From 1cb7a1ed7716ed9dc4248464524430c08b49ffb2 Mon Sep 17 00:00:00 2001
From: Dennis Eichhorn <spl1nes.com@googlemail.com>
Date: Wed, 7 Oct 2020 20:57:08 +0200
Subject: [PATCH] Add search and comparison functionality

---
 System/Search/StringSearch.php           | 186 +++++++++++++++++++++++
 Utils/MbStringUtils.php                  |   4 -
 Utils/StringCompare.php                  |  77 ++++++++++
 Utils/StringUtils.php                    |   4 -
 tests/System/CharsetTypeTest.php         |  54 +++++++
 tests/System/Search/StringSearchTest.php |  55 +++++++
 tests/Utils/StringCompareTest.php        |  14 ++
 7 files changed, 386 insertions(+), 8 deletions(-)
 create mode 100644 System/Search/StringSearch.php
 create mode 100644 tests/System/CharsetTypeTest.php
 create mode 100644 tests/System/Search/StringSearchTest.php

diff --git a/System/Search/StringSearch.php b/System/Search/StringSearch.php
new file mode 100644
index 000000000..bd7ef94f9
--- /dev/null
+++ b/System/Search/StringSearch.php
@@ -0,0 +1,186 @@
+<?php
+/**
+ * Orange Management
+ *
+ * PHP Version 7.4
+ *
+ * @package   phpOMS\System\Search
+ * @copyright Dennis Eichhorn
+ * @license   OMS License 1.0
+ * @version   1.0.0
+ * @link      https://orange-management.org
+ */
+declare(strict_types=1);
+
+namespace phpOMS\System\Search;
+
+/**
+ * Basic string search algorithms.
+ *
+ * @package phpOMS\System\Search
+ * @license OMS License 1.0
+ * @link    https://orange-management.org
+ * @since   1.0.0
+ */
+abstract class StringSearch
+{
+    /**
+     * @codeCoverageIgnore
+     */
+    private function __construct()
+    {
+    }
+
+    /**
+     * Find pattern in string
+     *
+     * @param string $pattern Pattern
+     * @param string $text    Text to search in
+     *
+     * @return int Match position
+     *
+     * @since 1.0.0
+     */
+    public static function knuthMorrisPrattSearch(string $pattern, string $text) : int
+    {
+        $patternSize = \strlen($pattern);
+        $textSize    = \strlen($text);
+
+        $shift = self::knuthMorrisPrattShift($pattern);
+
+        $i = 1;
+        $j = 0;
+        while ($i + $patternSize <= $textSize) {
+            while ($text[$i + $j] === $pattern[$j]) {
+                ++$j;
+                if ($j >= $patternSize) {
+                    return $i;
+                }
+            }
+
+            if ($j > 0) {
+                $i += $shift[$j - 1];
+                $j  = \max($j - $shift[$j - 1], 0);
+            } else {
+                ++$i;
+                $j = 0;
+            }
+        }
+
+        return -1;
+    }
+
+    /**
+     * Create shift array
+     *
+     * @param string $pattern Pattern
+     *
+     * @return int[]
+     *
+     * @since 1.0.0
+     */
+    private static function knuthMorrisPrattShift(string $pattern) : array
+    {
+        $patternSize = \strlen($pattern);
+        $shift       = [];
+        $shift[]     = 1;
+
+        $i = 1;
+        $j = 0;
+        while ($i + $j < $patternSize) {
+            if ($pattern[$i + $j] === $pattern[$j]) {
+                $shift[$i + $j] = $i;
+                ++$j;
+            } else {
+                if ($j === 0) {
+                    $shift[$i] = $i + 1;
+                }
+
+                if ($j > 0) {
+                    $i += $shift[$j - 1];
+                    $j  = \max($j - $shift[$j - 1], 0);
+                } else {
+                    ++$i;
+                    $j = 0;
+                }
+            }
+        }
+
+        return $shift;
+    }
+
+    /**
+     * Find pattern in string
+     *
+     * @param string $pattern Pattern
+     * @param string $text    Text to search in
+     *
+     * @return int Match position
+     *
+     * @since 1.0.0
+     */
+    public static function boyerMooreHorspoolSimpleSearch(string $pattern, string $text) : int
+    {
+        $patternSize = \strlen($pattern);
+        $textSize    = \strlen($text);
+
