     anonymize('Fax').using FieldStrategy::RandomPhoneNumber.new
     anonymize('Email').using FieldStrategy::StringTemplate.new('test+#{row_number}@gmail.com')
   end

data/examples/mongodb_blacklist_dsl.rb

@@ -4,20 +4,20 @@ require 'mongo'
 DataAnon::Utils::Logging.logger.level = Logger::INFO
 Mongo::Logger.logger.level = Logger::WARN
 
-Mongo::Client.new("mongodb://localhost/test").database.drop
-system "mongoimport --host=127.0.0.1 -d test --drop -c users --jsonArray ./sample-data/mongo/users.json"
-system "mongoimport --host=127.0.0.1 -d test --drop -c plans --jsonArray ./sample-data/mongo/plans.json"
+Mongo::Client.new('mongodb://localhost/test').database.drop
+system 'mongoimport --host=127.0.0.1 -d test --drop -c users --jsonArray ./sample-data/mongo/users.json'
+system 'mongoimport --host=127.0.0.1 -d test --drop -c plans --jsonArray ./sample-data/mongo/plans.json'
 
 
 database 'test' do
   strategy DataAnon::Strategy::MongoDB::Blacklist
-  source_db :mongodb_uri => "mongodb://localhost/test", :database => 'test'
+  source_db :mongodb_uri => 'mongodb://localhost/test', :database => 'test'
 
   collection 'users' do
     anonymize('date_of_birth').using FieldStrategy::TimeDelta.new(5,30)
     anonymize('user_id').using FieldStrategy::StringTemplate.new('user-#{row_number}')
     anonymize('email').using FieldStrategy::RandomMailinatorEmail.new
-    anonymize('password') { |field| "password" }
+    anonymize('password') { |field| 'password'}
     anonymize('first_name').using FieldStrategy::RandomFirstName.new
     anonymize('last_name').using FieldStrategy::RandomLastName.new
   end

data/examples/mongodb_whitelist_dsl.rb

@@ -5,22 +5,22 @@ DataAnon::Utils::Logging.logger.level = Logger::INFO
 Mongo::Logger.logger.level = Logger::WARN
 
 
-Mongo::Client.new("mongodb://localhost/test").database.drop
-Mongo::Client.new("mongodb://localhost/dest").database.drop
-system "mongoimport --host=127.0.0.1 -d test --drop -c users --jsonArray ./sample-data/mongo/users.json"
-system "mongoimport --host=127.0.0.1 -d test --drop -c plans --jsonArray ./sample-data/mongo/plans.json"
+Mongo::Client.new('mongodb://localhost/test').database.drop
+Mongo::Client.new('mongodb://localhost/dest').database.drop
+system 'mongoimport --host=127.0.0.1 -d test --drop -c users --jsonArray ./sample-data/mongo/users.json'
+system 'mongoimport --host=127.0.0.1 -d test --drop -c plans --jsonArray ./sample-data/mongo/plans.json'
 
 database 'test' do
   strategy DataAnon::Strategy::MongoDB::Whitelist
-  source_db :mongodb_uri => "mongodb://localhost/test", :database => 'test'
-  destination_db :mongodb_uri => "mongodb://localhost/dest", :database => 'dest'
+  source_db :mongodb_uri => 'mongodb://localhost/test', :database => 'test'
+  destination_db :mongodb_uri => 'mongodb://localhost/dest', :database => 'dest'
 
   collection 'users' do
     whitelist '_id','failed_attempts','updated_at'
     anonymize('date_of_birth').using FieldStrategy::TimeDelta.new(5,30)
     anonymize('user_id').using FieldStrategy::StringTemplate.new('user-#{row_number}')
     anonymize('email').using FieldStrategy::RandomMailinatorEmail.new
-    anonymize('password') { |field| "password" }
+    anonymize('password') { |field| 'password'}
     anonymize('first_name').using FieldStrategy::RandomFirstName.new
     anonymize('last_name').using FieldStrategy::RandomLastName.new
     anonymize 'password_reset_answer','password_reset_question'

data/examples/whitelist_dsl.rb

@@ -1,4 +1,4 @@
-system "rake empty_dest" # clean destination database on every call
+system 'rake empty_dest' # clean destination database on every call
 
 require 'data-anonymization'
 
@@ -15,7 +15,7 @@ database 'Chinook' do
     primary_key 'GenreId'
     whitelist 'GenreId'
     anonymize 'Name' do |field|
-      field.value + " test"
+      field.value + ' test'
     end
   end
 
