Prepare release 0.3.0 (Apache Tika integration)

README.md

@@ -0,0 +1,73 @@
+# Contao Meilisearch Bundle
+
+Eine schlanke Schnittstelle zwischen **Contao CMS (4.13 / 5.6 / 5.7 ready) unter PHP 8.4** und einer **selbst gehosteten Meilisearch-Instanz**.  
+Das Bundle erweitert den Contao-Suchindex um strukturierte Daten und ermöglicht eine performante, moderne Volltextsuche.
+Das Parsen von Dateien erfolgt über eine Apache-Tika-Instanz, welche extern bereitgestellt werden muss.
+
+---
+
+## ✨ Features
+
+- Integration von **Meilisearch** als externe Suchmaschine
+- Indexierung von:
+    - Contao-Seiten
+    - Inhaltselementen
+    - **PDF-Dateien**
+    - **Office-Dokumenten** (DOCX, XLSX, PPTX)
+- Unterstützung für:
+    - Seiten-Prioritäten
+    - Keywords
+    - Vorschaubild
+- Kompatibel mit:
+    - Contao **4.13**, **5.6** und **5.7**
+    - PHP **8.4**
+
+---
+
+## ⏱️ Scheduled Indexing (Cron setup)
+
+Das Bundle stellt eigene Commands zur Verfügung, um Dateien zu bereinigen und den Meilisearch-Index neu aufzubauen.  
+Für den produktiven Einsatz wird empfohlen, diese Commands regelmäßig per **System-Crontab** auszuführen.
+
+Das Bundle nutzt **keinen eigenen Contao-Cron**, sondern System-Cronjobs.
+
+## Verfügbare Commands
+
+### Datei-Cleanup
+
+```
+/vendor/bin/contao-console meilisearch:files:cleanup
+```
+
+### Datei-Parsing
+
+```
+/vendor/bin/contao-console meilisearch:files:parse
+```
+
+### Meilisearch-Index
+
+```
+/vendor/bin/contao-console meilisearch:index
+```
+
+
+
+## Beispiel Crontab
+
+```
+0 5 * * *  /usr/bin/php8.4 /path/to/project/vendor/bin/contao-console meilisearch:files:cleanup
+1 5 * * *  /usr/bin/php8.4 /path/to/project/vendor/bin/contao-console contao:crawl
+10 5 * * * /usr/bin/php8.4 /path/to/project/vendor/bin/contao-console meilisearch:files:parse
+20 5 * * * /usr/bin/php8.4 /path/to/project/vendor/bin/contao-console meilisearch:index
+```
+
+## Logging
+
+```
+>> var/logs/meilisearch_cron.log 2>&1
+```
+
+## Lizenz
+
+MIT

composer.json

@@ -8,11 +8,7 @@
     "contao/core-bundle": "^4.13 || ^5.6 || ^5.7",
     "contao/calendar-bundle": "^4.13 || ^5.6 || ^5.7",
     "contao/news-bundle": "^4.13 || ^5.6 || ^5.7",
-    "meilisearch/meilisearch-php": "^1.16",
+    "meilisearch/meilisearch-php": "^1.16"
-    "smalot/pdfparser": "^2.12",
-    "phpoffice/phpword": "^1.4",
-    "phpoffice/phpspreadsheet": "^3.0",
-    "phpoffice/phppresentation": "^1.2"
   },
   "autoload": {
     "psr-4": {

src/Command/MeilisearchFilesCleanupCommand.php

@@ -0,0 +1,100 @@
+<?php
+
+namespace MummertMedia\ContaoMeilisearchBundle\Command;
+
+use Contao\CoreBundle\Framework\ContaoFramework;
+use Doctrine\DBAL\Connection;
+use Symfony\Component\Console\Command\Command;
+use Symfony\Component\Console\Input\InputInterface;
+use Symfony\Component\Console\Input\InputOption;
+use Symfony\Component\Console\Output\OutputInterface;
+
+class MeilisearchFilesCleanupCommand extends Command
+{
+    public function __construct(
+        private readonly ContaoFramework $framework,
+        private readonly Connection $connection,
+    ) {
+        parent::__construct();
+    }
+
+    protected function configure(): void
+    {
+        $this
+            ->setName('meilisearch:files:cleanup')
+            ->setDescription('Remove stale indexed files from tl_search_files')
+            ->addOption(
+                'grace',
+                null,
+                InputOption::VALUE_OPTIONAL,
+                'Grace period in seconds (files newer than now-grace are kept)',
+                86400
+            )
+            ->addOption(
+                'dry-run',
+                null,
+                InputOption::VALUE_NONE,
+                'Show how many entries would be removed without deleting them'
+            );
+    }
+
+    protected function execute(InputInterface $input, OutputInterface $output): int
+    {
+        $this->framework->initialize();
+
+        $this->log('Cleaner gestartet');
+
+        try {
+            $grace  = max(0, (int) $input->getOption('grace'));
+            $dryRun = (bool) $input->getOption('dry-run');
+            $cutoff = time() - $grace;
+
+            if ($dryRun) {
+                $count = $this->connection->fetchOne(
+                    'SELECT COUNT(*) FROM tl_search_files WHERE last_seen < ?',
+                    [$cutoff]
+                );
+
+                $message = sprintf(
+                    '[DRY-RUN] %d stale file(s) would be removed (last_seen < %s)',
+                    $count,
+                    date('Y-m-d H:i:s', $cutoff)
+                );
+
+                $output->writeln('<comment>' . $message . '</comment>');
+                $this->log($message);
+
+                $this->log('Cleaner stopped (dry-run)');
+                return Command::SUCCESS;
+            }
+
+            $affected = $this->connection->executeStatement(
+                'DELETE FROM tl_search_files WHERE last_seen < ?',
+                [$cutoff]
+            );
+
+            $message = sprintf(
+                'Removed %d stale file(s) (last_seen < %s)',
+                $affected,
+                date('Y-m-d H:i:s', $cutoff)
+            );
+
+            $output->writeln('<info>' . $message . '</info>');
+            $this->log($message);
+
+            $this->log('Cleaner successfully stopped');
+            return Command::SUCCESS;
+
+        } catch (\Throwable $e) {
+            $this->log('Cleaner ERROR: ' . $e->getMessage());
+            $output->writeln('<error>' . $e->getMessage() . '</error>');
+
+            return Command::FAILURE;
+        }
+    }
+
+    private function log(string $message): void
+    {
+        error_log(sprintf('[%s] %s', date('Y-m-d H:i:s'), $message));
+    }
+}

