infong/php-readability
1
0
Fork 0
mirror of https://github.com/j0k3r/php-readability.git synced 2026-09-27 06:26:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

Skip empty (empty innerHTML) nodes when grabbing article

Browse Source
This commit is contained in:
nicofrand
2019-05-25 16:12:52 +02:00
parent f808c1b0a2
commit ff78c63e6d
1 changed files with 8 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/Readability.php
+8 -2
View File
@@ -1008,6 +1008,12 @@ class Readability implements LoggerAwareInterface
for ($nodeIndex = 0; ($node = $allElements->item($nodeIndex)); ++$nodeIndex) { for ($nodeIndex = 0; ($node = $allElements->item($nodeIndex)); ++$nodeIndex) {
$tagName = $node->tagName; $tagName = $node->tagName;
$nodeContent = $node->getInnerHTML();
if (empty($nodeContent)) {
$this->logger->debug('Skipping empty node');
continue;
}
// Some well known site uses sections as paragraphs. // Some well known site uses sections as paragraphs.
if (0 === strcasecmp($tagName, 'p') || 0 === strcasecmp($tagName, 'td') || 0 === strcasecmp($tagName, 'pre') || 0 === strcasecmp($tagName, 'section')) { if (0 === strcasecmp($tagName, 'p') || 0 === strcasecmp($tagName, 'td') || 0 === strcasecmp($tagName, 'pre') || 0 === strcasecmp($tagName, 'section')) {
$nodesToScore[] = $node; $nodesToScore[] = $node;
@@ -1016,11 +1022,11 @@ class Readability implements LoggerAwareInterface
// Turn divs into P tags where they have been used inappropriately // Turn divs into P tags where they have been used inappropriately
// (as in, where they contain no other block level elements). // (as in, where they contain no other block level elements).
if (0 === strcasecmp($tagName, 'div') || 0 === strcasecmp($tagName, 'article') || 0 === strcasecmp($tagName, 'section')) { if (0 === strcasecmp($tagName, 'div') || 0 === strcasecmp($tagName, 'article') || 0 === strcasecmp($tagName, 'section')) {
if (!preg_match($this->regexps['divToPElements'], $node->getInnerHTML())) { if (!preg_match($this->regexps['divToPElements'], $nodeContent)) {
$newNode = $this->dom->createElement('p'); $newNode = $this->dom->createElement('p');
try { try {
$newNode->setInnerHtml($node->getInnerHTML()); $newNode->setInnerHtml($nodeContent);
$node->parentNode->replaceChild($newNode, $node); $node->parentNode->replaceChild($newNode, $node);
--$nodeIndex; --$nodeIndex;