@@ -39,7 +39,7 @@ database 'Chinook' do
     anonymize('City').using FieldStrategy::RandomCity.region_US
     anonymize('State').using FieldStrategy::RandomProvince.region_US
     anonymize('PostalCode').using FieldStrategy::RandomZipcode.region_US
-    anonymize('Country') {|field| "USA" }
+    anonymize('Country') {|field| 'USA'}
     anonymize('Fax').using FieldStrategy::RandomPhoneNumber.new
     anonymize('Email').using FieldStrategy::StringTemplate.new('test+#{row_number}@gmail.com')
   end
@@ -56,7 +56,7 @@ database 'Chinook' do
     anonymize('City').using FieldStrategy::RandomCity.region_US
     anonymize('State').using FieldStrategy::RandomProvince.region_US
     anonymize('PostalCode').using FieldStrategy::RandomZipcode.region_US
-    anonymize('Country') {|field| "USA" }
+    anonymize('Country') {|field| 'USA'}
     anonymize('Phone').using FieldStrategy::RandomPhoneNumber.new
     anonymize('Fax').using FieldStrategy::RandomPhoneNumber.new
     anonymize('Email').using FieldStrategy::StringTemplate.new('test+#{row_number}@gmail.com')

data/examples/whitelist_dsl_threads.rb

@@ -0,0 +1,66 @@
+system 'rake empty_dest' # clean destination database on every call
+
+require 'data-anonymization'
+
+DataAnon::Utils::Logging.logger.level = Logger::INFO
+
+database 'Chinook' do
+  strategy DataAnon::Strategy::Whitelist
+  source_db :adapter => 'sqlite3', :database => 'sample-data/chinook.sqlite'
+  destination_db :adapter => 'sqlite3', :database => 'sample-data/chinook-empty.sqlite'
+
+  default_field_strategies :string => FieldStrategy::StringTemplate.new('Sunit #{row_number} Parekh')
+
+  table 'Genre' do
+    primary_key 'GenreId'
+    whitelist 'GenreId'
+    anonymize 'Name' do |field|
+      field.value + ' test'
+    end
+  end
+
+  table 'MediaType' do
+    primary_key 'MediaTypeId'
+    anonymize('MediaTypeId') { |field| field.value } # same as whitelist
+    anonymize('Name').using FieldStrategy::StringTemplate.new('Media Type #{row_number}')
+
+  end
+
+  table 'Customer' do
+    primary_key 'CustomerId'
+    thread_num 5  # thread_num
+
+    whitelist 'CustomerId', 'SupportRepId', 'Company'
+    anonymize('Phone').using FieldStrategy::RandomPhoneNumber.new
+    anonymize('FirstName').using FieldStrategy::RandomFirstName.new
+    anonymize('LastName').using FieldStrategy::RandomLastName.new
+    anonymize('Address').using FieldStrategy::RandomAddress.region_US
+    anonymize('City').using FieldStrategy::RandomCity.region_US
+    anonymize('State').using FieldStrategy::RandomProvince.region_US
+    anonymize('PostalCode').using FieldStrategy::RandomZipcode.region_US
+    anonymize('Country') {|field| 'USA'}
+    anonymize('Fax').using FieldStrategy::RandomPhoneNumber.new
+    anonymize('Email').using FieldStrategy::StringTemplate.new('test+#{row_number}@gmail.com')
+  end
+
+  table 'Employee' do
+    thread_num 5  # thread_num
+
+    whitelist 'EmployeeId', 'ReportsTo', 'Title'
+    anonymize('BirthDate').using FieldStrategy::DateTimeDelta.new(1, 1)
+    anonymize('FirstName').using FieldStrategy::RandomFirstName.new
+    anonymize('LastName').using FieldStrategy::RandomLastName.new
+    anonymize('HireDate').using FieldStrategy::DateTimeDelta.new(2, 0)
+    anonymize('Address').using FieldStrategy::RandomAddress.region_US
+    anonymize('City').using FieldStrategy::RandomCity.region_US
+    anonymize('State').using FieldStrategy::RandomProvince.region_US
+    anonymize('PostalCode').using FieldStrategy::RandomZipcode.region_US
+    anonymize('Country') {|field| 'USA'}
+    anonymize('Phone').using FieldStrategy::RandomPhoneNumber.new
+    anonymize('Fax').using FieldStrategy::RandomPhoneNumber.new
+    anonymize('Email').using FieldStrategy::StringTemplate.new('test+#{row_number}@gmail.com')
+  end
+
+
+
+end

data/lib/core/field.rb

@@ -3,7 +3,7 @@ module DataAnon
 
     class Field
 
-      def initialize name, value, row_number, ar_record, table_name = "unknown"
+      def initialize name, value, row_number, ar_record, table_name = 'unknown'
         @name = name
         @value = value
         @row_number = row_number

data/lib/core/table_errors.rb

@@ -11,7 +11,7 @@ module DataAnon
 
       def log_error record, exception
         @errors << { :record => record, :exception => exception}
-        raise "Reached limit of error for a table" if @errors.length > 100
+        raise 'Reached limit of error for a table' if @errors.length > 100
       end
 
       def errors

data/lib/data-anonymization.rb

@@ -1,28 +1,28 @@
-require "version"
+require 'version'
 