src/Command/MeilisearchFilesParseCommand.php

@@ -0,0 +1,217 @@
+<?php
+
+namespace MummertMedia\ContaoMeilisearchBundle\Command;
+
+use Contao\CoreBundle\Framework\ContaoFramework;
+use Contao\Database;
+use Symfony\Component\Console\Command\Command;
+use Symfony\Component\Console\Input\InputInterface;
+use Symfony\Component\Console\Input\InputOption;
+use Symfony\Component\Console\Output\OutputInterface;
+use Symfony\Component\HttpClient\HttpClient;
+
+class MeilisearchFilesParseCommand extends Command
+{
+    public function __construct(
+        private readonly ContaoFramework $framework,
+    ) {
+        parent::__construct();
+    }
+
+    protected function configure(): void
+    {
+        $this
+            ->setName('meilisearch:files:parse')
+            ->setDescription('Parse indexed files via Apache Tika and store extracted text')
+            ->addOption(
+                'limit',
+                null,
+                InputOption::VALUE_OPTIONAL,
+                'Maximum number of files to check per run (optional)'
+            )
+            ->addOption(
+                'dry-run',
+                null,
+                InputOption::VALUE_NONE,
+                'Do not send files to Tika, just show what would be parsed'
+            );
+    }
+
+    protected function execute(InputInterface $input, OutputInterface $output): int
+    {
+        $this->framework->initialize();
+        $this->log('Parser gestartet');
+
+        $dryRun = (bool) $input->getOption('dry-run');
+
+        // ---- LIMIT: nur wenn explizit gesetzt
+        $limitOption = $input->getOption('limit');
+        $limit = $limitOption !== null ? max(1, (int) $limitOption) : null;
+
+        // ---- Tika URL
+        $tikaUrl = rtrim((string) ($GLOBALS['TL_CONFIG']['meilisearch_tika_url'] ?? ''), '/');
+        if ($tikaUrl === '') {
+            $output->writeln('<error>Tika URL not configured</error>');
+            return Command::FAILURE;
+        }
+
+        $db = Database::getInstance();
+
+        // ---- Files laden
+        $sql = "SELECT * FROM tl_search_files ORDER BY tstamp ASC";
+        if ($limit !== null) {
+            $sql .= " LIMIT " . (int) $limit;
+        }
+
+        $files = $db->query($sql)->fetchAllAssoc();
+
+        if (!$files) {
+            $this->log('No files to parse');
+            return Command::SUCCESS;
+        }
+
+        $client = HttpClient::create([
+            'timeout' => 180,
+        ]);
+
+        foreach ($files as $file) {
+
+            $originalUrl = (string) $file['url'];
+            $normalized  = $originalUrl;
+
+            // -------------------------------------------------
+            // 1) ?file=files/…
+            // -------------------------------------------------
+            if (str_contains($normalized, '?')) {
+                $parts = parse_url($normalized);
+                if (!empty($parts['query'])) {
+                    parse_str($parts['query'], $query);
+                    if (!empty($query['file'])) {
+                        $normalized = (string) $query['file'];
+                    } else {
+                        $this->log('Not a direct file url, skip', ['url' => $originalUrl]);
+                        continue;
+                    }
+                }
+            }
+
+            // -------------------------------------------------
+            // 2) Fragment entfernen
+            // -------------------------------------------------
+            $normalized = strtok($normalized, '#');
+
+            // -------------------------------------------------
+            // 3) URL-Decoding
+            // -------------------------------------------------
+            $normalized = rawurldecode($normalized);
+
+            // -------------------------------------------------
+            // 4) Nur lokale files/
+            // -------------------------------------------------
+            $normalized = ltrim($normalized, '/');
+            if (!str_starts_with($normalized, 'files/')) {
+                $this->log('Not in files/, skip', ['url' => $originalUrl]);
+                continue;
+            }
+
+            $absolutePath = TL_ROOT . '/' . $normalized;
+
+            if (!is_file($absolutePath)) {
+                $this->log('File missing, skip', [
+                    'url'  => $originalUrl,
+                    'path' => $absolutePath,
+                ]);
+                continue;
+            }
+
+            $mtime    = filemtime($absolutePath) ?: 0;
+            $checksum = md5($normalized . '|' . $mtime);
+
+            // -------------------------------------------------
+            // 5) Skip unchanged
+            // -------------------------------------------------
+            if ($file['checksum'] === $checksum && !empty($file['text'])) {
+                continue;
+            }
+
+            if ($dryRun) {
+                $output->writeln('[DRY-RUN] Would parse: ' . $normalized);
+                continue;
+            }
+
+            // -------------------------------------------------
+            // 6) MIME-Type
+            // -------------------------------------------------
+            $ext = strtolower(pathinfo($normalized, PATHINFO_EXTENSION));
+
+            $mimeType = match ($ext) {
+                'pdf'  => 'application/pdf',
+                'docx' => 'application/vnd.openxmlformats-officedocument.wordprocessingml.document',
+                'xlsx' => 'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet',
+                'pptx' => 'application/vnd.openxmlformats-officedocument.presentationml.presentation',
+                default => null,
+            };
+
+            if ($mimeType === null) {
+                $this->log('Unsupported file type, skip', ['url' => $normalized]);
+                continue;
+            }
+
+            // -------------------------------------------------
+            // 7) Tika parse
+            // -------------------------------------------------
+            try {
+                $this->log('Parsing file', ['url' => $normalized]);
+
+                $response = $client->request(
+                    'PUT',
+                    $tikaUrl . '/tika/main',
+                    [
+                        'headers' => [
+                            'Accept'       => 'text/plain',
+                            'Content-Type' => $mimeType,
+                        ],
+                        'body' => fopen($absolutePath, 'rb'),
+                    ]
+                );
+
+                $text = trim((string) $response->getContent(false));
+
+                $db->prepare(
+                    "UPDATE tl_search_files
+                     SET text = ?, checksum = ?, file_mtime = ?, tstamp = ?
+                     WHERE id = ?"
+                )->execute(
+                    $text,
+                    $checksum,
+                    $mtime,
+                    time(),
+                    $file['id']
+                );
+
+                $this->log('File parsed', [
+                    'url'   => $normalized,
+                    'chars' => mb_strlen($text),
+                ]);
+
+            } catch (\Throwable $e) {
+                $this->log('Parse failed', [
+                    'url'   => $normalized,
+                    'error' => $e->getMessage(),
+                ]);
+            }
+        }
+
+        $this->log('Parser finished');
+        return Command::SUCCESS;
+    }
+
+    private function log(string $message, array $context = []): void
+    {
+        $ctx = $context
+            ? ' | ' . json_encode($context, JSON_UNESCAPED_SLASHES | JSON_UNESCAPED_UNICODE)
+            : '';
+
+        error_log('[MeilisearchFilesParse] ' . $message . $ctx);
+    }
+}

src/Command/MeilisearchIndexCommand.php

@@ -24,12 +24,34 @@ class MeilisearchIndexCommand extends Command
 
     protected function execute(InputInterface $input, OutputInterface $output): int
     {
+        $this->log('Meilisearch index gestartet');
         $output->writeln('<info>Meilisearch index started</info>');
 
-        $this->indexService->run();
+        try {
+            $this->indexService->run();
 
-        $output->writeln('<info>Meilisearch index finished</info>');
+            $this->log('Meilisearch index successfully stopped');
+            $output->writeln('<info>Meilisearch index finished</info>');
 
