prepare cross validation random splitter

src/Phpml/Dataset/CsvDataset.php

@@ -0,0 +1,65 @@
+<?php
+
+declare (strict_types = 1);
+
+namespace Phpml\Dataset;
+
+use Phpml\Exception\DatasetException;
+
+abstract class CsvDataset implements Dataset
+{
+    /**
+     * @var string
+     */
+    protected $filepath;
+
+    /**
+     * @var array
+     */
+    private $samples = [];
+
+    /**
+     * @var array
+     */
+    private $lables = [];
+
+    public function __construct()
+    {
+        $filepath = dirname(__FILE__).'/../../../data/'.$this->filepath;
+
+        if (!file_exists($filepath)) {
+            throw DatasetException::missingFile(basename($filepath));
+        }
+
+        $row = 0;
+        if (($handle = fopen($filepath, 'r')) !== false) {
+            while (($data = fgetcsv($handle, 1000, ',')) !== false) {
+                ++$row;
+                if ($row == 1) {
+                    continue;
+                }
+                $this->samples[] = array_slice($data, 0, 4);
+                $this->lables[] = $data[4];
+            }
+            fclose($handle);
+        } else {
+            throw DatasetException::cantOpenFile(basename($filepath));
+        }
+    }
+
+    /**
+     * @return array
+     */
+    public function getSamples(): array
+    {
+        return $this->samples;
+    }
+
+    /**
+     * @return array
+     */
+    public function getLabels(): array
+    {
+        return $this->lables;
+    }
+}

src/Phpml/Dataset/Dataset.php

@@ -4,64 +4,15 @@ declare (strict_types = 1);
 
 namespace Phpml\Dataset;
 
-use Phpml\Exception\DatasetException;
-
-abstract class Dataset
+interface Dataset
 {
     /**
-     * @var string
+     * @return array
      */
-    protected $filepath;
-
-    /**
-     * @var array
-     */
-    private $samples = [];
-
-    /**
-     * @var array
-     */
-    private $lables = [];
-    
-    public function __construct()
-    {
-        $filepath =  dirname(__FILE__) . '/../../../data/' . $this->filepath;
-
-        if(!file_exists($filepath)) {
-            throw DatasetException::missingFile(basename($filepath));
-        }
-
-        $row = 0;
-        if (($handle = fopen($filepath, "r")) !== FALSE) {
-            while (($data = fgetcsv($handle, 1000, ",")) !== FALSE) {
-                $row++;
-                if($row==1) {
-                    continue;
-                }
-                $this->samples[] = array_slice($data, 0, 4);
-                $this->lables[] = $data[4];
-            }
-            fclose($handle);
-        } else {
-            throw DatasetException::cantOpenFile(basename($filepath));
-        }
-
-    }
+    public function getSamples(): array;
 
     /**
      * @return array
      */
-    public function getSamples()
-    {
-        return $this->samples;
-    }
-
-    /**
-     * @return array
-     */
-    public function getLabels()
-    {
-        return $this->lables;
-    }
-
+    public function getLabels(): array;
 }

src/Phpml/Dataset/Iris.php

@@ -8,13 +8,12 @@ namespace Phpml\Dataset;
  * Classes: 3
  * Samples per class: 50
  * Samples total: 150
- * Features per sample: 4
+ * Features per sample: 4.
  */
-class Iris extends Dataset
+class Iris extends CsvDataset
 {
     /**
      * @var string
      */
     protected $filepath = 'iris.csv';
-
 }

src/Phpml/Exception/DatasetException.php

@@ -18,5 +18,4 @@ class DatasetException extends \Exception
     {
         return new self(sprintf('Dataset file %s can\'t be open.', $filepath));
     }
-
 }

src/Phpml/Exception/InvalidArgumentException.php

@@ -13,4 +13,14 @@ class InvalidArgumentException extends \Exception
     {
         return new self('Size of given arguments not match');
     }
+
+    /**
+     * @param $name
+     *
+     * @return InvalidArgumentException
+     */
+    public static function percentNotInRange($name)
+    {
+        return new self(sprintf('%s must be between 0.0 and 1.0', $name));
+    }
 }

tests/Phpml/Dataset/IrisTest.php

@@ -1,5 +1,6 @@
 <?php
-declare(strict_types = 1);
+
+declare (strict_types = 1);
 
 namespace tests\Phpml\Dataset;
 
@@ -7,7 +8,6 @@ use Phpml\Dataset\Iris;
 
 class IrisTest extends \PHPUnit_Framework_TestCase
 {
-
     public function testLoadingIrisDataset()
     {
         $iris = new Iris();
@@ -19,5 +19,4 @@ class IrisTest extends \PHPUnit_Framework_TestCase
         // one sample features count
         $this->assertEquals(4, count($iris->getSamples()[0]));
     }
-    
-}
+}