symspell 0.0.1 → 0.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 1460d117d70b607f1e3cd2a82e1775239dcf6409
-  data.tar.gz: 077eec846704a48345ffe305d4491ec85c40da57
+  metadata.gz: 4c6abf1253f5e16d82ae589bc6e35a6f5d72bc36
+  data.tar.gz: 6c9344ec4d40597700099e119b423ccfb85d3582
 SHA512:
-  metadata.gz: d60beee5581b45443ed882fce50add34f6a4b26cf6792996156d2e14b396d31246708a9a30ded49baf7f0f35b28e877fdd4409bbff7020b4006f15b97c84e372
-  data.tar.gz: 89497ab96606792069f18ca01bebdaf8e4aaeb68cc555053b5e0352a3656ef70faf6443803806630d73926a1abda41d74ed640366ffbe469f3cd12704bfcb541
+  metadata.gz: 2f71318a0ded808acf0f3eac41d8b569edd908ec0b81f126421fcfa7d206fce6969bc47e565fcdc324e1b4fb02f492e482e52e36b608559cf4364760a1847baf
+  data.tar.gz: 62215df1ac0378e26fe131ed8c730ebf3ad235107858bebf37802704bb8f8dd8a88c0c15f9a15da02518e7034cdeaa34802c24d03b38e31d8dc834c86254c03f

data/README.md

@@ -11,7 +11,7 @@ This is a straight port of SymSpell from C# to Ruby. I've started moving things
 
 Original source with inline comments and README is here: https://github.com/wolfgarbe/symspell.
 
-I've changed very little from the original source (apart from removing the commandline interface) but please note it has no test coverage at this time.
+I've changed very little from the original source (apart from removing the commandline interface) but please note it has only some very basic end to end tests at this time.
 
 
 ## Usage
@@ -20,13 +20,19 @@ I've changed very little from the original source (apart from removing the comma
 
     require 'symspell'
 
-    speller = SymSpell.new <EDIT_DISTANCE_MAX>
-    speller.create_dictionary('words.txt')
-    speller.lookup('something')
+    speller = SymSpell.new <EDIT_DISTANCE_MAX> <VERBOSE>
+    speller.create_dictionary %w(joe jo mark john peter mary andrew imogen)
+    speller.lookup 'jo'
 
-## EDIT_DISTANCE_MAX
+### EDIT_DISTANCE_MAX
 
 `EDIT_DISTANCE_MAX` is the number of operations needed to tranform one string into another.
 
 For example the edit distance between **CA** and **ABC** is 2 because **CA** => **AC** => **ABC**. Edit distances of 2-5 are normal. Note, however, increasing EDIT_DISTANCE_MAX exponentially increases the combinations and therefore the time it takes to create the dictionary.
 
+### VERBOSE
+
+* 0 - Return the top suggestion
+* 1 - Return the suggestions with the lowest edit distance
+* 2 - Return all suggestions
+

data/Rakefile

@@ -1,9 +1,9 @@
 require 'rake/testtask'
 
 desc 'Test, build and install the gem'
-task :default => [:spec, :install]
+task :default => [:test]
 
-Rake::TestTask.new(:spec) do |t|
+Rake::TestTask.new(:test) do |t|
   t.pattern = 'tests/*_test.rb'
 end
 

data/lib/symspell.rb

@@ -4,8 +4,9 @@ require 'set'
 class SymSpell
   MAX_INT = 2**30 - 1
 
-  def initialize(edit_distance_max)
+  def initialize(edit_distance_max, verbose)
     @edit_distance_max = edit_distance_max
+    @verbose = verbose
     @maxlength = 0
     @dictionary = {}
     @wordlist = []
@@ -14,7 +15,7 @@ class SymSpell
   def create_dictionary(corpus)
     word_count = 0
 
-    File.open(corpus, 'r').each_line do |word|
+    corpus.each do |word|
       word_count += 1 if create_dictionary_entry(word.strip)
     end
   end
@@ -35,7 +36,7 @@ class SymSpell
     while (candidates.count > 0)
       candidate = candidates.shift
 