-        return Command::SUCCESS;
+            return Command::SUCCESS;
+
+        } catch (\Throwable $e) {
+            $this->log('Meilisearch index ERROR: ' . $e->getMessage());
+            $output->writeln('<error>' . $e->getMessage() . '</error>');
+
+            return Command::FAILURE;
+        }
+    }
+
+    /**
+     * Einheitliches Logging mit Zeitstempel
+     */
+    private function log(string $message): void
+    {
+        error_log(sprintf(
+            '[%s] %s',
+            date('Y-m-d H:i:s'),
+            $message
+        ));
     }
 }

src/Cron/MeilisearchIndexCron.php

@@ -1,22 +0,0 @@
-<?php
-namespace MummertMedia\ContaoMeilisearchBundle\Cron;
-
-use Contao\CoreBundle\Framework\ContaoFramework;
-use MummertMedia\ContaoMeilisearchBundle\Service\MeilisearchIndexService;
-
-class MeilisearchIndexCron
-{
-    public function __construct(
-        private readonly MeilisearchIndexService $indexService,
-        private readonly ContaoFramework $framework,
-    ) {}
-
-    public function __invoke(): void
-    {
-        // Contao initialisieren (wichtig!)
-        $this->framework->initialize();
-
-        // einmal täglich indexieren
-        $this->indexService->run();
-    }
-}

src/EventListener/IndexPageListener.php

@@ -3,15 +3,13 @@
 namespace MummertMedia\ContaoMeilisearchBundle\EventListener;
 
 use Contao\Config;
-use MummertMedia\ContaoMeilisearchBundle\Service\PdfIndexService;
+use Contao\System;
-use MummertMedia\ContaoMeilisearchBundle\Service\OfficeIndexService;
 
 class IndexPageListener
 {
-    public function __construct(
+    public function __construct()
-        private readonly PdfIndexService $pdfIndexService,
+    {
-        private readonly OfficeIndexService $officeIndexService,
+    }
-    ) {}
 
     private function debug(string $message, array $context = []): void
     {
@@ -30,22 +28,6 @@ class IndexPageListener
             'set_keys'  => array_keys($set),
         ]);
 
-        /*
-         * =====================
-         * PDF: Reset genau 1× pro Crawl
-         * =====================
-         */
-        try {
-            $this->debug('PDF resetTableOnce(): call');
-            $this->pdfIndexService->resetTableOnce();
-            $this->debug('PDF resetTableOnce(): ok');
-        } catch (\Throwable $e) {
-            $this->debug('PDF resetTableOnce(): failed', [
-                'error' => $e->getMessage(),
-                'class' => $e::class,
-            ]);
-        }
-
         /*
          * =====================
          * SEITEN-METADATEN
@@ -119,7 +101,6 @@ class IndexPageListener
                 $this->debug('Meta: searchimage candidate', ['searchimage' => $searchImage]);
 
                 if (!empty($searchImage)) {
-                    // >>> HINWEIS: falls dein tl_search-Feld "image" heißt, hier auf $set['image'] ändern!
                     $set['imagepath'] = trim((string) $searchImage);
                 }
 
@@ -155,20 +136,12 @@ class IndexPageListener
                         'class' => $e::class,
                     ]);
                 }
-
-                $this->debug('Meta: final set snapshot', [
-                    'priority'  => $set['priority'] ?? null,
-                    'keywords'  => $set['keywords'] ?? null,
-                    'imagepath' => $set['imagepath'] ?? null,
-                    'startDate' => $set['startDate'] ?? null,
-                    'checksum'  => $set['checksum'] ?? null,
-                ]);
             }
         }
 
         /*
          * =====================
-         * DATEI-INDEXIERUNG (PDF / OFFICE)
+         * DATEI-ERKENNUNG + UPSERT
          * =====================
          */
         if ((int) ($data['protected'] ?? 0) !== 0) {
@@ -176,15 +149,13 @@ class IndexPageListener
             return;
         }
 
-        $indexPdfs   = (bool) Config::get('meilisearch_index_pdfs');
+        $indexFiles = (bool) Config::get('meilisearch_index_files');
-        $indexOffice = (bool) Config::get('meilisearch_index_office');
 
-        $this->debug('File indexing settings', [
+        $this->debug('File indexing setting', [
-            'meilisearch_index_pdfs'   => $indexPdfs,
+            'meilisearch_index_files' => $indexFiles,
-            'meilisearch_index_office' => $indexOffice,
         ]);
 
-        if (!$indexPdfs && !$indexOffice) {
+        if (!$indexFiles) {
             $this->debug('Abort: file indexing disabled');
             return;
         }
@@ -192,61 +163,85 @@ class IndexPageListener
         $links = $this->findAllLinks($content);
         $this->debug('Links found', ['count' => count($links)]);
 
-        $pdfLinks    = [];
+        $fileLinks = [];
-        $officeLinks = [];
 
         foreach ($links as $link) {
             $type = $this->detectIndexableFileType($link['url']);
-
+            if ($type !== null) {
-            if ($type === 'pdf' && $indexPdfs) {
+                $fileLinks[] = $link + ['type' => $type];
-                $pdfLinks[] = $link;
-                continue;
-            }
-
-            if (in_array($type, ['docx', 'xlsx', 'pptx'], true) && $indexOffice) {
-                $officeLinks[] = $link;
             }
         }
 
-        $this->debug('Indexable file links', [
+        $this->debug('Indexable file links found', [
-            'pdf'    => count($pdfLinks),
+            'count' => count($fileLinks),
-            'office' => count($officeLinks),
+            'types' => array_count_values(array_column($fileLinks, 'type')),
         ]);
 
-        try {
+        if ($fileLinks) {
-            if ($pdfLinks !== []) {
+            $db   = System::getContainer()->get('database_connection');
-                $this->debug('PDF handlePdfLinks(): call', ['count' => count($pdfLinks)]);
+            $time = time();
-                $this->pdfIndexService->handlePdfLinks($pdfLinks);
-                $this->debug('PDF handlePdfLinks(): ok');
-            }
 
-            if ($officeLinks !== []) {
+            foreach ($fileLinks as $file) {
-                $this->debug('Office handleOfficeLinks(): call', ['count' => count($officeLinks)]);
+                $url = strtok($file['url'], '#');
-                $this->officeIndexService->handleOfficeLinks($officeLinks);
+
-                $this->debug('Office handleOfficeLinks(): ok');
+                $path = parse_url($url, PHP_URL_PATH);
+                $abs  = $path ? TL_ROOT . '/' . ltrim($path, '/') : null;
+
+                $mtime = ($abs && is_file($abs)) ? filemtime($abs) : 0;
+                $checksum = md5($url . '|' . $mtime);
+
+                $existing = $db->fetchAssociative(
+                    'SELECT id, checksum FROM tl_search_files WHERE url = ?',
+                    [$url]
+                );
+
+                if ($existing) {
+                    $db->update(
+                        'tl_search_files',
+                        [
+                            'tstamp'      => $time,
+                            'last_seen'   => $time,
+                            'page_id'     => (int) ($data['pid'] ?? 0),
+                            'file_mtime'  => $mtime,
+                            'checksum'    => $checksum,
+                        ],
+                        ['id' => $existing['id']]
+                    );
+
+                    $this->debug('File updated', [
+                        'url'      => $url,
+                        'checksum' => $checksum,
+                    ]);
+                } else {
+                    $db->insert(
+                        'tl_search_files',
+                        [
+                            'tstamp'     => $time,
+                            'last_seen'  => $time,
+                            'type'       => $file['type'],
+                            'url'        => $url,
+                            'title'      => $file['linkText'] ?? basename($url),
+                            'page_id'    => (int) ($data['pid'] ?? 0),
+                            'file_mtime' => $mtime,
+                            'checksum'   => $checksum,
+                        ]
+                    );
+
+                    $this->debug('File inserted', [
+                        'url'      => $url,
+                        'checksum' => $checksum,
+                    ]);
+                }
             }
-        } catch (\Throwable $e) {
-            $this->debug('File indexing failed', [
-                'error' => $e->getMessage(),
-                'class' => $e::class,
-            ]);
         }
 
