edits 0.1.0 → 0.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: a437de45c8830ea622a9b57b84b3e92c54a58178
-  data.tar.gz: 36eff87bb3ba0e832f2cd908cb0bafe358e05a19
+  metadata.gz: a6cd424759a87d827084b94756bebd46a2cce133
+  data.tar.gz: 703ee761905db5db81c9462727c7812b4519ff8b
 SHA512:
-  metadata.gz: bf9474115aacaff6cb0411c85cf8f81409d6cfaff793ef656821fc1fc33ae0d5315aa6af98c8dbe17a68aa38514b8adef134824e2ea6fd8c1434aea372e37b6c
-  data.tar.gz: 71199e4c9e67f4b78b90e21368dfd1162cd660fa38b01e40ea412e1a37aa29529d57ee78f1e5584f5a6319a22b2134165b638335b0e75b163650aaadaa9d67ae
+  metadata.gz: ccee510e44ace7a1c88362dcd8a15fda256d652d9e65e78def3a28b95c3c30fded7ff2e1205b302f56d38f4365fe3f4a678dfec9b5bca2d441b08b00401bd390
+  data.tar.gz: 4b56c06fba560443d7aa2c057d959d271a013216d4e45867c688dda07bad18acce8e8730e6513384c3f03aa0b0985b9333c49e942d4e2bb0ee35b2239ea3abc1

data/.travis.yml

@@ -1,5 +1,14 @@
 sudo: false
 language: ruby
+cache: bundler
 rvm:
   - 2.4.0
 before_install: gem install bundler -v 1.15.4
+deploy:
+  provider: rubygems
+  api_key:
+    secure: NKUWS12yMpsMtPeKL+4X56xZ02p5y4R+HYQmgvxFNajOS42aTMXncx4eTnuHoNdKbh+x5yci+lbPr+j5t9qWo5GzJsxjqnFW4lO2V5O1ONBzcAB/g6BPDMQsWbX34eBVafCEidNoOY1HhLlKerWzlsGRq+P4q+3WKLJihrl99pyv+EQykqS11/YVCsmFjrPVa/aGk84njbnI/kIDOY5HV2gVBazARixRO5y2AUg3hRUf2+Tu/X1ke80YBD9LazW2kRUzz0Rs+1vc653JtYn6MeK/bEkGWvN2Qs/k6Q2nNr6ni4v39Y07yylu1EnYh1H/0OT/H2hehxsjMQCQhDlVKbN7NTmOV026aWXA2HdSxmVhxQCIKRGW3Nm81kBj1/edXLpJActnLeex2iCMcXJk8yAAzF0q+vSHLld0w9Jx95kIJB4tnaonWJxcaWaX58HWFbdOuYKvQhXqcflI4KmNH8xXm/O0FIM8VEJRg9dojZ5S8Us3fZpBFZxVJ3H3Fcb406AmoIqcHOsqJ1GvBM8EdWkwuaH9GsUWf8pydgKFgStYUaKk8DDmJonT748emG4yw+78uMGWPoFxf+Mc8jazRxIaRiQbjVFUcLjaRkRFlL8UT9BB9k7c1egjvKhUo/pqV7KIEsAJrrh6zZIkz4h9AwxWpZjVj9+z63Kh6NC1lUI=
+  gem: edits
+  on:
+    tags: true
+    repo: tcrouch/edits

data/README.md

@@ -1,8 +1,13 @@
 # Edits
 
+[![Build Status](https://travis-ci.org/tcrouch/edits.svg?branch=master)](https://travis-ci.org/tcrouch/edits)
+[![Code Climate](https://codeclimate.com/github/tcrouch/edits/badges/gpa.svg)](https://codeclimate.com/github/tcrouch/edits)
+[![Inline docs](http://inch-ci.org/github/tcrouch/edits.svg?branch=master)](http://inch-ci.org/github/tcrouch/edits)
+[![Yard Docs](http://img.shields.io/badge/yard-docs-blue.svg)](http://rubydoc.info/github/tcrouch/edits)
+
 A collection of edit distance algorithms in Ruby.
 