-      return sort(suggestions) if ((suggestions.count > 0) && (input.size - candidate.size > suggestions[0].distance))
+      return sort(suggestions) if @verbose < 2 && suggestions.count > 0 && (input.size - candidate.size) > suggestions[0].distance
 
       if valueo = @dictionary[candidate]
         value = DictionaryItem.new
@@ -51,7 +52,7 @@ class SymSpell
           si.count = value.count
           si.distance = input.size - candidate.size
           suggestions << si
-          return sort(suggestions) if input.size - candidate.size == 0
+          return sort(suggestions) if @verbose < 2 && input.size - candidate.size == 0
         end
 
         value2 = nil
@@ -85,12 +86,8 @@ class SymSpell
               end
             end
 
-            if suggestions.count > 0 && suggestions[0].distance > distance
-              suggestions.clear
-            end
-            if suggestions.count > 0 && distance > suggestions[0].distance
-              next
-            end
+            suggestions.clear if @verbose < 2  && suggestions.count > 0 && suggestions[0].distance > distance
+            next if @verbose < 2  && suggestions.count > 0 && distance > suggestions[0].distance
 
             if (distance <= @edit_distance_max)
               if value2 = @dictionary[suggestion]
@@ -200,9 +197,12 @@ class SymSpell
   end
 
   def add_lowest_distance(item, suggestion, suggestionint, delete)
-    if item.suggestions.count > 0 && @wordlist[item.suggestions[0]].size - delete.size > suggestion.size - delete.size
+    if @verbose < 2 && item.suggestions.count > 0 && @wordlist[item.suggestions[0]].size - delete.size > suggestion.size - delete.size
       item.suggestions.clear
     end
+    if @verbose == 2 || item.suggestions.size == 0 || (@wordlist[item.suggestions[0]].size - delete.size >= suggestion.size - delete.size)
+      item.suggestions << suggestionint
+    end
   end
 
   def edits(word, edit_distance, deletes)
@@ -221,7 +221,13 @@ class SymSpell
   end
 
   def sort(suggestions)
-    suggestions.sort! {|x, y| -x.count <=> y.count}
+    if @verbose < 2
+      suggestions.sort! {|x, y| -x.count <=> y.count}
+    else
+      suggestions.sort! {|x, y| (2 * x.distance <=> y.distance) - x.count <=> y.count}
+    end
+
+    @verbose == 0 ? suggestions[0..0] : suggestions
   end
 
   def damerau_levenshtein_distance(source, target)

data/symspell.gemspec

@@ -2,7 +2,7 @@ require 'base64'
 
 Gem::Specification.new do |s|
   s.name        = 'symspell'
-  s.version     = '0.0.1'
+  s.version     = '0.0.2'
   s.authors     = 'Phil Thompson'
   s.email       = Base64.decode64("cGhpbEBlbGVjdHJpY3Zpc2lvbnMuY29t\n")
   s.summary     = 'Ruby port of the symetric spell checking algorithm'

data/tests/symspell_test.rb

@@ -0,0 +1,39 @@
+require 'minitest/autorun'
+require_relative '../lib/symspell'
+
+class SymSpellTest < Minitest::Test
+  def setup
+    @edit_distance_max = 2
+    @verbose = 0
+  end
+
+  def subject
+    @subject ||= SymSpell.new(@edit_distance_max, @verbose).tap do |subject|
+      words = %w(joe mark john peter mary andrew imogen)
+      subject.create_dictionary words
+    end
+  end
+  def test_lookup_correctly_spelled_word
+    assert_equal 'andrew', subject.lookup('andrew').first.term
+  end
+
+  def test_lookup_misspelt_word
+    assert_equal 'andrew', subject.lookup('andre').first.term
+  end
+
+  def test_lookup_fails_to_find_match
+    assert_equal nil, subject.lookup('amigon').first
+  end
+
+  def test_lookup_finds_match_after_turning_up_edit_distance
+    @edit_distance_max = 3
+    assert_equal ['imogen'], subject.lookup('amigon').map(&:term)
+  end
+
+  def test_lookup_returns_multiple_suggestions
+    @edit_distance_max = 2
+    @verbose = 2
+    assert_equal ['joe', 'john'], subject.lookup('jo').map(&:term)
+  end
+end
+

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: symspell
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.0.2
 platform: ruby
 authors:
 - Phil Thompson
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-07-31 00:00:00.000000000 Z
+date: 2015-08-03 00:00:00.000000000 Z
 dependencies: []
 description: 
 email: phil@electricvisions.com
@@ -20,6 +20,7 @@ files:
 - Rakefile
 - lib/symspell.rb
 - symspell.gemspec
+- tests/symspell_test.rb
 homepage: https://github.com/PhilT/symspell
 licenses: []
 metadata: {}