From 6f6cff092dd682a60065ad4990dc11aff90da893 Mon Sep 17 00:00:00 2001
From: Jan Tojnar <jtojnar@gmail.com>
Date: Sat, 16 Mar 2024 17:15:53 +0100
Subject: [PATCH 01/12] Remove deprecated debug property

Breaking BC for Readability 3.0.
---
 src/Readability.php       |  2 --
 tests/ReadabilityTest.php | 21 ---------------------
 2 files changed, 23 deletions(-)

diff --git a/src/Readability.php b/src/Readability.php
index 836a333..aa49727 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -37,8 +37,6 @@ class Readability implements LoggerAwareInterface
     public $url = null;
     // preserves more content (experimental)
     public $lightClean = true;
-    // no more used, keept to avoid BC
-    public $debug = false;
     public $tidied = false;
 
     /**
diff --git a/tests/ReadabilityTest.php b/tests/ReadabilityTest.php
index 3fb9dc0..2fee95f 100644
--- a/tests/ReadabilityTest.php
+++ b/tests/ReadabilityTest.php
@@ -115,7 +115,6 @@ public function testInitDivP(): void
     public function testInitDiv(): void
     {
         $readability = $this->getReadability('<div>' . str_repeat('This is the awesome content :)', 7) . '</div>', 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -129,7 +128,6 @@ public function testInitDiv(): void
     public function testWithFootnotes(): void
     {
         $readability = $this->getReadability('<div>' . str_repeat('<p>This is an awesome text with some links, here there are: <a href="http://0.0.0.0/test.html">the awesome</a></p>', 7) . '</div>', 'http://0.0.0.0');
-        $readability->debug = true;
         $readability->convertLinksToFootnotes = true;
         $res = $readability->init();
 
@@ -146,7 +144,6 @@ public function testWithFootnotes(): void
     public function testStandardClean(): void
     {
         $readability = $this->getReadability('<div><h2>Title</h2>' . str_repeat('<p>This is an awesome text with some links, here there are: <a href="http://0.0.0.0/test.html">the awesome</a></p>', 7) . '<a href="#nofollow" rel="nofollow">will NOT be removed</a></div>', 'http://0.0.0.0');
-        $readability->debug = true;
         $readability->lightClean = false;
         $res = $readability->init();
 
@@ -163,7 +160,6 @@ public function testStandardClean(): void
     public function testWithIframe(): void
     {
         $readability = $this->getReadability('<div><h2>Title</h2>' . str_repeat('<p>This is an awesome text with some links, here there are: <a href="http://0.0.0.0/test.html">the awesome</a></p>', 7) . '<p>This is an awesome text with some links, here there are <iframe src="http://youtube.com/test" href="#nofollow" rel="nofollow"></iframe><iframe>http://soundcloud.com/test</iframe></p></div>', 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -178,7 +174,6 @@ public function testWithIframe(): void
     public function testWithArticle(): void
     {
         $readability = $this->getReadability('<article><p>' . str_repeat('This is an awesome text with some links, here there are: the awesome', 20) . '</p><p>This is an awesome text with some links, here there are <iframe src="http://youtube.com/test" href="#nofollow" rel="nofollow"></iframe></p></article>', 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -193,7 +188,6 @@ public function testWithArticle(): void
     public function testWithAside(): void
     {
         $readability = $this->getReadability('<article>' . str_repeat('<p>This is an awesome text with some links, here there are: <a href="http://0.0.0.0/test.html">the awesome</a></p>', 7) . '<footer><aside>' . str_repeat('<p>This is an awesome text with some links, here there are</p>', 8) . '</aside></footer></article>', 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -208,7 +202,6 @@ public function testWithAside(): void
     public function testWithClasses(): void
     {
         $readability = $this->getReadability('<article>' . str_repeat('<p>This is an awesome text with some links, here there are: <a href="http://0.0.0.0/test.html">the awesome</a></p>', 7) . '<div style="display:none">' . str_repeat('<p class="clock">This text should be removed</p>', 10) . '</div></article>', 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -223,7 +216,6 @@ public function testWithClasses(): void
     public function testWithClassesWithoutLightClean(): void
     {
         $readability = $this->getReadability('<article>' . str_repeat('<p>This is an awesome text with some links, here there are: <a href="http://0.0.0.0/test.html">the awesome</a></p>', 7) . '<div style="display:none">' . str_repeat('<p class="clock">This text should be removed</p>', 10) . '</div></article>', 'http://0.0.0.0');
-        $readability->debug = true;
         $readability->lightClean = false;
         $res = $readability->init();
 
@@ -239,7 +231,6 @@ public function testWithClassesWithoutLightClean(): void
     public function testWithTd(): void
     {
         $readability = $this->getReadability('<table><tr>' . str_repeat('<td><p>This is an awesome text with some links, here there are the awesome</td>', 7) . '</tr></table>', 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -252,7 +243,6 @@ public function testWithTd(): void
     public function testWithSameClasses(): void
     {
         $readability = $this->getReadability('<article class="awesomecontent">' . str_repeat('<p>This is an awesome text with some links, here there are the awesome</p>', 7) . '<div class="awesomecontent">This text is also an awesome text and you should know that !</div></article>', 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -266,7 +256,6 @@ public function testWithSameClasses(): void
     public function testWithScript(): void
     {
         $readability = $this->getReadability('<article class="awesomecontent">' . str_repeat('<p>This is an awesome text with some links, here there are the awesome</p>', 7) . '<p><script>This text is also an awesome text and you should know that !</script></p></article>', 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -280,7 +269,6 @@ public function testWithScript(): void
     public function testTitle(): void
     {
         $readability = $this->getReadability('<title>this is my title</title><article class="awesomecontent">' . str_repeat('<p>This is an awesome text with some links, here there are the awesome</p>', 7) . '<p></p></article>', 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -294,7 +282,6 @@ public function testTitle(): void
     public function testTitleWithDash(): void
     {
         $readability = $this->getReadability('<title> title2 - title3 </title><article class="awesomecontent">' . str_repeat('<p>This is an awesome text with some links, here there are the awesome</p>', 7) . '<p></p></article>', 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -308,7 +295,6 @@ public function testTitleWithDash(): void
     public function testTitleWithDoubleDot(): void
     {
         $readability = $this->getReadability('<title> title2 : title3 </title><article class="awesomecontent">' . str_repeat('<p>This is an awesome text with some links, here there are the awesome</p>', 7) . '<p></p></article>', 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -322,7 +308,6 @@ public function testTitleWithDoubleDot(): void
     public function testTitleTooShortUseH1(): void
     {
         $readability = $this->getReadability('<title>too short</title><h1>this is my h1 title !</h1><article class="awesomecontent">' . str_repeat('<p>This is an awesome text with some links, here there are the awesome</p>', 7) . '<p></p></article>', 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -369,7 +354,6 @@ public function testAutoClosingIframeNotThrowingException(): void
                 </html>';
 
             $readability = $this->getReadability($data, 'http://iosgames.ru/?p=22030');
-            $readability->debug = true;
 
             $res = $readability->init();
 
@@ -437,7 +421,6 @@ public function testAppendIdAlreadyHere(): void
             </html>';
 
         $readability = $this->getReadability($data, 'http://0.0.0.0');
-        $readability->debug = true;
 
         $res = $readability->init();
 
@@ -474,7 +457,6 @@ public function testChildNodeGoneNull(): void
         $html = (string) file_get_contents('tests/fixtures/childNodeGoesNull.html');
 
         $readability = $this->getReadability($html, 'http://0.0.0.0');
-        $readability->debug = true;
         $readability->convertLinksToFootnotes = true;
         $res = $readability->init();
 
@@ -487,7 +469,6 @@ public function testKeepFootnotes(): void
         $html = (string) file_get_contents('tests/fixtures/keepFootnotes.html');
 
         $readability = $this->getReadability($html, 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -501,7 +482,6 @@ public function testWithWipedBody(): void
         $html = (string) file_get_contents('tests/fixtures/wipedBody.html');
 
         $readability = $this->getReadability($html, 'http://0.0.0.0', 'libxml', false);
-        $readability->debug = true;
         $res = $readability->init();
 
         $this->assertTrue($res);
@@ -540,7 +520,6 @@ public function dataForVisibleNode(): array
     public function testVisibleNode(string $content, bool $shouldBeVisible): void
     {
         $readability = $this->getReadability($content, 'http://0.0.0.0');
-        $readability->debug = true;
         $res = $readability->init();
 
         if ($shouldBeVisible) {

From dfa7cb08cdcb141b8bb75dd1ee019c422ce3d4e7 Mon Sep 17 00:00:00 2001
From: Jan Tojnar <jtojnar@gmail.com>
Date: Sat, 16 Mar 2024 17:57:49 +0100
Subject: [PATCH 02/12] Fix missing return value in grabArticle

Not sure if this is expected but at least it works the same as before.
---
 src/Readability.php | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/Readability.php b/src/Readability.php
index aa49727..22b2ee1 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -1039,7 +1039,7 @@ protected function grabArticle(?\DOMElement $page = null)
 
             foreach ($ancestors as $level => $ancestor) {
                 if (!$ancestor->nodeName || !$ancestor->parentNode) {
-                    return;
+                    return false;
                 }
 
                 if (!$ancestor->hasAttribute('readability')) {

From c83d33738aa84752321c57401dcb4d2a6ed8269d Mon Sep 17 00:00:00 2001
From: Jan Tojnar <jtojnar@gmail.com>
Date: Sat, 16 Mar 2024 18:33:56 +0100
Subject: [PATCH 03/12] Use helpers for content score manipulation
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

`DOMAttr::$value` must be a `string`.

