From ab890d2cc595a4fac0afef4d1b691c60bb8d515e Mon Sep 17 00:00:00 2001
From: Johannes Pichler <fetzi@users.noreply.github.com>
Date: Tue, 3 Oct 2017 09:40:58 +0200
Subject: [PATCH] add crawl_profile config to set custom CrawlProfile
 implementation (#94)

* add crawl_profile config to set custom CrawlProfile implementation, fixes #90

* apply style ci fixes

* add tests for custom crawl profile

* apply fixes from style ci

* remove empty line
---
 README.md                      |  6 ++++
 config/sitemap.php             |  7 ++++
 src/SitemapGenerator.php       |  7 ++--
 src/SitemapServiceProvider.php |  5 +++
 tests/CrawlProfileTest.php     | 58 ++++++++++++++++++++++++++++++++++
 tests/CustomCrawlProfile.php   | 21 ++++++++++++
 6 files changed, 102 insertions(+), 2 deletions(-)
 create mode 100644 tests/CrawlProfileTest.php
 create mode 100644 tests/CustomCrawlProfile.php

diff --git a/README.md b/README.md
index c100420..2a7ef09 100644
--- a/README.md
+++ b/README.md
@@ -130,6 +130,12 @@ return [
      */
     'chrome_binary_path' => '',
 
+    /*
+     * The sitemap generator uses a CrawlProfile implementation to determine
+     * which urls should be crawled for the sitemap.
+     */
+    'crawl_profile' => Profile::class,
+    
 ];
 ```
 
diff --git a/config/sitemap.php b/config/sitemap.php
index 2ad425a..8ffec69 100644
--- a/config/sitemap.php
+++ b/config/sitemap.php
@@ -1,6 +1,7 @@
 <?php
 
 use GuzzleHttp\RequestOptions;
+use Spatie\Sitemap\Crawler\Profile;
 
 return [
 
@@ -47,4 +48,10 @@
      */
     'chrome_binary_path' => null,
 
+    /*
+     * The sitemap generator uses a CrawlProfile implementation to determine
+     * which urls should be crawled for the sitemap.
+     */
+    'crawl_profile' => Profile::class,
+
 ];
diff --git a/src/SitemapGenerator.php b/src/SitemapGenerator.php
index 3c4d5a3..b065ad4 100644
--- a/src/SitemapGenerator.php
+++ b/src/SitemapGenerator.php
@@ -4,6 +4,7 @@
 
 use Spatie\Crawler\Crawler;
 use Spatie\Sitemap\Tags\Url;
+use Spatie\Crawler\CrawlProfile;
 use Spatie\Sitemap\Crawler\Profile;
 use Spatie\Sitemap\Crawler\Observer;
 use Spatie\Crawler\Url as CrawlerUrl;
@@ -103,7 +104,7 @@ public function writeToFile(string $path)
         return $this;
     }
 
-    protected function getCrawlProfile(): Profile
+    protected function getCrawlProfile(): CrawlProfile
     {
         $shouldCrawl = function (CrawlerUrl $url) {
             if ($url->host !== CrawlerUrl::create($this->urlToBeCrawled)->host) {
@@ -117,7 +118,9 @@ protected function getCrawlProfile(): Profile
             return ($this->shouldCrawl)($url);
         };
 
-        return new Profile($shouldCrawl);
+        $profileClass = config('sitemap.crawl_profile', Profile::class);
+
+        return app($profileClass, [$shouldCrawl]);
     }
 
     protected function getCrawlObserver(): Observer
diff --git a/src/SitemapServiceProvider.php b/src/SitemapServiceProvider.php
index a7204e9..446c745 100644
--- a/src/SitemapServiceProvider.php
+++ b/src/SitemapServiceProvider.php
@@ -3,6 +3,7 @@
 namespace Spatie\Sitemap;
 
 use Spatie\Crawler\Crawler;
+use Spatie\Sitemap\Crawler\Profile;
 use Illuminate\Support\ServiceProvider;
 
 class SitemapServiceProvider extends ServiceProvider
@@ -27,6 +28,10 @@ public function boot()
             ->give(function () {
                 return Crawler::create(config('sitemap.guzzle_options'));
             });
+
+        $this->app->bind(Profile::class, function ($app, $params) {
+            return new Profile(reset($params));
+        });
     }
 
     /**
diff --git a/tests/CrawlProfileTest.php b/tests/CrawlProfileTest.php
new file mode 100644
index 0000000..7dcb9b0
--- /dev/null
+++ b/tests/CrawlProfileTest.php
@@ -0,0 +1,58 @@
+<?php
+
+namespace Spatie\Sitemap\Test;
+
+use Spatie\Crawler\Crawler;
+use Spatie\Sitemap\Sitemap;
+use Spatie\Sitemap\Crawler\Profile;
+use Spatie\Sitemap\SitemapGenerator;
+
+class CrawlProfileTest extends TestCase
+{
+    /**
+     * @var Crawler
+     */
+    private $crawler;
+
+    public function setUp()
+    {
+        parent::setUp();
+
+        $this->crawler = $this->createMock(Crawler::class);
+
+        $this->crawler->method('setCrawlObserver')->willReturn($this->crawler);
+        $this->crawler->method('setConcurrency')->willReturn($this->crawler);
+    }
+
+    /** @test */
+    public function it_should_use_the_default_crawl_profile()
+    {
+        $this->crawler
+            ->method('setCrawlProfile')
+            ->with($this->isInstanceOf(Profile::class))
+            ->willReturn($this->crawler);
+
+        $sitemapGenerator = new SitemapGenerator($this->crawler);
+
+        $sitemap = $sitemapGenerator->getSitemap();
+
+        $this->assertInstanceOf(Sitemap::class, $sitemap);
+    }
+
+    /** @test */
+    public function it_should_use_a_custom_crawl_profile()
+    {
+        config(['sitemap.crawl_profile' => CustomCrawlProfile::class]);
+
+        $this->crawler
+            ->method('setCrawlProfile')
+            ->with($this->isInstanceOf(CustomCrawlProfile::class))
+            ->willReturn($this->crawler);
+
+        $sitemapGenerator = new SitemapGenerator($this->crawler);
+
+        $sitemap = $sitemapGenerator->getSitemap();
+
+        $this->assertInstanceOf(Sitemap::class, $sitemap);
+    }
+}
diff --git a/tests/CustomCrawlProfile.php b/tests/CustomCrawlProfile.php
new file mode 100644
index 0000000..a6f1326
--- /dev/null
+++ b/tests/CustomCrawlProfile.php
@@ -0,0 +1,21 @@
+<?php
+
+namespace Spatie\Sitemap\Test;
+
+use Spatie\Crawler\Url;
+use Spatie\Crawler\CrawlProfile;
+
+class CustomCrawlProfile implements CrawlProfile
+{
+    /**
+     * Determine if the given url should be crawled.
+     *
+     * @param \Spatie\Crawler\Url $url
+     *
+     * @return bool
+     */
+    public function shouldCrawl(Url $url): bool
+    {
+        return true;
+    }
+}