jaro 0.0.2 → 0.0.3

data/jaro.rb

@@ -1,7 +1,13 @@
+# encoding: utf-8
+
 class String
-  # Jaro distance
+  # Jaro-Winkler distance
+  # @param [String] other string
+  # @return [Float] distance, normalized between 0.0 (no match) and 1.0 (perfect match)
   def ^(other)
-    return 0 if self.empty? or other.empty?
+    return 1.0 if self == other
+    return 0.0 if self.empty? or other.empty?
+
     s1 = self.codepoints.to_a
     s2 = other.codepoints.to_a
     s1, s2 = s2, s1 if s1.size > s2.size
@@ -10,7 +16,7 @@ class String
     m, t = 0.0, 0
     max_dist = s2s/2 - 1
 
-    m1 = Array.new(s1s, nil)
+    m1 = Array.new(s1s, -1)
     m2 = Array.new(s2s, false)
 
     # find m
@@ -25,12 +31,14 @@ class String
         end
       end
     end
+
+    return 0.0 if m.zero?
     
     m1.reduce do |a, b|
       # if either a or b are nil, that means there was no match
       # if a > b, that means the previous value is greater than the current
       # which means it went down
-      if a != nil and b != nil and a > b
+      if a > -1 and b > -1 and a > b
         t += (a-b > 1 ? 1 : 2)
       end
       b
@@ -41,7 +49,10 @@ class String
     # winkler adjustment
     l = 0
     for i in 0..3
-      s1[i] == s2[i] ? l += 1 : break
+      if s1[i] != s2[i]
+        l = i
+        break
+      end
     end
 
     # standard weight (p) for winkler == 0.1

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: jaro
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: 0.0.3
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-09-28 00:00:00.000000000 Z
+date: 2012-09-29 00:00:00.000000000 Z
 dependencies: []
 description: Adds String#^ for measuring string similarity
 email: moshee@displaynone.us