Let’s add helpers for manipulating the `readability` attribute
so that we do not have to keep casting it from and to `string`
in order to appease `strict_types`.
---
 src/Readability.php | 59 ++++++++++++++++++++++++++++++---------------
 1 file changed, 40 insertions(+), 19 deletions(-)

diff --git a/src/Readability.php b/src/Readability.php
index 22b2ee1..e7e1db5 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -616,7 +616,7 @@ public function cleanConditionally(\DOMElement $e, string $tag): void
         for ($i = $curTagsLength - 1; $i >= 0; --$i) {
             $node = $tagsList->item($i);
             $weight = $this->getWeight($node);
-            $contentScore = ($node->hasAttribute('readability')) ? (int) $node->getAttribute('readability') : 0;
+            $contentScore = self::getContentScore($node);
             $this->logger->debug('Start conditional cleaning of ' . $node->getNodePath() . ' (class=' . $node->getAttribute('class') . '; id=' . $node->getAttribute('id') . ')' . (($node->hasAttribute('readability')) ? (' with score ' . $node->getAttribute('readability')) : ''));
 
             // XXX Incomplete implementation
@@ -829,29 +829,26 @@ protected function initializeNode(\DOMElement $node): void
             return;
         }
 
-        $readability = $this->dom->createAttribute('readability');
-        // this is our contentScore
-        $readability->value = 0;
-        $node->setAttributeNode($readability);
+        $contentScore = 0;
 
         // using strtoupper just in case
         switch (strtoupper($node->tagName)) {
             case 'ARTICLE':
-                $readability->value += 15;
+                $contentScore += 15;
                 // no break
             case 'DIV':
-                $readability->value += 5;
+                $contentScore += 5;
                 break;
             case 'PRE':
             case 'CODE':
             case 'TD':
             case 'BLOCKQUOTE':
             case 'FIGURE':
-                $readability->value += 3;
+                $contentScore += 3;
                 break;
             case 'SECTION':
                 // often misused
-                // $readability->value += 2;
+                // $contentScore += 2;
                 break;
             case 'OL':
             case 'UL':
@@ -859,7 +856,7 @@ protected function initializeNode(\DOMElement $node): void
             case 'DD':
             case 'DT':
             case 'LI':
-                $readability->value -= 3;
+                $contentScore -= 3;
                 break;
             case 'ASIDE':
             case 'FOOTER':
@@ -870,7 +867,7 @@ protected function initializeNode(\DOMElement $node): void
             case 'TEXTAREA':
             case 'INPUT':
             case 'NAV':
-                $readability->value -= 3;
+                $contentScore -= 3;
                 break;
             case 'H1':
             case 'H2':
@@ -880,11 +877,15 @@ protected function initializeNode(\DOMElement $node): void
             case 'H6':
             case 'TH':
             case 'HGROUP':
-                $readability->value -= 5;
+                $contentScore -= 5;
                 break;
         }
 
-        $readability->value += $this->getWeight($node);
+        $contentScore += $this->getWeight($node);
+
+        $readability = $this->dom->createAttribute('readability');
+        $readability->value = (string) $contentScore;
+        $node->setAttributeNode($readability);
     }
 
     /**
@@ -1054,7 +1055,8 @@ protected function grabArticle(?\DOMElement $page = null)
                 } else {
                     $scoreDivider = $level * 3;
                 }
-                $ancestor->getAttributeNode('readability')->value += $contentScore / $scoreDivider;
+
+                self::updateContentScore($ancestor, fn ($prevScore) => $prevScore + $contentScore / $scoreDivider);
             }
         }
 
@@ -1069,7 +1071,7 @@ protected function grabArticle(?\DOMElement $page = null)
                 $node = $candidates->item($c);
                 // node should be readable but not inside of an article otherwise it's probably non-readable block
                 if ($node->hasAttribute('readability') && (int) $node->getAttributeNode('readability')->value < 40 && ($node->parentNode ? 0 !== strcasecmp($node->parentNode->tagName, 'article') : true)) {
-                    $this->logger->debug('Removing unlikely candidate (using note) ' . $node->getNodePath() . ' by "' . $node->tagName . '" with readability ' . ($node->hasAttribute('readability') ? (int) $node->getAttributeNode('readability')->value : 0));
+                    $this->logger->debug('Removing unlikely candidate (using note) ' . $node->getNodePath() . ' by "' . $node->tagName . '" with readability ' . self::getContentScore($node));
                     $node->parentNode->removeChild($node);
                 }
             }
@@ -1098,14 +1100,13 @@ protected function grabArticle(?\DOMElement $page = null)
                 // Scale the final candidates score based on link density. Good content should have a
                 // relatively small link density (5% or less) and be mostly unaffected by this operation.
                 // If not for this we would have used XPath to find maximum @readability.
-                $readability = $item->getAttributeNode('readability');
-                $readability->value = round($readability->value * (1 - $this->getLinkDensity($item)), 0, \PHP_ROUND_HALF_UP);
+                self::updateContentScore($item, fn ($prevScore) => round($prevScore * (1 - $this->getLinkDensity($item)), 0, \PHP_ROUND_HALF_UP));
 
                 for ($t = 0; $t < 5; ++$t) {
                     $aTopCandidate = $topCandidates[$t];
 
-                    if (!$aTopCandidate || $readability->value > (int) $aTopCandidate->getAttribute('readability')) {
-                        $this->logger->debug('Candidate: ' . $item->getNodePath() . ' (' . $item->getAttribute('class') . ':' . $item->getAttribute('id') . ') with score ' . $readability->value);
+                    if (!$aTopCandidate || self::getContentScore($item) > self::getContentScore($aTopCandidate)) {
+                        $this->logger->debug('Candidate: ' . $item->getNodePath() . ' (' . $item->getAttribute('class') . ':' . $item->getAttribute('id') . ') with score ' . self::getContentScore($item));
                         array_splice($topCandidates, $t, 0, [$item]);
                         if (\count($topCandidates) > 5) {
                             array_pop($topCandidates);
@@ -1376,6 +1377,26 @@ protected function reinitBody(): void
         }
     }
 
+    /**
+     * Updates the content score for the given element using the provided function.
+     *
+     * @param callable(float): float $f
+     */
+    private static function updateContentScore(\DOMElement $element, callable $f): void
+    {
+        $readabilityAttr = $element->getAttributeNode('readability');
+        $prevScore = (float) $readabilityAttr->value;
+        $readabilityAttr->value = (string) $f($prevScore);
+    }
+
+    /**
+     * Gets the content score for given element.
+     */
+    private static function getContentScore(\DOMElement $element): float
+    {
+        return $element->hasAttribute('readability') ? (float) $element->getAttribute('readability') : 0;
+    }
+
     /**
      * Load HTML in a DOMDocument.
      * Apply Pre filters

From 7a2ae65f7289fb2226d1b48c386f979df86c62fa Mon Sep 17 00:00:00 2001
From: Jan Tojnar <jtojnar@gmail.com>
Date: Sat, 16 Mar 2024 17:15:31 +0100
Subject: [PATCH 04/12] Use PHP type hints instead of PHPDoc

PHP 7.4 supports property type hints.

- `getInnerText` actually accepts `DOMNode`, not `DOMElement`, since it can be e.g. `DOMText`.
  Really, we just need the `textContent` property.
- Since we now use property type hints,
  we can drop the `no_null_property_initialization` PHP-CS-Fixer rule.
---
 .php-cs-fixer.php         |   2 -
 src/Readability.php       | 135 ++++++++++++++++++++++++--------------
 tests/ReadabilityTest.php |   8 +--
 3 files changed, 88 insertions(+), 57 deletions(-)

diff --git a/.php-cs-fixer.php b/.php-cs-fixer.php
index 5f09a0c..d0c927c 100644
--- a/.php-cs-fixer.php
+++ b/.php-cs-fixer.php
@@ -28,8 +28,6 @@
         'concat_space' => ['spacing' => 'one'],
         // Pulled in by @Symfony:risky but we still support PHP 7.4
         'modernize_strpos' => false,
-        // Pulled in by @Symfony, we cannot add property types until we bump PHP to ≥ 7.4
-        'no_null_property_initialization' => false,
     ])
     ->setFinder($finder)
 ;
diff --git a/src/Readability.php b/src/Readability.php
index e7e1db5..e87aec6 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -24,26 +24,36 @@ class Readability implements LoggerAwareInterface
     public const MIN_ARTICLE_LENGTH = 200;
     public const MIN_NODE_LENGTH = 80;
     public const MAX_LINK_DENSITY = 0.25;
-    public $convertLinksToFootnotes = false;
-    public $revertForcedParagraphElements = false;
-    public $articleTitle;
-    public $articleContent;
-    public $original_html;
+
+    public bool $convertLinksToFootnotes = false;
+    public bool $revertForcedParagraphElements = false;
+
+    public ?\DOMElement $articleTitle;
+
+    public ?\DOMElement $articleContent;
+
+    public ?string $original_html;
+
+    public ?\DOMDocument $dom;
+
     /**
-     * @var \DOMDocument
+     * @var ?string URL where HTML was retrieved
      */
-    public $dom;
-    // optional - URL where HTML was retrieved
-    public $url = null;
-    // preserves more content (experimental)
-    public $lightClean = true;
-    public $tidied = false;
+    public ?string $url = null;
 
