mirror of
https://github.com/j0k3r/php-readability.git
synced 2026-09-27 06:26:17 +00:00
Update coding style for upcoming PHP-CS-Fixer changes
Once we bump minimum PHP version, we will get newer PHP-CS-Fixer, which will try to apply this cleanups. Also manually tweak anonymous functions so that they are cleanly formatted once we switch to `fn` syntax.
This commit is contained in:
committed by
Jérémy Benoist
parent
f28191a728
commit
648d8c605b
@@ -26,6 +26,10 @@ return (new PhpCsFixer\Config())
|
|||||||
'strict_comparison' => true,
|
'strict_comparison' => true,
|
||||||
'strict_param' => true,
|
'strict_param' => true,
|
||||||
'concat_space' => ['spacing' => 'one'],
|
'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($finder)
|
->setFinder($finder)
|
||||||
;
|
;
|
||||||
|
|||||||
+38
-23
@@ -142,7 +142,7 @@ class Readability implements LoggerAwareInterface
|
|||||||
* @param string $parser Which parser to use for turning raw HTML into a DOMDocument
|
* @param string $parser Which parser to use for turning raw HTML into a DOMDocument
|
||||||
* @param bool $useTidy Use tidy
|
* @param bool $useTidy Use tidy
|
||||||
*/
|
*/
|
||||||
public function __construct(string $html, string $url = null, string $parser = 'libxml', bool $useTidy = true)
|
public function __construct(string $html, ?string $url = null, string $parser = 'libxml', bool $useTidy = true)
|
||||||
{
|
{
|
||||||
$this->url = $url;
|
$this->url = $url;
|
||||||
$this->html = $html;
|
$this->html = $html;
|
||||||
@@ -739,7 +739,7 @@ class Readability implements LoggerAwareInterface
|
|||||||
*/
|
*/
|
||||||
public function addFlag(int $flag): void
|
public function addFlag(int $flag): void
|
||||||
{
|
{
|
||||||
$this->flags = $this->flags | $flag;
|
$this->flags |= $flag;
|
||||||
}
|
}
|
||||||
|
|
||||||
/**
|
/**
|
||||||
@@ -747,7 +747,7 @@ class Readability implements LoggerAwareInterface
|
|||||||
*/
|
*/
|
||||||
public function removeFlag(int $flag): void
|
public function removeFlag(int $flag): void
|
||||||
{
|
{
|
||||||
$this->flags = $this->flags & ~$flag;
|
$this->flags &= ~$flag;
|
||||||
}
|
}
|
||||||
|
|
||||||
/**
|
/**
|
||||||
@@ -893,11 +893,9 @@ class Readability implements LoggerAwareInterface
|
|||||||
* Using a variety of metrics (content score, classname, element types), find the content that is
|
* 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.
|
* most likely to be the stuff a user wants to read. Then return it wrapped up in a div.
|
||||||
*
|
*
|
||||||
* @param \DOMElement $page
|
|
||||||
*
|
|
||||||
* @return \DOMElement|false
|
* @return \DOMElement|false
|
||||||
*/
|
*/
|
||||||
protected function grabArticle(\DOMElement $page = null)
|
protected function grabArticle(?\DOMElement $page = null)
|
||||||
{
|
{
|
||||||
if (!$page) {
|
if (!$page) {
|
||||||
$page = $this->dom;
|
$page = $this->dom;
|
||||||
@@ -933,9 +931,9 @@ class Readability implements LoggerAwareInterface
|
|||||||
// Remove unlikely candidates
|
// Remove unlikely candidates
|
||||||
$unlikelyMatchString = $node->getAttribute('class') . ' ' . $node->getAttribute('id') . ' ' . $node->getAttribute('style');
|
$unlikelyMatchString = $node->getAttribute('class') . ' ' . $node->getAttribute('id') . ' ' . $node->getAttribute('style');
|
||||||
|
|
||||||
if (mb_strlen($unlikelyMatchString) > 3 && // don't process "empty" strings
|
if (mb_strlen($unlikelyMatchString) > 3 // don't process "empty" strings
|
||||||
preg_match($this->regexps['unlikelyCandidates'], $unlikelyMatchString) &&
|
&& preg_match($this->regexps['unlikelyCandidates'], $unlikelyMatchString)
|
||||||
!preg_match($this->regexps['okMaybeItsACandidate'], $unlikelyMatchString)
|
&& !preg_match($this->regexps['okMaybeItsACandidate'], $unlikelyMatchString)
|
||||||
) {
|
) {
|
||||||
$this->logger->debug('Removing unlikely candidate (using conf) ' . $node->getNodePath() . ' by "' . $unlikelyMatchString . '"');
|
$this->logger->debug('Removing unlikely candidate (using conf) ' . $node->getNodePath() . ' by "' . $unlikelyMatchString . '"');
|
||||||
$node->parentNode->removeChild($node);
|
$node->parentNode->removeChild($node);
|
||||||
