add pdf reader modul

src/Knowledge/Ingest/DocumentLoader.php

@@ -5,6 +5,8 @@ declare(strict_types=1);
 
 namespace App\Knowledge\Ingest;
 
+use Smalot\PdfParser\Parser;
+
 final class DocumentLoader
 {
     public function load(string $path): string
@@ -17,8 +19,10 @@ final class DocumentLoader
 
         return match ($ext) {
             'txt', 'md' => $this->loadText($path),
-            // 'pdf' => $this->loadPdf($path),
-            // 'docx' => $this->loadDocx($path),
+            'pdf'       => $this->loadPdf($path),
+
+            // vorbereitet für später:
+            // 'docx'   => $this->loadDocx($path),
 
             default => throw new \RuntimeException("Unsupported file type: .{$ext}"),
         };
@@ -30,7 +34,41 @@ final class DocumentLoader
         if ($content === false) {
             throw new \RuntimeException("Could not read file: {$path}");
         }
-        return $content;
+
+        return $this->normalize($content);
     }
 
+    private function loadPdf(string $path): string
+    {
+        $parser = new Parser();
+
+        try {
+            $pdf = $parser->parseFile($path);
+            $text = $pdf->getText();
+        } catch (\Throwable $e) {
+            throw new \RuntimeException("Failed to parse PDF: {$path}", 0, $e);
+        }
+
+        return $this->normalize($text);
+    }
+
+    /**
+     * Zentraler Normalizer für alle Dokumenttypen
+     */
+    private function normalize(string $text): string
+    {
+        // Silbentrennung entfernen
+        $text = preg_replace('/-\n/', '', $text);
+
+        // Windows-Zeilenumbrüche
+        $text = str_replace("\r\n", "\n", $text);
+
+        // Mehrfache Leerzeichen
+        $text = preg_replace('/[ \t]+/', ' ', $text);
+
+        // Mehrfache Leerzeilen
+        $text = preg_replace('/\n{3,}/', "\n\n", $text);
+
+        return trim($text);
+    }
 }