         $this->debug('Hook end', [
             'final_set_keys' => array_keys($set),
-            'final_set'      => [
-                'priority'  => $set['priority'] ?? null,
-                'keywords'  => $set['keywords'] ?? null,
-                'imagepath' => $set['imagepath'] ?? null,
-                'startDate' => $set['startDate'] ?? null,
-                'checksum'  => $set['checksum'] ?? null,
-            ],
         ]);
     }
 
-    /**
+    /* === Hilfsmethoden unverändert === */
-     * Extrahiert MEILISEARCH_JSON aus HTML-Kommentar
+
-     */
     private function extractMeilisearchJson(string $content): ?array
     {
         if (!preg_match('/<!--\s*MEILISEARCH_JSON\s*(\{.*?\})\s*-->/s', $content, $m)) {
@@ -261,9 +256,6 @@ class IndexPageListener
             : null;
     }
 
-    /**
-     * Sammle alle <a href="…"> Links
-     */
     private function findAllLinks(string $content): array
     {
         if (!preg_match_all(
@@ -286,12 +278,8 @@ class IndexPageListener
         return $result;
     }
 
-    /**
-     * Ermittelt indexierbaren Dateityp (pdf|docx|xlsx|pptx) oder null
-     */
     private function detectIndexableFileType(string $url): ?string
     {
-        // Hash entfernen
         $url = strtok($url, '#');
 
         $parts = parse_url($url);
@@ -299,7 +287,6 @@ class IndexPageListener
             return null;
         }
 
-        // direkter Pfad (/files/…)
         if (!empty($parts['path'])) {
             $ext = strtolower(pathinfo($parts['path'], PATHINFO_EXTENSION));
             if (in_array($ext, ['pdf', 'docx', 'xlsx', 'pptx'], true)) {
@@ -307,18 +294,12 @@ class IndexPageListener
             }
         }
 
-        // Query-Parameter (Contao 4 + 5)
         if (!empty($parts['query'])) {
             parse_str($parts['query'], $query);
 
             foreach (['file', 'p', 'f'] as $param) {
                 if (!empty($query[$param])) {
-                    $candidate = (string) $query[$param];
+                    $candidate = rawurldecode(html_entity_decode((string) $query[$param], ENT_QUOTES));
-
-                    // sicher decodieren (Contao 4 + 5)
-                    $candidate = html_entity_decode($candidate, ENT_QUOTES);
-                    $candidate = rawurldecode($candidate);
-
                     $ext = strtolower(pathinfo($candidate, PATHINFO_EXTENSION));
 
                     if (in_array($ext, ['pdf', 'docx', 'xlsx', 'pptx'], true)) {

src/EventListener/MeilisearchPageMarkerListener.php

@@ -16,6 +16,13 @@ class MeilisearchPageMarkerListener
             return $buffer;
         }
 
+        // ⛔ Marker bereits vorhanden → nichts mehr tun
+        if (str_contains($buffer, '⟦MEILISEARCH_META⟧')
+            || str_contains($buffer, 'MEILISEARCH_JSON')
+        ) {
+            return $buffer;
+        }
+
         $data = [];
 
         /*

src/Resources/config/services.yaml

@@ -3,22 +3,22 @@ services:
   Psr\Container\ContainerInterface: '@service_container'
 
   MummertMedia\ContaoMeilisearchBundle\:
-    resource: '../../{Command,Cron,EventListener,Service}'
+    resource: '../../{Command,EventListener,Service}'
     autowire: true
     autoconfigure: true
 
+  MummertMedia\ContaoMeilisearchBundle\EventListener\MeilisearchPageMarkerListener:
+    autowire: true
+    autoconfigure: false
+    tags:
+      - { name: contao.hook, hook: outputFrontendTemplate, method: onOutputFrontendTemplate }
+
   MummertMedia\ContaoMeilisearchBundle\EventListener\IndexPageListener:
     autowire: true
     autoconfigure: false
     tags:
       - { name: contao.hook, hook: indexPage, method: onIndexPage }
 
-  MummertMedia\ContaoMeilisearchBundle\Cron\MeilisearchIndexCron:
-    autowire: true
-    autoconfigure: false
-    tags:
-      - { name: contao.cron, interval: daily, method: __invoke }
-
   MummertMedia\ContaoMeilisearchBundle\Controller\FrontendModule\MeilisearchSearchController:
     autowire: true
     autoconfigure: false

src/Resources/contao/dca/tl_search_files.php

@@ -2,16 +2,17 @@
 
 use Contao\DC_Table;
 
-$GLOBALS['TL_DCA']['tl_search_pdf'] = [
+$GLOBALS['TL_DCA']['tl_search_files'] = [
     'config' => [
         'dataContainer' => DC_Table::class,
         'sql' => [
             'keys' => [
-                'id'       => 'primary',
+                'id'        => 'primary',
-                'checksum' => 'unique',
+                'page_id'   => 'index',
-                'page_id'  => 'index',
+                'url'       => 'unique',
-                'url'      => 'index',
+                'type'      => 'index',
-                'type'     => 'index', // ⬅️ NEU
+                'checksum'  => 'index',
+                'last_seen' => 'index', // ⬅️ NEU (für Cleanup-Performance)
             ],
         ],
     ],
@@ -25,10 +26,18 @@ $GLOBALS['TL_DCA']['tl_search_pdf'] = [
             'sql' => "int(10) unsigned NOT NULL default 0",
         ],
 
+        /*
+         * Zeitpunkt, wann die Datei zuletzt beim Crawl gesehen wurde
+         * → Basis für Cleanup
+         */
+        'last_seen' => [ // ⬅️ NEU
+            'sql' => "int(10) unsigned NOT NULL default 0",
+        ],
+
         /*
          * Dateityp: pdf | docx | xlsx | pptx
          */
-        'type' => [ // ⬅️ NEU
+        'type' => [
             'sql' => "varchar(16) NOT NULL default 'pdf'",
         ],
 