     /**
-     * All of the regular expressions in use within readability.
+     * @var bool preserves more content (experimental)
+     */
+    public bool $lightClean = true;
+
+    public bool $tidied = false;
+
+    /**
+     * @var array<string, string> All of the regular expressions in use within readability.
+     *
      * Defined up here so we don't instantiate them repeatedly in loops.
      */
-    public $regexps = [
+    public array $regexps = [
         'unlikelyCandidates' => '/-ad-|ai2html|banner|breadcrumbs|combx|comment|community|cover-wrap|disqus|extra|footer|gdpr|header|legends|menu|related|remark|replies|rss|shoutbox|sidebar|skyscraper|social|sponsor|supplemental|ad-break|agegate|pagination|pager|popup|yom-remote/i',
         'okMaybeItsACandidate' => '/article\b|contain|\bcontent|column|general|detail|shadow|lightbox|blog|body|entry|main|page|footnote|element/i',
         'positive' => '/read|full|article|body|\bcontent|contain|entry|main|markdown|media|page|attach|pagination|post|text|blog|story/i',
@@ -55,10 +65,18 @@ class Readability implements LoggerAwareInterface
         'hasContent' => '/\S$/',
         'isNotVisible' => '/display\s*:\s*none/',
     ];
-    public $defaultTagsToScore = ['section', 'h2', 'h3', 'h4', 'h5', 'h6', 'p', 'td', 'pre'];
-    // The commented out elements qualify as phrasing content but tend to be
-    // removed by readability when put into paragraphs, so we ignore them here.
-    public $phrasingElements = [
+
+    /**
+     * @var array<string>
+     */
+    public array $defaultTagsToScore = ['section', 'h2', 'h3', 'h4', 'h5', 'h6', 'p', 'td', 'pre'];
+
+    /**
+     * @var array<string>
+     */
+    public array $phrasingElements = [
+        // The commented out elements qualify as phrasing content but tend to be
+        // removed by readability when put into paragraphs, so we ignore them here.
         // "CANVAS", "IFRAME", "SVG", "VIDEO",
         'ABBR', 'AUDIO', 'B', 'BDO', 'BR', 'BUTTON', 'CITE', 'CODE', 'DATA',
         'DATALIST', 'DFN', 'EM', 'EMBED', 'I', 'IMG', 'INPUT', 'KBD', 'LABEL',
@@ -66,7 +84,11 @@ class Readability implements LoggerAwareInterface
         'RUBY', 'SAMP', 'SCRIPT', 'SELECT', 'SMALL', 'SPAN', 'STRONG', 'SUB',
         'SUP', 'TEXTAREA', 'TIME', 'VAR', 'WBR',
     ];
-    public $tidy_config = [
+
+    /**
+     * @var array<string, bool|string>
+     */
+    public array $tidy_config = [
         'tidy-mark' => false,
         'vertical-space' => false,
         'doctype' => 'omit',
@@ -90,21 +112,41 @@ class Readability implements LoggerAwareInterface
         'output-encoding' => 'utf8',
         'hide-comments' => true,
     ];
-    // article domain regexp for calibration
-    protected $domainRegExp = null;
-    protected $body = null;
-    // Cache the body HTML in case we need to re-use it later
-    protected $bodyCache = null;
-    // 1 | 2 | 4;   // Start with all processing flags set.
-    protected $flags = 7;
-    // indicates whether we were able to extract or not
-    protected $success = false;
-    protected $logger;
-    protected $parser;
-    protected $html;
-    protected $useTidy;
-    // raw HTML filters
-    protected $pre_filters = [
+
+    /**
+     * @var ?string article domain regexp for calibration
+     */
+    protected ?string $domainRegExp = null;
+
+    protected ?\DOMElement $body = null;
+
+    /**
+     * @var ?string Cache the body HTML in case we need to re-use it later
+     */
+    protected ?string $bodyCache = null;
+
+    /**
+     * @var int-mask-of<self::FLAG_*> start with all processing flags set
+     */
+    protected int $flags = self::FLAG_STRIP_UNLIKELYS | self::FLAG_WEIGHT_ATTRIBUTES | self::FLAG_CLEAN_CONDITIONALLY;
+
+    /**
+     * @var bool indicates whether we were able to extract or not
+     */
+    protected bool $success = false;
+
+    protected LoggerInterface $logger;
+
+    protected string $parser;
+
+    protected string $html;
+
+    protected bool $useTidy;
+
+    /**
+     * @var array<string, string> raw HTML filters
+     */
+    protected array $pre_filters = [
         // remove spans as we redefine styles and they're probably special-styled
         '!</?span[^>]*>!is' => '',
         // HACK: firewall-filtered content
@@ -116,8 +158,11 @@ class Readability implements LoggerAwareInterface
         // replace fonts to spans
         '!<(/?)font[^>]*>!is' => '<\\1span>',
     ];
-    // output HTML filters
-    protected $post_filters = [
+
+    /**
+     * @var array<string, string> output HTML filters
+     */
+    protected array $post_filters = [
         // replace excessive br's
         '/<br\s*\/?>\s*<p/i' => '<p',
         // replace empty tags that break layouts
@@ -157,20 +202,16 @@ public function setLogger(LoggerInterface $logger): void
 
     /**
      * Get article title element.
-     *
-     * @return \DOMElement
      */
-    public function getTitle()
+    public function getTitle(): \DOMElement
     {
         return $this->articleTitle;
     }
 
     /**
      * Get article content element.
-     *
-     * @return \DOMElement
      */
-    public function getContent()
+    public function getContent(): \DOMElement
     {
         return $this->articleContent;
     }
@@ -452,12 +493,8 @@ public function prepArticle(\DOMNode $articleContent): void
     /**
      * Get the inner text of a node.
      * This also strips out any excess whitespace to be found.
-     *
-     * @param \DOMElement $e
-     * @param bool        $normalizeSpaces (default: true)
-     * @param bool        $flattenLines    (default: false)
      */
-    public function getInnerText($e, bool $normalizeSpaces = true, bool $flattenLines = false): string
+    public function getInnerText(?\DOMNode $e, bool $normalizeSpaces = true, bool $flattenLines = false): string
     {
         if (null === $e || !isset($e->textContent) || '' === $e->textContent) {
             return '';
@@ -750,10 +787,8 @@ public function removeFlag(int $flag): void
 
     /**
      * Get the article title as an H1.
-     *
-     * @return \DOMElement
      */
-    protected function getArticleTitle()
+    protected function getArticleTitle(): \DOMElement
     {
         try {
             $curTitle = $origTitle = $this->getInnerText($this->dom->getElementsByTagName('title')->item(0));
diff --git a/tests/ReadabilityTest.php b/tests/ReadabilityTest.php
index 2fee95f..cce4568 100644
--- a/tests/ReadabilityTest.php
+++ b/tests/ReadabilityTest.php
@@ -10,10 +10,8 @@
 
 class ReadabilityTest extends \PHPUnit\Framework\TestCase
 {
-    /** @var TestHandler */
-    public $logHandler;
-    /** @var LoggerInterface */
-    public $logger;
+    public TestHandler $logHandler;
+    public LoggerInterface $logger;
 
     /**
      * @requires extension tidy
@@ -323,7 +321,7 @@ public function testAutoClosingIframeNotThrowingException(): void
         $oldErrorReporting = error_reporting(\E_ALL | \E_STRICT);
         $oldDisplayErrors = ini_set('display_errors', '1');
         // dummy function to be used to the next test
-        set_error_handler(function (int $errno, string $errstr, string $errfile, int $errline, array $errcontext) {
+        set_error_handler(function (int $errno, string $errstr, string $errfile, int $errline, array $errcontext): bool {
             throw new \Exception($errstr, $errno);
         }, \E_ALL | \E_STRICT);
 

From 1acfc6fede1e256a6a2e3c1ecc777cdd3168e68e Mon Sep 17 00:00:00 2001
From: Jan Tojnar <jtojnar@gmail.com>
Date: Sat, 16 Mar 2024 21:51:43 +0100
Subject: [PATCH 05/12] Use JSLikeHTMLElement in type hints

It is more specific then DOMElement.

This allows us to get rid of the assertions in tests.
---
 src/Readability.php       | 12 ++++++------
 tests/ReadabilityTest.php | 41 ---------------------------------------
 2 files changed, 6 insertions(+), 47 deletions(-)

diff --git a/src/Readability.php b/src/Readability.php
index e87aec6..8cf7202 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -28,9 +28,9 @@ class Readability implements LoggerAwareInterface
     public bool $convertLinksToFootnotes = false;
     public bool $revertForcedParagraphElements = false;
 
-    public ?\DOMElement $articleTitle;
+    public ?JSLikeHTMLElement $articleTitle;
 
-    public ?\DOMElement $articleContent;
+    public ?JSLikeHTMLElement $articleContent;
 
     public ?string $original_html;
 
@@ -203,7 +203,7 @@ public function setLogger(LoggerInterface $logger): void
     /**
      * Get article title element.
      */
-    public function getTitle(): \DOMElement
+    public function getTitle(): JSLikeHTMLElement
     {
         return $this->articleTitle;
     }
@@ -211,7 +211,7 @@ public function getTitle(): \DOMElement
     /**
      * Get article content element.
      */
-    public function getContent(): \DOMElement
+    public function getContent(): JSLikeHTMLElement
     {
         return $this->articleContent;
     }
@@ -394,7 +394,7 @@ public function addFootnotes(\DOMElement $articleContent): void
      */
     public function prepArticle(\DOMNode $articleContent): void
     {
-        if (!$articleContent instanceof \DOMElement) {
+        if (!$articleContent instanceof JSLikeHTMLElement) {
             return;
         }
 
@@ -590,7 +590,7 @@ public function getWeight(\DOMElement $e): int
     /**
      * Remove extraneous break tags from a node.
      */
-    public function killBreaks(\DOMElement $node): void
+    public function killBreaks(JSLikeHTMLElement $node): void
     {
         $html = $node->getInnerHTML();
         $html = preg_replace($this->regexps['killBreaks'], '<br />', $html);
diff --git a/tests/ReadabilityTest.php b/tests/ReadabilityTest.php
index cce4568..21f9bf7 100644
--- a/tests/ReadabilityTest.php
+++ b/tests/ReadabilityTest.php
@@ -5,7 +5,6 @@
 use Monolog\Handler\TestHandler;
 use Monolog\Logger;
 use Psr\Log\LoggerInterface;
-use Readability\JSLikeHTMLElement;
 use Readability\Readability;
 
 class ReadabilityTest extends \PHPUnit\Framework\TestCase
@@ -78,8 +77,6 @@ public function testInitNoContent(): void
         $res = $readability->init();
 
