infong
/
php-readability
mirror of https://github.com/j0k3r/php-readability
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Use `SetList::CODE_QUALITY` with Rector

Browse Source
pull/72/head
Jeremy Benoist 4 years ago
parent 8af69ad68c
commit 28f1b08fac
No known key found for this signature in database
GPG Key ID: BCA73962457ACC3C
3 changed files with 49 additions and 49 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 4
      rector.php
  2. 4
      src/JSLikeHTMLElement.php
  3. 50
      src/Readability.php

4
rector.php
Unescape View File

@ -4,7 +4,9 @@ declare(strict_types=1);
use Rector\Core\Configuration\Option;
use Rector\Core\ValueObject\PhpVersion;
use Rector\PHPUnit\Set\PHPUnitSetList;
use Rector\Set\ValueObject\LevelSetList;
use Rector\Set\ValueObject\SetList;
use Symfony\Component\DependencyInjection\Loader\Configurator\ContainerConfigurator;
return static function (ContainerConfigurator $containerConfigurator): void {
@ -26,6 +28,8 @@ return static function (ContainerConfigurator $containerConfigurator): void {
// Define what rule sets will be applied
$containerConfigurator->import(LevelSetList::UP_TO_PHP_72);
$containerConfigurator->import(PHPUnitSetList::PHPUNIT_80);
$containerConfigurator->import(SetList::CODE_QUALITY);
// is your PHP version different from the one your refactor to?
$parameters->set(Option::PHP_VERSION_FEATURES, PhpVersion::PHP_72);

4
src/JSLikeHTMLElement.php
Unescape View File

@ -53,7 +53,7 @@ class JSLikeHTMLElement extends \DOMElement
}
// first, empty the element
if (isset($this->childNodes)) {
if (null !== $this->childNodes) {
for ($x = $this->childNodes->length - 1; $x >= 0; --$x) {
$this->removeChild($this->childNodes->item($x));
}
@ -114,7 +114,7 @@ class JSLikeHTMLElement extends \DOMElement
if ('innerHTML' === $name) {
$inner = '';
if (isset($this->childNodes)) {
if (null !== $this->childNodes) {
foreach ($this->childNodes as $child) {
$inner .= $this->ownerDocument->saveXML($child);
}