@@ -64,7 +73,7 @@ $GLOBALS['TL_DCA']['tl_search_pdf'] = [
 
         /*
          * Herkunftsseite (tl_page.id)
-         * → Cleanup / Referenz
+         * → optional, Debug / Referenz
          */
         'page_id' => [
             'sql' => "int(10) unsigned NOT NULL default 0",

src/Resources/contao/dca/tl_settings.php

@@ -4,14 +4,17 @@ use Contao\CoreBundle\DataContainer\PaletteManipulator;
 use Contao\System;
 
 /**
+ * -------------------------------------------------
  * Fields
+ * -------------------------------------------------
  */
+
 $GLOBALS['TL_DCA']['tl_settings']['fields']['meilisearch_host'] = [
     'inputType' => 'text',
     'eval' => [
         'mandatory' => true,
-        'rgxp' => 'url',
+        'rgxp'      => 'url',
-        'tl_class' => 'w50',
+        'tl_class'  => 'w50',
     ],
 ];
 
@@ -19,7 +22,7 @@ $GLOBALS['TL_DCA']['tl_settings']['fields']['meilisearch_index'] = [
     'inputType' => 'text',
     'eval' => [
         'mandatory' => true,
-        'tl_class' => 'w50',
+        'tl_class'  => 'w50',
     ],
 ];
 
@@ -27,7 +30,7 @@ $GLOBALS['TL_DCA']['tl_settings']['fields']['meilisearch_api_write'] = [
     'inputType' => 'text',
     'eval' => [
         'mandatory' => true,
-        'tl_class' => 'w50',
+        'tl_class'  => 'w50',
         'hideInput' => true,
     ],
 ];
@@ -36,7 +39,7 @@ $GLOBALS['TL_DCA']['tl_settings']['fields']['meilisearch_api_search'] = [
     'inputType' => 'text',
     'eval' => [
         'mandatory' => true,
-        'tl_class' => 'w50',
+        'tl_class'  => 'w50',
         'hideInput' => true,
     ],
 ];
@@ -55,50 +58,71 @@ $GLOBALS['TL_DCA']['tl_settings']['fields']['meilisearch_imagesize'] = [
         return $options;
     },
     'eval' => [
-        'tl_class' => 'w50',
+        'tl_class'           => 'w50',
-        'chosen' => true,
+        'chosen'             => true,
         'includeBlankOption' => true,
     ],
-    // 🔥 DAS HAT GEFEHLT
     'sql' => "int(10) unsigned NOT NULL default 0",
 ];
 
-$GLOBALS['TL_DCA']['tl_settings']['fields']['meilisearch_index_past_events'] = [
-    'inputType' => 'checkbox',
-    'eval'      => [
-        'tl_class' => 'w50 clr',
-    ],
-];
-
 $GLOBALS['TL_DCA']['tl_settings']['fields']['meilisearch_fallback_image'] = [
     'inputType' => 'fileTree',
     'eval' => [
         'filesOnly' => true,
         'fieldType' => 'radio',
-        'tl_class' => 'w50',
+        'tl_class'  => 'w50',
     ],
     'sql' => "varbinary(16) NULL",
 ];
 
-$GLOBALS['TL_DCA']['tl_settings']['fields']['meilisearch_index_pdfs'] = [
+$GLOBALS['TL_DCA']['tl_settings']['fields']['meilisearch_index_past_events'] = [
-    'label'     => &$GLOBALS['TL_LANG']['tl_settings']['meilisearch_index_pdfs'],
     'inputType' => 'checkbox',
-    'eval'      => [
+    'eval' => [
-        'tl_class' => 'w50',
+        'tl_class' => 'w50 clr',
     ],
-    'sql'       => "char(1) NOT NULL default '1'",
-];
-
-$GLOBALS['TL_DCA']['tl_settings']['fields']['meilisearch_index_office'] = [
-    'label'     => &$GLOBALS['TL_LANG']['tl_settings']['meilisearch_index_office'],
-    'inputType' => 'checkbox',
-    'eval'      => ['tl_class' => 'w50'],
-    'sql'       => "char(1) NOT NULL default '0'",
 ];
 
 /**
- * Palette
+ * -------------------------------------------------
+ * Datei-Indexierung (Tika)
+ * -------------------------------------------------
  */
+
+$GLOBALS['TL_DCA']['tl_settings']['fields']['meilisearch_index_files'] = [
+    'inputType' => 'checkbox',
+    'eval' => [
+        'tl_class'       => 'w50',
+        'submitOnChange' => true,
+    ],
+    'sql' => "char(1) NOT NULL default '0'",
+];
+
+$GLOBALS['TL_DCA']['tl_settings']['fields']['meilisearch_tika_url'] = [
+    'inputType' => 'text',
+    'eval' => [
+        'rgxp'      => 'url',
+        'mandatory' => true,
+        'tl_class'  => 'w50 clr',
+    ],
+];
+
+/**
+ * -------------------------------------------------
+ * Selector / Subpalette
+ * -------------------------------------------------
+ */
+
+$GLOBALS['TL_DCA']['tl_settings']['palettes']['__selector__'][] = 'meilisearch_index_files';
+
+$GLOBALS['TL_DCA']['tl_settings']['subpalettes']['meilisearch_index_files']
+    = 'meilisearch_tika_url';
+
+/**
+ * -------------------------------------------------
+ * Palette
+ * -------------------------------------------------
+ */
+
 PaletteManipulator::create()
     ->addLegend('meilisearch_legend', null, PaletteManipulator::POSITION_AFTER, true)
     ->addField('meilisearch_host', 'meilisearch_legend')
@@ -108,6 +132,5 @@ PaletteManipulator::create()
     ->addField('meilisearch_imagesize', 'meilisearch_legend')
     ->addField('meilisearch_fallback_image', 'meilisearch_legend')
     ->addField('meilisearch_index_past_events', 'meilisearch_legend')
-    ->addField('meilisearch_index_pdfs', 'meilisearch_legend')
+    ->addField('meilisearch_index_files', 'meilisearch_legend')
-    ->addField('meilisearch_index_office', 'meilisearch_legend')
     ->applyToPalette('default', 'tl_settings');

src/Resources/contao/languages/de/tl_settings.php

@@ -28,10 +28,10 @@ $GLOBALS['TL_LANG']['tl_settings']['meilisearch_index_past_events'][0]
 $GLOBALS['TL_LANG']['tl_settings']['meilisearch_index_past_events'][1]
     = 'Vergangene Kalender-Events werden ebenfalls in Meilisearch indexiert.';
 
-$GLOBALS['TL_LANG']['tl_settings']['meilisearch_index_pdfs'] = [
+$GLOBALS['TL_LANG']['tl_settings']['meilisearch_index_files'] = [
-    'PDFs indexieren',
+    'Dateien indexieren',
-    'Aktiviert die Indexierung von PDF-Dateien für die Suche.',
+    'Aktiviert die Indexierung von PDF-Dateien sowie DOCX, XLSX und PPTX.',
 ];
 
