Dhuvas intergrated
This commit is contained in:
		@@ -50,10 +50,10 @@ class DhenScraper
 | 
			
		||||
            'date'       =>  Carbon::parse(str_replace("- ", "", $crawler->filter('.story-item--meta time')->first()->text()))->format("Y-m-d H:i:s"),
 | 
			
		||||
            'guid'       => str_replace("https://dhen.mv/","",$url),
 | 
			
		||||
            'author'     => $this->author,
 | 
			
		||||
            'topics'       => [
 | 
			
		||||
            'topics'       =>  [
 | 
			
		||||
                [
 | 
			
		||||
                    "name" => "ވަކި މަޢުލޫއެއް ނޭންގެ",
 | 
			
		||||
                    "slug" => "uncategorized"
 | 
			
		||||
                    "slug" => "no-specific-topic"
 | 
			
		||||
                ]
 | 
			
		||||
            ]
 | 
			
		||||
        ];
 | 
			
		||||
 
 | 
			
		||||
							
								
								
									
										62
									
								
								app/Services/Scrapers/DhuvasScraper.php
									
									
									
									
									
										Normal file
									
								
							
							
						
						
									
										62
									
								
								app/Services/Scrapers/DhuvasScraper.php
									
									
									
									
									
										Normal file
									
								
							@@ -0,0 +1,62 @@
 | 
			
		||||
<?php
 | 
			
		||||
 | 
			
		||||
namespace App\Services\Scrapers;
 | 
			
		||||
 | 
			
		||||
use Goutte\Client;
 | 
			
		||||
use Illuminate\Support\Carbon;
 | 
			
		||||
 | 
			
		||||
class DhuvasScraper
 | 
			
		||||
{
 | 
			
		||||
    protected $client;
 | 
			
		||||
 | 
			
		||||
    protected $title;
 | 
			
		||||
    protected $content;
 | 
			
		||||
    protected $author;
 | 
			
		||||
 | 
			
		||||
    public function __construct()
 | 
			
		||||
    {
 | 
			
		||||
        $this->client = new Client;
 | 
			
		||||
    }
 | 
			
		||||
 | 
			
		||||
    public function extract($url, $date = null)
 | 
			
		||||
    {
 | 
			
		||||
 | 
			
		||||
        $crawler = $this->client->request('GET', $url);
 | 
			
		||||
 | 
			
		||||
        $crawler->filter('h1')->each(function ($node) {
 | 
			
		||||
            $this->title = $node->text();
 | 
			
		||||
        });
 | 
			
		||||
 | 
			
		||||
        $crawler->filter('.elementor-widget-container > p')->each(function ($node) {
 | 
			
		||||
            $this->content[] = preg_replace("/[a-zA-Z]/","",$node->text());
 | 
			
		||||
        });
 | 
			
		||||
 | 
			
		||||
        if($crawler->filter('span[class*="elementor-icon-list-text elementor-post-info__item elementor-post-info__item--type-author"]')->count() == 1)
 | 
			
		||||
        {
 | 
			
		||||
            $this->author = $crawler->filter('span[class*="elementor-icon-list-text elementor-post-info__item elementor-post-info__item--type-author"]')->first()->text();
 | 
			
		||||
        }
 | 
			
		||||
 | 
			
		||||
 | 
			
		||||
     
 | 
			
		||||
 | 
			
		||||
        //Remove all the alphabets from string
 | 
			
		||||
        //preg_replace("/[a-zA-Z]/", "",$string);
 | 
			
		||||
       return [
 | 
			
		||||
            'source'    => 'Dhuvas',
 | 
			
		||||
            'title'      => $this->title,
 | 
			
		||||
            'og_title'   => $crawler->filter('meta[property*="og:title"]')->first()->attr('content'),
 | 
			
		||||
            'image'      => $crawler->filter('.elementor-widget-theme-post-featured-image .elementor-image a img')->first()->attr('src'),
 | 
			
		||||
            'content'    => $this->content,
 | 
			
		||||
            'url'        => $url,
 | 
			
		||||
            'date'       => Carbon::parse($date)->format("Y-m-d H:i:s"),
 | 
			
		||||
            'guid'       => str_replace("https://dhuvas.mv/","",$url),
 | 
			
		||||
            'author'     => $this->author,
 | 
			
		||||
            'topics'       =>  [
 | 
			
		||||
                [
 | 
			
		||||
                    "name" => "ވަކި މަޢުލޫއެއް ނޭންގެ",
 | 
			
		||||
                    "slug" => "no-specific-topic"
 | 
			
		||||
                ]
 | 
			
		||||
            ]
 | 
			
		||||
        ];
 | 
			
		||||
    }
 | 
			
		||||
}
 | 
			
		||||
		Reference in New Issue
	
	Block a user