From 9f140d5b6ff611b36a33af620d156b9f5c317cda Mon Sep 17 00:00:00 2001
From: Arkadiusz Kondas <arkadiusz.kondas@gmail.com>
Date: Thu, 14 Jul 2016 13:25:11 +0200
Subject: [PATCH] fix problem with token count vectorizer array order

---
 .../FeatureExtraction/TokenCountVectorizer.php     |  2 ++
 .../FeatureExtraction/TokenCountVectorizerTest.php | 14 +++++++-------
 2 files changed, 9 insertions(+), 7 deletions(-)

diff --git a/src/Phpml/FeatureExtraction/TokenCountVectorizer.php b/src/Phpml/FeatureExtraction/TokenCountVectorizer.php
index 3ec6af1..56d63e2 100644
--- a/src/Phpml/FeatureExtraction/TokenCountVectorizer.php
+++ b/src/Phpml/FeatureExtraction/TokenCountVectorizer.php
@@ -116,6 +116,8 @@ class TokenCountVectorizer implements Transformer
             }
         }
 
+        ksort($counts);
+
         $sample = $counts;
     }
 
diff --git a/tests/Phpml/FeatureExtraction/TokenCountVectorizerTest.php b/tests/Phpml/FeatureExtraction/TokenCountVectorizerTest.php
index b18db60..22ff1a9 100644
--- a/tests/Phpml/FeatureExtraction/TokenCountVectorizerTest.php
+++ b/tests/Phpml/FeatureExtraction/TokenCountVectorizerTest.php
@@ -40,10 +40,10 @@ class TokenCountVectorizerTest extends \PHPUnit_Framework_TestCase
         $vectorizer = new TokenCountVectorizer(new WhitespaceTokenizer());
 
         $vectorizer->fit($samples);
-        $this->assertEquals($vocabulary, $vectorizer->getVocabulary());
+        $this->assertSame($vocabulary, $vectorizer->getVocabulary());
 
         $vectorizer->transform($samples);
-        $this->assertEquals($tokensCounts, $samples);
+        $this->assertSame($tokensCounts, $samples);
     }
 
     public function testTransformationWithMinimumDocumentTokenCountFrequency()
@@ -74,10 +74,10 @@ class TokenCountVectorizerTest extends \PHPUnit_Framework_TestCase
         $vectorizer = new TokenCountVectorizer(new WhitespaceTokenizer(), null, 0.5);
 
         $vectorizer->fit($samples);
-        $this->assertEquals($vocabulary, $vectorizer->getVocabulary());
+        $this->assertSame($vocabulary, $vectorizer->getVocabulary());
 
         $vectorizer->transform($samples);
-        $this->assertEquals($tokensCounts, $samples);
+        $this->assertSame($tokensCounts, $samples);
 
         // word at least once in all samples
         $samples = [
@@ -96,7 +96,7 @@ class TokenCountVectorizerTest extends \PHPUnit_Framework_TestCase
         $vectorizer->fit($samples);
         $vectorizer->transform($samples);
 
-        $this->assertEquals($tokensCounts, $samples);
+        $this->assertSame($tokensCounts, $samples);
     }
 
     public function testTransformationWithStopWords()
@@ -131,9 +131,9 @@ class TokenCountVectorizerTest extends \PHPUnit_Framework_TestCase
         $vectorizer = new TokenCountVectorizer(new WhitespaceTokenizer(), $stopWords);
 
         $vectorizer->fit($samples);
-        $this->assertEquals($vocabulary, $vectorizer->getVocabulary());
+        $this->assertSame($vocabulary, $vectorizer->getVocabulary());
 
         $vectorizer->transform($samples);
-        $this->assertEquals($tokensCounts, $samples);
+        $this->assertSame($tokensCounts, $samples);
     }
 }