-$GLOBALS['TL_LANG']['tl_settings']['meilisearch_index_office']
+$GLOBALS['TL_LANG']['tl_settings']['meilisearch_tika_url']
-    = ['Office-Dateien indexieren', 'DOCX, XLSX und PPTX in die Suche aufnehmen.'];
+    = ['Apache Tika URL', 'URL der Apache Tika Instanz (z. B. https://tika.domain.tld).'];

src/Resources/public/icons/crawl_search-index_log.csv

@@ -1,3 +0,0 @@
-Time,Source,URI,"Found on URI","Found on level",Tags,Message
-"2025-12-28 10:44:12.357228","Contao\CoreBundle\Crawl\Escargot\Subscriber\SearchIndexSubscriber",https://flowerpowerabi.de/,https://flowerpowerabi.de/sitemap.xml,3,,"Forwarded to the search indexer. Was indexed successfully."
-"2025-12-28 10:44:12.413689","Contao\CoreBundle\Crawl\Escargot\Subscriber\SearchIndexSubscriber",https://flowerpowerabi.de/testseite-fuer-meilisearch,https://flowerpowerabi.de/sitemap.xml,3,,"Forwarded to the search indexer. Was indexed successfully."

src/Service/MeilisearchIndexService.php

@@ -41,8 +41,8 @@ class MeilisearchIndexService
             return;
         }
 
-        $host = (string) Config::get('meilisearch_host');
+        $host      = (string) Config::get('meilisearch_host');
-        $apiKey = (string) Config::get('meilisearch_api_write');
+        $apiKey    = (string) Config::get('meilisearch_api_write');
         $this->indexName = (string) Config::get('meilisearch_index');
 
         if ($host === '' || $this->indexName === '') {
@@ -72,7 +72,7 @@ class MeilisearchIndexService
         }
 
         $this->indexTlSearch($index);
-        $this->indexTlSearchPdf($index);
+        $this->indexTlSearchFiles($index);
     }
 
     private function ensureIndexSettings(Indexes $index): void
@@ -80,6 +80,7 @@ class MeilisearchIndexService
         $index->updateSettings([
             'searchableAttributes' => ['title', 'keywords', 'text'],
             'sortableAttributes'   => ['priority', 'startDate'],
+            'filterableAttributes' => ['type', 'filetype'],
         ]);
     }
 
@@ -132,7 +133,7 @@ class MeilisearchIndexService
     }
 
     /**
-     * tl_search indexieren
+     * tl_search indexieren (Seiten / News / Events)
      */
     private function indexTlSearch(Indexes $index): void
     {
@@ -164,13 +165,11 @@ class MeilisearchIndexService
                     }
                 }
 
-                $cleanText = $this->stripMeilisearchMeta((string) $row['text']);
-
                 $doc = [
                     'id'        => $type . '_' . $row['id'],
                     'type'      => $type,
                     'title'     => $row['title'],
-                    'text'      => $cleanText,
+                    'text'      => $this->stripMeilisearchMeta((string) $row['text']),
                     'url'       => $row['url'],
                     'protected' => (bool) $row['protected'],
                     'checksum'  => $row['checksum'],
@@ -192,31 +191,24 @@ class MeilisearchIndexService
                 $documents[] = $doc;
 
             } catch (\Throwable $e) {
-                error_log(
+                error_log('[ContaoMeilisearch] Failed to build tl_search document: ' . $e->getMessage());
-                    '[ContaoMeilisearch] Failed to build document for tl_search ID '
-                    . ($row['id'] ?? '?') . ': ' . $e->getMessage()
-                );
             }
         }
 
         if ($documents !== []) {
-            try {
+            $index->addDocuments($documents);
-                $index->addDocuments($documents);
-            } catch (\Throwable $e) {
-                error_log('[ContaoMeilisearch] Failed to add tl_search documents: ' . $e->getMessage());
-            }
         }
     }
 
     /**
-     * tl_search_pdf indexieren
+     * tl_search_files indexieren (PDF / Office)
      */
-    private function indexTlSearchPdf(Indexes $index): void
+    private function indexTlSearchFiles(Indexes $index): void
     {
         try {
-            $rows = $this->connection->fetchAllAssociative('SELECT * FROM tl_search_pdf');
+            $rows = $this->connection->fetchAllAssociative('SELECT * FROM tl_search_files');
         } catch (\Throwable $e) {
-            error_log('[ContaoMeilisearch] Failed to read tl_search_pdf: ' . $e->getMessage());
+            error_log('[ContaoMeilisearch] Failed to read tl_search_files: ' . $e->getMessage());
             return;
         }
 
@@ -233,30 +225,24 @@ class MeilisearchIndexService
                     : 'pdf';
 
                 $documents[] = [
-                    'id'       => $fileType . '_' . $row['id'],
+                    'id'        => 'file_' . $row['id'],
-                    'type'     => $fileType,
+                    'type'      => 'file',
-                    'title'    => $row['title'],
+                    'filetype'  => $fileType,
-                    'text'     => $this->stripMeilisearchMeta((string) $row['text']),
+                    'title'     => $row['title'] ?: basename($row['url']),
-                    'url'      => $row['url'],
+                    'text'      => (string) $row['text'],
-                    'checksum' => $row['checksum'],
+                    'url'       => $row['url'],
-                    'poster'   => self::FILETYPE_ICON_MAP[$fileType]
+                    'checksum'  => $row['checksum'],
+                    'poster'    => self::FILETYPE_ICON_MAP[$fileType]
                         ?? self::FILETYPE_ICON_MAP['pdf'],
                 ];
 
             } catch (\Throwable $e) {
-                error_log(
+                error_log('[ContaoMeilisearch] Failed to build file document: ' . $e->getMessage());
-                    '[ContaoMeilisearch] Failed to build PDF document for ID '
-                    . ($row['id'] ?? '?') . ': ' . $e->getMessage()
-                );
             }
         }
 
         if ($documents !== []) {
-            try {
+            $index->addDocuments($documents);
-                $index->addDocuments($documents);
-            } catch (\Throwable $e) {
-                error_log('[ContaoMeilisearch] Failed to add tl_search_pdf documents: ' . $e->getMessage());
-            }
         }
     }
 

