This commit is contained in:
Jürgen Mummert
2025-12-28 11:29:13 +01:00
parent e4c8473999
commit 9edb1e4713
2 changed files with 241 additions and 160 deletions
+112 -66
View File
@@ -15,18 +15,92 @@ class IndexPageListener
public function onIndexPage(string $content, array &$data, array &$set): void
{
fwrite(STDERR, "\n[Meili DEBUG] onIndexPage() called\n");
/*
* =====================
* PDF: Reset genau 1× pro Crawl
* =====================
*/
try {
fwrite(STDERR, "[Meili DEBUG] resetTableOnce()\n");
$this->pdfIndexService->resetTableOnce();
} catch (\Throwable $e) {
fwrite(STDERR, "[Meili DEBUG] PDF reset failed: {$e->getMessage()}\n");
error_log('[ContaoMeilisearch] PDF reset failed: ' . $e->getMessage());
}
/*
* =====================
* SEITEN-METADATEN
* =====================
*/
if (str_contains($content, 'MEILISEARCH_JSON')) {
try {
$parsed = $this->extractMeilisearchJson($content);
} catch (\Throwable $e) {
error_log('[ContaoMeilisearch] Failed to extract MEILISEARCH_JSON: ' . $e->getMessage());
$parsed = null;
}
if (is_array($parsed)) {
// PRIORITY
$priority =
$parsed['event']['priority']
?? $parsed['news']['priority']
?? $parsed['page']['priority']
?? null;
if ($priority !== null && $priority !== '') {
$set['priority'] = (int) $priority;
}
// KEYWORDS
$keywordSources = [
$parsed['event']['keywords'] ?? null,
$parsed['news']['keywords'] ?? null,
$parsed['page']['keywords'] ?? null,
];
$keywords = [];
foreach ($keywordSources as $src) {
if (!is_string($src) || trim($src) === '') {
continue;
}
foreach (preg_split('/\s+/', trim($src)) as $word) {
$keywords[] = $word;
}
}
if ($keywords) {
$set['keywords'] = implode(' ', array_unique($keywords));
}
// IMAGEPATH
if (!empty($parsed['page']['searchimage'])) {
$set['imagepath'] = trim((string) $parsed['page']['searchimage']);
}
// STARTDATE
$startDate =
$parsed['event']['startDate']
?? $parsed['news']['startDate']
?? null;
if (is_numeric($startDate) && (int) $startDate > 0) {
$set['startDate'] = (int) $startDate;
}
// CHECKSUM
try {
$checksumSeed = (string) ($data['checksum'] ?? '');
$checksumSeed .= '|' . ($set['keywords'] ?? '');
$checksumSeed .= '|' . ($set['priority'] ?? '');
$checksumSeed .= '|' . ($set['imagepath'] ?? '');
$checksumSeed .= '|' . ($set['startDate'] ?? '');
$set['checksum'] = md5($checksumSeed);
} catch (\Throwable $e) {
error_log('[ContaoMeilisearch] Failed to generate checksum: ' . $e->getMessage());
}
}
}
/*
@@ -35,91 +109,67 @@ class IndexPageListener
* =====================
*/
if ((int) ($data['protected'] ?? 0) !== 0) {
fwrite(STDERR, "[Meili DEBUG] Page is protected → skip files\n");
return;
}
$indexPdfs = (bool) Config::get('meilisearch_index_pdfs');
$indexOffice = (bool) Config::get('meilisearch_index_office');
fwrite(
STDERR,
"[Meili DEBUG] Settings: pdfs="
. ($indexPdfs ? '1' : '0')
. " office="
. ($indexOffice ? '1' : '0')
. "\n"
);
if (!$indexPdfs && !$indexOffice) {
fwrite(STDERR, "[Meili DEBUG] No file indexing enabled → return\n");
return;
}
$links = $this->findAllLinks($content);
fwrite(STDERR, "[Meili DEBUG] Found " . count($links) . " <a> links\n");
$pdfLinks = [];
