Change minimal length word

Assisted-Mindfulness · May 10, 2024 · 5a89eaf · 5a89eaf
1 parent 2f2edfe
commit 5a89eaf
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 1 deletion.
diff --git a/src/Classifier.php b/src/Classifier.php
@@ -63,7 +63,8 @@ public function tokenize(string $string): Collection
 
         return Str::of($string)
             ->lower()
-            ->matchAll('/[[:alpha:]]+/u');
+            ->matchAll('/[[:alpha:]]+/u')
+            ->filter(fn (string $word) => Str::length($word) > 3);
     }
 
     /**

diff --git a/tests/ClassifierTest.php b/tests/ClassifierTest.php
@@ -5,6 +5,7 @@
 namespace AssistedMindfulness\NaiveBayes\Tests;
 
 use AssistedMindfulness\NaiveBayes\Classifier;
+use Illuminate\Support\Str;
 use PHPUnit\Framework\TestCase;
 
 class ClassifierTest extends TestCase
@@ -13,6 +14,12 @@ public function testTokenizeClassifier(): void
     {
         $classifier = new Classifier();
 
+        $classifier->setTokenizer(function (string $string) {
+            return Str::of($string)
+                ->lower()
+                ->matchAll('/[[:alpha:]]+/u');
+        });
+
         $this->assertEquals(
             ['hello', 'how', 'are', 'you'],
             $classifier->tokenize('Hello, how are you?')->toArray()
@@ -198,6 +205,12 @@ public function testSimpleSpam(): void
     {
         $classifier = new Classifier();
 
+        $classifier->setTokenizer(function (string $string) {
+            return Str::of($string)
+                ->lower()
+                ->matchAll('/[[:alpha:]]+/u');
+        });
+
         $classifier
             ->learn('Learn how to grow your business with these proven strategies', 'ham')
             ->learn('Unlock the secrets of successful investing in our latest guide', 'ham')