         $this->assertFalse($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('Sorry, Readability was unable to parse this page for content.', $readability->getContent()->getInnerHtml());
     }
@@ -90,8 +87,6 @@ public function testInitP(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertStringContainsString('<div readability=', $readability->getContent()->getInnerHtml());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is the awesome content :)', $readability->getContent()->getInnerHtml());
@@ -103,8 +98,6 @@ public function testInitDivP(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertStringContainsString('<div readability=', $readability->getContent()->getInnerHtml());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is the awesome content :)', $readability->getContent()->getInnerHtml());
@@ -116,8 +109,6 @@ public function testInitDiv(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertStringContainsString('<div readability=', $readability->getContent()->getInnerHtml());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is the awesome content :)', $readability->getContent()->getInnerHtml());
@@ -130,8 +121,6 @@ public function testWithFootnotes(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertStringContainsString('<div readability=', $readability->getContent()->getInnerHtml());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
@@ -146,8 +135,6 @@ public function testStandardClean(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertStringContainsString('<div readability=', $readability->getContent()->getInnerHtml());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
@@ -161,8 +148,6 @@ public function testWithIframe(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertStringContainsString('<div readability=', $readability->getContent()->getInnerHtml());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
@@ -175,8 +160,6 @@ public function testWithArticle(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertStringContainsString('alt="article"', $readability->getContent()->getInnerHtml());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
@@ -189,8 +172,6 @@ public function testWithAside(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
         $this->assertStringNotContainsString('<aside>', $readability->getContent()->getInnerHtml());
@@ -203,8 +184,6 @@ public function testWithClasses(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertStringContainsString('alt="article"', $readability->getContent()->getInnerHtml());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
@@ -218,8 +197,6 @@ public function testWithClassesWithoutLightClean(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertStringContainsString('alt="article"', $readability->getContent()->getInnerHtml());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
@@ -232,8 +209,6 @@ public function testWithTd(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
     }
@@ -244,8 +219,6 @@ public function testWithSameClasses(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
         $this->assertStringContainsString('This text is also an awesome text and you should know that', $readability->getContent()->getInnerHtml());
@@ -257,8 +230,6 @@ public function testWithScript(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertEmpty($readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
         $this->assertStringNotContainsString('This text is also an awesome text and you should know that', $readability->getContent()->getInnerHtml());
@@ -270,8 +241,6 @@ public function testTitle(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertSame('this is my title', $readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
         $this->assertStringNotContainsString('This text is also an awesome text and you should know that', $readability->getContent()->getInnerHtml());
@@ -283,8 +252,6 @@ public function testTitleWithDash(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertSame('title2 - title3', $readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
         $this->assertStringNotContainsString('This text is also an awesome text and you should know that', $readability->getContent()->getInnerHtml());
@@ -296,8 +263,6 @@ public function testTitleWithDoubleDot(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertSame('title2 : title3', $readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
         $this->assertStringNotContainsString('This text is also an awesome text and you should know that', $readability->getContent()->getInnerHtml());
@@ -309,8 +274,6 @@ public function testTitleTooShortUseH1(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
         $this->assertSame('this is my h1 title !', $readability->getTitle()->getInnerHtml());
         $this->assertStringContainsString('This is an awesome text with some links, here there are', $readability->getContent()->getInnerHtml());
         $this->assertStringNotContainsString('This text is also an awesome text and you should know that', $readability->getContent()->getInnerHtml());
@@ -356,8 +319,6 @@ public function testAutoClosingIframeNotThrowingException(): void
             $res = $readability->init();
 
             $this->assertTrue($res);
-            $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-            $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
             $this->assertStringContainsString('<iframe src="https://www.youtube.com/embed/PUep6xNeKjA" width="560" height="315" frameborder="0" allowfullscreen="allowfullscreen"> </iframe>', $readability->getContent()->getInnerHtml());
             $this->assertStringContainsString('3D Touch', $readability->getTitle()->getInnerHtml());
         } finally {
@@ -423,8 +384,6 @@ public function testAppendIdAlreadyHere(): void
         $res = $readability->init();
 
         $this->assertTrue($res);
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getContent());
-        $this->assertInstanceOf(JSLikeHTMLElement::class, $readability->getTitle());
     }
 
     public function testPostFilters(): void

From e9c9dbcc2567c26e78fd07bc718bc79ba8b082d6 Mon Sep 17 00:00:00 2001
From: Jan Tojnar <jtojnar@gmail.com>
Date: Sat, 16 Mar 2024 22:05:09 +0100
Subject: [PATCH 06/12] JSLikeHTMLElement: Get rid of magic properties

Will make it easier to statically analyze with PHPStan.

This is a BC break.
---
 src/JSLikeHTMLElement.php | 98 ++++++++++++---------------------------
 1 file changed, 29 insertions(+), 69 deletions(-)

diff --git a/src/JSLikeHTMLElement.php b/src/JSLikeHTMLElement.php
index 18116a6..23585ff 100644
--- a/src/JSLikeHTMLElement.php
+++ b/src/JSLikeHTMLElement.php
@@ -1,57 +1,41 @@
 <?php
 
+// SPDX-FileCopyrightText: 2011 Keyvan Minoukadeh - http://www.keyvan.net - keyvan@keyvan.net
+// SPDX-License-Identifier: Apache-2.0
+
 namespace Readability;
 
 /**
- * JavaScript-like HTML DOM Element.
+ * Wrapper for DOMElement adding methods for accessing string representation of inner HTML contents.
  *
- * This class extends PHP's DOMElement to allow
- * users to get and set the innerHTML property of
- * HTML elements in the same way it's done in
- * JavaScript.
+ * Inspired by JavaScript innerHTML property.
+ * https://developer.mozilla.org/en-US/docs/Web/API/Element/innerHTML
  *
  * Example usage:
- *     require_once 'JSLikeHTMLElement.php';
- *     header('Content-Type: text/plain');
- *     $doc = new DOMDocument();
- *     $doc->registerNodeClass('DOMElement', 'JSLikeHTMLElement');
- *     $doc->loadHTML('<div><p>Para 1</p><p>Para 2</p></div>');
- *     $elem = $doc->getElementsByTagName('div')->item(0);
- *
- *     // print innerHTML
- *     echo $elem->innerHTML; // prints '<p>Para 1</p><p>Para 2</p>'
- *     echo "\n\n";
  *
- *     // set innerHTML
- *     $elem->innerHTML = '<a href="http://fivefilters.org">FiveFilters.org</a>';
- *     echo $elem->innerHTML; // prints '<a href="http://fivefilters.org">FiveFilters.org</a>'
- *     echo "\n\n";
+ * ```php
+ * $doc = new DOMDocument();
+ * $doc->loadHTML('<div><p>Para 1</p><p>Para 2</p></div>');
+ * $elem = $doc->getElementsByTagName('div')->item(0);
  *
- *     // print document (with our changes)
- *     echo $doc->saveXML();
+ * // Get inner HTML
+ * assert($elem->getInnerHtml() === '<p>Para 1</p><p>Para 2</p>');
  *
- * @author Keyvan Minoukadeh - http://www.keyvan.net - keyvan@keyvan.net
+ * // Set inner HTML
+ * $elem->setInnerHtml('<a href="http://fivefilters.org">FiveFilters.org</a>');
+ * assert($elem->getInnerHtml() === '<a href="http://fivefilters.org">FiveFilters.org</a>');
  *
- * @see http://fivefilters.org (the project this was written for)
+ * // print document (with our changes)
+ * echo $doc->saveXML();
+ * ```
  */
-class JSLikeHTMLElement extends \DOMElement
+final class JSLikeHTMLElement extends \DOMElement
 {
     /**
-     * Used for setting innerHTML like it's done in JavaScript:.
-     *
-     * ```php
-     * $div->innerHTML = '<h2>Chapter 2</h2><p>The story begins...</p>';
-     * ```
+     * Sets inner HTML.
      */
-    public function __set($name, $value)
+    public function setInnerHtml(string $value): void
     {
-        if ('innerHTML' !== $name) {
-            $trace = debug_backtrace();
-            trigger_error('Undefined property via __set(): ' . $name . ' in ' . $trace[0]['file'] . ' on line ' . $trace[0]['line'], \E_USER_NOTICE);
-
-            return;
-        }
-
         // first, empty the element
         if (isset($this->childNodes)) {
             for ($x = $this->childNodes->length - 1; $x >= 0; --$x) {
@@ -81,7 +65,7 @@ public function __set($name, $value)
             $f = new \DOMDocument();
 
             // Using <htmlfragment> will generate a warning, but so will bad HTML
-            // (and by this point, bad HTML is what we've got).
+            // (and by element point, bad HTML is what we've got).
             // We use it (and suppress the warning) because an HTML fragment will
             // be wrapped around <html><body> tags which we don't really want to keep.
             // Note: despite the warning, if loadHTML succeeds it will return true.
@@ -102,42 +86,18 @@ public function __set($name, $value)
     }
 
     /**
-     * Used for getting innerHTML like it's done in JavaScript:.
-     *
-     * ```php
-     * $string = $div->innerHTML;
-     * ```
+     * Gets inner HTML.
      */
-    public function __get($name)
+    public function getInnerHtml(): string
     {
-        if ('innerHTML' === $name) {
-            $inner = '';
+        $inner = '';
 
