fixup! Add ability to process prefetched content

Signed-off-by: Kevin Decherf <kevin@kdecherf.com>
j0k3r · Jan 15, 2022 · 7a9b7a2 · 7a9b7a2
1 parent 0208102
commit 7a9b7a2
Show file tree

Hide file tree

Showing 2 changed files with 19 additions and 3 deletions.
diff --git a/src/Graby.php b/src/Graby.php
@@ -266,7 +266,7 @@ public function cleanupHtml($contentBlock, string $url): string
         return trim($this->cleanupXss((string) $html));
     }
 
-    private function fakeFetch($url, $prefetchedContent) {
+    private function getResponseForPrefetchedContent($url, $prefetchedContent) {
         return [
             'body' => $prefetchedContent,
             'effective_url' => $url,
@@ -291,8 +291,8 @@ private function doFetchContent(string $url, string $prefetchedContent = null):
             $this->logger->info('Fetching url: {url}', ['url' => $url]);
             $response = $this->httpClient->fetch($url, false, $siteConfig->http_header);
         } else {
-            $this->logger->info('Fake fetching url: {url}', ['url' => $url]);
-            $response = $this->fakeFetch($url, $prefetchedContent);
+            $this->logger->info('Content provided, does not fetch url: {url}', ['url' => $url]);
+            $response = $this->getResponseForPrefetchedContent($url, $prefetchedContent);
         }
 
         $effectiveUrl = $response['effective_url'];

diff --git a/tests/GrabyTest.php b/tests/GrabyTest.php
@@ -1396,6 +1396,22 @@ public function testWithTooLongHtmlJitFail(): void
         $this->assertNotSame('No title found', $res['title']);
     }
 
+    public function testPrefetchedContent()
+    {
+        $httpMockClient = new HttpMockClient();
+        $graby = new Graby([
+            'debug' => true
+        ], $httpMockClient);
+
+        $input = '<html><body><h1>This is my awesome article</h1><article><p>' . str_repeat('This is an awesome text with some links, here there are the awesome', 7) . '</p></article></body></html>';
+
+        $res = $graby->fetchContent('https://example.com/prefetched-content', $input);
+
+        $this->assertSame('This is my awesome article', $res['title']);
+        $this->assertSame('https://example.com/prefetched-content', $res['url']);
+        $this->assertStringContainsString('here there are the awesome', $res['html']);
+    }
+
     /**
      * Return an instance of graby with a mocked Guzzle client returning data from a predefined file.
      */