-require "utils/logging"
-require "utils/random_int"
-require "utils/random_float"
-require "utils/random_string"
-require "utils/random_string_chars_only"
-require "utils/geojson_parser"
-require "utils/progress_bar"
-require "utils/parallel_progress_bar"
-require "utils/resource"
-require "utils/template_helper"
-require "parallel/table"
-require "core/database"
-require "core/fields_missing_strategy"
-require "thor/helpers/rdbms_dsl_generator"
-require "core/field"
-require "core/table_errors"
-require "strategy/strategies"
-require "utils/database"
-require "core/dsl"
+require 'utils/logging'
+require 'utils/random_int'
+require 'utils/random_float'
+require 'utils/random_string'
+require 'utils/random_string_chars_only'
+require 'utils/geojson_parser'
+require 'utils/progress_bar'
+require 'utils/parallel_progress_bar'
+require 'utils/resource'
+require 'utils/template_helper'
+require 'parallel/table'
+require 'core/database'
+require 'core/fields_missing_strategy'
+require 'thor/helpers/rdbms_dsl_generator'
+require 'core/field'
+require 'core/table_errors'
+require 'strategy/strategies'
+require 'utils/database'
+require 'core/dsl'
 
 begin
   require 'mongo'
-  require "thor/helpers/mongodb_dsl_generator"
+  require 'thor/helpers/mongodb_dsl_generator'
 rescue LoadError
-  "Ignoring the mongodb specific libraries if monog driver is not specified in gem"
+  'Ignoring the mongodb specific libraries if monog driver is not specified in gem'
 end

data/lib/strategy/base.rb

@@ -37,6 +37,10 @@ module DataAnon
         @limit = limit
       end
 
+      def thread_num thread_num
+        @thread_num = thread_num
+      end
+
       def whitelist *fields
         fields.each { |f| @fields[f] = DataAnon::Strategy::Field::Whitelist.new }
       end
@@ -95,13 +99,15 @@ module DataAnon
           progress = progress_bar.new(@name, total)
           if @primary_keys.empty? || !@batch_size.present?
             process_table progress
+          elsif @thread_num.present?
+            process_table_in_threads progress
           else
             process_table_in_batches progress
           end
           progress.close
         end
         if source_table.respond_to?('clear_all_connections!')
-	        source_table.clear_all_connections!
+          source_table.clear_all_connections!
         end
       end
 
@@ -134,6 +140,40 @@ module DataAnon
         end
       end
 
+      def process_table_in_threads progress
+        logger.info "Processing table #{@name} records in batch size of #{@batch_size} [THREADS]"
+
+        index = 0
+        threads = []
+
+        source_table.find_in_batches(batch_size: @batch_size) do |records|
+          until threads.count(&:alive?) <= @thread_num
+            thr = threads.delete_at 0
+            thr.join
+            progress.show index
+          end
+
+          thr = Thread.new {
+            records.each do |record|
+              begin
+                process_record_if index, record
+                index += 1
+              rescue => exception
+                puts exception.inspect
+                @errors.log_error record, exception
+              end
+            end
+          }
+          threads << thr
+        end
+
+        until threads.empty?
+          thr = threads.delete_at 0
+          thr.join
+          progress.show index
+        end
+      end
+
       def source_table_limited
         @source_table_limited ||= begin
           if @limit.present?

data/lib/strategy/field/default_anon.rb

@@ -5,6 +5,7 @@ module DataAnon
       class DefaultAnon
 
         DEFAULT_STRATEGIES = {:string => FieldStrategy::RandomString.new,
+                              :integer => FieldStrategy::RandomIntegerDelta.new(5),
                               :fixnum => FieldStrategy::RandomIntegerDelta.new(5),
                               :bignum => FieldStrategy::RandomIntegerDelta.new(5000),
                               :float => FieldStrategy::RandomFloatDelta.new(5.0),

data/lib/strategy/strategies.rb

@@ -9,5 +9,5 @@ begin
   require 'strategy/mongodb/whitelist'
   require 'strategy/mongodb/blacklist'
 rescue LoadError
-  "Ignoring the mongodb specific libraries if monog driver is not specified in gem"
+  'Ignoring the mongodb specific libraries if monog driver is not specified in gem'
 end

data/lib/thor/helpers/mongodb_dsl_generator.rb

@@ -18,13 +18,13 @@ module DataAnon
 
       def generate
 
-        db = Mongo::Connection.from_uri(@mongodb_uri)[@configuration_hash[:database]]
+        db = Mongo::Client.new(@mongodb_uri, :database => @configuration_hash[:database])
         collections = db.collections
         collections.each do |collection|
           unless collection.name.start_with?('system.')
             depth = 2
             @output << "\tcollection '#{collection.name}' do"
-            document = collection.find_one
+            document = collection.find({}).first
             process_document(depth, document)
             @output << "\tend\n"
           end
@@ -63,4 +63,3 @@ module DataAnon
     end
   end
 end
-

data/lib/utils/template_helper.rb

@@ -5,9 +5,11 @@ module DataAnon
       def self.source_connection_specs_rdbms config_hash
 
