FUnadhoo times support

app/Console/Commands/ScrapeFunadhooTimesCommand.php

@@ -0,0 +1,74 @@
+<?php
+
+namespace App\Console\Commands;
+
+use Illuminate\Console\Command;
+use App\Source;
+use App\Services\FunadhooTimesService;
+use App\Topic;
+use Illuminate\Support\Carbon;
+
+class ScrapeFunadhooTimesCommand extends Command
+{
+    /**
+     * The name and signature of the console command.
+     *
+     * @var string
+     */
+    protected $signature = 'scrape:funadhoo-times';
+
+    /**
+     * The console command description.
+     *
+     * @var string
+     */
+    protected $description = 'Scrape Funadhoo times';
+
+    /**
+     * 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', 'funadhoo-times')->first();
+
+        $articles = (new FunadhooTimesService)->scrape();
+
+        foreach ($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/Services/FunadhooTimesService.php

@@ -0,0 +1,30 @@
+<?php
+
+namespace App\Services;
+
+use App\Services\Scrapers\FunadhooTimesScraper;
+use Illuminate\Support\Str;
+
+
+class FunadhooTimesService extends Client
+{
+    /**
+     * Scrap all the rss articles from Press
+     *
+     * @return array
+     */
+    public function scrape(): array
+    {
+        //Return only the rss that contains "news" keyboard in its url
+        $articles = $this->get("https://funadhootimes.com/rss.xml")["channel"]["article"];
+        $articlesitems = [];
+        //Looping through the articles and scraping and while scraping it creates a new instance of the scraper.
+        foreach ($articles as $article) {
+            $link = $article['link'];
+            $title = $article['title'];
+            $articlesitems[] = (new FunadhooTimesScraper)->extract($link, $title);
+        }
+
+        return $articlesitems;
+    }
+}

app/Services/Scrapers/FunadhooTimesScraper.php

@@ -0,0 +1,60 @@
+<?php
+
+namespace App\Services\Scrapers;
+
+use Exception;
+use Goutte\Client;
+use Illuminate\Support\Carbon;
+
+class FunadhooTimesScraper
+{
+    protected $client;
+
+    protected $title;
+    protected $content;
+    protected $topics;
+
+    public function __construct()
+    {
+        $this->client = new Client;
+    }
+
+    public function extract($url,$title)
+    {
+            $crawler = $this->client->request('GET', $url);
+
+            $crawler->filter('.card-body > p')->each(function ($node) {
+                $this->content[] = $node->text();
+            });
+
+
+            $crawler->filter('.article-tags a')->each(function ($node) {
+                $this->topics[] = [
+                    "name" => $node->text(),
+                    "slug" => str_replace("https://hama.mv/", "", $node->attr('href'))
+                ];
+            });
+
+
+            //Remove all the alphabets from string
+            //preg_replace("/[a-zA-Z]/", "",$string);
+            return [
+                'source'    => 'Funadhoo times',
+                'title'      => $title,
+                'og_title'   => $crawler->filter('meta[property*="og:title"]')->first()->attr('content'),
+               'image'      =>  "https://funadhootimes.com/".$crawler->filter(".card img")->first()->attr('src'),
+                'content'    => $this->content,
+                'url'        => $url,
+                'date'       => Carbon::parse(explode('|',$crawler->filter('div[class*="d-flex flex-row-reverse justify-content-between"] small')->first()->text())[0])->format("Y-m-d H:i:s"),
+                'guid'       => str_replace("https://www.funadhootimes.com/p.php?id=","",$url),
+                'author'     => explode('|',$crawler->filter('div[class*="d-flex flex-row-reverse justify-content-between"] small')->first()->text())[1],
+                'topics'       =>    [
+                    [
+                        "name" => "ވަކި މަޢުލޫއެއް ނޭންގެ",
+                        "slug" => "no-specific-topic"
+                    ]
+                ]
+            ];
+        
+    }
+}