Resolve Respekt redirects when getting URL content

remp2020 · Mar 27, 2024 · ede45c6 · ede45c6
1 parent 29acc1b
commit ede45c6
Show file tree

Hide file tree

Showing 2 changed files with 77 additions and 11 deletions.
diff --git a/Mailer/app/Models/PageMeta/Content/RespektContent.php b/Mailer/app/Models/PageMeta/Content/RespektContent.php
@@ -2,10 +2,13 @@
 
 namespace Remp\Mailer\Models\PageMeta\Content;
 
+use GuzzleHttp\Exception\RequestException;
 use Nette\Utils\Json;
 use Nette\Utils\JsonException;
 use Remp\Mailer\Models\PageMeta\RespektMeta;
+use Remp\Mailer\Models\PageMeta\Transport\RespektApiTransport;
 use Remp\MailerModule\Models\PageMeta\Content\ContentInterface;
+use Remp\MailerModule\Models\PageMeta\Content\InvalidUrlException;
 use Remp\MailerModule\Models\PageMeta\Meta;
 use Remp\MailerModule\Models\PageMeta\Transport\TransportInterface;
 use Tracy\Debugger;
@@ -27,19 +30,34 @@ public function __construct(private TransportInterface $transport)
 
     public function fetchUrlMeta(string $url): ?Meta
     {
-        $data = $this->transport->getContent($url);
-        if ($data === null) {
-            return null;
-        }
-
         try {
-            $data = Json::decode($data, true);
-        } catch (JsonException $e) {
-            Debugger::log($e->getMessage(), ILogger::ERROR);
-            return null;
-        }
+            $data = $this->transport->getContent($url);
+            if ($data === null) {
+                return null;
+            }
 
-        $article = $data['data']['getArticle'];
+            try {
+                $data = Json::decode($data, true);
+            } catch (JsonException $e) {
+                Debugger::log($e->getMessage(), ILogger::ERROR);
+                return null;
+            }
+
+            $article = $data['data']['getArticle'];
+            if ($article === null) {
+                // URL has changed, we need to follow redirects and determine new URL
+                if ($this->transport instanceof RespektApiTransport) {
+                    $resolvedUrl = $this->transport->resolveRedirects($url);
+                    if ($url && $resolvedUrl !== $url) {
+                        return $this->fetchUrlMeta($resolvedUrl);
+                    }
+                }
+
+                return null;
+            }
+        } catch (RequestException $e) {
+            throw new InvalidUrlException("Invalid URL: {$url}", 0, $e);
+        }
 
         // get article title
         $title = $article['title'];

diff --git a/Mailer/app/Models/PageMeta/Transport/RespektApiTransport.php b/Mailer/app/Models/PageMeta/Transport/RespektApiTransport.php
@@ -6,6 +6,7 @@
 use GuzzleHttp\Client;
 use GuzzleHttp\Exception\GuzzleException;
 use GuzzleHttp\Psr7\Uri;
+use Nette\Utils\Json;
 use Remp\MailerModule\Models\PageMeta\Transport\TransportInterface;
 use Tracy\Debugger;
 use Tracy\ILogger;
@@ -60,4 +61,51 @@ public function getContent(string $url): ?string
 
         return $response->getBody()->getContents();
     }
+
+    public function resolveRedirects(string $url): ?string
+    {
+        $uri = new Uri($url);
+        $contentPath = trim($uri->getPath(), '/');
+
+        $client = new Client();
+        try {
+            $query = <<<'GRAPHQL'
+query GetRedirectUrlForMailer($articleUrl: String) {
+  getUrl(by: { url: $articleUrl}) {
+    redirectType
+    target {
+      internalTarget {
+        url
+      }
+    }
+  }
+}
+GRAPHQL;
+
+            $response = $client->post($this->respektContentUrl, [
+                'headers' => ['Authorization' => 'Bearer ' . $this->respektContentToken],
+                'json' => [
+                    'query' => $query,
+                    'variables' => [
+                        'articleUrl' => $contentPath
+                    ]
+                ]
+            ]);
+        } catch (GuzzleException $e) {
+            Debugger::log($e->getMessage(), ILogger::ERROR);
+            return null;
+        }
+
+        $response = $response->getBody()->getContents();
+        $json = Json::decode($response);
+
+        $path = $json->data->getUrl->target->internalTarget->url ?? null;
+        if (!$path) {
+            return null;
+        }
+
+        $resolvedUrl = new Uri($url);
+        $resolvedUrl = $resolvedUrl->withPath($path);
+        return (string) $resolvedUrl;
+    }
 }