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

Update AvasScraper.php

Browse Source
This commit is contained in:
Mohamed Jinas
2020-09-24 13:37:36 +05:00
parent b8641509a5
commit 4c72042aeb
1 changed files with 11 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
app/Services/Scrapers/AvasScraper.php
View File

@@ -1,4 +1,5 @@
<?php <?php
namespace App\Services\Scrapers; namespace App\Services\Scrapers;
use Goutte\Client; use Goutte\Client;
@@ -11,7 +12,7 @@ class AvasScraper
protected $content; protected $content;
protected $author; protected $author;
protected $topics = []; protected $topics = [];
public function __construct() public function __construct()
{ {
$this->client = new Client; $this->client = new Client;
@@ -26,17 +27,16 @@ class AvasScraper
$image = $crawler->filter('figure img')->first()->attr('src'); $image = ($crawler->filter('figure img')->first()->attr('src')) ? $crawler->filter('figure img')->first()->attr('src') : "/images/noimg.jpg";
$crawler->filter('.post_content p')->each(function ($node) { $crawler->filter('.post_content p')->each(function ($node) {
$this->content[] = preg_replace("/[a-zA-Z]/","",$node->text()); $this->content[] = preg_replace("/[a-zA-Z]/", "", $node->text());
}); });
$crawler->filter('div[class*="border-t border-grey-light border-dotted mt-7 py-3"] a')->each(function ($node) { $crawler->filter('div[class*="border-t border-grey-light border-dotted mt-7 py-3"] a')->each(function ($node) {
//Removing the show more tags button //Removing the show more tags button
if($node->text() == "+") if ($node->text() == "+") {
{
return; return;
} }
$this->topics[] = [ $this->topics[] = [
@@ -45,17 +45,16 @@ class AvasScraper
]; ];
}); });
if($crawler->filter('div[class*="font-waheed text-grey ml-3 pl-3 text-lg border-l border-grey border-dotted"] a')->count() == 1)
{ if ($crawler->filter('div[class*="font-waheed text-grey ml-3 pl-3 text-lg border-l border-grey border-dotted"] a')->count() == 1) {
$this->author = $crawler->filter('div[class*="font-waheed text-grey ml-3 pl-3 text-lg border-l border-grey border-dotted"] a')->first()->text(); $this->author = $crawler->filter('div[class*="font-waheed text-grey ml-3 pl-3 text-lg border-l border-grey border-dotted"] a')->first()->text();
} }
//Remove all the alphabets from string //Remove all the alphabets from string
//preg_replace("/[a-zA-Z]/", "",$string); //preg_replace("/[a-zA-Z]/", "",$string);
return [ return [
'source' => 'Avas', 'source' => 'Avas',
'title' => $title, 'title' => $title,
'og_title' => $crawler->filter('meta[property*="og:title"]')->first()->attr('content'), 'og_title' => $crawler->filter('meta[property*="og:title"]')->first()->attr('content'),
@@ -63,9 +62,9 @@ class AvasScraper
'content' => $this->content, 'content' => $this->content,
'url' => $url, 'url' => $url,
'date' => Carbon::parse($crawler->filter('timeago')->first()->attr('datetime'))->format("Y-m-d H:i:s"), 'date' => Carbon::parse($crawler->filter('timeago')->first()->attr('datetime'))->format("Y-m-d H:i:s"),
'guid' => str_replace("https://avas.mv/","",$url), 'guid' => str_replace("https://avas.mv/", "", $url),
'author' => $this->author, 'author' => $this->author,
'topics' => $this->topics 'topics' => $this->topics
]; ];
} }
} }