src/Service/OfficeIndexService.php

@@ -1,273 +0,0 @@
-<?php
-
-namespace MummertMedia\ContaoMeilisearchBundle\Service;
-
-use Contao\Database;
-use PhpOffice\PhpWord\IOFactory as WordIOFactory;
-use PhpOffice\PhpSpreadsheet\IOFactory as SpreadsheetIOFactory;
-use PhpOffice\PhpPresentation\IOFactory as PresentationIOFactory;
-use Symfony\Component\DependencyInjection\ParameterBag\ParameterBagInterface;
-
-class OfficeIndexService
-{
-    private string $projectDir;
-
-    // pro Crawl-Durchlauf: doppelte Verarbeitung vermeiden
-    private array $seenThisCrawl = [];
-
-    public function __construct(ParameterBagInterface $params)
-    {
-        $this->projectDir = rtrim((string) $params->get('kernel.project_dir'), '/');
-    }
-
-    /**
-     * @param array<int,array{url:string,linkText:?string}> $officeLinks
-     */
-    public function handleOfficeLinks(array $officeLinks): void
-    {
-        foreach ($officeLinks as $row) {
-            $url = (string) ($row['url'] ?? '');
-            $linkText = $row['linkText'] ?? null;
-
-            if ($url === '') {
-                continue;
-            }
-
-            try {
-                // innerhalb des Crawls gleiche URL nicht mehrfach parsen
-                $seenKey = md5($url);
-                if (isset($this->seenThisCrawl[$seenKey])) {
-                    continue;
-                }
-                $this->seenThisCrawl[$seenKey] = true;
-
-                $normalized = $this->normalizeOfficeUrl($url);
-                if ($normalized === null) {
-                    continue;
-                }
-
-                [$relativePath, $type] = $normalized;
-
-                $absolutePath = $this->getAbsolutePath($relativePath);
-                if (!is_file($absolutePath)) {
-                    continue;
-                }
-
-                $mtime = (int) (filemtime($absolutePath) ?: 0);
-                $checksum = md5($relativePath . '|' . $mtime);
-
-                $title = $linkText ?: basename($absolutePath);
-
-                $text = $this->parseOfficeFile($absolutePath, $type);
-                if ($text === '') {
-                    continue;
-                }
-
-                $this->upsertOffice(
-                    $relativePath,
-                    $title,
-                    $text,
-                    $checksum,
-                    $mtime,
-                    $type
-                );
-
-            } catch (\Throwable $e) {
-                error_log(
-                    '[ContaoMeilisearch] Office indexing failed for "' . $url . '": ' . $e->getMessage()
-                );
-            }
-        }
-    }
-
-    /**
-     * @return array{string,string}|null [relativePath, type]
-     */
-    private function normalizeOfficeUrl(string $url): ?array
-    {
-        $decoded = html_entity_decode($url);
-        $parts = parse_url($decoded);
-
-        // 1) files/... (ohne führenden Slash)
-        if (!empty($parts['path']) && str_starts_with($parts['path'], 'files/')) {
-            $ext = strtolower(pathinfo($parts['path'], PATHINFO_EXTENSION));
-            if (in_array($ext, ['docx', 'xlsx', 'pptx'], true)) {
-                return ['/' . $parts['path'], $ext];
-            }
-        }
-
-        // 2) /files/...
-        if (!empty($parts['path']) && str_starts_with($parts['path'], '/files/')) {
-            $ext = strtolower(pathinfo($parts['path'], PATHINFO_EXTENSION));
-            if (in_array($ext, ['docx', 'xlsx', 'pptx'], true)) {
-                return [$parts['path'], $ext];
-            }
-        }
-
-        if (empty($parts['query'])) {
-            return null;
-        }
-
-        parse_str($parts['query'], $query);
-
-        // 3) Contao 4: ?file=files/...
-        if (!empty($query['file'])) {
-            $file = urldecode((string) $query['file']);
-            $file = ltrim($file, '/');
-            $ext = strtolower(pathinfo($file, PATHINFO_EXTENSION));
-
-            if (
-                str_starts_with($file, 'files/')
-                && in_array($ext, ['docx', 'xlsx', 'pptx'], true)
-            ) {
-                return ['/' . $file, $ext];
-            }
-        }
-
-        // 4) Contao 5: ?p=...
-        if (!empty($query['p'])) {
-            $p = urldecode((string) $query['p']);
-            $ext = strtolower(pathinfo($p, PATHINFO_EXTENSION));
-
-            if (in_array($ext, ['docx', 'xlsx', 'pptx'], true)) {
-                return ['/files/' . ltrim($p, '/'), $ext];
-            }
-        }
-
-        return null;
-    }
-
-    private function getAbsolutePath(string $relativePath): string
-    {
-        return $this->projectDir . '/' . ltrim($relativePath, '/');
-    }
-
-    private function upsertOffice(
-        string $url,
-        string $title,
-        string $text,
-        string $checksum,
-        int $mtime,
-        string $type
-    ): void {
-        try {
-            Database::getInstance()
-                ->prepare('
-                    INSERT INTO tl_search_pdf
-                        (tstamp, type, url, title, text, checksum, file_mtime)
-                    VALUES
-                        (?, ?, ?, ?, ?, ?, ?)
-                    ON DUPLICATE KEY UPDATE
-                        tstamp=VALUES(tstamp),
-                        type=VALUES(type),
-                        url=VALUES(url),
-                        title=VALUES(title),
-                        text=VALUES(text),
-                        file_mtime=VALUES(file_mtime)
-                ')
-                ->execute(
-                    time(),
-                    $type,
-                    $url,
-                    $title,
-                    $text,
-                    $checksum,
-                    $mtime
-                );
-        } catch (\Throwable $e) {
-            error_log(
-                '[ContaoMeilisearch] Failed to write Office index entry (' . $url . '): ' . $e->getMessage()
-            );
-        }
-    }
-
-    private function parseOfficeFile(string $absolutePath, string $type): string
-    {
-        return match ($type) {
-            'docx' => $this->parseDocx($absolutePath),
-            'xlsx' => $this->parseXlsx($absolutePath),
-            'pptx' => $this->parsePptx($absolutePath),
-            default => '',
-        };
-    }
-
-    private function parseDocx(string $absolutePath): string
-    {
-        try {
-            $phpWord = WordIOFactory::load($absolutePath);
-            $text = '';
-
-            foreach ($phpWord->getSections() as $section) {
-                foreach ($section->getElements() as $element) {
-                    if (method_exists($element, 'getText')) {
-                        $text .= ' ' . $element->getText();
-                    }
-                }
-            }
-
-            return $this->cleanText($text);
-        } catch (\Throwable $e) {
-            error_log(
-                '[ContaoMeilisearch] Failed to parse DOCX "' . $absolutePath . '": ' . $e->getMessage()
-            );
-            return '';
-        }
-    }
-
-    private function parseXlsx(string $absolutePath): string
-    {
-        try {
-            $spreadsheet = SpreadsheetIOFactory::load($absolutePath);
-            $text = '';
-
-            foreach ($spreadsheet->getAllSheets() as $sheet) {
-                foreach ($sheet->toArray() as $row) {
-                    $text .= ' ' . implode(' ', array_filter($row, 'is_scalar'));
-                }
-            }
-
-            return $this->cleanText($text);
-        } catch (\Throwable $e) {
-            error_log(
-                '[ContaoMeilisearch] Failed to parse XLSX "' . $absolutePath . '": ' . $e->getMessage()
-            );
-            return '';
-        }
-    }
-
-    private function parsePptx(string $absolutePath): string
-    {
-        try {
-            $presentation = PresentationIOFactory::load($absolutePath);
-            $text = '';
-
-            foreach ($presentation->getAllSlides() as $slide) {
-                foreach ($slide->getShapeCollection() as $shape) {
-                    if (method_exists($shape, 'getPlainText')) {
-                        $text .= ' ' . $shape->getPlainText();
-                    }
-                }
-            }
-
-            return $this->cleanText($text);
-        } catch (\Throwable $e) {
-            error_log(
-                '[ContaoMeilisearch] Failed to parse PPTX "' . $absolutePath . '": ' . $e->getMessage()
-            );
-            return '';
-        }
-    }
-
-    private function cleanText(string $text): string
-    {
-        if (class_exists(\Normalizer::class)) {
-            $text = \Normalizer::normalize($text, \Normalizer::FORM_C) ?? $text;
-        }
-
-        $text = str_replace(["\r\n", "\r"], "\n", $text);
-        $text = preg_replace('/[^\p{L}\p{N}\p{P}\p{Z}\n]/u', ' ', $text);
-        $text = preg_replace('/\s+/u', ' ', $text);
-
-        return trim(mb_substr($text, 0, 20000));
-    }
-}

