improving fuzzy match

teamtnt · Jun 29, 2018 · 8447a36 · staabm · Jul 18, 2018 · nticaric
1 parent 8d95869
commit 8447a36
Show file tree

Hide file tree

Showing 2 changed files with 28 additions and 18 deletions.
diff --git a/src/TNTFuzzyMatch.php b/src/TNTFuzzyMatch.php
@@ -94,6 +94,8 @@ public function fuzzyMatchFromFile($pattern, $path)
         if ($lines) {
             while (!feof($lines)) {
                 $line = fgets($lines, 4096);
+                $line = str_replace("\r", "", $line);
+                $line = str_replace("\n", "", $line);
                 if ($this->hasCommonSubsequence($pattern, $line)) {
                     $res[] = $line;
                 }
@@ -118,7 +120,6 @@ public function fuzzyMatchFromFile($pattern, $path)
         }
 
         arsort($sorted);
-
         return $sorted;
     }
 
@@ -140,7 +141,12 @@ public function fuzzyMatch($pattern, $items)
             $wordVector             = $this->wordToVector($word);
             $normalizedPaternVector = $this->makeVectorSameLength($wordVector, $paternVector);
 
-            $angle         = $this->angleBetweenVectors($wordVector, $normalizedPaternVector);
+            $angle = $this->angleBetweenVectors($wordVector, $normalizedPaternVector);
+
+            if (strpos($word, $pattern) !== false) {
+                $angle += 0.2;
+            }
+
             $sorted[$word] = $angle;
         }
 

diff --git a/tests/TNTFuzzyMatchTest.php b/tests/TNTFuzzyMatchTest.php
@@ -72,39 +72,43 @@ public function testFuzzyMatchFromFile()
     {
         $res = $this->fm->fuzzyMatchFromFile('search', __DIR__.'/_files/english_wordlist_2k.txt');
 
-        $this->assertEquals([
-            'search'   => 1,
-            'research' => 0.86602345529065
-        ], $res);
+        $equal = bccomp($res['search'], 1.2, 2);
+        $this->assertEquals(0, $equal);
+
+        $equal = bccomp($res['research'], 1.06, 2);
+        $this->assertEquals(0, $equal);
     }
 
     public function testFuzzyMatchFromFileFunction()
     {
         $res = fuzzyMatchFromFile('search', __DIR__.'/_files/english_wordlist_2k.txt');
 
-        $this->assertEquals([
-            'search'   => 1,
-            'research' => 0.86602345529065
-        ], $res);
+        $equal = bccomp($res['search'], 1.2, 2);
+        $this->assertEquals(0, $equal);
+
+        $equal = bccomp($res['research'], 1.06, 2);
+        $this->assertEquals(0, $equal);
     }
 
     public function testFuzzyMatch()
     {
         $res = $this->fm->fuzzyMatch('search', ['search', 'research', 'something']);
 
-        $this->assertEquals([
-            'search'   => 1,
-            'research' => 0.86602345529065
-        ], $res);
+        $equal = bccomp($res['search'], 1.2, 2);
+        $this->assertEquals(0, $equal);
+
+        $equal = bccomp($res['research'], 1.06, 2);
+        $this->assertEquals(0, $equal);
     }
 
     public function testFuzzyMatchFunction()
     {
         $res = fuzzyMatch('search', ['search', 'research', 'something']);
 
-        $this->assertEquals([
-            'search'   => 1,
-            'research' => 0.86602345529065
-        ], $res);
+        $equal = bccomp($res['search'], 1.2, 2);
+        $this->assertEquals(0, $equal);
+
+        $equal = bccomp($res['research'], 1.06, 2);
+        $this->assertEquals(0, $equal);
     }
 }