         config_hash.keys.reject{|key| config_hash[key].nil? }.collect { |key|
-          if ((config_hash[key].class.to_s.downcase == "string"))
+          if ((config_hash[key].class.to_s.downcase == 'string'))
             ":#{key} => '#{config_hash[key]}'"
-          elsif ((config_hash[key].class.to_s.downcase == "fixnum"))
+          elsif ((config_hash[key].class.to_s.downcase == 'integer'))
+            ":#{key} => #{config_hash[key]}"
+          elsif ((config_hash[key].class.to_s.downcase == 'fixnum'))
             ":#{key} => #{config_hash[key]}"
           end
         }.join ', '

data/lib/version.rb

@@ -1,3 +1,3 @@
 module DataAnonymization
-  VERSION = '0.8.0'
+  VERSION = '0.8.1'
 end

data/spec/acceptance/mongodb_blacklist_spec.rb

@@ -1,56 +1,56 @@
-require "spec_helper"
+require 'spec_helper'
 require 'mongo'
 
-describe "End 2 End MongoDB Blacklist Acceptance Test" do
+describe 'End 2 End MongoDB Blacklist Acceptance Test' do
 
   before(:each) do
-    Mongo::Client.new("mongodb://localhost/test").database().drop()
+    Mongo::Client.new('mongodb://localhost/test').database().drop()
     users = [
         {
-            "_id" => 1,
-            "USER_ID" => "sunitparekh",
-            "date_of_birth" => Time.new(2012, 7, 14, 13, 1, 0),
-            "email" => "parekh-sunit@mailinator.com",
-            "password" => "TfqIK8Pd8GlbMDFZCX4l/5EtnOkfLCeynOL85tJQuxum&382knaflk@@",
-            "failed_attempts" => 0,
-            "first_name" => "Sunit",
-            "last_name" => "Parekh",
-            "password_reset_answer" => "manza",
-            "password_reset_question" => "My new car modal?",
-            "updated_at" => Time.new(2012, 8, 15, 13, 1, 0),
-            "alternate_emails" => ["abc@test.com","abc2@test.com"]
+            '_id' => 1,
+            'USER_ID' => 'sunitparekh',
+            'date_of_birth' => Time.new(2012, 7, 14, 13, 1, 0),
+            'email' => 'parekh-sunit@mailinator.com',
+            'password' => 'TfqIK8Pd8GlbMDFZCX4l/5EtnOkfLCeynOL85tJQuxum&382knaflk@@',
+            'failed_attempts' => 0,
+            'first_name' => 'Sunit',
+            'last_name' => 'Parekh',
+            'password_reset_answer' => 'manza',
+            'password_reset_question' => 'My new car modal?',
+            'updated_at' => Time.new(2012, 8, 15, 13, 1, 0),
+            'alternate_emails' => ['abc@test.com', 'abc2@test.com']
 
         },
         {
-            "_id" => 2,
-            "USER_ID" => "anandagrawal",
-            "date_of_birth" => Time.new(2011, 8, 11, 13, 1, 0),
-            "email" => "anand-agrawal@mailinator.com",
-            "password" => "Tz548O0RWusldVAWkwqfzO3jK/X4l/5EtnOkfLCeynOL85tJQuxum",
-            "failed_attempts" => 0,
-            "first_name" => "Anand",
-            "last_name" => "Agrawal",
-            "password_reset_answer" => "android",
-            "password_reset_question" => "My phone?",
-            "updated_at" => Time.new(2012, 2, 11, 13, 1, 0),
-            "alternate_emails" => ["abc@test.com","abc2@test.com"]
+            '_id' => 2,
+            'USER_ID' => 'anandagrawal',
+            'date_of_birth' => Time.new(2011, 8, 11, 13, 1, 0),
+            'email' => 'anand-agrawal@mailinator.com',
+            'password' => 'Tz548O0RWusldVAWkwqfzO3jK/X4l/5EtnOkfLCeynOL85tJQuxum',
+            'failed_attempts' => 0,
+            'first_name' => 'Anand',
+            'last_name' => 'Agrawal',
+            'password_reset_answer' => 'android',
+            'password_reset_question' => 'My phone?',
+            'updated_at' => Time.new(2012, 2, 11, 13, 1, 0),
+            'alternate_emails' => ['abc@test.com', 'abc2@test.com']
         }
     ]
-    users_coll = Mongo::Client.new("mongodb://localhost/test").database().collection('users')
+    users_coll = Mongo::Client.new('mongodb://localhost/test').database().collection('users')
     users.each { |p| users_coll.insert_one p }
   end
 