src/Service/PdfIndexService.php

@@ -1,224 +0,0 @@
-<?php
-
-namespace MummertMedia\ContaoMeilisearchBundle\Service;
-
-use Contao\Database;
-use Smalot\PdfParser\Parser;
-use Symfony\Component\DependencyInjection\ParameterBag\ParameterBagInterface;
-
-class PdfIndexService
-{
-    private string $projectDir;
-
-    private bool $didReset = false;
-    private array $seenThisCrawl = [];
-
-    public function __construct(ParameterBagInterface $params)
-    {
-        $this->projectDir = rtrim((string) $params->get('kernel.project_dir'), '/');
-    }
-
-    /**
-     * Wird aus dem Listener beim ersten Hook-Call pro Crawl aufgerufen.
-     */
-    public function resetTableOnce(): void
-    {
-        if ($this->didReset) {
-            return;
-        }
-
-        $this->didReset = true;
-        $this->seenThisCrawl = [];
-
-        Database::getInstance()->execute('TRUNCATE tl_search_pdf');
-    }
-
-    /**
-     * @param array<int,array{url:string,linkText:?string}> $pdfLinks
-     */
-    public function handlePdfLinks(array $pdfLinks): void
-    {
-        foreach ($pdfLinks as $row) {
-            $url = (string) ($row['url'] ?? '');
-            $linkText = $row['linkText'] ?? null;
-
-            if ($url === '') {
-                continue;
-            }
-
-            // innerhalb eines Crawls doppelte URLs vermeiden
-            $seenKey = md5($url);
-            if (isset($this->seenThisCrawl[$seenKey])) {
-                continue;
-            }
-            $this->seenThisCrawl[$seenKey] = true;
-
-            $normalizedPath = $this->normalizePdfUrl($url);
-            if ($normalizedPath === null) {
-                continue;
-            }
-
-            $absolutePath = $this->getAbsolutePath($normalizedPath);
-            if (!is_file($absolutePath)) {
-                continue;
-            }
-
-            $mtime = (int) (filemtime($absolutePath) ?: 0);
-            $checksum = md5($normalizedPath . '|' . $mtime);
-
-            // Titel-Priorität:
-            // 1) Linktext
-            // 2) PDF-Metadaten
-            // 3) Dateiname
-            $pdfMetaTitle = $this->readPdfMetaTitle($absolutePath);
-            $title = $linkText ?: ($pdfMetaTitle ?: basename($absolutePath));
-
-            $text = $this->parsePdf($absolutePath);
-            if ($text === '') {
-                continue;
-            }
-
-            $this->upsertPdf(
-                $normalizedPath,
-                $title,
-                $text,
-                $checksum,
-                $mtime
-            );
-        }
-    }
-
-    private function normalizePdfUrl(string $url): ?string
-    {
-        $decoded = html_entity_decode($url);
-        $parts = parse_url($decoded);
-
-        // 1) files/...pdf (ohne führenden Slash)
-        if (
-            !empty($parts['path'])
-            && str_starts_with($parts['path'], 'files/')
-            && str_ends_with(strtolower($parts['path']), '.pdf')
-        ) {
-            return '/' . $parts['path'];
-        }
-
-        // 2) /files/...pdf
-        if (
-            !empty($parts['path'])
-            && str_starts_with($parts['path'], '/files/')
-            && str_ends_with(strtolower($parts['path']), '.pdf')
-        ) {
-            return $parts['path'];
-        }
-
-        if (empty($parts['query'])) {
-            return null;
-        }
-
-        parse_str($parts['query'], $query);
-
-        // 3) Contao 4: ?file=files/...
-        if (!empty($query['file'])) {
-            $file = urldecode((string) $query['file']);
-            $file = ltrim($file, '/');
-
-            if (
-                str_starts_with($file, 'files/')
-                && str_ends_with(strtolower($file), '.pdf')
-            ) {
-                return '/' . $file;
-            }
-        }
-
-        // 4) Contao 5: ?p=...
-        if (!empty($query['p'])) {
-            $p = urldecode((string) $query['p']);
-            return '/files/' . ltrim($p, '/');
-        }
-
-        return null;
-    }
-
-    private function getAbsolutePath(string $relativePath): string
-    {
-        return $this->projectDir . '/' . ltrim($relativePath, '/');
-    }
-
-    private function upsertPdf(
-        string $url,
-        string $title,
-        string $text,
-        string $checksum,
-        int $mtime
-    ): void {
-        Database::getInstance()
-            ->prepare('
-                INSERT INTO tl_search_pdf
-                    (tstamp, url, title, text, checksum, file_mtime)
-                VALUES
-                    (?, ?, ?, ?, ?, ?)
-                ON DUPLICATE KEY UPDATE
-                    tstamp=VALUES(tstamp),
-                    url=VALUES(url),
-                    title=VALUES(title),
-                    text=VALUES(text),
-                    file_mtime=VALUES(file_mtime)
-            ')
-            ->execute(
-                time(),
-                $url,
-                $title,
-                $text,
-                $checksum,
-                $mtime
-            );
-    }
-
-    private function parsePdf(string $absolutePath): string
-    {
-        try {
-            $parser = new Parser();
-            $pdf = $parser->parseFile($absolutePath);
-            $text = $this->cleanPdfContent($pdf->getText());
-
-            return mb_substr($text, 0, 20000);
-        } catch (\Throwable) {
-            return '';
-        }
-    }
-
-    private function readPdfMetaTitle(string $absolutePath): ?string
-    {
-        try {
-            $parser = new Parser();
-            $pdf = $parser->parseFile($absolutePath);
-            $details = $pdf->getDetails();
-
-            foreach (['Title', 'title'] as $key) {
-                if (!empty($details[$key]) && is_string($details[$key])) {
-                    $t = trim($details[$key]);
-                    if ($t !== '') {
-                        return $t;
-                    }
-                }
-            }
-        } catch (\Throwable) {
-        }
-
-        return null;
-    }
-
-    private function cleanPdfContent(string $text): string
-    {
-        if (class_exists(\Normalizer::class)) {
-            $text = \Normalizer::normalize($text, \Normalizer::FORM_C) ?? $text;
-        }
-
-        $text = str_replace(["\r\n", "\r"], "\n", $text);
-        $text = preg_replace('/[^\p{L}\p{N}\p{P}\p{Z}\n]/u', ' ', $text);
-        $text = preg_replace('/(?<=\p{L})\s+(?=\p{L})/u', ' ', $text);
-        $text = preg_replace('/\s+/u', ' ', $text);
-
-        return trim($text);
-    }
-}