-Includes Levenshtein, Restricted Edit (Optimal Alignment) and Damerau-Levenshtein distances, and Jaro and Jaro-Winkler similarity.
+Includes Levenshtein, Restricted Edit (Optimal Alignment) and Damerau-Levenshtein distances, and Jaro & Jaro-Winkler similarity.
 
 ## Installation
 
@@ -37,14 +42,15 @@ Edits::Levenshtein.distance "acer", "earn"
 # Max distance
 Edits::Levenshtein.distance_with_max "iota", "atom", 2
 # => 2
-Edits::Levenshtein.most_similar "atom", %w[tram atlas rota racer]
-# => "atlas"
+Edits::Levenshtein.most_similar "atom", %w[tree rota toes racer]
+# => "toes"
 ```
 
 ### Restricted Edit (Optimal Alignment)
 
-Edit distance, accounting for deletion, addition, substitution and swapped
-characters.
+Edit distance, accounting for deletion, addition, substitution and
+transposition (two adjacent characters are swapped). This variant is
+restricted by the condition that no sub-string is edited more than once.
 
 ```ruby
 Edits::RestrictedEdit.distance "raked", "bakers"
@@ -53,12 +59,18 @@ Edits::RestrictedEdit.distance "iota", "atom"
 # => 3
 Edits::RestrictedEdit.distance "acer", "earn"
 # => 4
+
+# Max distance
+Edits::RestrictedEdit.distance_with_max "iota", "atom", 2
+# => 2
+Edits::RestrictedEdit.most_similar "atom", %w[tree rota toes racer]
+# => "rota"
 ```
 
 ### Damerau-Levenshtein
 
 Edit distance, accounting for deletions, additions, substitution and
-transposition.
+transposition (two adjacent characters are swapped).
 
 ```ruby
 Edits::DamerauLevenshtein.distance "raked", "bakers"

data/lib/edits.rb

@@ -2,6 +2,7 @@
 
 require "edits/version"
 
+require "edits/compare"
 require "edits/damerau_levenshtein"
 require "edits/hamming"
 require "edits/jaro"

data/lib/edits/compare.rb

@@ -0,0 +1,36 @@
+# frozen_string_literal: true
+
+module Edits
+  # Comparison helpers
+  module Compare
+    # Given a prototype string and an array of strings, determines which
+    # string is most similar to the prototype.
+    #
+    # `most_similar("foo", strings)` is functionally equivalent to
+    # `strings.min_by { |s| distance("foo", s) }`, leveraging
+    # {.distance_with_max}.
+    #
+    # @example
+    #   most_similar("atom", %w[tram atlas rota racer])
+    #   # => "atlas"
+    # @param prototype [String]
+    # @param strings [<String>]
+    # @return [String, nil] most similar string, or nil for empty array
+    def most_similar(prototype, strings)
+      return nil if strings.empty?
+      min_s = strings[0]
+      min_d = distance(prototype, min_s)
+
+      strings[1..-1].each do |s|
+        return min_s if min_d.zero?
+        d = distance_with_max(prototype, s, min_d)
+        if d < min_d
+          min_d = d
+          min_s = s
+        end
+      end
+
+      min_s
+    end
+  end
+end

data/lib/edits/damerau_levenshtein.rb

@@ -1,13 +1,13 @@
 # frozen_string_literal: true
 
 module Edits
-  # Implemention of the Damerau/Levenshtein distance algorithm.
+  # Implements the Damerau/Levenshtein distance algorithm.
   #
   # Determines distance between two strings by counting edits, identifying:
   # * Insertion
   # * Deletion
   # * Substitution
-  # * Transposition
+  # * Adjacent transposition
   module DamerauLevenshtein
     # Calculate the Damerau/Levenshtein distance of two sequences.
     #
@@ -18,13 +18,9 @@ module Edits
     # @param seq2 [String, Array]
     # @return [Integer]
     def self.distance(seq1, seq2)
-      if seq1.length > seq2.length
-        temp = seq1
-        seq1 = seq2
-        seq2 = temp
-      end
+      seq1, seq2 = seq2, seq1 if seq1.length > seq2.length
 
-      # array of Integer codepoints outperforms String
+      # array of codepoints outperforms String
       seq1 = seq1.codepoints if seq1.is_a? String
       seq2 = seq2.codepoints if seq2.is_a? String
 
@@ -34,7 +30,7 @@ module Edits
       return rows if cols.zero?
 
