teebo 0.0.3 → 0.0.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: c08b25e131c864ae6ff3222cf665dc2359ad7849
-  data.tar.gz: 7533533ef81909ca782e0d3acfb53dc616592946
+  metadata.gz: 4d25e0acff23d8791b96d50d72b36d326143a2c1
+  data.tar.gz: deb705f45ee7dfc04e34b6b6021a32640ff08c5a
 SHA512:
-  metadata.gz: 61eff7754a95eea9831e654aeee41539062d2b3b77efbf97e684ce0f3248bb196d1a236c5593a85edb25b13873dedbf3049aeb8ab5f0a714cd613ab245e64da1
-  data.tar.gz: 4f96222ca2bfe0dafa081e3cc34f7dfe6c6590db374fd0e82f0c8cc790ea7058d3b79feb6e8b4f410c9a2c71351428e20f6023ba3555167f40897cc31d55f615
+  metadata.gz: bed5a9899872dc9f634824f8614838af9b70c74e4c232041484bd40889376470e8aabadb992854d87343e632839a6c1db4b7773b21db628bfb50cab3a60fda66
+  data.tar.gz: c941cb7f2fc2f4cf7fc6c98b8a2d5a85a8c69aeeb26adf1d74787c914cf6ade5f6ac1e6e585798a394646837c4465ff8ac5f5776192635926517e6cf4cd4a442

data/lib/data/en-us.yml

@@ -0,0 +1,38 @@
+---
+credit-card-issuers:
+  - 
+    name: American Express
+    iin-prefixes:
+      - 34
+      - 37
+    lengths:
+      - 15
+    validation: yes
+    probability: .092
+    cvc-length: 4
+  - 
+    name: Discover Card
+    iin-prefixes: [ 6011, 622126, 622127, 622128, 622129, 622130, 622131, 622132, 622133, 622134, 622135, 622136, 622137, 622138, 622139, 622140, 622141, 622142, 622143, 622144, 622145, 622146, 622147, 622148, 622149, 622150, 622151, 622152, 622153, 622154, 622155, 622156, 622157, 622158, 622159, 622160, 622161, 622162, 622163, 622164, 622165, 622166, 622167, 622168, 622169, 622170, 622171, 622172, 622173, 622174, 622175, 622176, 622177, 622178, 622179, 622180, 622181, 622182, 622183, 622184, 622185, 622186, 622187, 622188, 622189, 622190, 622191, 622192, 622193, 622194, 622195, 622196, 622197, 622198, 622199, 622200, 6222, 6223, 6224, 6225, 6226, 6227, 6228, 622901, 622902, 622903, 622904, 622905, 622906, 622907, 622908, 622909, 622910, 622911, 622912, 622913, 622914, 622915, 622916, 622917, 622918, 622919, 622920, 622921, 622922, 622923, 622924, 622925, 644, 645, 646, 647, 648, 649, 65 ]
+    lengths:
+      - 16
+    validation: yes
+    probability: .107
+    cvc-length: 3
+  - 
+    name: MasterCard
+    iin-prefixes: [ 51, 52, 53, 54, 55 ]
+    lengths:
+      - 16
+    validation: yes
+    probability: .313
+    cvc-length: 3
+  - 
+    name: Visa
+    iin-prefixes:
+      - 4
+    lengths:
+      - 13
+      - 16
+    validation: yes
+    probability: .488
+    cvc-length: 3

data/lib/teebo.rb

@@ -1,16 +1,18 @@
 require 'sqlite3'
+require 'yaml'
+require 'randexp'
 
 module Teebo
 
-  class Base
-
+  class TeeboGenerator
     def initialize
-      # Initialize the database
-      @database = SQLite3::Database.new "lib/data/seed-data.db"
-      @database.results_as_hash = true
+      @db_connection = Teebo::DatabaseHandler.new
+      @yaml_mapping = YAML::load(File.open('lib/data/en-us.yml'))
     end
-
   end
 end
 