50
src/Readability.php
Unescape View File

@ -220,7 +220,7 @@ class Readability implements LoggerAwareInterface
{
$this->loadHtml();
if (!isset($this->dom->documentElement)) {
if (!(property_exists($this->dom, 'documentElement') && null !== $this->dom->documentElement)) {
return false;
}
@ -236,7 +236,7 @@ class Readability implements LoggerAwareInterface
}
}
if ($bodyElems->length > 0 && null === $this->body) {
if ($bodyElems->length > 0 && !$this->body instanceof \DOMElement) {
$this->body = $bodyElems->item(0);
}
@ -264,7 +264,7 @@ class Readability implements LoggerAwareInterface
$overlay->appendChild($innerDiv);
// without tidy the body can (sometimes) be wiped, so re-create it
if (false === isset($this->body->childNodes)) {
if (!(property_exists($this->body, 'childNodes') && null !== $this->body->childNodes)) {
$this->body = $this->dom->createElement('body');
}
@ -313,7 +313,7 @@ class Readability implements LoggerAwareInterface
$refLink = $this->dom->createElement('a');
$footnote = $this->dom->createElement('li');
$linkDomain = @parse_url($footnoteLink->getAttribute('href'), \PHP_URL_HOST);
if (!$linkDomain && isset($this->url)) {
if (!$linkDomain && null !== $this->url) {
$linkDomain = @parse_url($this->url, \PHP_URL_HOST);
}
@ -433,7 +433,7 @@ class Readability implements LoggerAwareInterface
}
// add extra text to iframe tag to avoid an auto-closing iframe and then break the html code
if ($iframeCount) {
if (0 !== $iframeCount) {
$iframe = $item->getElementsByTagName('iframe');
$iframe->item(0)->nodeValue = ' ';
@ -463,7 +463,7 @@ class Readability implements LoggerAwareInterface
*/
public function getInnerText($e, bool $normalizeSpaces = true, bool $flattenLines = false): string
{
if (null === $e || !isset($e->textContent) || '' === $e->textContent) {
if (!$e instanceof \DOMNode || !(property_exists($e, 'textContent') && null !== $e->textContent) || '' === $e->textContent) {
return '';
}
@ -679,8 +679,7 @@ class Readability implements LoggerAwareInterface
$this->logger->debug(' more than 3 embeds');
$toRemove = true;
}
} else {
if ($img > $p) {
} elseif ($img > $p) {
$this->logger->debug(' more image elements than paragraph elements');
$toRemove = true;
} elseif (!$isList && $li > $p) {
@ -702,7 +701,6 @@ class Readability implements LoggerAwareInterface
$this->logger->debug(' 1 embed and content length smaller than 75 chars, or more than one embed');
$toRemove = true;
}
}
if ($toRemove) {
$this->logger->debug('Removing...');
@ -741,7 +739,7 @@ class Readability implements LoggerAwareInterface
*/
public function addFlag(int $flag): void
{
$this->flags = $this->flags | $flag;
$this->flags |= $flag;
}
/**
@ -749,7 +747,7 @@ class Readability implements LoggerAwareInterface
*/
public function removeFlag(int $flag): void
{
$this->flags = $this->flags & ~$flag;
$this->flags &= ~$flag;
}
/**
@ -829,7 +827,7 @@ class Readability implements LoggerAwareInterface
*/
protected function initializeNode(DOMElement $node): void
{
if (!isset($node->tagName)) {
if (!(property_exists($node, 'tagName') && null !== $node->tagName)) {
return;
}
@ -901,14 +899,14 @@ class Readability implements LoggerAwareInterface
*/
protected function grabArticle(DOMElement $page = null)
{
if (!$page) {
if (null === $page) {
$page = $this->dom;
}
$xpath = null;
$nodesToScore = [];
if ($page instanceof \DOMDocument && isset($page->documentElement)) {
if ($page instanceof \DOMDocument && (property_exists($page, 'documentElement') && null !== $page->documentElement)) {
$xpath = new \DOMXPath($page);
}
@ -1019,16 +1017,13 @@ class Readability implements LoggerAwareInterface
* A score is determined by things like number of commas, class names, etc.
* Maybe eventually link density.
*/
for ($pt = 0, $scored = \count($nodesToScore); $pt < $scored; ++$pt) {
$ancestors = $this->getAncestors($nodesToScore[$pt], 5);
foreach ($nodesToScore as $pt => $singleNodesToScore) {
$ancestors = $this->getAncestors($singleNodesToScore, 5);
// No parent node? Move on...
if (0 === \count($ancestors)) {
if ([] === $ancestors) {
continue;
}
$innerText = $this->getInnerText($nodesToScore[$pt]);
$innerText = $this->getInnerText($singleNodesToScore);
// If this paragraph is less than MIN_PARAGRAPH_LENGTH (default:20) characters, don't even count it.
if (mb_strlen($innerText) < self::MIN_PARAGRAPH_LENGTH) {
continue;
@ -1074,7 +1069,7 @@ class Readability implements LoggerAwareInterface
for ($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 && (null !== $node->parentNode ? 0 !== strcasecmp($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));
$node->parentNode->removeChild($node);
}
@ -1093,7 +1088,7 @@ class Readability implements LoggerAwareInterface
* and find the one with the highest score.
*/
$topCandidates = array_fill(0, 5, null);
if ($xpath) {
if (null !== $xpath) {
// Using array of DOMElements after deletion is a path to DOOMElement.
$candidates = $xpath->query('.//*[@data-candidate]', $page->documentElement);
$this->logger->debug('Candidates: ' . $candidates->length);
@ -1135,7 +1130,7 @@ class Readability implements LoggerAwareInterface
$topCandidate = $this->dom->createElement('div');
if ($page instanceof \DOMDocument) {
if (!isset($page->documentElement)) {
if (!(property_exists($page, 'documentElement') && null !== $page->documentElement)) {
// we don't have a body either? what a mess! :)
$this->logger->debug('The page has no body!');
} else {
@ -1165,7 +1160,8 @@ class Readability implements LoggerAwareInterface
$parentOfTopCandidate = $topCandidate->parentNode;
while ('body' !== $parentOfTopCandidate->nodeName) {
$listsContainingThisAncestor = 0;
for ($ancestorIndex = 0; $ancestorIndex < \count($alternativeCandidateAncestors) && $listsContainingThisAncestor < 3; ++$ancestorIndex) {
$alternativeCandidateAncestorsCount = \count($alternativeCandidateAncestors);
for ($ancestorIndex = 0; $ancestorIndex < $alternativeCandidateAncestorsCount && $listsContainingThisAncestor < 3; ++$ancestorIndex) {
$listsContainingThisAncestor += (int) \in_array($parentOfTopCandidate, $alternativeCandidateAncestors[$ancestorIndex], true);
}
if ($listsContainingThisAncestor >= 3) {
@ -1374,7 +1370,7 @@ class Readability implements LoggerAwareInterface
*/
protected function reinitBody(): void
{
if (!isset($this->body->childNodes)) {
if (!(property_exists($this->body, 'childNodes') && null !== $this->body->childNodes)) {
$this->body = $this->dom->createElement('body');
$this->body->setInnerHtml($this->bodyCache);
}
@ -1482,7 +1478,7 @@ class Readability implements LoggerAwareInterface
preg_match($this->regexps['hasContent'], $this->getInnerText($childNode));
});
return 0 === \count($a);
return [] === $a;
}
/**

Write Preview
Loading…
Cancel
Save