-            if (isset($this->childNodes)) {
-                foreach ($this->childNodes as $child) {
-                    $inner .= $this->ownerDocument->saveXML($child);
-                }
+        if (isset($this->childNodes)) {
+            foreach ($this->childNodes as $child) {
+                $inner .= $this->ownerDocument->saveXML($child);
             }
-
-            return $inner;
         }
 
-        $trace = debug_backtrace();
-        trigger_error('Undefined property via __get(): ' . $name . ' in ' . $trace[0]['file'] . ' on line ' . $trace[0]['line'], \E_USER_NOTICE);
-    }
-
-    public function __toString()
-    {
-        return '[' . $this->tagName . ']';
-    }
-
-    public function getInnerHtml()
-    {
-        return $this->__get('innerHTML');
-    }
-
-    public function setInnerHtml($value)
-    {
-        return $this->__set('innerHTML', $value);
+        return $inner;
     }
 }

From b9fdf04da713319be6ebce7a5f28ac999503dd49 Mon Sep 17 00:00:00 2001
From: Jan Tojnar <jtojnar@gmail.com>
Date: Sat, 16 Mar 2024 17:16:21 +0100
Subject: [PATCH 07/12] Enable strict_types

---
 src/JSLikeHTMLElement.php | 2 ++
 src/Readability.php       | 2 ++
 tests/ReadabilityTest.php | 2 ++
 3 files changed, 6 insertions(+)

diff --git a/src/JSLikeHTMLElement.php b/src/JSLikeHTMLElement.php
index 23585ff..18fe9f7 100644
--- a/src/JSLikeHTMLElement.php
+++ b/src/JSLikeHTMLElement.php
@@ -1,5 +1,7 @@
 <?php
 
+declare(strict_types=1);
+
 // SPDX-FileCopyrightText: 2011 Keyvan Minoukadeh - http://www.keyvan.net - keyvan@keyvan.net
 // SPDX-License-Identifier: Apache-2.0
 
diff --git a/src/Readability.php b/src/Readability.php
index 8cf7202..c0ab7e4 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -1,5 +1,7 @@
 <?php
 
+declare(strict_types=1);
+
 namespace Readability;
 
 use Masterminds\HTML5;
diff --git a/tests/ReadabilityTest.php b/tests/ReadabilityTest.php
index 21f9bf7..390f85a 100644
--- a/tests/ReadabilityTest.php
+++ b/tests/ReadabilityTest.php
@@ -1,5 +1,7 @@
 <?php
 
+declare(strict_types=1);
+
 namespace Tests\Readability;
 
 use Monolog\Handler\TestHandler;

From a12629ef7855e235ca1b395579eca85d0b810279 Mon Sep 17 00:00:00 2001
From: Jan Tojnar <jtojnar@gmail.com>
Date: Sat, 16 Mar 2024 20:24:17 +0100
Subject: [PATCH 08/12] Raise PHPStan to level 4

Nothing affecting correctness, just stuff making it easier for PHPStan to reason about the code.

Remove `$errcontext` argument in `set_error_handler` since it is removed in PHP 8.
---
 phpstan.neon              |  2 +-
 src/Readability.php       | 14 +++++++++-----
 tests/ReadabilityTest.php |  2 +-
 3 files changed, 11 insertions(+), 7 deletions(-)

diff --git a/phpstan.neon b/phpstan.neon
index 895d58d..252ad62 100644
--- a/phpstan.neon
+++ b/phpstan.neon
@@ -1,5 +1,5 @@
 parameters:
-    level: 1
+    level: 4
     paths:
         - src
         - tests
diff --git a/src/Readability.php b/src/Readability.php
index c0ab7e4..5221f51 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -423,6 +423,7 @@ public function prepArticle(\DOMNode $articleContent): void
         }
 
         // Remove service data-candidate attribute.
+        /** @var \DOMNodeList<\DOMElement> */
         $elems = $xpath->query('.//*[@data-candidate]', $articleContent);
         for ($i = $elems->length - 1; $i >= 0; --$i) {
             $elems->item($i)->removeAttribute('data-candidate');
@@ -1102,12 +1103,13 @@ protected function grabArticle(?\DOMElement $page = null)
          * This is faster to do before scoring but safer after.
          */
         if ($this->flagIsActive(self::FLAG_STRIP_UNLIKELYS) && $xpath) {
+            /** @var \DOMNodeList<\DOMElement> */
             $candidates = $xpath->query('.//*[(self::footer and count(//footer)<2) or (self::aside and count(//aside)<2)]', $page->documentElement);
 
             for ($c = $candidates->length - 1; $c >= 0; --$c) {
                 $node = $candidates->item($c);
                 // node should be readable but not inside of an article otherwise it's probably non-readable block
-                if ($node->hasAttribute('readability') && (int) $node->getAttributeNode('readability')->value < 40 && ($node->parentNode ? 0 !== strcasecmp($node->parentNode->tagName, 'article') : true)) {
+                if ($node->hasAttribute('readability') && (int) $node->getAttributeNode('readability')->value < 40 && ($node->parentNode instanceof \DOMElement ? 0 !== strcasecmp($node->parentNode->tagName, 'article') : true)) {
                     $this->logger->debug('Removing unlikely candidate (using note) ' . $node->getNodePath() . ' by "' . $node->tagName . '" with readability ' . self::getContentScore($node));
                     $node->parentNode->removeChild($node);
                 }
@@ -1128,6 +1130,7 @@ protected function grabArticle(?\DOMElement $page = null)
         $topCandidates = array_fill(0, 5, null);
         if ($xpath) {
             // Using array of DOMElements after deletion is a path to DOOMElement.
+            /** @var \DOMNodeList<\DOMElement> */
             $candidates = $xpath->query('.//*[@data-candidate]', $page->documentElement);
             $this->logger->debug('Candidates: ' . $candidates->length);
 
@@ -1154,6 +1157,7 @@ protected function grabArticle(?\DOMElement $page = null)
             }
         }
 
+        /** @var \DOMNodeList<\DOMElement> */
         $topCandidates = array_filter(
             $topCandidates,
             fn ($v, $idx) => 0 === $idx || null !== $v,
@@ -1276,7 +1280,7 @@ protected function grabArticle(?\DOMElement $page = null)
             $siblingNode = $siblingNodes->item($s);
             $siblingNodeName = $siblingNode->nodeName;
             $append = false;
-            $this->logger->debug('Looking at sibling node: ' . $siblingNode->getNodePath() . ((\XML_ELEMENT_NODE === $siblingNode->nodeType && $siblingNode->hasAttribute('readability')) ? (' with score ' . $siblingNode->getAttribute('readability')) : ''));
+            $this->logger->debug('Looking at sibling node: ' . $siblingNode->getNodePath() . (($siblingNode instanceof \DOMElement && $siblingNode->hasAttribute('readability')) ? (' with score ' . $siblingNode->getAttribute('readability')) : ''));
 
             if ($siblingNode->isSameNode($topCandidate)) {
                 $append = true;
@@ -1284,11 +1288,11 @@ protected function grabArticle(?\DOMElement $page = null)
                 $contentBonus = 0;
 
                 // Give a bonus if sibling nodes and top candidates have the same classname.
-                if (\XML_ELEMENT_NODE === $siblingNode->nodeType && $siblingNode->getAttribute('class') === $topCandidate->getAttribute('class') && '' !== $topCandidate->getAttribute('class')) {
+                if ($siblingNode instanceof \DOMElement && $siblingNode->getAttribute('class') === $topCandidate->getAttribute('class') && '' !== $topCandidate->getAttribute('class')) {
                     $contentBonus += ((int) $topCandidate->getAttribute('readability')) * 0.2;
                 }
 
-                if (\XML_ELEMENT_NODE === $siblingNode->nodeType && $siblingNode->hasAttribute('readability') && (((int) $siblingNode->getAttribute('readability')) + $contentBonus) >= $siblingScoreThreshold) {
+                if ($siblingNode instanceof \DOMElement && $siblingNode->hasAttribute('readability') && (((int) $siblingNode->getAttribute('readability')) + $contentBonus) >= $siblingScoreThreshold) {
                     $append = true;
                 } elseif (0 === strcasecmp($siblingNodeName, 'p')) {
                     $linkDensity = (int) $this->getLinkDensity($siblingNode);
@@ -1518,7 +1522,7 @@ private function getAncestors(\DOMElement $node, int $maxDepth = 0): array
 
     private function isPhrasingContent($node): bool
     {
-        return \XML_TEXT_NODE === $node->nodeType
+        return $node instanceof \DOMText
             || \in_array(strtoupper($node->nodeName), $this->phrasingElements, true)
             || (
                 \in_array(strtoupper($node->nodeName), ['A', 'DEL', 'INS'], true)
diff --git a/tests/ReadabilityTest.php b/tests/ReadabilityTest.php
index 390f85a..cfc865d 100644
--- a/tests/ReadabilityTest.php
+++ b/tests/ReadabilityTest.php
@@ -286,7 +286,7 @@ public function testAutoClosingIframeNotThrowingException(): void
         $oldErrorReporting = error_reporting(\E_ALL | \E_STRICT);
         $oldDisplayErrors = ini_set('display_errors', '1');
         // dummy function to be used to the next test
-        set_error_handler(function (int $errno, string $errstr, string $errfile, int $errline, array $errcontext): bool {
+        set_error_handler(function (int $errno, string $errstr, string $errfile, int $errline): bool {
             throw new \Exception($errstr, $errno);
         }, \E_ALL | \E_STRICT);
 

From d2d7dcd476f9d76aaa0c48decfc64b102f6e014d Mon Sep 17 00:00:00 2001
From: Jan Tojnar <jtojnar@gmail.com>
Date: Sat, 16 Mar 2024 22:17:47 +0100
Subject: [PATCH 09/12] JSHtml