       # 'infinite' edit distance for padding cost matrix.
-      # Can be any value greater than max[rows, cols]
+      # Can be any value > max[rows, cols]
       inf = rows + cols
 
       # Initialize first two rows of cost matrix.
@@ -71,14 +67,14 @@ module Edits
 
           # TODO: do insertion/deletion need to be considered when
           # seq1_item == seq2_item ?
-          deletion = matrix[row][col + 1] + 1
-          insertion = matrix[row + 1][col] + 1
-          substitution = matrix[row][col] + sub_cost
-
-          # step cost is min of operation costs
-          cost = substitution < insertion ? substitution : insertion
-          cost = deletion if deletion < cost
-          cost = transposition if transposition < cost
+          #
+          # substitution, deletion, insertion, transposition
+          cost = [
+            matrix[row][col] + sub_cost,
+            matrix[row][col + 1] + 1,
+            matrix[row + 1][col] + 1,
+            transposition
+          ].min
 
           matrix[row + 1][col + 1] = cost
 

data/lib/edits/hamming.rb

@@ -1,6 +1,8 @@
 # frozen_string_literal: true
 
 module Edits
+  # Implements Hamming distance algorithm
+  #
   # @see https://en.wikipedia.org/wiki/Hamming_distance
   module Hamming
     # Calculate the Hamming distance between two sequences.

data/lib/edits/jaro.rb

@@ -1,6 +1,8 @@
 # frozen_string_literal: true
 
 module Edits
+  # Implements Jaro similarity algorithm.
+  #
   # @see https://en.wikipedia.org/wiki/Jaro-Winkler_distance
   module Jaro
     # Calculate Jaro similarity
@@ -14,7 +16,7 @@ module Edits
     #   # => 0.9023569023569024
     # @param seq1 [String, Array]
     # @param seq2 [String, Array]
-    # @return [Float]
+    # @return [Float] similarity, between 0.0 (none) and 1.0 (identical)
     def self.similarity(seq1, seq2)
       return 1.0 if seq1 == seq2
       return 0.0 if seq1.empty? || seq2.empty?
@@ -26,16 +28,18 @@ module Edits
       return 0.0 if m.zero?
 
       m = m.to_f
-      (1.0 / 3) * ((m / seq1.length) + (m / seq2.length) + ((m - t) / m))
+      ((m / seq1.length) + (m / seq2.length) + ((m - t) / m)) / 3
     end
 
     # Calculate Jaro distance
     #
+    # `Dj = 1 - Sj`
+    #
     # @example
     #   Edits::Jaro.distance("information", "informant")
     #   # => 0.09764309764309764
     # @param (see #distance)
-    # @return [Float]
+    # @return [Float] distance, between 0.0 (identical) and 1.0 (distant)
     def self.distance(str1, str2)
       1.0 - similarity(str1, str2)
     end
@@ -45,11 +49,7 @@ module Edits
     # @param (see #distance)
     # @return [(Integer, Integer)] matches and transpositions
     def self.jaro_matches(seq1, seq2)
-      if seq1.length > seq2.length
-        temp = seq1
-        seq1 = seq2
-        seq2 = temp
-      end
+      seq1, seq2 = seq2, seq1 if seq1.length > seq2.length
 
       # search range: (max(|A|, |B|) / 2) - 1
       range = (seq2.length / 2) - 1

data/lib/edits/jaro_winkler.rb

@@ -1,6 +1,8 @@
 # frozen_string_literal: true
 
 module Edits
+  # Implements Jaro-Winkler similarity algorithm.
+  #
   # @see https://en.wikipedia.org/wiki/Jaro-Winkler_distance
   module JaroWinkler
     # Prefix scaling factor for jaro-winkler metric. Default is 0.1
@@ -13,9 +15,9 @@ module Edits
 
