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

Fix mihaaru scraper

Browse Source
This commit is contained in:
Mohamed jinas
2024-01-13 04:53:31 +05:00
parent 252d750d7d
commit 736181d461
5 changed files with 96 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

34
app/Services/Scrapers/MihaaruScraper.php
View File

@@ -3,6 +3,7 @@
namespace App\Services\Scrapers;
use Goutte\Client;
use Symfony\Component\HttpClient\HttpClient;
class MihaaruScraper
{
@@ -16,7 +17,11 @@ class MihaaruScraper
public function __construct()
{
$this->client = new Client;
$this->client = new Client(
HttpClient::create([
"proxy" => config('karudhaas.proxy.host')
])
);
}
public function extract($url, $date = null)
@@ -28,9 +33,7 @@ class MihaaruScraper
$this->title = $node->text();
});
$crawler->filter('.container img')->eq(3)->each(function ($node) {
$this->image = $node->attr('src');
});
$this->image = $crawler->filter('.w-full.flex.flex-col.items-end.max-w-3xl.mb-10.relative img')->attr('src');
$crawler->filter('.by-line address')->each(function ($node) {
$author = $node->text();
@@ -41,21 +44,28 @@ class MihaaruScraper
$this->author = $cleaneddata;
});
$crawler->filter('article p')->each(function ($node) {
$this->content[] = preg_replace("/[a-zA-Z]/","",$node->text());
$crawler->filter('.text-faseyha')->each(function ($node) {
$this->content[] = $node->text();
});
$crawler->filter('.article-tags')->each(function ($node) {
$crawler->filter('.items-end a')->each(function ($node) {
try {
$topicName = $node->filter('span')->text();
$topicSlug = ltrim($node->attr('href'), '/');
} catch (\Throwable $th) {
return;
}
$this->topics[] = [
"name" => $node->text(),
"slug" => str_replace("https://mihaaru.com/", "", $node->attr('href'))
"name" => $topicName,
"slug" => $topicSlug
];
});
//Remove all the alphabets from string
//preg_replace("/[a-zA-Z]/", "",$string);
return [
return [
'source' => 'Mihaaru',
'title' => $this->title,
'og_title' => $crawler->filter('meta[property*="og:title"]')->first()->attr('content'),
@@ -63,7 +73,7 @@ class MihaaruScraper
'content' => $this->content,
'url' => $url,
'date' => $date,
'guid' => str_replace("https://mihaaru.com/news/","",$url),
'guid' => str_replace("https://mihaaru.com/news/", "", $url),
'author' => $this->author,
'topics' => $this->topics
];