mirror/RSS-Bridge.rss-bridge
1
0
Fork 0
mirror of https://github.com/RSS-Bridge/rss-bridge.git synced 2025-08-30 22:59:46 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 3

fix GolemBridge pagination (and avoid having identical content on multiple pages - especially images)
Some checks are pending
Build Image on Commit and Release / bake (push) Waiting to run
Details
Lint / phpcs (7.4) (push) Waiting to run
Details
Lint / phpcompatibility (7.4) (push) Waiting to run
Details
Lint / executable_php_files_check (push) Waiting to run
Details
Tests / phpunit8 (8.1) (push) Waiting to run
Details
Tests / phpunit8 (8.2) (push) Waiting to run
Details
Tests / phpunit8 (8.3) (push) Waiting to run
Details
Tests / phpunit8 (8.4) (push) Waiting to run
Details
Tests / phpunit8 (7.4) (push) Waiting to run
Details
Tests / phpunit8 (8.0) (push) Waiting to run
Details

Browse source
This commit is contained in:
Micha 2025-08-28 13:24:11 +02:00 • committed by Mynacol
parent 32f324dbb5
commit 3f33d0e312
1 changed files with 31 additions and 17 deletions
Download patch file Download diff file Expand all files Collapse all files

22
bridges/GolemBridge.php
View file

@ -82,6 +82,7 @@ class GolemBridge extends FeedExpander
// URI without RSS feed reference // URI without RSS feed reference
$item['uri'] = $articlePage->find('head meta[name="twitter:url"]', 0)->content; $item['uri'] = $articlePage->find('head meta[name="twitter:url"]', 0)->content;
if (!array_key_exists('categories', $item)) {
$categories = $articlePage->find('div.go-tag-list__tags a.go-tag'); $categories = $articlePage->find('div.go-tag-list__tags a.go-tag');
foreach ($categories as $category) { foreach ($categories as $category) {
$trimmedcategories[] = trim(html_entity_decode($category->plaintext)); $trimmedcategories[] = trim(html_entity_decode($category->plaintext));
@ -89,18 +90,27 @@ class GolemBridge extends FeedExpander
if (isset($trimmedcategories)) { if (isset($trimmedcategories)) {
$item['categories'] = array_unique($trimmedcategories); $item['categories'] = array_unique($trimmedcategories);
} }
}
$item['content'] .= $this->extractContent($articlePage); $item['content'] .= $this->extractContent($articlePage, $item['content']);
// next page // next page
$nextUri = $articlePage->find('link[rel="next"]', 0); $nextUri = $articlePage->find('li.go-pagination__item--next>a', 0);
$uri = $nextUri ? static::URI . $nextUri->href : null; if ($nextUri) {
$nextUri = $nextUri->href;
if (str_starts_with($nextUri, '/')) {
$nextUri = substr($nextUri, 1);
}
$uri = static::URI . $nextUri;
} else {
$uri = null;
}
} }
return $item; return $item;
} }
private function extractContent($page) private function extractContent($page, $prevcontent)
{ {
$item = ''; $item = '';
@ -150,6 +160,7 @@ class GolemBridge extends FeedExpander
} }
$header = $article->find('header', 0); $header = $article->find('header', 0);
if (isset($header)) {
foreach ($header->find('p, figure') as $element) { foreach ($header->find('p, figure') as $element) {
$item .= $element; $item .= $element;
} }
@ -158,10 +169,13 @@ class GolemBridge extends FeedExpander
foreach ($article->find('img[data-src-full][src*="."]') as $img) { foreach ($article->find('img[data-src-full][src*="."]') as $img) {
$img->src = $img->getAttribute('data-src-full'); $img->src = $img->getAttribute('data-src-full');
} }
}
foreach ($article->find('div.go-article-header__intro, p, h1, h2, h3, pre, img[src*="."], div[class*="golem_tablediv"], iframe, video') as $element) { foreach ($article->find('div.go-article-header__intro, p, h1, h2, h3, pre, img[src*="."], div[class*="golem_tablediv"], iframe, video') as $element) {
if (!str_contains($prevcontent, $element)) {
$item .= $element; $item .= $element;
} }
}
return $item; return $item;
} }