     # Calculate Jaro-Winkler similarity of given strings
     #
-    # Adds weight to Jaro distance according to the length of a common prefix
+    # Adds weight to Jaro similarity according to the length of a common prefix
     # of up to 4 letters, where exists. The additional weighting is only
-    # applied when the original distance passes a threshold.
+    # applied when the original similarity passes a threshold.
     #
     # `Sw = Sj + (l * p * (1 - Dj))`
     #
@@ -29,7 +31,7 @@ module Edits
     # @param seq2 [String, Array]
     # @param threshold [Float] threshold for applying Winkler prefix weighting
     # @param weight [Float] weighting for common prefix, should not exceed 0.25
-    # @return [Float]
+    # @return [Float] similarity, between 0.0 (none) and 1.0 (identical)
     def self.similarity(
       seq1, seq2,
       threshold: WINKLER_THRESHOLD,
@@ -60,7 +62,7 @@ module Edits
     #   Edits::JaroWinkler.distance("information", "informant")
     #   # => 0.05858585858585863
     # @param (see #distance)
-    # @return [Float]
+    # @return [Float] distance, between 0.0 (identical) and 1.0 (distant)
     def self.distance(
       seq1, seq2,
       threshold: WINKLER_THRESHOLD,

data/lib/edits/levenshtein.rb

@@ -1,13 +1,15 @@
 # frozen_string_literal: true
 
 module Edits
-  # Implementation of Levenshtein distance algorithm.
+  # Implements Levenshtein distance algorithm.
   #
   # Determines distance between two string by counting edits, identifying:
-  # - Insertion
-  # - Deletion
-  # - Substitution
+  # * Insertion
+  # * Deletion
+  # * Substitution
   module Levenshtein
+    extend Compare
+
     # Calculate the Levenshtein (edit) distance of two sequences.
     #
     # @note A true distance metric, satisfies triangle inequality.
@@ -18,13 +20,9 @@ module Edits
     # @param seq2 [String, Array]
     # @return [Integer]
     def self.distance(seq1, seq2)
-      if seq1.length > seq2.length
-        temp = seq1
-        seq1 = seq2
-        seq2 = temp
-      end
+      seq1, seq2 = seq2, seq1 if seq1.length > seq2.length
 
-      # array of Integer codepoints outperforms String
+      # array of codepoints outperforms String
       seq1 = seq1.codepoints if seq1.is_a? String
       seq2 = seq2.codepoints if seq2.is_a? String
 
@@ -41,24 +39,25 @@ module Edits
       last_row = 0.upto(cols).to_a
 
       rows.times do |row|
-        last_col = row + 1
-
+        prev_col_cost = row + 1
         seq1_item = seq1[row]
 
         cols.times do |col|
-          deletion = last_row[col + 1] + 1
-          insertion = last_col + 1
-          substitution = last_row[col] + (seq1_item == seq2[col] ? 0 : 1)
-
+          # | Xs | Xd |
+          # | Xi | ?  |
           # step cost is min of operation costs
-          cost = deletion < insertion ? deletion : insertion
-          cost = substitution if substitution < cost
+          # substitution, deletion, insertion
+          cost = [
+            last_row[col] + (seq1_item == seq2[col] ? 0 : 1),
+            last_row[col + 1] + 1,
+            prev_col_cost + 1
+          ].min
 
           # overwrite previous row as we progress
-          last_row[col] = last_col
-          last_col = cost
+          last_row[col] = prev_col_cost
+          prev_col_cost = cost
         end
-        last_row[cols] = last_col
+        last_row[cols] = prev_col_cost
       end
 
       last_row[cols]
@@ -77,85 +76,56 @@ module Edits
     # @param max [Integer] maximum distance
     # @return [Integer]
     def self.distance_with_max(seq1, seq2, max)
-      if seq1.length > seq2.length
-        temp = seq1
-        seq1 = seq2
-        seq2 = temp
-      end
+      seq1, seq2 = seq2, seq1 if seq1.length > seq2.length
 
       rows = seq1.length
       cols = seq2.length
-      return cols if rows.zero?
-      return rows if cols.zero?
-      return max if (rows - cols).abs >= max
+      return cols > max ? max : cols if rows.zero?
+      return rows > max ? max : rows if cols.zero?
+      return max if (cols - rows) >= max
 
