OneOnline Support

app/Console/Commands/ScrapeOneOnlineCommand.php

@@ -0,0 +1,72 @@
+<?php
+
+namespace App\Console\Commands;
+
+use Illuminate\Console\Command;
+use App\Services\OneOnlineService;
+use App\Topic;
+use App\Source;
+
+class ScrapeOneOnlineCommand extends Command
+{
+    /**
+     * The name and signature of the console command.
+     *
+     * @var string
+     */
+    protected $signature = 'scrape:oneonline';
+
+    /**
+     * The console command description.
+     *
+     * @var string
+     */
+    protected $description = 'Scrape OneOnline';
+
+    /**
+     * 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', 'oneonline')->first();
+
+        $articles = (new OneOnlineService)->scrape();
+
+        foreach ($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" => $article["og_title"]
+                ]
+
+            ]);
+
+            collect($article["topics"])->each(function($topic) use ($articleModel) {
+                $topicModel = Topic::firstOrCreate(["slug" => $topic["slug"]],["name" => $topic["name"]]);
+
+                $topicModel->articles()->syncWithoutDetaching($articleModel);
+            });
+
+        }
+    }
+}

app/Console/Kernel.php

@@ -54,6 +54,9 @@ class Kernel extends ConsoleKernel
 
         $schedule->command('scrape:psm')->everyFiveMinutes()
             ->pingOnSuccess(env('APP_URL') . "/api/ping/psm");
+
+        $schedule->command('scrape:oneonline')->everyFiveMinutes()
+            ->pingOnSuccess(env('APP_URL') . "/api/ping/oneonline");
     }
 
     /**

app/Services/Feeds/OneOnlineFeed.php

@@ -0,0 +1,37 @@
+<?php
+namespace App\Services\Feeds;
+
+use Goutte\Client;
+
+class OneOnlineFeed
+{
+    protected $client;
+
+    public function __construct()
+    {
+        $this->client = new Client();        
+    }    
+    /**
+     * Return the latest articles from avas
+     *
+     * @return array
+     */
+    public function get() : array
+    {
+      
+        $crawler = $this->client->request('GET', "https://oneonline.mv/");
+
+        $feeds = [];
+
+        $crawler->filter('.latest-listing a')->each(function($node) use (&$feeds){
+            $feeds[] = [
+                "title" => $node->text(),
+                "link" => $node->attr('href')
+            ];
+        });
+
+
+        return $feeds;
+
+    }
+}

app/Services/OneOnlineService.php

@@ -0,0 +1,26 @@
+<?php
+namespace App\Services;
+
+use App\Services\Feeds\OneOnlineFeed;
+use App\Services\Scrapers\OneOnlineScraper;
+
+class OneOnlineService
+{
+    /**
+     * Scrap all the rss articles from Sun
+     *
+     * @return array
+     */
+    public function scrape(): array
+    {
+        //Return only the rss that contains "news" keyboard in its url
+        $articles = (new OneOnlineFeed)->get();
+        $articlesitems = [];
+        //Looping through the articles and scraping and while scraping it creates a new instance of the scraper.
+        foreach ($articles as $article) {
+            $articlesitems[] = (new OneOnlineScraper)->extract($article["link"]);
+        }
+
+        return $articlesitems;
+    }
+}

app/Services/Scrapers/OneOnlineScraper.php

@@ -0,0 +1,66 @@
+<?php
+
+namespace App\Services\Scrapers;
+
+use Goutte\Client;
+use Illuminate\Support\Carbon;
+
+class OneOnlineScraper
+{
+    protected $client;
+
+    protected $title;
+    protected $content;
+    protected $topics = [];
+    protected $author = "unknown";
+
+    public function __construct()
+    {
+        $this->client = new Client;
+    }
+
+    public function extract($url)
+    {
+
+        $crawler = $this->client->request('GET', $url);
+
+        $crawler->filter('h1')->each(function ($node) {
+            $this->title = $node->text();
+        });
+
+        $crawler->filter('.content p')->each(function ($node) {
+            $this->content[] = $node->text();
+        });
+
+        $crawler->filter('.tags a')->each(function ($node) {
+            //Removing the show more tags button
+            if ($node->text() == "ގުޅޭ ޓެގު") {
+                return;
+            }
+
+            $this->topics[] = [
+                "name" => $node->text(),
+                "slug" => str_replace("https://oneonline.mv/", "", $node->attr('href'))
+            ];
+        });
+
+
+        if ($crawler->filter('div[class*="text-grey-dark font-waheed flex flex-row items-center"] span')->count() == 1) {
+            $this->author = $crawler->filter('div[class*="text-grey-dark font-waheed flex flex-row items-center"] span')->first()->text();
+        }
+        //Remove all the alphabets from string
+        //preg_replace("/[a-zA-Z]/", "",$string);
+        return [
+            'source'    => 'OneOnline',
+            'title'      => $this->title,
+            'og_title'   => $crawler->filter('meta[property*="og:title"]')->first()->attr('content'),
+            'image'      => $crawler->filter("figure img")->first()->attr('src'),
+            'content'    => $this->content,
+            'url'        => $url,
+            'date'       => Carbon::parse($crawler->filter('time')->first()->attr('datetime'))->format("Y-m-d H:i:s"),
+            'guid'       => str_replace("https://oneonline.mv/", "", $url),
+            'author'     => $this->author,
+            'topics'       => $this->topics
+        ];
+    }
+}