Page MenuHomeDevCentral
Files F27327249

No OneTemporary
Actions

View Options

diff --git a/src/Strings/SorensenDiceCoefficient.php b/src/Strings/SorensenDiceCoefficient.php
index 040f352..12913b4 100644
--- a/src/Strings/SorensenDiceCoefficient.php
+++ b/src/Strings/SorensenDiceCoefficient.php
@@ -1,55 +1,74 @@
<?php
declare(strict_types=1);
namespace Keruald\OmniTools\Strings;
use Keruald\OmniTools\Strings\Multibyte\OmniString;
+/**
+ * Computes the Sørensen–Dice coefficient, a statistic used to evaluate
+ * the similarity between two strings.
+ */
class SorensenDiceCoefficient {
/**
* @var string[]
*/
- private $x;
+ private array $x;
/**
* @var string[]
*/
- private $y;
+ private array $y;
///
/// Constructors
///
+ /**
+ * @param string $left The first string to compare
+ * @param string $right The second string to compare
+ */
public function __construct (string $left, string $right) {
$this->x = (new OmniString($left))->getBigrams();
$this->y = (new OmniString($right))->getBigrams();
}
+ /**
+ * Allows to directly compute the coefficient between two strings.
+ *
+ * @param string $left The first string to compare
+ * @param string $right The second string to compare
+ *
+ * @return float The Sørensen–Dice coefficient for the two specified strings.
+ */
public static function computeFor(string $left, string $right) : float {
$instance = new self($left, $right);
return $instance->compute();
}
///
/// Sørensen formula
///
+ /**
+ * @return float The Sørensen–Dice coefficient.
+ */
public function compute() : float {
return 2 * $this->countIntersect()
/
$this->countCharacters();
}
private function countIntersect () : int {
$intersect = array_intersect($this->x, $this->y);
return count($intersect);
}
private function countCharacters () : int {
return count($this->x) + count($this->y);
}
}
diff --git a/tests/Strings/SorensenDiceCoefficientTest.php b/tests/Strings/SorensenDiceCoefficientTest.php
index 2e7d17c..c69aafa 100644
--- a/tests/Strings/SorensenDiceCoefficientTest.php
+++ b/tests/Strings/SorensenDiceCoefficientTest.php
@@ -1,17 +1,24 @@
<?php
declare(strict_types=1);
namespace Keruald\OmniTools\Tests\Strings;
use Keruald\OmniTools\Strings\SorensenDiceCoefficient;
use PHPUnit\Framework\TestCase;
class SorensenDiceCoefficientTest extends TestCase {
public function testCoefficient () : void {
$actual = new SorensenDiceCoefficient('night', 'nacht');
$this->assertEquals(0.25, $actual->compute());
}
+ public function testComputeFor () : void {
+ $score = SorensenDiceCoefficient::computeFor('night', 'nacht');
+
+ $this->assertGreaterThan(0, $score);
+ $this->assertLessThan(1, $score);
+ }
+
}

File Metadata

Mime Type
text/x-diff
Expires
Tue, May 5, 11:49 (15 h, 43 m)
Storage Engine
blob
Storage Format
Raw Data
Storage Handle
3684471
Default Alt Text
(2 KB)

Event Timeline

Nasqueron DevCentral · If it had been much bigger the moon would have had a core of ice. · Powered by Phabricator