+      # array of codepoints outperforms String
       seq1 = seq1.codepoints if seq1.is_a? String
       seq2 = seq2.codepoints if seq2.is_a? String
 
+      # 'infinite' edit distance for padding cost matrix.
+      # Can be any value > max[rows, cols]
+      inf = cols + 1
+
+      # retain previous row of cost matrix
       last_row = 0.upto(cols).to_a
 
       rows.times do |row|
-        last_col_cost = row + 1
-        seq1_item = seq1[row]
-
+        # Ukkonen cut-off
         min_col = row > max ? row - max : 0
         max_col = row + max
         max_col = cols - 1 if max_col > cols - 1
+
+        prev_col_cost = min_col.zero? ? row + 1 : inf
+        seq1_item = seq1[row]
         diagonal = cols - rows + row
 
-        cols.times do |col|
+        min_col.upto(max_col) do |col|
           return max if diagonal == col && last_row[col] >= max
-          col_cost =
-            if col < min_col || col > max_col
-              max + 1
-            else
-              # step cost is min of operation costs
-              deletion = last_row[col + 1] + 1
-              insertion = last_col_cost + 1
-              substitution = last_row[col] + (seq1_item == seq2[col] ? 0 : 1)
-
-              cost = deletion < insertion ? deletion : insertion
-              substitution < cost ? substitution : cost
-            end
-
-          last_row[col] = last_col_cost
-          last_col_cost = col_cost
-        end
-
-        last_row[cols] = last_col_cost
-      end
 
-      last_row[cols] > max ? max : last_row[cols]
-    end
+          # | Xs | Xd |
+          # | Xi | ?  |
+          # substitution, deletion, insertion
+          cost = [
+            last_row[col] + (seq1_item == seq2[col] ? 0 : 1),
+            last_row[col + 1] + 1,
+            prev_col_cost + 1
+          ].min
 
-    # Given a prototype string and an array of strings, determines which
-    # string is most similar to the prototype.
-    #
-    # `Levenshtein.most_similar("foo", strings)` is functionally equivalent to
-    # `strings.min_by { |s| Levenshtein.distance("foo", s) }`, leveraging
-    # {.distance_with_max}.
-    #
-    # @example
-    #   Edits::Levenshtein.most_similar("atom", %w[tram atlas rota racer])
-    #   # => "atlas"
-    # @param prototype [String]
-    # @param strings [<String>]
-    # @return [String, nil] most similar string, or nil for empty array
-    def self.most_similar(prototype, strings)
-      return nil if strings.empty?
-      min_s = strings[0]
-      min_d = distance(prototype, min_s)
-
-      strings[1..-1].each do |s|
-        return min_s if min_d.zero?
-        d = distance_with_max(prototype, s, min_d)
-        if d < min_d
-          min_d = d
-          min_s = s
+          # overwrite previous row as we progress
+          last_row[col] = prev_col_cost
+          prev_col_cost = cost
         end
+
+        last_row[cols] = prev_col_cost
       end
 
-      min_s
+      last_row[cols] > max ? max : last_row[cols]
     end
   end
 end

data/lib/edits/restricted_edit.rb

@@ -8,8 +8,13 @@ module Edits
   # * Insertion
   # * Deletion
   # * Substitution
-  # * Swapped items
+  # * Adjacent transposition
+  #
+  # This variant is restricted by the condition that no sub-string is edited
+  # more than once.
   module RestrictedEdit
+    extend Compare
+
     # Calculate the Restricted Damerau-Levenshtein distance (Optimal Alignment)
     # of two sequences.
     #
@@ -21,13 +26,9 @@ module Edits
     # @param seq2 [String, Array]
     # @return [Integer]
     def self.distance(seq1, seq2)
-      if seq1.length > seq2.length
-        temp = seq1
-        seq1 = seq2
-        seq2 = temp
-      end
+      seq1, seq2 = seq2, seq1 if seq1.length > seq2.length
 
-      # array of Integer codepoints outperforms String
+      # array of codepoints outperforms String
       seq1 = seq1.codepoints if seq1.is_a? String
       seq2 = seq2.codepoints if seq2.is_a? String
 
@@ -36,9 +37,10 @@ module Edits
       return cols if rows.zero?
       return rows if cols.zero?
 
