From 5904418e8d78c0657598d5dceba69822cbfdf698 Mon Sep 17 00:00:00 2001
From: Charles Teh <charlesteh90@gmail.com>
Date: Wed, 29 Nov 2023 23:02:23 +0800
Subject: [PATCH] Add (custom crawl profile)

---
 .../SitemapCrawler/CustomCrawlProfile.php     | 28 +++++++++++++++++++
 config/sitemap.php                            |  3 +-
 2 files changed, 30 insertions(+), 1 deletion(-)
 create mode 100644 app/Helpers/FirstParty/SitemapCrawler/CustomCrawlProfile.php

diff --git a/app/Helpers/FirstParty/SitemapCrawler/CustomCrawlProfile.php b/app/Helpers/FirstParty/SitemapCrawler/CustomCrawlProfile.php
new file mode 100644
index 0000000..5b231f5
--- /dev/null
+++ b/app/Helpers/FirstParty/SitemapCrawler/CustomCrawlProfile.php
@@ -0,0 +1,28 @@
+<?php
+
+namespace App\Helpers\FirstParty\SitemapCrawler;
+
+use Spatie\Crawler\CrawlProfiles\CrawlProfile;
+use Psr\Http\Message\UriInterface;
+
+class CustomCrawlProfile extends CrawlProfile
+{
+    /** @var callable */
+    protected $callback;
+
+    public function shouldCrawlCallback(callable $callback): void
+    {
+        $this->callback = $callback;
+    }
+
+    public function shouldCrawl(UriInterface $url): bool
+    {
+        if ($url->getQuery() !== '') {
+            return false;
+        }
+        
+
+        return ($this->callback)($url);
+    }
+
+}
diff --git a/config/sitemap.php b/config/sitemap.php
index 69be0f3..2007c5e 100644
--- a/config/sitemap.php
+++ b/config/sitemap.php
@@ -1,5 +1,6 @@
 <?php
 
+use App\Helpers\FirstParty\SitemapCrawler\CustomCrawlProfile;
 use GuzzleHttp\RequestOptions;
 use Spatie\Sitemap\Crawler\Profile;
 
@@ -52,6 +53,6 @@
      * The sitemap generator uses a CrawlProfile implementation to determine
      * which urls should be crawled for the sitemap.
      */
-    'crawl_profile' => Profile::class,
+    'crawl_profile' => CustomCrawlProfile::class,
 
 ];