-  it "should anonymize plans collection" do
+  it 'should anonymize plans collection' do
 
     database 'test' do
       strategy DataAnon::Strategy::MongoDB::Blacklist
-      source_db :mongodb_uri => "mongodb://localhost/test", :database => 'test'
+      source_db :mongodb_uri => 'mongodb://localhost/test', :database => 'test'
 
       collection 'users' do
         anonymize('date_of_birth').using FieldStrategy::TimeDelta.new(5,30)
         anonymize('USER_ID').using FieldStrategy::StringTemplate.new('user-#{row_number}')
         anonymize('email').using FieldStrategy::RandomMailinatorEmail.new
-        anonymize('password') { |field| "password" }
+        anonymize('password') { |field| 'password'}
         anonymize('first_name').using FieldStrategy::RandomFirstName.new
         anonymize('last_name').using FieldStrategy::RandomLastName.new
         anonymize('alternate_emails').using FieldStrategy::AnonymizeArray.new(FieldStrategy::RandomMailinatorEmail.new)
@@ -58,20 +58,20 @@ describe "End 2 End MongoDB Blacklist Acceptance Test" do
 
     end
 
-    users_coll = Mongo::Client.new("mongodb://localhost/test").database().collection('users')
+    users_coll = Mongo::Client.new('mongodb://localhost/test').database().collection('users')
     users_coll.find.count.to_int.should be 2
     user = users_coll.find({'_id' => 1}).to_a[0]
 
     user['_id'].should == 1
-    user['USER_ID'].should == "user-1"
+    user['USER_ID'].should == 'user-1'
     user['date_of_birth'].to_i.should_not == Time.new(2012, 7, 14, 13, 1, 0).to_i
-    user['email'].should_not == "parekh-sunit@mailinator.com"
-    user['password'].should == "password"
+    user['email'].should_not == 'parekh-sunit@mailinator.com'
+    user['password'].should == 'password'
     user['failed_attempts'].should == 0
-    user['first_name'].should_not be "Sunit"
-    user['last_name'].should_not be "Parekh"
-    user['password_reset_answer'].should == "manza"
-    user['password_reset_question'].should == "My new car modal?"
+    user['first_name'].should_not be 'Sunit'
+    user['last_name'].should_not be 'Parekh'
+    user['password_reset_answer'].should == 'manza'
+    user['password_reset_question'].should == 'My new car modal?'
     user['updated_at'].to_i.should == Time.new(2012, 8, 15, 13, 1, 0).to_i
     user['alternate_emails'].length.should == 2
     user['alternate_emails'][0].should_not == 'abc@test.com'

data/spec/acceptance/mongodb_whitelist_spec.rb

@@ -1,65 +1,65 @@
-require "spec_helper"
+require 'spec_helper'
 require 'mongo'
 
-describe "End 2 End MongoDB Whitelist Acceptance Test" do
+describe 'End 2 End MongoDB Whitelist Acceptance Test' do
 
   before(:each) do
-    Mongo::Client.new("mongodb://localhost/test").database.drop
-    Mongo::Client.new("mongodb://localhost/dest").database.drop
+    Mongo::Client.new('mongodb://localhost/test').database.drop
+    Mongo::Client.new('mongodb://localhost/dest').database.drop
     plans = [
         {
-            "_id" => 1,
-            "name" => "Free",
-            "nick_names" => ["Name1","Name2"],
-            "features" => [
+            '_id' => 1,
+            'name' => 'Free',
+            'nick_names' => ['Name1', 'Name2'],
+            'features' => [
                 {
-                    "max_storage" => 21474836480,
-                    "type" => "AmazonS3",
-                    "users" => {"max" => 1, "additional" => false}
+                    'max_storage' => 21474836480,
+                    'type' => 'AmazonS3',
+                    'users' => {'max' => 1, 'additional' => false}
                 },
                 {
-                    "max_storage" => 21474836480,
-                    "type" => "DropBox",
-                    "users" => {"max" => 1, "additional" => false}
+                    'max_storage' => 21474836480,
+                    'type' => 'DropBox',
+                    'users' => {'max' => 1, 'additional' => false}
                 }
             ],
-            "term" => "month",
-            "public_sharing" => false,
-            "photo_sharing" => true,
-            "created_at" => Time.new(2012, 6, 21, 13, 30, 0)
+            'term' => 'month',
+            'public_sharing' => false,
+            'photo_sharing' => true,
+            'created_at' => Time.new(2012, 6, 21, 13, 30, 0)
         },
         {
-            "_id" => 2,
-            "name" => "Team",
-            "plan_aliases" => ["Business", "Paid"],
-            "features" => [
+            '_id' => 2,
+            'name' => 'Team',
+            'plan_aliases' => ['Business', 'Paid'],
+            'features' => [
                 {
-                    "max_storage" => 53687091200,
-                    "type" => "AmazonS3",
-                    "users" => {"max" => 5, "additional" => true}
+                    'max_storage' => 53687091200,
+                    'type' => 'AmazonS3',
+                    'users' => {'max' => 5, 'additional' => true}
                 },
                 {
-                    "max_storage" => 53687091200,
-                    "type" => "DropBox",
-                    "users" => {"max" => 5, "additional" => true}
+                    'max_storage' => 53687091200,
+                    'type' => 'DropBox',
+                    'users' => {'max' => 5, 'additional' => true}
                 }
             ],
-            "term" => "month",
-            "public_sharing" => true,
-            "photo_sharing" => true,
-            "created_at" => Time.new(2012, 8, 11, 13, 1, 0)
+            'term' => 'month',
+            'public_sharing' => true,
+            'photo_sharing' => true,
+            'created_at' => Time.new(2012, 8, 11, 13, 1, 0)
         }
     ]
-    plans_coll =  Mongo::Client.new("mongodb://localhost/test").database.collection('plans')
+    plans_coll =  Mongo::Client.new('mongodb://localhost/test').database.collection('plans')
     plans.each { |p| plans_coll.insert_one p }
   end
 
