fix some bugs, update README, turn into a gem, add some tests

PhilT · Jul 31, 2015 · d495600 · d495600
1 parent 3814c18
commit d495600
Show file tree

Hide file tree

Showing 5 changed files with 52 additions and 12 deletions.
diff --git a/README.md b/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
@@ -26,5  26,7 @@ I've changed very little from the original source (apart from removing the comma
 
 ## EDIT_DISTANCE_MAX
 
-`EDIT_DISTANCE_MAX` is the number of letters to remove to find a match. Standard text should be around 2-3 if you have a smaller dictionary you could try larger numbers to catch drastically misspelt words. Note that creating the dictionary will take a lot longer as the combinations go up exponentially.
 `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.
 
diff --git a/Rakefile b/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
 

diff --git a/lib/symspell.rb b/lib/symspell.rb
@@ -15,7  15,7 @@ def create_dictionary(corpus)
     word_count = 0
 
     File.open(corpus, 'r').each_line do |word|
-      word_count  = 1 if create_dictionary_entry(word.strip, language)
       word_count  = 1 if create_dictionary_entry(word.strip)
     end
   end
 
@@ -156,10  156,10 @@ def parse_words(text)
     text.downcase.scan(/[\w-[\d_]] /).first
   end
 
-  def create_dictionary_entry(key, language)
   def create_dictionary_entry(key)
     result = false
     value = nil
-    if valueo = @dictionary[language   key]
     if valueo = @dictionary[key]
       if valueo.is_a?(Fixnum)
         tmp = valueo
         value = DictionaryItem.new
@@ -171,7  171,7 @@ def create_dictionary_entry(key, language)
     elsif @wordlist.count < MAX_INT
       value = DictionaryItem.new
       value.count  = 1
-      @dictionary[language   key] = value
       @dictionary[key] = value
 
       @maxlength = key.size if key.size > @maxlength
     end
@@ -182,17  182,17 @@ def create_dictionary_entry(key, language)
       result = true
 
       edits(key, 0, Set.new).each do |delete|
-        if value2 = @dictionary[language   delete]
         if value2 = @dictionary[delete]
           if value2.is_a?(Fixnum)
             tmp = value2
             di = DictionaryItem.new
             di.suggestions << tmp
-            @dictionary[language   delete] = di
             @dictionary[delete] = di
             add_lowest_distance(di, key, keyint, delete) unless di.suggestions.include?(keyint)
           elsif !value2.suggestions.include?(keyint)
           end
         else
-          @dictionary[language   delete] = keyint
           @dictionary[delete] = keyint
         end
       end
     end
@@ -239,7  239,7 @@ def damerau_levenshtein_distance(source, target)
       sd[letter] = 0 unless sd[letter]
     end
 
-    (0..m).each do |i|
     (1..m).each do |i|
       db = 0
       (0..n).each do |j|
         i1 = sd[target[j - 1]]

diff --git a/tests/symspell_test.rb b/tests/symspell_test.rb
@@ -0,0  1,31 @@
 require 'minitest/autorun'
 require_relative '../lib/symspell'
 
 class SymSpellTest < Minitest::Test
   def setup
     @edit_distance_max = 2
   end
 
   def subject
     @subject ||= SymSpell.new(@edit_distance_max).tap do |subject|
       subject.create_dictionary 'tests/words.txt'
     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').first.term
   end
 end
 
diff --git a/tests/words.txt b/tests/words.txt
@@ -0,0  1,7 @@
 mark
 john
 peter
 mary
 andrew
 imogen