---
 src/Readability.php | 61 ++++++++++++++++++++++-----------------------
 1 file changed, 30 insertions(+), 31 deletions(-)

diff --git a/src/Readability.php b/src/Readability.php
index 5221f51..81f223f 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -120,7 +120,7 @@ class Readability implements LoggerAwareInterface
      */
     protected ?string $domainRegExp = null;
 
-    protected ?\DOMElement $body = null;
+    protected ?JSLikeHTMLElement $body = null;
 
     /**
      * @var ?string Cache the body HTML in case we need to re-use it later
@@ -262,6 +262,7 @@ public function init(): bool
 
         // Assume successful outcome
         $this->success = true;
+        /** @var \DOMNodeList<JSLikeHTMLElement> */
         $bodyElems = $this->dom->getElementsByTagName('body');
 
         // WTF multiple body nodes?
@@ -284,7 +285,7 @@ public function init(): bool
         $articleTitle = $this->getArticleTitle();
         $articleContent = $this->grabArticle();
 
-        if (!$articleContent) {
+        if (null === $articleContent) {
             $this->success = false;
             $articleContent = $this->dom->createElement('div');
             $articleContent->setAttribute('class', 'readability-content');
@@ -423,7 +424,7 @@ public function prepArticle(\DOMNode $articleContent): void
         }
 
         // Remove service data-candidate attribute.
-        /** @var \DOMNodeList<\DOMElement> */
+        /** @var \DOMNodeList<JSLikeHTMLElement> */
         $elems = $xpath->query('.//*[@data-candidate]', $articleContent);
         for ($i = $elems->length - 1; $i >= 0; --$i) {
             $elems->item($i)->removeAttribute('data-candidate');
@@ -519,7 +520,7 @@ public function getInnerText(?\DOMNode $e, bool $normalizeSpaces = true, bool $f
     /**
      * Remove the style attribute on every $e and under.
      */
-    public function cleanStyles(\DOMElement $e): void
+    public function cleanStyles(JSLikeHTMLElement $e): void
     {
         if (\is_object($e)) {
             $elems = $e->getElementsByTagName('*');
@@ -552,7 +553,7 @@ public function getWordCount(string $text): int
      * This is the amount of text that is inside a link divided by the total text in the node.
      * Can exclude external references to differentiate between simple text and menus/infoblocks.
      */
-    public function getLinkDensity(\DOMElement $e, bool $excludeExternal = false): float
+    public function getLinkDensity(JSLikeHTMLElement $e, bool $excludeExternal = false): float
     {
         $links = $e->getElementsByTagName('a');
         $textLength = mb_strlen($this->getInnerText($e, true, true));
@@ -575,7 +576,7 @@ public function getLinkDensity(\DOMElement $e, bool $excludeExternal = false): f
     /**
      * Get an element relative weight.
      */
-    public function getWeight(\DOMElement $e): int
+    public function getWeight(JSLikeHTMLElement $e): int
     {
         if (!$this->flagIsActive(self::FLAG_WEIGHT_ATTRIBUTES)) {
             return 0;
@@ -606,7 +607,7 @@ public function killBreaks(JSLikeHTMLElement $node): void
      *
      * Updated 2012-09-18 to preserve youtube/vimeo iframes
      */
-    public function clean(\DOMElement $e, string $tag): void
+    public function clean(JSLikeHTMLElement $e, string $tag): void
     {
         $targetList = $e->getElementsByTagName($tag);
         $isEmbed = ('audio' === $tag || 'video' === $tag || 'iframe' === $tag || 'object' === $tag || 'embed' === $tag);
@@ -638,7 +639,7 @@ public function clean(\DOMElement $e, string $tag): void
      * "Fishy" is an algorithm based on content length, classnames,
      * link density, number of images & embeds, etc.
      */
-    public function cleanConditionally(\DOMElement $e, string $tag): void
+    public function cleanConditionally(JSLikeHTMLElement $e, string $tag): void
     {
         if (!$this->flagIsActive(self::FLAG_CLEAN_CONDITIONALLY)) {
             return;
@@ -751,7 +752,7 @@ public function cleanConditionally(\DOMElement $e, string $tag): void
     /**
      * Clean out spurious headers from an Element. Checks things like classnames and link density.
      */
-    public function cleanHeaders(\DOMElement $e): void
+    public function cleanHeaders(JSLikeHTMLElement $e): void
     {
         for ($headerIndex = 1; $headerIndex < 3; ++$headerIndex) {
             $headers = $e->getElementsByTagName('h' . $headerIndex);
@@ -791,7 +792,7 @@ public function removeFlag(int $flag): void
     /**
      * Get the article title as an H1.
      */
-    protected function getArticleTitle(): \DOMElement
+    protected function getArticleTitle(): JSLikeHTMLElement
     {
         try {
             $curTitle = $origTitle = $this->getInnerText($this->dom->getElementsByTagName('title')->item(0));
@@ -861,7 +862,7 @@ protected function prepDocument(): void
      * Initialize a node with the readability object. Also checks the
      * className/id for special names to add to its score.
      */
-    protected function initializeNode(\DOMElement $node): void
+    protected function initializeNode(JSLikeHTMLElement $node): void
     {
         if (!isset($node->tagName)) {
             return;
@@ -929,10 +930,8 @@ protected function initializeNode(\DOMElement $node): void
     /**
      * Using a variety of metrics (content score, classname, element types), find the content that is
      * most likely to be the stuff a user wants to read. Then return it wrapped up in a div.
-     *
-     * @return \DOMElement|false
      */
-    protected function grabArticle(?\DOMElement $page = null)
+    protected function grabArticle(?JSLikeHTMLElement $page = null): ?JSLikeHTMLElement
     {
         if (!$page) {
             $page = $this->dom;
@@ -1078,7 +1077,7 @@ protected function grabArticle(?\DOMElement $page = null)
 
             foreach ($ancestors as $level => $ancestor) {
                 if (!$ancestor->nodeName || !$ancestor->parentNode) {
-                    return false;
+                    return null;
                 }
 
                 if (!$ancestor->hasAttribute('readability')) {
@@ -1103,13 +1102,13 @@ protected function grabArticle(?\DOMElement $page = null)
          * This is faster to do before scoring but safer after.
          */
         if ($this->flagIsActive(self::FLAG_STRIP_UNLIKELYS) && $xpath) {
-            /** @var \DOMNodeList<\DOMElement> */
+            /** @var \DOMNodeList<JSLikeHTMLElement> */
             $candidates = $xpath->query('.//*[(self::footer and count(//footer)<2) or (self::aside and count(//aside)<2)]', $page->documentElement);
 
             for ($c = $candidates->length - 1; $c >= 0; --$c) {
                 $node = $candidates->item($c);
                 // node should be readable but not inside of an article otherwise it's probably non-readable block
-                if ($node->hasAttribute('readability') && (int) $node->getAttributeNode('readability')->value < 40 && ($node->parentNode instanceof \DOMElement ? 0 !== strcasecmp($node->parentNode->tagName, 'article') : true)) {
+                if ($node->hasAttribute('readability') && (int) $node->getAttributeNode('readability')->value < 40 && ($node->parentNode instanceof JSLikeHTMLElement ? 0 !== strcasecmp($node->parentNode->tagName, 'article') : true)) {
                     $this->logger->debug('Removing unlikely candidate (using note) ' . $node->getNodePath() . ' by "' . $node->tagName . '" with readability ' . self::getContentScore($node));
                     $node->parentNode->removeChild($node);
                 }
@@ -1130,7 +1129,7 @@ protected function grabArticle(?\DOMElement $page = null)
         $topCandidates = array_fill(0, 5, null);
         if ($xpath) {
             // Using array of DOMElements after deletion is a path to DOOMElement.
-            /** @var \DOMNodeList<\DOMElement> */
+            /** @var \DOMNodeList<JSLikeHTMLElement> */
             $candidates = $xpath->query('.//*[@data-candidate]', $page->documentElement);
             $this->logger->debug('Candidates: ' . $candidates->length);
 
@@ -1157,7 +1156,7 @@ protected function grabArticle(?\DOMElement $page = null)
             }
         }
 
-        /** @var \DOMNodeList<\DOMElement> */
+        /** @var \DOMNodeList<JSLikeHTMLElement> */
         $topCandidates = array_filter(
             $topCandidates,
             fn ($v, $idx) => 0 === $idx || null !== $v,
@@ -1250,7 +1249,7 @@ protected function grabArticle(?\DOMElement $page = null)
         if (0 === strcasecmp($tagName, 'td') || 0 === strcasecmp($tagName, 'tr')) {
             $up = $topCandidate;
 
-            if ($up->parentNode instanceof \DOMElement) {
+            if ($up->parentNode instanceof JSLikeHTMLElement) {
                 $up = $up->parentNode;
 
                 if (0 === strcasecmp($up->tagName, 'table')) {
@@ -1280,7 +1279,7 @@ protected function grabArticle(?\DOMElement $page = null)
             $siblingNode = $siblingNodes->item($s);
             $siblingNodeName = $siblingNode->nodeName;
             $append = false;
-            $this->logger->debug('Looking at sibling node: ' . $siblingNode->getNodePath() . (($siblingNode instanceof \DOMElement && $siblingNode->hasAttribute('readability')) ? (' with score ' . $siblingNode->getAttribute('readability')) : ''));
+            $this->logger->debug('Looking at sibling node: ' . $siblingNode->getNodePath() . (($siblingNode instanceof JSLikeHTMLElement && $siblingNode->hasAttribute('readability')) ? (' with score ' . $siblingNode->getAttribute('readability')) : ''));
 
             if ($siblingNode->isSameNode($topCandidate)) {
                 $append = true;
@@ -1288,11 +1287,11 @@ protected function grabArticle(?\DOMElement $page = null)
                 $contentBonus = 0;
 
