limit common prefix in jaro-winkler

CHANGELOG.md

@@ -17,9 +17,12 @@ This project attempts to adhere to [Semantic Versioning](http://semver.org).
   - reduce runtime in our own benchmark by more than `70%`
   - reduce binary size by more than `25%`
 
+- only boost similarity in Jaro-Winkler once the Jaro similarity exceeds 0.7
+
 ### Fixed
 
 - Fix transposition counting in Jaro and Jaro-Winkler.
+- Limit common prefix in Jaro-Winkler to 4 characters
 
 ## [0.10.0] - (2020-01-31)
 

README.md

@@ -10,7 +10,7 @@
   - [Levenshtein] - distance & normalized
   - [Optimal string alignment]
   - [Damerau-Levenshtein] - distance & normalized
-  - [Jaro and Jaro-Winkler] - this implementation of Jaro-Winkler does not limit the common prefix length
+  - [Jaro and Jaro-Winkler]
   - [Sørensen-Dice]
 
 The normalized versions return values between `0.0` and `1.0`, where `1.0` means

src/lib.rs

@@ -194,22 +194,19 @@ where
     &'b Iter2: IntoIterator<Item = Elem2>,
     Elem1: PartialEq<Elem2>,
 {
-    let jaro_distance = generic_jaro(a, b);
+    let sim = generic_jaro(a, b);
 
-    // Don't limit the length of the common prefix
-    let prefix_length = a
-        .into_iter()
-        .zip(b)
-        .take_while(|(a_elem, b_elem)| a_elem == b_elem)
-        .count();
+    if sim > 0.7 {
+        let prefix_length = a
+            .into_iter()
+            .take(4)
+            .zip(b)
+            .take_while(|(a_elem, b_elem)| a_elem == b_elem)
+            .count();
 
-    let jaro_winkler_distance =
-        jaro_distance + (0.1 * prefix_length as f64 * (1.0 - jaro_distance));
-
-    if jaro_winkler_distance <= 1.0 {
-        jaro_winkler_distance
+        sim + 0.1 * prefix_length as f64 * (1.0 - sim)
     } else {
-        1.0
+        sim
     }
 }
 
@@ -218,7 +215,7 @@ where
 /// ```
 /// use strsim::jaro_winkler;
 ///
-/// assert!((0.911 - jaro_winkler("cheeseburger", "cheese fries")).abs() <
+/// assert!((0.866 - jaro_winkler("cheeseburger", "cheese fries")).abs() <
 ///         0.001);
 /// ```
 pub fn jaro_winkler(a: &str, b: &str) -> f64 {
@@ -960,15 +957,15 @@ mod tests {
     #[test]
     fn jaro_winkler_names() {
         assert_delta!(
-            0.562,
+            0.452,
             jaro_winkler("Friedrich Nietzsche", "Fran-Paul Sartre"),
             0.001
         );
     }
 
     #[test]
     fn jaro_winkler_long_prefix() {
-        assert_delta!(0.911, jaro_winkler("cheeseburger", "cheese fries"), 0.001);
+        assert_delta!(0.866, jaro_winkler("cheeseburger", "cheese fries"), 0.001);
     }
 
     #[test]
@@ -984,7 +981,7 @@ mod tests {
     #[test]
     fn jaro_winkler_very_long_prefix() {
         assert_delta!(
-            1.0,
+            0.98519,
             jaro_winkler("thequickbrownfoxjumpedoverx", "thequickbrownfoxjumpedovery")
         );
     }

tests/lib.rs

@@ -67,5 +67,5 @@ fn jaro_works() {
 
 #[test]
 fn jaro_winkler_works() {
-    assert_delta!(0.911, jaro_winkler("cheeseburger", "cheese fries"), 0.001);
+    assert_delta!(0.866, jaro_winkler("cheeseburger", "cheese fries"), 0.001);
 }