-  it "should anonymize plans collection" do
+  it 'should anonymize plans collection' do
 
     database 'dest' do
       strategy DataAnon::Strategy::MongoDB::Whitelist
-      source_db :mongodb_uri => "mongodb://localhost/test", :database => 'test'
-      destination_db :mongodb_uri => "mongodb://localhost/dest", :database => 'dest'
+      source_db :mongodb_uri => 'mongodb://localhost/test', :database => 'test'
+      destination_db :mongodb_uri => 'mongodb://localhost/dest', :database => 'dest'
 
       collection 'plans' do
         whitelist '_id', 'name', 'term', 'created_at'
@@ -78,15 +78,15 @@ describe "End 2 End MongoDB Whitelist Acceptance Test" do
 
     end
 
-    plans_coll = Mongo::Client.new("mongodb://localhost/dest").database.collection('plans')
+    plans_coll = Mongo::Client.new('mongodb://localhost/dest').database.collection('plans')
     plans_coll.find.count.to_int.should be 2
     plan = plans_coll.find({ '_id' => 1}).to_a[0]
 
     plan['_id'].should == 1
-    plan['name'].should == "Free"
-    plan['nick_names'][0].should_not == "Name1"
-    plan['nick_names'][1].should_not == "Name2"
-    plan['term'].should == "month"
+    plan['name'].should == 'Free'
+    plan['nick_names'][0].should_not == 'Name1'
+    plan['nick_names'][1].should_not == 'Name2'
+    plan['term'].should == 'month'
     plan['created_at'].should == Time.new(2012, 6, 21, 13, 30, 0)
     plan['plan_aliases'].should be_nil
     [true,false].should include(plan['public_sharing'])
@@ -94,14 +94,14 @@ describe "End 2 End MongoDB Whitelist Acceptance Test" do
     plan['features'].length.should == 2
     feature1 = plan['features'][0]
     [10737418240, 21474836480, 53687091200].should include(feature1['max_storage'])
-    feature1['type'].should == "AmazonS3"
+    feature1['type'].should == 'AmazonS3'
     feature1['users']['max'].should be_kind_of(Fixnum)
     [true,false].should include(feature1['users']['additional'])
 
 
     plan = plans_coll.find({ '_id' => 2}).to_a[0]
     plan['plan_aliases'].length.should == 2
-    ["Free", "Team", "Business", "Paid"].should include(plan['plan_aliases'][0])
-    ["Free", "Team", "Business", "Paid"].should include(plan['plan_aliases'][1])
+    ['Free', 'Team', 'Business', 'Paid'].should include(plan['plan_aliases'][0])
+    ['Free', 'Team', 'Business', 'Paid'].should include(plan['plan_aliases'][1])
   end
 end

data/spec/acceptance/rdbms_whitelist_spec.rb

@@ -1,4 +1,4 @@
-require "spec_helper"
+require 'spec_helper'
 
 describe 'End 2 End RDBMS Whitelist Acceptance Test using SQLite database' do
 

data/spec/acceptance/rdbms_whitelist_with_primary_key_spec.rb

@@ -1,6 +1,6 @@
-require "spec_helper"
+require 'spec_helper'
 
-describe "End 2 End RDBMS Whitelist Acceptance Test using SQLite database" do
+describe 'End 2 End RDBMS Whitelist Acceptance Test using SQLite database' do
 
   source_connection_spec = {:adapter => 'sqlite3', :database => 'tmp/customer.sqlite'}
   dest_connection_spec = {:adapter => 'sqlite3', :database => 'tmp/customer-dest.sqlite'}
@@ -13,9 +13,9 @@ describe "End 2 End RDBMS Whitelist Acceptance Test using SQLite database" do
     CustomerSample.create_schema dest_connection_spec
   end
 
-  it "should anonymize customer table record " do
+  it 'should anonymize customer table record ' do
 
-    database "Customer" do
+    database 'Customer' do
       strategy DataAnon::Strategy::Whitelist
       source_db source_connection_spec
       destination_db dest_connection_spec
@@ -37,14 +37,14 @@ describe "End 2 End RDBMS Whitelist Acceptance Test using SQLite database" do
 