@@ -1120,9 +1118,13 @@ class Readability implements LoggerAwareInterface
|
|||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
$topCandidates = array_filter($topCandidates, function ($v, $idx) {
|
$topCandidates = array_filter(
|
||||||
return 0 === $idx || null !== $v;
|
$topCandidates,
|
||||||
}, \ARRAY_FILTER_USE_BOTH);
|
function ($v, $idx) {
|
||||||
|
return 0 === $idx || null !== $v;
|
||||||
|
},
|
||||||
|
\ARRAY_FILTER_USE_BOTH
|
||||||
|
);
|
||||||
$topCandidate = $topCandidates[0];
|
$topCandidate = $topCandidates[0];
|
||||||
|
|
||||||
/*
|
/*
|
||||||
@@ -1442,7 +1444,7 @@ class Readability implements LoggerAwareInterface
|
|||||||
libxml_use_internal_errors(false);
|
libxml_use_internal_errors(false);
|
||||||
}
|
}
|
||||||
|
|
||||||
$this->dom->registerNodeClass(\DOMElement::class, \Readability\JSLikeHTMLElement::class);
|
$this->dom->registerNodeClass(\DOMElement::class, JSLikeHTMLElement::class);
|
||||||
}
|
}
|
||||||
|
|
||||||
private function getAncestors(\DOMElement $node, int $maxDepth = 0): array
|
private function getAncestors(\DOMElement $node, int $maxDepth = 0): array
|
||||||
@@ -1464,9 +1466,19 @@ class Readability implements LoggerAwareInterface
|
|||||||
{
|
{
|
||||||
return \XML_TEXT_NODE === $node->nodeType
|
return \XML_TEXT_NODE === $node->nodeType
|
||||||
|| \in_array(strtoupper($node->nodeName), $this->phrasingElements, true)
|
|| \in_array(strtoupper($node->nodeName), $this->phrasingElements, true)
|
||||||
|| (\in_array(strtoupper($node->nodeName), ['A', 'DEL', 'INS'], true) && !\in_array(false, array_map(function ($c) {
|
|| (
|
||||||
return $this->isPhrasingContent($c);
|
\in_array(strtoupper($node->nodeName), ['A', 'DEL', 'INS'], true)
|
||||||
}, iterator_to_array($node->childNodes)), true));
|
&& !\in_array(
|
||||||
|
false,
|
||||||
|
array_map(
|
||||||
|
function ($c) {
|
||||||
|
return $this->isPhrasingContent($c);
|
||||||
|
},
|
||||||
|
iterator_to_array($node->childNodes)
|
||||||
|
),
|
||||||
|
true
|
||||||
|
)
|
||||||
|
);
|
||||||
}
|
}
|
||||||
|
|
||||||
private function hasSingleTagInsideElement(\DOMElement $node, string $tag): bool
|
private function hasSingleTagInsideElement(\DOMElement $node, string $tag): bool
|
||||||
@@ -1475,10 +1487,12 @@ class Readability implements LoggerAwareInterface
|
|||||||
return false;
|
return false;
|
||||||
}
|
}
|
||||||
|
|
||||||
$a = array_filter(iterator_to_array($node->childNodes), function ($childNode) {
|
$a = array_filter(
|
||||||
return $childNode instanceof \DOMText &&
|
iterator_to_array($node->childNodes),
|
||||||
preg_match($this->regexps['hasContent'], $this->getInnerText($childNode));
|
function ($childNode) {
|
||||||
});
|
return $childNode instanceof \DOMText && preg_match($this->regexps['hasContent'], $this->getInnerText($childNode));
|
||||||
|
}
|
||||||
|
);
|
||||||
|
|
||||||
return 0 === \count($a);
|
return 0 === \count($a);
|
||||||
}
|
}
|
||||||
@@ -1491,9 +1505,10 @@ class Readability implements LoggerAwareInterface
|
|||||||
*/
|
*/
|
||||||
private function isNodeVisible(\DOMElement $node): bool
|
private function isNodeVisible(\DOMElement $node): bool
|
||||||
{
|
{
|
||||||
return !($node->hasAttribute('style')
|
return !(
|
||||||
&& preg_match($this->regexps['isNotVisible'], $node->getAttribute('style'))
|
$node->hasAttribute('style')
|
||||||
|
&& preg_match($this->regexps['isNotVisible'], $node->getAttribute('style'))
|
||||||
)
|
)
|
||||||
&& !$node->hasAttribute('hidden');
|
&& !$node->hasAttribute('hidden');
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|||||||
@@ -550,7 +550,7 @@ class ReadabilityTest extends \PHPUnit\Framework\TestCase
|
|||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
private function getReadability(string $html, string $url = null, string $parser = 'libxml', bool $useTidy = true): Readability
|
private function getReadability(string $html, ?string $url = null, string $parser = 'libxml', bool $useTidy = true): Readability
|
||||||
{
|
{
|
||||||
$readability = new Readability($html, $url, $parser, $useTidy);
|
$readability = new Readability($html, $url, $parser, $useTidy);
|
||||||
|
|
||||||
|
|||||||
Reference in New Issue
Block a user