+require 'teebo/database_handler'
 require 'teebo/name'
+require 'teebo/number'
+require 'teebo/credit_card'

data/lib/teebo/credit_card.rb

@@ -0,0 +1,41 @@
+module Teebo
+  #
+  # Helper class for generating credit card numbers. Categorizes cards according to their card
+  # issuer, while attempting to maintain a realistic distribution between the issuers. Also ensures
+  # that the credit card numbers are actually valid.
+  #
+  class CreditCard < TeeboGenerator
+
+    def initialize
+      super
+      @cc_issuers = @yaml_mapping['credit-card-issuers']
+    end
+
+    #
+    # Returns a credit card issuer according to the likelihood that it would be seen in the wild.
+    #
+    def get_issuer
+      random_choice = Random.rand
+      full_weight = 0
+      @cc_issuers.each do |issuer|
+        full_weight += issuer['probability']
+        if random_choice < full_weight
+          return issuer
+        end
+      end
+    end
+
+    #
+    # Generates a credit card number according to the pattern specified in the 'issuer' passed in.
+    #
+    def generate_number(issuer)
+      # TODO: Sample according to realistic distribution - numbers w/long prefixes are prioritized too highly right now.
+      prefix = issuer['iin-prefixes'].sample
+      length = issuer['lengths'].sample
+      puts prefix, length
+    end
+
+
+
+  end
+end

data/lib/teebo/database_handler.rb

@@ -0,0 +1,49 @@
+module Teebo
+  # This class facilitates the use of the database that accompanies this gem. Its primary goal is to
+  # simplify querying the database for required data, without having to write raw SQL statements
+  # throughout the codebase.
+  #
+  # Author:: Russ Taylor (mailto:russ@russt.me)
+  class DatabaseHandler
+
+    def initialize
+      @database = SQLite3::Database.new 'lib/data/seed-data.db'
+      @database.results_as_hash = true
+    end
+
+    #
+    # Retrieves the sum of the specified column in the database. Optionally, a 'where' clause may
+    # be specified to restrict the sum to rows meeting a specified condition. Note that at present,
+    # only where clauses with one condition and an '=' comparison are supported.
+    #
+    def get_sum(table_name, column_name, where_clause=nil)
+      where_statement = ''
+      unless where_clause.nil?
+        where_statement = "where #{where_clause[:column]} = '#{where_clause[:condition]}'"
+      end
+      statement = <<-SQL
+        select sum(#{column_name}) from #{table_name} #{where_statement}
+      SQL
+
+      @database.execute(statement)[0][0]
+    end
+
+    #
+    # Retrieves the one row from the database where the specified number fits into this row's
+    # 'count_to' range.
+    #
+    def get_row_for_count(table_name, count_column, count_value, where_clause=nil)
+      where_statement = ''
+      unless where_clause.nil?
+        where_statement = "and #{where_clause[:column]} = '#{where_clause[:condition]}'"
+      end
+      statement = <<-SQL
+        select * from #{table_name} where (#{count_column} - #{count_value}) >= 0
+        #{where_statement}
+        order by id limit 1
+      SQL
+
+      @database.execute(statement)[0]
+    end
+  end
+end

data/lib/teebo/name.rb

@@ -3,17 +3,20 @@ module Teebo
   # Generates names in accordance with their frequency in the United States
   # population.
   #
-  class Name < Base
+  class Name < TeeboGenerator
+
+    GIVEN_NAMES_TABLE = 'given_names'
+    SURNAMES_TABLE = 'surnames'
 
     #
     # Picks a random first & last name, selecting a random gender if it's not
     # specified.
     #
-    def name sex=nil
+    def name (sex=nil)
       if sex.nil?
-        sex = ['M', 'F'].sample
+        sex = %w(M F).sample
       end
-      given_name(sex) + " " + surname
+      given_name(sex) + ' ' + surname
     end
 
