shihaam/karudhaas
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

fix thiladhun

Browse Source
This commit is contained in:
Mohamed jinas
2024-01-07 21:42:31 +05:00
parent c8329962b7
commit 0589b19843
3 changed files with 29 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files

29
app/Services/Scrapers/ThiladhunScraper.php
View File

@@ -34,7 +34,7 @@ class ThiladhunScraper
*
* @return array
*/
public function extract($url, $date)
public function extract($url)
{
$crawler = $this->client->request('GET', $url);
@@ -51,20 +51,23 @@ class ThiladhunScraper
$this->image = $crawler->filter('div[data-widget_type*="theme-post-featured-image.default"] figure img')->first()->attr('src');
$crawler->filter('.entry-tags a')->each(function ($node) {
if(!preg_match('/[^A-Za-z0-9-]/', basename($node->attr('href'))))
{
$this->topics[] = [
"name" => $node->text(),
"slug" => basename($node->attr('href'))
];
if (!preg_match('/[^A-Za-z0-9-]/', basename($node->attr('href')))) {
$this->topics[] = [
"name" => $node->text(),
"slug" => basename($node->attr('href'))
];
}
});
});
$crawler->filter('a > .elementor-post-info__item--type-author')->each(function ($node) {
$this->author = $node->text();
});
$urlComponents = parse_url($url);
$cleanUrl = $urlComponents['scheme'] . '://' . $urlComponents['host'] . $urlComponents['path'];
// Extract and modify the guid
$guid = basename($urlComponents['path']);
return [
'source' => 'Thiladhun News',
@@ -72,11 +75,11 @@ class ThiladhunScraper
'og_title' => str_replace(" | Thiladhun", "", $crawler->filter('title')->first()->text('content')),
'image' => $this->image,
'content' => $this->content,
'date' => Carbon::parse($date)->format("Y-m-d H:i:s"),
'url' => $url,
'date' => Carbon::now(),
'url' => $cleanUrl,
'author' => $this->author,
'guid' => basename($url),
'topics' => $this->topics ? : [
'guid' => $guid,
'topics' => $this->topics ?: [
[
"name" => "ވަކި މަޢުލޫއެއް ނޭންގެ",
"slug" => "no-specific-topic"