Improving levenshtein performance. Refs #84

.npmignore

@@ -1,4 +1,6 @@
 node_modules
+experiments
+benchmark
 .gitignore
 .npmignore
 *.yml

benchmark/levenshtein.js

@@ -0,0 +1,32 @@
+require('babel-core/register');
+
+var levenshtein = require('../src/metrics/distance/levenshtein');
+var leven = require('leven');
+
+function run(fn) {
+  fn('a', 'b');
+  fn('ab', 'ac');
+  fn('ac', 'bc');
+  fn('abc', 'axc');
+  fn('kitten', 'sitting');
+  fn('xabxcdxxefxgx', '1ab2cd34ef5g6');
+  fn('cat', 'cow');
+  fn('xabxcdxxefxgx', 'abcdefg');
+  fn('javawasneat', 'scalaisgreat');
+  fn('example', 'samples');
+  fn('sturgeon', 'urgently');
+  fn('levenshtein', 'frankenstein');
+  fn('distance', 'difference');
+  fn('abcde', 'tes');
+  fn('因為我是中國人所以我會說中文', '因為我是英國人所以我會說英文');
+}
+
+suite('Levenshtein', function() {
+  bench('talisman', function() {
+    run(levenshtein);
+  });
+
+  bench('leven', function() {
+    run(leven);
+  });
+});

package.json

@@ -80,6 +80,8 @@
     "chai": "^3.5.0",
     "csv": "^1.1.0",
     "eslint": "^3.2.2",
+    "leven": "^2.0.0",
+    "matcha": "^0.7.0",
     "mocha": "^3.0.2",
     "rimraf": "^2.5.2",
     "seedrandom": "^2.4.2"

src/metrics/distance/levenshtein.js

@@ -21,7 +21,95 @@
  * @param  {mixed}  b - The second sequence to process.
  * @return {number}   - The Levenshtein distance between a & b.
  */
+function levenshteinForStrings(a, b) {
+  if (a === b)
+    return 0;
+
+  var tmp;
+
+  // Swapping the strings so that the shorter string is the first one.
+  if (a.length > b.length) {
+    tmp = a;
+    a = b;
+    b = tmp;
+  }
+
+  let la = a.length,
+      lb = b.length;
+
+  if (!la)
+    return lb;
+  if (!lb)
+    return la;
+
+  // Ignoring common suffix
+  while (la > 0 && (a.charCodeAt(la - 1) === b.charCodeAt(lb - 1))) {
+    la--;
+    lb--;
+  }
+
+  if (!la)
+    return lb;
+
+  let start = 0;
+
+  // Ignoring common prefix
+  while (start < la && (a.charCodeAt(start) === b.charCodeAt(start)))
+    start++;
+
+  la -= start;
+  lb -= start;
+
+  if (!la)
+    return lb;
+
+  const v0 = new Array(lb);
+
+  let i = 0;
+
+  while (i < lb)
+    v0[i] = ++i;
+
+  let current = 0;
+
+  // Starting the nested loops
+  for (i = 0; i < la; i++) {
+    let left = i;
+
+    current = i + 1;
+    const charA = a.charCodeAt(start + i);
+
+    for (let j = 0; j < lb; j++) {
+      const above = current;
+
+      const charB = b.charCodeAt(j);
+      current = left;
+      left = v0[j];
+
+      if (charA !== charB) {
+
+        // Insertion
+        if (left < current)
+          current = left;
+
+        // Deletion
+        if (above < current)
+          current = above;
+
+        current++;
+      }
+
+      v0[j] = current;
+    }
+  }
+
+  return current;
+};
+
 export default function levenshtein(a, b) {
+  if (typeof a === 'string')
+    return levenshteinForStrings(a, b);
+
   if (a === b)
     return 0;
 

src/structures/bk-tree.js

@@ -44,6 +44,7 @@ function searchWord(distance, tree, word, n, acc) {
   if (d <= n)
     acc.push(tree.word);
 
+  // NOTE: this can be improved by for in iteration!
   for (let i = d - n, l = d + n + 1; i < l; i++) {
     const children = tree.children[i];