+        $i = 0;
+        $j = 0;
+        while ($i + $patternSize <= $textSize) {
+            $j = $patternSize - 1;
+
+            while ($text[$i + $j] === $pattern[$j]) {
+                --$j;
+                if ($j < 0) {
+                    return $i;
+                }
+            }
+
+            ++$i;
+        }
+
+        return -1;
+    }
+
+    /**
+     * Find pattern in string
+     *
+     * @param string $pattern Pattern
+     * @param string $text    Text to search in
+     *
+     * @return int Match position
+     *
+     * @since 1.0.0
+     */
+    public static function boyerMooreHorspoolSearch(string $pattern, string $text) : int
+    {
+        $patternSize = \strlen($pattern);
+        $textSize    = \strlen($text);
+
+        $shift = [];
+        for ($k = 0; $k < 256; ++$k) {
+            $shift[$k] = $patternSize;
+        }
+
+        for ($k = 0; $k < $patternSize - 1; ++$k) {
+            $shift[\ord($pattern[$k])] = $patternSize - 1 - $k;
+        }
+
+        $i = 0;
+        $j = 0;
+        while ($i + $patternSize <= $textSize) {
+            $j = $patternSize - 1;
+
+            while ($text[$i + $j] === $pattern[$j]) {
+                --$j;
+                if ($j < 0) {
+                    return $i;
+                }
+            }
+
+            $i += $shift[\ord($text[$i + $patternSize - 1])];
+        }
+
+        return -1;
+    }
+}
diff --git a/Utils/MbStringUtils.php b/Utils/MbStringUtils.php
index f3d3b09ad..519a6809e 100644
--- a/Utils/MbStringUtils.php
+++ b/Utils/MbStringUtils.php
@@ -26,10 +26,6 @@ use phpOMS\System\CharsetType;
  * @link    https://orange-management.org
  * @since   1.0.0
  *
- * @todo Orange-Management/phpOMS#119
- *  Create jaro winkler distance
- *  https://en.wikipedia.org/wiki/Jaro%E2%80%93Winkler_distance
- *
  * @SuppressWarnings(PHPMD.CamelCaseMethodName)
  */
 final class MbStringUtils
diff --git a/Utils/StringCompare.php b/Utils/StringCompare.php
index f17ee1b9b..aed4220b8 100644
--- a/Utils/StringCompare.php
+++ b/Utils/StringCompare.php
@@ -86,6 +86,83 @@ final class StringCompare
         return $bestMatch;
     }
 
+    /**
+     * Jaro string distance
+     *
+     * @param string $s1 String1
+     * @param string $s2 String2
+     *
+     * @return float
+     *
+     * @since 1.0.0
+     */
+    public static function jaro(string $s1, string $s2) : float
+    {
+        $s1Size = \strlen($s1);
+        $s2Size = \strlen($s2);
+
+        if ($s1Size === 0) {
+            return $s2Size === 0 ? 1.0 : 0.0;
+        }
+
+        $mDistance = (int) (\max($s1Size, $s2Size) / 2 - 1);
+
+        $matches        = 0;
+        $transpositions = 0.0;
+
+        $s1Matches = [];
+        $s2Matches = [];
+
+        for ($i = 0; $i < $s1Size; ++$i) {
+            $start = \max(0, $i - $mDistance);
+            $end   = \min($i + $mDistance + 1, $s2Size);
+
+            for ($j = $start; $j < $end; ++$j) {
+                if (isset($s2Matches[$j])) {
+                    continue;
+                }
+
+                if ($s1[$i] !== $s2[$j]) {
+                    continue;
+                }
+
+                $s1Matches[$i] = true;
+                $s2Matches[$j] = true;
+
+                ++$matches;
+                break;
+            }
+        }
+
+        if ($matches === 0) {
+            return 0.0;
+        }
+
+        $j = 0;
+        for ($i = 0; $i < $s1Size; ++$i) {
+            if (!isset($s1Matches[$i])) {
+                continue;
+            }
+
+            while (!isset($s2Matches[$j])) {
+                ++$j;
+            }
+
+            if ($s1[$i] !== $s2[$j]) {
+                ++$transpositions;
+            }
+
+            ++$j;
+        }
+
+        $transpositions /= 2.0;
+
+        return ($matches / $s1Size
+            + $matches / $s2Size
+            + ($matches - $transpositions) / $matches)
+            / 3.0;
+    }
+
     /**
      * Calculate word match score.