                 // Give a bonus if sibling nodes and top candidates have the same classname.
-                if ($siblingNode instanceof \DOMElement && $siblingNode->getAttribute('class') === $topCandidate->getAttribute('class') && '' !== $topCandidate->getAttribute('class')) {
+                if ($siblingNode instanceof JSLikeHTMLElement && $siblingNode->getAttribute('class') === $topCandidate->getAttribute('class') && '' !== $topCandidate->getAttribute('class')) {
                     $contentBonus += ((int) $topCandidate->getAttribute('readability')) * 0.2;
                 }
 
-                if ($siblingNode instanceof \DOMElement && $siblingNode->hasAttribute('readability') && (((int) $siblingNode->getAttribute('readability')) + $contentBonus) >= $siblingScoreThreshold) {
+                if ($siblingNode instanceof JSLikeHTMLElement && $siblingNode->hasAttribute('readability') && (((int) $siblingNode->getAttribute('readability')) + $contentBonus) >= $siblingScoreThreshold) {
                     $append = true;
                 } elseif (0 === strcasecmp($siblingNodeName, 'p')) {
                     $linkDensity = (int) $this->getLinkDensity($siblingNode);
@@ -1369,7 +1368,7 @@ protected function grabArticle(?\DOMElement $page = null)
                 return $this->grabArticle($this->body);
             }
 
-            return false;
+            return null;
         }
 
         return $articleContent;
@@ -1379,7 +1378,7 @@ protected function grabArticle(?\DOMElement $page = null)
      * Get an element weight by attribute.
      * Uses regular expressions to tell if this element looks good or bad.
      */
-    protected function weightAttribute(\DOMElement $element, string $attribute): int
+    protected function weightAttribute(JSLikeHTMLElement $element, string $attribute): int
     {
         if (!$element->hasAttribute($attribute)) {
             return 0;
@@ -1423,7 +1422,7 @@ protected function reinitBody(): void
      *
      * @param callable(float): float $f
      */
-    private static function updateContentScore(\DOMElement $element, callable $f): void
+    private static function updateContentScore(JSLikeHTMLElement $element, callable $f): void
     {
         $readabilityAttr = $element->getAttributeNode('readability');
         $prevScore = (float) $readabilityAttr->value;
@@ -1433,7 +1432,7 @@ private static function updateContentScore(\DOMElement $element, callable $f): v
     /**
      * Gets the content score for given element.
      */
-    private static function getContentScore(\DOMElement $element): float
+    private static function getContentScore(JSLikeHTMLElement $element): float
     {
         return $element->hasAttribute('readability') ? (float) $element->getAttribute('readability') : 0;
     }
@@ -1505,11 +1504,11 @@ private function loadHtml(): void
         $this->dom->registerNodeClass(\DOMElement::class, JSLikeHTMLElement::class);
     }
 
-    private function getAncestors(\DOMElement $node, int $maxDepth = 0): array
+    private function getAncestors(JSLikeHTMLElement $node, int $maxDepth = 0): array
     {
         $ancestors = [];
         $i = 0;
-        while ($node->parentNode instanceof \DOMElement) {
+        while ($node->parentNode instanceof JSLikeHTMLElement) {
             $ancestors[] = $node->parentNode;
             if (++$i === $maxDepth) {
                 break;
@@ -1537,7 +1536,7 @@ private function isPhrasingContent($node): bool
             );
     }
 
-    private function hasSingleTagInsideElement(\DOMElement $node, string $tag): bool
+    private function hasSingleTagInsideElement(JSLikeHTMLElement $node, string $tag): bool
     {
         if (1 !== $node->childNodes->length || $node->childNodes->item(0)->nodeName !== $tag) {
             return false;
@@ -1557,7 +1556,7 @@ private function hasSingleTagInsideElement(\DOMElement $node, string $tag): bool
      * Tidy must be configured to not clean the input for this function to
      * work as expected, see $this->tidy_config['clean']
      */
-    private function isNodeVisible(\DOMElement $node): bool
+    private function isNodeVisible(JSLikeHTMLElement $node): bool
     {
         return !(
             $node->hasAttribute('style')

From 273a95ae9e24975ebbc0e26637cec47c55a37687 Mon Sep 17 00:00:00 2001
From: Jan Tojnar <jtojnar@gmail.com>
Date: Sat, 16 Mar 2024 21:22:18 +0100
Subject: [PATCH 10/12] Fix hasSingleTagInsideElement method
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

It would fail for e.g. `<div> <p>foo</p> </div>`.

mozilla/readability uses children for the tag lookup, which return only elements.
PHP does not have children property so b580cf216d9001f82c866bb9a6c8bcad1cc862d8
mistakenly used `childNodes` instead, but that can return any node type.

Let’s filter the children ourselves.

Also add comments from mozilla/readability’s `_hasSingleTagInsideElement`.
---
 src/Readability.php | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

diff --git a/src/Readability.php b/src/Readability.php
index 81f223f..82f09ae 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -1536,16 +1536,23 @@ private function isPhrasingContent($node): bool
             );
     }
 
+    /**
+     * Checks if `$node` has only whitespace and a single element with `$tag` for the tag name.
+     * Returns false if `$node` contains non-empty text nodes
+     * or if it contains no element with given tag or more than 1 element.
+     */
     private function hasSingleTagInsideElement(JSLikeHTMLElement $node, string $tag): bool
     {
-        if (1 !== $node->childNodes->length || $node->childNodes->item(0)->nodeName !== $tag) {
-            return false;
+        $childNodes = iterator_to_array($node->childNodes);
+        $children = array_filter($childNodes, fn ($childNode) => $childNode instanceof \DOMElement);
+
+        // There should be exactly 1 element child with given tag
+        if (1 !== \count($children) || $children[0]->nodeName !== $tag) {
+            return null;
         }
 
-        $a = array_filter(
-            iterator_to_array($node->childNodes),
-            fn ($childNode) => $childNode instanceof \DOMText && preg_match($this->regexps['hasContent'], $this->getInnerText($childNode))
-        );
+        // And there should be no text nodes with real content
+        $a = array_filter($childNodes, fn ($childNode) => $childNode instanceof \DOMText && preg_match($this->regexps['hasContent'], $this->getInnerText($childNode)));
 
         return 0 === \count($a);
     }

From ccdd16649a6a9921f72ab20b4cd036320e2eba1c Mon Sep 17 00:00:00 2001
From: Jan Tojnar <jtojnar@gmail.com>
Date: Sat, 16 Mar 2024 21:33:29 +0100
Subject: [PATCH 11/12] Raise PHPStan version to 5

Converting `hasSingleTagInsideElement` into a type-safe getter
will allow PHPStan to know the type of `newNode` is `DOMElement`.
---
 phpstan.neon        |  2 +-
 src/Readability.php | 10 +++++-----
 2 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/phpstan.neon b/phpstan.neon
index 252ad62..cb60e7e 100644
--- a/phpstan.neon
+++ b/phpstan.neon
@@ -1,5 +1,5 @@
 parameters:
-    level: 4
+    level: 5
     paths:
         - src
         - tests
diff --git a/src/Readability.php b/src/Readability.php
index 82f09ae..6b53dc1 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -1035,8 +1035,8 @@ protected function grabArticle(?JSLikeHTMLElement $page = null): ?JSLikeHTMLElem
                         }
                     }
 
-                    if ($this->hasSingleTagInsideElement($node, 'p') && $this->getLinkDensity($node) < 0.25) {
-                        $newNode = $node->childNodes->item(0);
+                    $newNode = $this->getSingleTagInsideElement($node, 'p');
+                    if (null !== $newNode && $this->getLinkDensity($node) < 0.25) {
                         $node->parentNode->replaceChild($newNode, $node);
                         $nodesToScore[] = $newNode;
                     }
@@ -1538,10 +1538,10 @@ private function isPhrasingContent($node): bool
 
     /**
      * Checks if `$node` has only whitespace and a single element with `$tag` for the tag name.
-     * Returns false if `$node` contains non-empty text nodes
+     * Returns the matched element, or `null` if `$node` contains non-empty text nodes
      * or if it contains no element with given tag or more than 1 element.
      */
-    private function hasSingleTagInsideElement(JSLikeHTMLElement $node, string $tag): bool
+    private function getSingleTagInsideElement(JSLikeHTMLElement $node, string $tag): ?JSLikeHTMLElement
     {
         $childNodes = iterator_to_array($node->childNodes);
         $children = array_filter($childNodes, fn ($childNode) => $childNode instanceof \DOMElement);
@@ -1554,7 +1554,7 @@ private function hasSingleTagInsideElement(JSLikeHTMLElement $node, string $tag)
         // And there should be no text nodes with real content
         $a = array_filter($childNodes, fn ($childNode) => $childNode instanceof \DOMText && preg_match($this->regexps['hasContent'], $this->getInnerText($childNode)));
 
-        return 0 === \count($a);
+        return 0 === \count($a) ? $children[0] : null;
     }
 
     /**

From 8cbb8a8589d7abe29b0866472b8ffebe74cbd199 Mon Sep 17 00:00:00 2001
From: Jan Tojnar <jtojnar@gmail.com>
Date: Sun, 17 Mar 2024 23:36:41 +0100
Subject: [PATCH 12/12] fixup! JSHtml

---
 src/Readability.php | 44 +++++++++++++++++++++++++++++++++-----------
 1 file changed, 33 insertions(+), 11 deletions(-)

diff --git a/src/Readability.php b/src/Readability.php
index 6b53dc1..18bd88b 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -287,6 +287,7 @@ public function init(): bool
 
         if (null === $articleContent) {
             $this->success = false;
+            /** @var JSLikeHTMLElement */
             $articleContent = $this->dom->createElement('div');
             $articleContent->setAttribute('class', 'readability-content');
             $articleContent->setInnerHtml('<p>Sorry, Readability was unable to parse this page for content.</p>');
@@ -302,7 +303,9 @@ public function init(): bool
 
