mirrors/rss-bridge
0
0
Fork 0
mirror of https://github.com/RSS-Bridge/rss-bridge.git synced 2025-07-01 11:02:56 +00:00
Code

Merge 583e5b144f into 354cea09a7

Browse Source
This commit is contained in:
Quentin B. 2025-06-27 19:33:26 +00:00 committed by GitHub
commit 9e59011719
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 20 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

38
bridges/CentreFranceBridge.php
View File

@ -72,15 +72,9 @@ class CentreFranceBridge extends BridgeAbstract
$newspaperUrl = 'https://www.' . $this->getInput('newspaper') . '/' . $localitySlug . '/'; $newspaperUrl = 'https://www.' . $this->getInput('newspaper') . '/' . $localitySlug . '/';
$html = getSimpleHTMLDOM($newspaperUrl); $html = getSimpleHTMLDOM($newspaperUrl);
// Articles are detected through their titles // Articles are detected through a standard tag
foreach ($html->find('.c-titre') as $articleTitleDOMElement) { foreach ($html->find('article') as $articleDOMElement) {
$articleLinkDOMElement = $articleTitleDOMElement->find('a', 0); $articleLinkDOMElement = $articleDOMElement->find('a', 0);
// Ignore articles in the « Les + partagés » block
if (strpos($articleLinkDOMElement->id, 'les_plus_partages') !== false) {
continue;
}
$articleURI = $articleLinkDOMElement->href; $articleURI = $articleLinkDOMElement->href;
// If the URI has already been processed, ignore it // If the URI has already been processed, ignore it
@ -96,7 +90,7 @@ class CentreFranceBridge extends BridgeAbstract
$articleTitle = ''; $articleTitle = '';
// If article is reserved for subscribers // If article is reserved for subscribers
if ($articleLinkDOMElement->find('span.premium-picto', 0)) { if ($articleLinkDOMElement->find('span.premium-icon', 0)) {
if ($this->getInput('remove-reserved-for-subscribers-articles') === true) { if ($this->getInput('remove-reserved-for-subscribers-articles') === true) {
continue; continue;
} }
@ -104,18 +98,22 @@ class CentreFranceBridge extends BridgeAbstract
$articleTitle .= '🔒 '; $articleTitle .= '🔒 ';
} }
$articleTitleDOMElement = $articleLinkDOMElement->find('span[data-tb-title]', 0);
if ($articleTitleDOMElement === null) {
continue;
}
if ($limit > 0 && count($this->items) === $limit) { if ($limit > 0 && count($this->items) === $limit) {
break; break;
} }
$articleTitle .= $articleLinkDOMElement->find('span[data-tb-title]', 0)->innertext; // Loop through each possible title class name
$articleFullURI = urljoin('https://www.' . $this->getInput('newspaper') . '/', $articleURI); for ($i = 1; $i <= 3; $i++) {
$articleTitleDOMElement = $articleLinkDOMElement->find('.typo-card-title-' . $i, 0);
if (!$articleTitleDOMElement instanceof \simple_html_dom_node) {
continue;
}
$articleTitle .= $articleTitleDOMElement->innertext;
break;
}
$articleFullURI = urljoin('https://www.' . $this->getInput('newspaper') . '/', $articleURI);
$item = [ $item = [
'title' => $articleTitle, 'title' => $articleTitle,
'uri' => $articleFullURI, 'uri' => $articleFullURI,
@ -184,7 +182,7 @@ class CentreFranceBridge extends BridgeAbstract
$articleTags = $html->find('#content>div.flex+div.grid section>.bg-gray-light>a.border-gray-dark'); $articleTags = $html->find('#content>div.flex+div.grid section>.bg-gray-light>a.border-gray-dark');
if (is_array($articleTags)) { if (is_array($articleTags)) {
$item['categories'] = array_map(static fn ($articleTag) => $articleTag->innertext, $articleTags); $item['categories'] = array_map(static fn ($articleTag) => html_entity_decode($articleTag->innertext), $articleTags);
} }
$explode = explode('_', $uri); $explode = explode('_', $uri);
@ -195,6 +193,10 @@ class CentreFranceBridge extends BridgeAbstract
$item['uid'] = $uid; $item['uid'] = $uid;
} }
if (!isset($item['content'])) {
$item['content'] = '';
}
// If the article is a "grand format", we use another parsing strategy // If the article is a "grand format", we use another parsing strategy
if ($item['content'] === '' && $html->find('article') !== []) { if ($item['content'] === '' && $html->find('article') !== []) {
$articleContent = $html->find('article > section'); $articleContent = $html->find('article > section');