-      # previous two rows of cost matrix are retained
+      # retain previous two rows of cost matrix
       lastlast_row = []
       last_row = []
+
       # Initialize first row of cost matrix.
       # The full initial state where cols=3, rows=2 would be:
       #   [[0, 1, 2, 3],
@@ -47,29 +49,29 @@ module Edits
       curr_row = 0.upto(cols).to_a
 
       rows.times do |row|
-        lastlast_row = last_row
-        last_row = curr_row
+        # rotate row arrays
+        curr_row, last_row, lastlast_row = lastlast_row, curr_row, last_row
 
-        # generate next row of cost matrix
-        curr_row = Array.new(cols + 1, 0)
         curr_row[0] = row + 1
-
-        curr_item = seq1[row]
+        seq1_item = seq1[row]
 
         cols.times do |col|
-          sub_cost = curr_item == seq2[col] ? 0 : 1
-          is_swap = sub_cost == 1 &&
+          sub_cost = seq1_item == seq2[col] ? 0 : 1
+          is_swap = sub_cost.positive? &&
             row.positive? && col.positive? &&
-            curr_item == seq2[col - 1] &&
+            seq1_item == seq2[col - 1] &&
             seq1[row - 1] == seq2[col]
 
-          deletion = last_row[col + 1] + 1
-          insertion = curr_row[col] + 1
-          substitution = last_row[col] + sub_cost
-
+          # | Xt |    |    |
+          # |    | Xs | Xd |
+          # |    | Xi | ?  |
           # step cost is min of operation costs
-          cost = deletion < insertion ? deletion : insertion
-          cost = substitution if substitution < cost
+          # substitution, deletion, insertion, transposition
+          cost = [
+            last_row[col] + sub_cost,
+            last_row[col + 1] + 1,
+            curr_row[col] + 1
+          ].min
 
           if is_swap
             swap = lastlast_row[col - 1] + 1
@@ -82,5 +84,84 @@ module Edits
 
       curr_row[cols]
     end
+
+    # Calculate the Restricted Damerau-Levenshtein distance (Optimal Alignment)
+    # of two sequences, bounded by a maximum value.
+    #
+    # @example
+    #   Edits::RestrictedEdit.distance("cloud", "crayon")
+    #   # => 5
+    #   Edits::RestrictedEdit.distance_with_max("cloud", "crayon", 2)
+    #   # => 2
+    # @param seq1 [String, Array]
+    # @param seq2 [String, Array]
+    # @param max [Integer] maximum distance
+    # @return [Integer]
+    def self.distance_with_max(seq1, seq2, max)
+      seq1, seq2 = seq2, seq1 if seq1.length > seq2.length
+
+      rows = seq1.length
+      cols = seq2.length
+      return cols > max ? max : cols if rows.zero?
+      return rows > max ? max : rows if cols.zero?
+      return max if (cols - rows) >= max
+
+      # array of codepoints outperforms String
+      seq1 = seq1.codepoints if seq1.is_a? String
+      seq2 = seq2.codepoints if seq2.is_a? String
+
+      # 'infinite' edit distance for padding cost matrix.
+      # Can be any value > max[rows, cols]
+      inf = cols + 1
+
+      # retain previous two rows of cost matrix,
+      # padded with "inf" as matrix is not fully evaluated
+      lastlast_row = Array.new(inf, inf)
+      last_row = Array.new(inf, inf)
+      curr_row = 0.upto(cols).to_a
+
+      rows.times do |row|
+        # rotate row arrays
+        curr_row, last_row, lastlast_row = lastlast_row, curr_row, last_row
+
+        # Ukkonen cut-off
+        min_col = row > max ? row - max : 0
+        max_col = row + max
+        max_col = cols - 1 if max_col > cols - 1
+
+        curr_row[min_col] = min_col.zero? ? row + 1 : inf
+        seq1_item = seq1[row]
+        diagonal = cols - rows + row
+
+        min_col.upto(max_col) do |col|
+          return max if diagonal == col && last_row[col] >= max
+
+          sub_cost = seq1_item == seq2[col] ? 0 : 1
+          is_swap = sub_cost.positive? &&
+            row.positive? && col.positive? &&
+            seq1_item == seq2[col - 1] &&
+            seq1[row - 1] == seq2[col]
+
+          # | Xt |    |    |
+          # |    | Xs | Xd |
+          # |    | Xi | ?  |
+          # substitution, deletion, insertion, transposition
+          cost = [
+            last_row[col] + sub_cost,
+            last_row[col + 1] + 1,
+            curr_row[col] + 1
+          ].min
+
+          if is_swap
+            swap = lastlast_row[col - 1] + 1
+            cost = swap if swap < cost
+          end
+
+          curr_row[col + 1] = cost
+        end
+      end
+
+      curr_row[cols] > max ? max : curr_row[cols]
+    end
   end
 end