     DataAnon::Utils::DestinationDatabase.establish_connection dest_connection_spec
     dest_table = DataAnon::Utils::DestinationTable.create 'customers'
-    new_rec = dest_table.where("cust_id" => CustomerSample::SAMPLE_DATA[0][:cust_id]).first
-    new_rec.first_name.should_not be("Sunit")
-    new_rec.last_name.should_not be("Parekh")
+    new_rec = dest_table.where('cust_id' => CustomerSample::SAMPLE_DATA[0][:cust_id]).first
+    new_rec.first_name.should_not be('Sunit')
+    new_rec.last_name.should_not be('Parekh')
     new_rec.birth_date.should_not be(Date.new(1977,7,8))
     new_rec.address.should == 'F 501 Shanti Nagar'
     ['Gujrat','Karnataka'].should include(new_rec.state)
     new_rec.zipcode.should == '411048'
-    new_rec.phone.should_not be "9923700662"
+    new_rec.phone.should_not be '9923700662'
     new_rec.email.should == 'test+1@gmail.com'
     [true,false].should include(new_rec.terms_n_condition)
     new_rec.age.should be_between(0,100)

data/spec/core/fields_missing_strategy_spec.rb

@@ -1,26 +1,26 @@
-require "spec_helper"
+require 'spec_helper'
 
 describe DataAnon::Core::FieldsMissingStrategy do
 
   FMS = DataAnon::Core::FieldsMissingStrategy
 
-  it "should be able to add field for new table that doesn't exist" do
-    users = FMS.new("users")
-    users.missing("confirm_email")
-    users.fields_missing_strategy.should == ["confirm_email"]
+  it 'should be able to add field for new table that doesnot exist' do
+    users = FMS.new('users')
+    users.missing('confirm_email')
+    users.fields_missing_strategy.should == ['confirm_email']
   end
 
-  it "should be able to take care for same field appearing multiple time" do
-    users = FMS.new("users")
-    users.missing("confirm_email")
-    users.missing("confirm_email")
-    users.fields_missing_strategy.should == ["confirm_email"]
+  it 'should be able to take care for same field appearing multiple time' do
+    users = FMS.new('users')
+    users.missing('confirm_email')
+    users.missing('confirm_email')
+    users.fields_missing_strategy.should == ['confirm_email']
   end
 
-  it "should be able to add multiple fields for table" do
-    users = FMS.new("users")
-    users.missing("confirm_email")
-    users.missing("password_reset")
-    users.fields_missing_strategy.should == ["confirm_email","password_reset"]
+  it 'should be able to add multiple fields for table' do
+    users = FMS.new('users')
+    users.missing('confirm_email')
+    users.missing('password_reset')
+    users.fields_missing_strategy.should == %w(confirm_email password_reset)
   end
 end

data/spec/strategy/field/contact/random_address_spec.rb

@@ -1,4 +1,4 @@
-require "spec_helper"
+require 'spec_helper'
 
 describe DataAnon::Strategy::Field::RandomAddress do
 
@@ -7,6 +7,6 @@ describe DataAnon::Strategy::Field::RandomAddress do
 
   describe 'anonymized address should be different from original address' do
     let(:anonymized_address) {RandomAddress.region_US.anonymize(field)}
-    it {anonymized_address.should_not eq("1 Infinite Loop")}
+    it {anonymized_address.should_not eq('1 Infinite Loop')}
   end
 end

data/spec/strategy/field/default_anon_spec.rb

@@ -1,4 +1,4 @@
-require "spec_helper"
+require 'spec_helper'
 
 describe FieldStrategy::DefaultAnon do
 

data/spec/strategy/field/random_boolean_spec.rb

@@ -1,4 +1,4 @@
-require "spec_helper"
+require 'spec_helper'
 
 describe FieldStrategy::RandomBoolean do
 

data/spec/strategy/field/whitelist_spec.rb

@@ -1,4 +1,4 @@
-require "spec_helper"
+require 'spec_helper'
 
 describe FieldStrategy::Whitelist do
 

data/spec/strategy/mongodb/anonymize_field_spec.rb

@@ -1,4 +1,4 @@
-require "spec_helper"
+require 'spec_helper'
 
 describe DataAnon::Strategy::MongoDB::AnonymizeField do
 
@@ -6,30 +6,30 @@ describe DataAnon::Strategy::MongoDB::AnonymizeField do
   it 'should do callback recursive in case of sub document' do
     sub_document = {'key' => 'value'}
     field_strategy = {'key' => FieldStrategy::LoremIpsum.new}
-    anonymization_strategy = double("AnonymizationStrategy")
-    anonymization_strategy.should_receive(:anonymize_document).with(sub_document,1,field_strategy).and_return({'key' => "anonymized_value"})
+    anonymization_strategy = double('AnonymizationStrategy')
+    anonymization_strategy.should_receive(:anonymize_document).with(sub_document,1,field_strategy).and_return({'key' => 'anonymized_value'})
     field = DataAnon::Core::Field.new('sub_document_field', sub_document,1,nil)
     anonymize_field = DataAnon::Strategy::MongoDB::AnonymizeField.new(field, field_strategy,anonymization_strategy)
     anonymized_value = anonymize_field.anonymize
-    anonymized_value['key'].should == "anonymized_value"
+    anonymized_value['key'].should == 'anonymized_value'
   end
 