         // without tidy the body can (sometimes) be wiped, so re-create it
         if (false === isset($this->body->childNodes)) {
-            $this->body = $this->dom->createElement('body');
+            /** @var JSLikeHTMLElement */
+            $body = $this->dom->createElement('body');
+            $this->body = $body;
         }
 
         // Clear the old HTML, insert the new content.
@@ -335,19 +338,23 @@ public function postProcessContent(\DOMElement $articleContent): void
      */
     public function addFootnotes(\DOMElement $articleContent): void
     {
+        /** @var JSLikeHTMLElement */
         $footnotesWrapper = $this->dom->createElement('footer');
         $footnotesWrapper->setAttribute('class', 'readability-footnotes');
         $footnotesWrapper->setInnerHtml('<h3>References</h3>');
         $articleFootnotes = $this->dom->createElement('ol');
         $articleFootnotes->setAttribute('class', 'readability-footnotes-list');
         $footnotesWrapper->appendChild($articleFootnotes);
+        /** @var \DOMNodeList<JSLikeHTMLElement> */
         $articleLinks = $articleContent->getElementsByTagName('a');
         $linkCount = 0;
 
         for ($i = 0; $i < $articleLinks->length; ++$i) {
             $articleLink = $articleLinks->item($i);
             $footnoteLink = $articleLink->cloneNode(true);
+            /** @var JSLikeHTMLElement */
             $refLink = $this->dom->createElement('a');
+            /** @var JSLikeHTMLElement */
             $footnote = $this->dom->createElement('li');
             $linkDomain = @parse_url($footnoteLink->getAttribute('href'), \PHP_URL_HOST);
             if (!$linkDomain && isset($this->url)) {
@@ -609,6 +616,7 @@ public function killBreaks(JSLikeHTMLElement $node): void
      */
     public function clean(JSLikeHTMLElement $e, string $tag): void
     {
+        /** @var \DOMNodeList<JSLikeHTMLElement> */
         $targetList = $e->getElementsByTagName($tag);
         $isEmbed = ('audio' === $tag || 'video' === $tag || 'iframe' === $tag || 'object' === $tag || 'embed' === $tag);
 
@@ -645,6 +653,7 @@ public function cleanConditionally(JSLikeHTMLElement $e, string $tag): void
             return;
         }
 
+        /** @var \DOMNodeList<JSLikeHTMLElement> */
         $tagsList = $e->getElementsByTagName($tag);
         $curTagsLength = $tagsList->length;
 
@@ -755,6 +764,7 @@ public function cleanConditionally(JSLikeHTMLElement $e, string $tag): void
     public function cleanHeaders(JSLikeHTMLElement $e): void
     {
         for ($headerIndex = 1; $headerIndex < 3; ++$headerIndex) {
+            /** @var \DOMNodeList<JSLikeHTMLElement> */
             $headers = $e->getElementsByTagName('h' . $headerIndex);
 
             for ($i = $headers->length - 1; $i >= 0; --$i) {
@@ -823,6 +833,7 @@ protected function getArticleTitle(): JSLikeHTMLElement
             $curTitle = $origTitle;
         }
 
+        /** @var JSLikeHTMLElement */
         $articleTitle = $this->dom->createElement('h1');
         $articleTitle->setInnerHtml($curTitle);
 
@@ -840,7 +851,9 @@ protected function prepDocument(): void
          * so we create a new body node and append it to the document.
          */
         if (null === $this->body) {
-            $this->body = $this->dom->createElement('body');
+            /** @var JSLikeHTMLElement */
+            $body = $this->dom->createElement('body');
+            $this->body = $body;
             $this->dom->documentElement->appendChild($this->body);
         }
 
@@ -944,6 +957,7 @@ protected function grabArticle(?JSLikeHTMLElement $page = null): ?JSLikeHTMLElem
             $xpath = new \DOMXPath($page);
         }
 
+        /** @var \DOMNodeList<JSLikeHTMLElement> */
         $allElements = $page->getElementsByTagName('*');
 
         for ($nodeIndex = 0; $allElements->item($nodeIndex); ++$nodeIndex) {
@@ -986,6 +1000,7 @@ protected function grabArticle(?JSLikeHTMLElement $page = null): ?JSLikeHTMLElem
             //  (as in, where they contain no other block level elements).
             if ('div' === $tagName) {
                 if (!preg_match($this->regexps['divToPElements'], $nodeContent)) {
+                    /** @var JSLikeHTMLElement */
                     $newNode = $this->dom->createElement('p');
 
                     try {
@@ -1156,7 +1171,7 @@ protected function grabArticle(?JSLikeHTMLElement $page = null): ?JSLikeHTMLElem
             }
         }
 
-        /** @var \DOMNodeList<JSLikeHTMLElement> */
+        /** @var non-empty-array<JSLikeHTMLElement|null> */
         $topCandidates = array_filter(
             $topCandidates,
             fn ($v, $idx) => 0 === $idx || null !== $v,
@@ -1169,18 +1184,21 @@ protected function grabArticle(?JSLikeHTMLElement $page = null): ?JSLikeHTMLElem
          * We also have to copy the body node so it is something we can modify.
          */
         if (null === $topCandidate || 0 === strcasecmp($topCandidate->tagName, 'body')) {
+            /** @var JSLikeHTMLElement */
             $topCandidate = $this->dom->createElement('div');
 
             if ($page instanceof \DOMDocument) {
-                if (!isset($page->documentElement)) {
+                /** @var ?JSLikeHTMLElement */
+                $documentElement = $page->documentElement;
+                if (null === $documentElement) {
                     // we don't have a body either? what a mess! :)
                     $this->logger->debug('The page has no body!');
                 } else {
                     $this->logger->debug('Setting body to a raw HTML of original page!');
-                    $topCandidate->setInnerHtml($page->documentElement->getInnerHTML());
-                    $page->documentElement->setInnerHtml('');
+                    $topCandidate->setInnerHtml($documentElement->getInnerHTML());
+                    $documentElement->setInnerHtml('');
                     $this->reinitBody();
-                    $page->documentElement->appendChild($topCandidate);
+                    $documentElement->appendChild($topCandidate);
                 }
             } else {
                 $topCandidate->setInnerHtml($page->getInnerHTML());
@@ -1189,7 +1207,7 @@ protected function grabArticle(?JSLikeHTMLElement $page = null): ?JSLikeHTMLElem
             }
 
             $this->initializeNode($topCandidate);
-        } elseif ($topCandidate) {
+        } elseif (null !== $topCandidate) {
             $alternativeCandidateAncestors = [];
             foreach ($topCandidates as $candidate) {
                 if ((int) $candidate->getAttribute('readability') / (int) $topCandidate->getAttribute('readability') >= 0.75) {
@@ -1200,7 +1218,7 @@ protected function grabArticle(?JSLikeHTMLElement $page = null): ?JSLikeHTMLElem
             }
             if (\count($alternativeCandidateAncestors) >= 3) {
                 $parentOfTopCandidate = $topCandidate->parentNode;
-                while ('body' !== $parentOfTopCandidate->nodeName) {
+                while ('body' !== $parentOfTopCandidate->nodeName && $parentOfTopCandidate instanceof JSLikeHTMLElement) {
                     $listsContainingThisAncestor = 0;
                     for ($ancestorIndex = 0; $ancestorIndex < \count($alternativeCandidateAncestors) && $listsContainingThisAncestor < 3; ++$ancestorIndex) {
                         $listsContainingThisAncestor += (int) \in_array($parentOfTopCandidate, $alternativeCandidateAncestors[$ancestorIndex], true);
@@ -1264,6 +1282,7 @@ protected function grabArticle(?JSLikeHTMLElement $page = null): ?JSLikeHTMLElem
          * Now that we have the top candidate, look through its siblings for content that might also be related.
          * Things like preambles, content split by ads that we removed, etc.
          */
+        /** @var JSLikeHTMLElement */
         $articleContent = $this->dom->createElement('div');
         $articleContent->setAttribute('class', 'readability-content');
         $siblingScoreThreshold = max(10, ((int) $topCandidate->getAttribute('readability')) * 0.2);
@@ -1311,6 +1330,7 @@ protected function grabArticle(?JSLikeHTMLElement $page = null): ?JSLikeHTMLElem
                 if (0 !== strcasecmp($siblingNodeName, 'div') && 0 !== strcasecmp($siblingNodeName, 'p')) {
                     // We have a node that isn't a common block level element, like a form or td tag. Turn it into a div so it doesn't get filtered out later by accident.
                     $this->logger->debug('Altering siblingNode "' . $siblingNodeName . '" to "div".');
+                    /** @var JSLikeHTMLElement */
                     $nodeToAppend = $this->dom->createElement('div');
 
                     try {
@@ -1412,7 +1432,9 @@ protected function weightAttribute(JSLikeHTMLElement $element, string $attribute
     protected function reinitBody(): void
     {
         if (!isset($this->body->childNodes)) {
-            $this->body = $this->dom->createElement('body');
+            /** @var JSLikeHTMLElement */
+            $body = $this->dom->createElement('body');
+            $this->body = $body;
             $this->body->setInnerHtml($this->bodyCache);
         }
     }
@@ -1544,7 +1566,7 @@ private function isPhrasingContent($node): bool
     private function getSingleTagInsideElement(JSLikeHTMLElement $node, string $tag): ?JSLikeHTMLElement
     {
         $childNodes = iterator_to_array($node->childNodes);
-        $children = array_filter($childNodes, fn ($childNode) => $childNode instanceof \DOMElement);
+        $children = array_filter($childNodes, fn ($childNode) => $childNode instanceof JSLikeHTMLElement);
 
         // There should be exactly 1 element child with given tag
         if (1 !== \count($children) || $children[0]->nodeName !== $tag) {