      *
diff --git a/Utils/StringUtils.php b/Utils/StringUtils.php
index 40946959f..e5d498c9d 100644
--- a/Utils/StringUtils.php
+++ b/Utils/StringUtils.php
@@ -26,10 +26,6 @@ use phpOMS\Contract\RenderableInterface;
  * @link    https://orange-management.org
  * @since   1.0.0
  *
- * @todo Orange-Management/phpOMS#119
- *  Create jaro winkler distance
- *  https://en.wikipedia.org/wiki/Jaro%E2%80%93Winkler_distance
- *
  * @SuppressWarnings(PHPMD.CamelCaseMethodName)
  */
 final class StringUtils
diff --git a/tests/System/CharsetTypeTest.php b/tests/System/CharsetTypeTest.php
new file mode 100644
index 000000000..1f85abc48
--- /dev/null
+++ b/tests/System/CharsetTypeTest.php
@@ -0,0 +1,54 @@
+<?php
+/**
+ * Orange Management
+ *
+ * PHP Version 7.4
+ *
+ * @package   tests
+ * @copyright Dennis Eichhorn
+ * @license   OMS License 1.0
+ * @version   1.0.0
+ * @link      https://orange-management.org
+ */
+declare(strict_types=1);
+
+namespace phpOMS\tests\System;
+
+require_once __DIR__ . '/../Autoloader.php';
+
+use phpOMS\System\CharsetType;
+
+/**
+ * @internal
+ */
+class CharsetTypeTest extends \PHPUnit\Framework\TestCase
+{
+    /**
+     * @group framework
+     * @coversNothing
+     */
+    public function testEnumCount() : void
+    {
+        self::assertCount(3, CharsetType::getConstants());
+    }
+
+    /**
+     * @group framework
+     * @coversNothing
+     */
+    public function testUnique() : void
+    {
+        self::assertEquals(CharsetType::getConstants(), \array_unique(CharsetType::getConstants()));
+    }
+
+    /**
+     * @group framework
+     * @coversNothing
+     */
+    public function testEnums() : void
+    {
+        self::assertEquals('us-ascii', CharsetType::ASCII);
+        self::assertEquals('iso-8859-1', CharsetType::ISO_8859_1);
+        self::assertEquals('utf-8', CharsetType::UTF_8);
+    }
+}
diff --git a/tests/System/Search/StringSearchTest.php b/tests/System/Search/StringSearchTest.php
new file mode 100644
index 000000000..5623842bd
--- /dev/null
+++ b/tests/System/Search/StringSearchTest.php
@@ -0,0 +1,55 @@
+<?php
+/**
+ * Orange Management
+ *
+ * PHP Version 7.4
+ *
+ * @package   tests
+ * @copyright Dennis Eichhorn
+ * @license   OMS License 1.0
+ * @version   1.0.0
+ * @link      https://orange-management.org
+ */
+declare(strict_types=1);
+
+namespace phpOMS\tests\System\Search;
+
+use phpOMS\System\Search\StringSearch;
+
+/**
+ * @testdox phpOMS\tests\System\Search\StringSearchTest: Search utilities
+ *
+ * @internal
+ */
+class StringSearchTest extends \PHPUnit\Framework\TestCase
+{
+    public function testKnutMorrisPrattSearch() : void
+    {
+        self::assertEquals(15, StringSearch::knuthMorrisPrattSearch('ABCDABD', 'ABC ABCDAB ABCDABCDABDE'));
+    }
+
+    public function testInvalidKnutMorrisPrattSearch() : void
+    {
+        self::assertEquals(-1, StringSearch::knuthMorrisPrattSearch('ABCDABDZ', 'ABC ABCDAB ABCDABCDABDE'));
+    }
+
+    public function testBoyerMooreHorspoolSimpleSearch() : void
+    {
+        self::assertEquals(15, StringSearch::boyerMooreHorspoolSimpleSearch('ABCDABD', 'ABC ABCDAB ABCDABCDABDE'));
+    }
+
+    public function testInvalidBoyerMooreHorspoolSimpleSearch() : void
+    {
+        self::assertEquals(-1, StringSearch::boyerMooreHorspoolSimpleSearch('ABCDABDZ', 'ABC ABCDAB ABCDABCDABDE'));
+    }
+
+    public function testBoyerMooreHorspoolSearch() : void
+    {
+        self::assertEquals(15, StringSearch::boyerMooreHorspoolSearch('ABCDABD', 'ABC ABCDAB ABCDABCDABDE'));
+    }
+
+    public function testInvalidBoyerMooreHorspoolSearch() : void
+    {
+        self::assertEquals(-1, StringSearch::boyerMooreHorspoolSearch('ABCDABDZ', 'ABC ABCDAB ABCDABCDABDE'));
+    }
+}
diff --git a/tests/Utils/StringCompareTest.php b/tests/Utils/StringCompareTest.php
index d5f2d5a03..bec92253e 100644
--- a/tests/Utils/StringCompareTest.php
+++ b/tests/Utils/StringCompareTest.php
@@ -94,4 +94,18 @@ class StringCompareTest extends \PHPUnit\Framework\TestCase
         // a is compared to is which has a distance of 2
         self::assertEquals(2, StringCompare::valueWords('This is a test', 'This is not test'));
     }
+
+    public function testJaro() : void
+    {
+        self::assertEqualsWithDelta(0.944444, StringCompare::jaro('MARTHA', 'MARHTA'), 0.01);
+        self::assertEqualsWithDelta(0.766667, StringCompare::jaro('DIXON', 'DICKSONX'), 0.01);
+        self::assertEqualsWithDelta(0.896296, StringCompare::jaro('JELLYFISH', 'SMELLYFISH'), 0.01);
+    }
+
+    public function testJaroEmpty() : void
+    {
+        self::assertEquals(1.0, StringCompare::jaro('', ''));
+        self::assertEquals(0.0, StringCompare::jaro('', 'test'));
+        self::assertEquals(0.0, StringCompare::jaro('test', ''));
+    }
 }