   it 'should do callback recursive multiple time in case of array of sub document' do
     sub_documents = [{'key' => 'value1'},{'key' => 'value2'}]
     field_strategy = {'key' => FieldStrategy::LoremIpsum.new}
-    anonymization_strategy = double("AnonymizationStrategy")
-    anonymization_strategy.should_receive(:anonymize_document).with({'key' => 'value1'},1,field_strategy).and_return({'key' => "anonymized_value1"})
-    anonymization_strategy.should_receive(:anonymize_document).with({'key' => 'value2'},1,field_strategy).and_return({'key' => "anonymized_value2"})
+    anonymization_strategy = double('AnonymizationStrategy')
+    anonymization_strategy.should_receive(:anonymize_document).with({'key' => 'value1'},1,field_strategy).and_return({'key' => 'anonymized_value1'})
+    anonymization_strategy.should_receive(:anonymize_document).with({'key' => 'value2'},1,field_strategy).and_return({'key' => 'anonymized_value2'})
     field = DataAnon::Core::Field.new('sub_document_field', sub_documents,1,nil)
     anonymize_field = DataAnon::Strategy::MongoDB::AnonymizeField.new(field, field_strategy,anonymization_strategy)
     anonymized_value = anonymize_field.anonymize
     anonymized_value.length.should == 2
-    anonymized_value[0]['key'].should == "anonymized_value1"
-    anonymized_value[1]['key'].should == "anonymized_value2"
+    anonymized_value[0]['key'].should == 'anonymized_value1'
+    anonymized_value[1]['key'].should == 'anonymized_value2'
   end
 
   it 'should anonymize array field data type' do
-    anonymization_strategy = double("AnonymizationStrategy")
+    anonymization_strategy = double('AnonymizationStrategy')
     anonymization_strategy.should_not_receive(:anonymize_document)
     field = DataAnon::Core::Field.new('tags',['tag1','tag2'],1,nil)
     anonymize_field = DataAnon::Strategy::MongoDB::AnonymizeField.new(field,FieldStrategy::SelectFromList.new(['tag4','tag5','tag6','tag7','tag8']),anonymization_strategy)
@@ -40,7 +40,7 @@ describe DataAnon::Strategy::MongoDB::AnonymizeField do
   end
 
   it 'should anonymize field and return anonymized value using passed strategy' do
-    anonymization_strategy = double("AnonymizationStrategy")
+    anonymization_strategy = double('AnonymizationStrategy')
     anonymization_strategy.should_not_receive(:anonymize_document)
     field = DataAnon::Core::Field.new('boolean_field',false,1,nil)
     anonymize_field = DataAnon::Strategy::MongoDB::AnonymizeField.new(field,FieldStrategy::RandomBoolean.new,anonymization_strategy)

data/spec/utils/template_helper_spec.rb

@@ -1,14 +1,14 @@
-require "spec_helper"
+require 'spec_helper'
 
-describe "Template Helper" do
+describe 'Template Helper' do
 
-  it "should return a correctly formatted string based on input connection hash for source" do
-    connection_hash = {:adapter => "test_adapter", :port => 5000}
+  it 'should return a correctly formatted string based on input connection hash for source' do
+    connection_hash = {adapter: 'test_adapter', port: 5000}
     DataAnon::Utils::TemplateHelper.source_connection_specs_rdbms(connection_hash).should eq(":adapter => 'test_adapter', :port => 5000")
   end
 
-  it "should return a correctly formatted string based on input connection hash for destination" do
-    connection_hash = {:adapter => "test_adapter", :port => 5000}
+  it 'should return a correctly formatted string based on input connection hash for destination' do
+    connection_hash = {adapter: 'test_adapter', port: 5000}
     DataAnon::Utils::TemplateHelper.destination_connection_specs_rdbms(connection_hash).should eq(":adapter => '<enter_value>', :port => '<enter_value>'")
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: data-anonymization
 version: !ruby/object:Gem::Version
-  version: 0.8.0
+  version: 0.8.1
 platform: ruby
 authors:
 - Sunit Parekh
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-10-31 00:00:00.000000000 Z
+date: 2017-08-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activerecord
@@ -102,28 +102,28 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.9'
+        version: '1.12'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.9'
+        version: '1.12'
 - !ruby/object:Gem::Dependency
   name: thor
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.19'
+        version: '0.20'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.19'
+        version: '0.20'
 description: Data anonymization tool for RDBMS and MongoDB databases
 email:
 - parekh.sunit@gmail.com
@@ -151,6 +151,7 @@ files:
 - examples/mongodb_blacklist_dsl.rb
 - examples/mongodb_whitelist_dsl.rb
 - examples/whitelist_dsl.rb
+- examples/whitelist_dsl_threads.rb
 - lib/core/database.rb
 - lib/core/dsl.rb
 - lib/core/field.rb
@@ -297,7 +298,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.5.1
+rubygems_version: 2.6.12
 signing_key: 
 specification_version: 4
 summary: Tool to create anonymized production data dump to use for performance and