     #
@@ -23,52 +26,38 @@ module Teebo
     # simply be another given name of the same gender is almost certainly less
     # than 100%.
     #
-    # TODO: Make this take into account different probablities of different
-    #       types of middle names.
-    #
-    def full_name sex=nil
+    def full_name(sex=nil)
+      # TODO: Take into account different probabilities of different types of middle names.
       if sex.nil?
-        sex = ['M', 'F'].sample
+        sex = %w(M F).sample
       end
-      given_name(sex) + " " + given_name(sex) + " " + surname
+      given_name(sex) + ' ' + given_name(sex) + ' ' + surname
     end
 
     #
     # Finds the total count for the number of names in the database.
     #
-    def sum_count sex
-      find_count = <<-SQL
-        select sum(count) from 'given_names' where sex = ?
-      SQL
-      @database.execute(find_count, sex)[0][0]
+    def sum_count(sex)
+      @db_connection.get_sum(GIVEN_NAMES_TABLE, 'count', {column: 'sex', condition: sex})
     end
 
     #
     # Selects a random (weighted) given name from the database.
     #
-    def given_name sex
-      find_range_query = <<-SQL
-        select * from given_names where sex = ?
-          and (count_to - ?) >= 0
-          order by id limit 1
-      SQL
-
-      count = sum_count sex
+    def given_name(sex)
+      count = sum_count(sex)
       selection = rand(count)
-      @database.execute(find_range_query, sex, selection)[0]['name']
+      @db_connection.get_row_for_count(GIVEN_NAMES_TABLE, 'count_to', selection,
+                                {column: 'sex', condition: sex})['name']
     end
 
     #
     # Selects a random (weighted) surname from the database.
     #
     def surname
-      find_range_query = <<-SQL
-        select * from surnames where (count_to - ?) >= 0 order by id limit 1
-      SQL
-
-      count = @database.execute('select sum(count) from surnames')[0][0]
+      count = @db_connection.get_sum(SURNAMES_TABLE, 'count')
       selection = rand(count)
-      @database.execute(find_range_query, selection)[0]['name']
+      @db_connection.get_row_for_count(SURNAMES_TABLE, 'count_to', selection)['name']
     end
   end
 end

data/lib/teebo/number.rb

@@ -0,0 +1,28 @@
+module Teebo
+  #
+  # Helper for generating  numbers according to various distributions.
+  #
+  class Number
+    #
+    # Generates a number using a normal distribution.
+    #
+    # A basic implementation of the Box-Muller transform. Adapted from an answer by antonakos on
+    # Stack Overflow here: http://stackoverflow.com/questions/5825680
+    #
+    def normal_dist(mean, std_deviation, rand = lambda { Kernel.rand })
+      theta = 2 * Math::PI * rand.call
+      rho = Math.sqrt(-2 * Math.log(1 - rand.call))
+      scale = std_deviation * rho
+      mean + scale * Math.cos(theta)
+    end
+
+    #
+    # Generates a number according to Benford's law, meaning that it is more indicative of numbers
+    # encountered in real life.
+    #
+    def benford_dist(upper_bound, decimals=0)
+      (upper_bound**Random.rand).round(decimals)
+    end
+
+  end
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: teebo
 version: !ruby/object:Gem::Version
-  version: 0.0.3
+  version: 0.0.4
 platform: ruby
 authors:
 - Russ Taylor
@@ -30,15 +30,33 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: 1.3.9
+- !ruby/object:Gem::Dependency
+  name: randexp
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.1.7
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.1.7
 description: Contains basic functionality for data generation.
 email: russ@russt.me
 executables: []
 extensions: []
 extra_rdoc_files: []
 files:
+- lib/data/en-us.yml
 - lib/data/seed-data.db
 - lib/teebo.rb
+- lib/teebo/credit_card.rb
+- lib/teebo/database_handler.rb
 - lib/teebo/name.rb
+- lib/teebo/number.rb
 homepage: http://github.com/russtaylor/teebo
 licenses:
 - MIT