Add Minoos support

This commit is contained in:
Mohamed jinas
2024-01-12 04:34:28 +05:00
parent fbaa43748f
commit 46e2e849bb
5 changed files with 194 additions and 0 deletions

View File

@@ -0,0 +1,73 @@
<?php
namespace App\Console\Commands;
use Illuminate\Console\Command;
use App\Services\MinoosService;
use App\Topic;
use App\Source;
class ScrapeMinoosCommand extends Command
{
/**
* The name and signature of the console command.
*
* @var string
*/
protected $signature = 'scrape:minoos';
/**
* The console command description.
*
* @var string
*/
protected $description = 'Scrape Minoos';
/**
* Create a new command instance.
*
* @return void
*/
public function __construct()
{
parent::__construct();
}
/**
* Execute the console command.
*
* @return int
*/
public function handle()
{
$source = Source::where('slug', 'minoos')->first();
$articles = (new MinoosService)->scrape();
foreach ($articles as $article) {
// Attach the relationship between source and article and return the curren article instance
$articleModel = $source->articles()->updateOrCreate(["guid" => $article["guid"]],
[
"title" => $article["title"],
"url" => $article["url"],
"author" => $article["author"],
"featured_image" => $article["image"],
"body" => $article["content"],
"published_date" => $article["date"],
"meta" => [
"title" => $article["title"],
"highlights" => $article["highlights"]
]
]);
collect($article["topics"])->each(function($topic) use ($articleModel) {
$topicModel = Topic::firstOrCreate(["slug" => $topic["slug"]],["name" => $topic["name"]]);
$topicModel->articles()->syncWithoutDetaching($articleModel);
});
}
}
}

View File

@@ -105,6 +105,10 @@ class Kernel extends ConsoleKernel
$schedule->command('scrape:adhadhu')->everyFiveMinutes()
->runInBackground()
->pingOnSuccess(config('app.url') . "/api/ping/adhadhu");
$schedule->command('scrape:minoos')->everyFiveMinutes()
->runInBackground()
->pingOnSuccess(config('app.url') . "/api/ping/minoos");
}

View File

@@ -0,0 +1,38 @@
<?php
namespace App\Services\Feeds;
use GuzzleHttp\Client;
class MinoosFeed implements Feed
{
protected $client;
public function __construct()
{
$this->client = new Client();
}
/**
* Get all the latest news
*
* @return array
*/
public function get() : array
{
$response = $this->client->request('GET', "https://fili.minoos.mv/api/category/news/posts");
$data = json_decode($response->getBody(), true);
$feeds = [];
foreach ($data['data'] as $item) {
$feeds[] = [
"title" => $item['heading'],
"link" => "https://minoos.mv/" . $item['id'],
"date" => $item['published_at'],
"highlights" => $item['highlights'],
];
}
return $feeds;
}
}

View File

@@ -0,0 +1,31 @@
<?php
namespace App\Services;
use App\Services\Feeds\MinoosFeed;
use App\Services\Feeds\OneOnlineFeed;
use App\Services\Scrapers\MinoosScraper;
use App\Services\Scrapers\OneOnlineScraper;
class MinoosService
{
/**
*
* @return array
*/
public function scrape(): array
{
//Return only the rss that contains "news" keyboard in its url
$articles = (new MinoosFeed)->get();
$articlesitems = [];
//Looping through the articles and scraping and while scraping it creates a new instance of the scraper.
foreach ($articles as $article) {
$scraped_article = (new MinoosScraper)->extract($article["link"], $article["date"], $article["highlights"]);
if (!is_null($scraped_article)) {
$articlesitems[] = $scraped_article;
}
}
return $articlesitems;
}
}

View File

@@ -0,0 +1,48 @@
<?php
namespace App\Services\Scrapers;
use Goutte\Client;
use Symfony\Component\DomCrawler\Crawler;
use Illuminate\Support\Carbon;
class MinoosScraper
{
protected $client;
public function __construct()
{
$this->client = new Client();
}
public function extract($url, $date = null, $highlights = null)
{
$crawler = $this->client->request('GET', $url);
$title = $crawler->filter('h1')->first()->text();
$content = $crawler->filter('.doc-text')->each(function (Crawler $node) {
return $node->text();
});
$image = $crawler->filter('meta[property="og:image"]')->attr('content');
$topics = $crawler->filter('a[href^="/tags/"]')->each(function (Crawler $node) {
return [
'name' => $node->text(),
'slug' => str_replace("/tags/", "", $node->attr('href'))
];
});
return [
'source' => 'minoos',
'title' => $title,
'guid' => basename($url),
'content' => $content,
'author' => 'unknown',
'image' => $image,
'highlights' => $highlights,
'url' => $url,
'date' => Carbon::parse($date)->format("Y-m-d H:i:s"),
'topics' => $topics
];
}
}