data/lib/edits/version.rb

@@ -1,5 +1,6 @@
 # frozen_string_literal: true
 
 module Edits
-  VERSION = "0.1.0"
+  # Current gem version
+  VERSION = "0.2.0"
 end

data/tasks/benchmark/levenshtein.rake

@@ -5,7 +5,7 @@ require "benchmark/ips"
 require "edits"
 
 namespace :benchmark do
-  desc "distance vs. distance_with_max (x100)"
+  desc "levenshtein distance vs. distance_with_max (x100)"
   task :lev_max do
     words = File.read("/usr/share/dict/words")
       .split(/\n/).compact.shuffle(random: Random.new(1))
@@ -64,6 +64,65 @@ namespace :benchmark do
     end
   end
 
+  desc "restricted distance vs. distance_with_max (x100)"
+  task :restricted_max do
+    words = File.read("/usr/share/dict/words")
+      .split(/\n/).compact.shuffle(random: Random.new(1))
+      .take(101)
+
+    Benchmark.ips do |x|
+      x.report("distance") do
+        words.each_cons(2) do |a, b|
+          Edits::RestrictedEdit.distance a, b
+        end
+      end
+
+      x.report("with max 1") do
+        words.each_cons(2) do |a, b|
+          Edits::RestrictedEdit.distance_with_max a, b, 1
+        end
+      end
+
+      x.report("with max 2") do
+        words.each_cons(2) do |a, b|
+          Edits::RestrictedEdit.distance_with_max a, b, 2
+        end
+      end
+
+      x.report("with max 3") do
+        words.each_cons(2) do |a, b|
+          Edits::RestrictedEdit.distance_with_max a, b, 3
+        end
+      end
+
+      x.report("with max 4") do
+        words.each_cons(2) do |a, b|
+          Edits::RestrictedEdit.distance_with_max a, b, 4
+        end
+      end
+
+      x.report("with max 6") do
+        words.each_cons(2) do |a, b|
+          Edits::RestrictedEdit.distance_with_max a, b, 6
+        end
+      end
+
+      x.report("with max 8") do
+        words.each_cons(2) do |a, b|
+          Edits::RestrictedEdit.distance_with_max a, b, 8
+        end
+      end
+
+      x.report("with max 50") do
+        words.each_cons(2) do |a, b|
+          Edits::RestrictedEdit.distance_with_max a, b, 100
+        end
+      end
+
+      x.compare!
+    end
+  end
+
   desc "most_similar vs. min_by (100 words)"
   task :lev_similar do
     words = File.read("/usr/share/dict/words")

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: edits
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.2.0
 platform: ruby
 authors:
 - Tom Crouch
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2017-09-22 00:00:00.000000000 Z
+date: 2017-10-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -114,6 +114,7 @@ files:
 - bin/setup
 - edits.gemspec
 - lib/edits.rb
+- lib/edits/compare.rb
 - lib/edits/damerau_levenshtein.rb
 - lib/edits/hamming.rb
 - lib/edits/jaro.rb
@@ -143,7 +144,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.6.8
+rubygems_version: 2.6.13
 signing_key: 
 specification_version: 4
 summary: A collection of edit distance algorithms.