$officeLinks = [];
foreach ($links as $link) {
fwrite(STDERR, "[Meili DEBUG] URL: {$link['url']}\n");
$type = $this->detectIndexableFileType($link['url']);
fwrite(
STDERR,
"[Meili DEBUG] → detected type: "
. ($type ?? 'none')
. "\n"
);
if ($type === 'pdf') {
if ($indexPdfs) {
fwrite(STDERR, "[Meili DEBUG] → add to PDF queue\n");
$pdfLinks[] = $link;
} else {
fwrite(STDERR, "[Meili DEBUG] → PDF indexing disabled\n");
}
if ($type === 'pdf' && $indexPdfs) {
$pdfLinks[] = $link;
continue;
}
if (in_array($type, ['docx', 'xlsx', 'pptx'], true)) {
if ($indexOffice) {
fwrite(STDERR, "[Meili DEBUG] → add to OFFICE queue\n");
$officeLinks[] = $link;
} else {
fwrite(STDERR, "[Meili DEBUG] → Office indexing disabled\n");
}
continue;
if (
in_array($type, ['docx', 'xlsx', 'pptx'], true)
&& $indexOffice
) {
$officeLinks[] = $link;
}
fwrite(STDERR, "[Meili DEBUG] → ignored\n");
}
fwrite(
STDERR,
"[Meili DEBUG] Final queues: pdf="
. count($pdfLinks)
. " office="
. count($officeLinks)
. "\n"
);
try {
if ($pdfLinks !== []) {
fwrite(STDERR, "[Meili DEBUG] Calling handlePdfLinks()\n");
$this->pdfIndexService->handlePdfLinks($pdfLinks);
}
if ($officeLinks !== []) {
fwrite(STDERR, "[Meili DEBUG] Calling handleOfficeLinks()\n");
$this->officeIndexService->handleOfficeLinks($officeLinks);
}
} catch (\Throwable $e) {
fwrite(STDERR, "[Meili DEBUG] File indexing failed: {$e->getMessage()}\n");
error_log('[ContaoMeilisearch] File indexing failed: ' . $e->getMessage());
}
}
/**
* Extrahiert MEILISEARCH_JSON aus HTML-Kommentar
*/
private function extractMeilisearchJson(string $content): ?array
{
if (!preg_match('/<!--\s*MEILISEARCH_JSON\s*(\{.*?\})\s*-->/s', $content, $m)) {
return null;
}
$json = preg_replace('/^\xEF\xBB\xBF/', '', trim($m[1]));
$data = json_decode($json, true);
return json_last_error() === JSON_ERROR_NONE && is_array($data)
? $data
: null;
}
/**
* Sammle alle <a href="…"> Links
*/
@@ -150,39 +200,35 @@ class IndexPageListener
*/
private function detectIndexableFileType(string $url): ?string
{
fwrite(STDERR, "[Meili DEBUG] detectIndexableFileType(): $url\n");
// Hash entfernen
$url = strtok($url, '#');
$parts = parse_url($url);
$parts = parse_url($url);
if (!$parts) {
fwrite(STDERR, "[Meili DEBUG] → parse_url failed\n");
return null;
}
// direkter Pfad (/files/…)
if (!empty($parts['path'])) {
$ext = strtolower(pathinfo($parts['path'], PATHINFO_EXTENSION));
fwrite(STDERR, "[Meili DEBUG] → path ext: $ext\n");
if (in_array($ext, ['pdf', 'docx', 'xlsx', 'pptx'], true)) {
return $ext;
}
}
// Query-Parameter (Contao 4 + 5)
if (!empty($parts['query'])) {
parse_str($parts['query'], $query);
foreach (['file', 'p', 'f'] as $param) {
if (!empty($query[$param])) {
$candidate = rawurldecode(
html_entity_decode((string) $query[$param], ENT_QUOTES)
);
$candidate = (string) $query[$param];
// sicher decodieren (Contao 4 + 5)
$candidate = html_entity_decode($candidate, ENT_QUOTES);
$candidate = rawurldecode($candidate);
$ext = strtolower(pathinfo($candidate, PATHINFO_EXTENSION));
fwrite(
STDERR,
"[Meili DEBUG] → query $param=$candidate ext=$ext\n"
);
if (in_array($ext, ['pdf', 'docx', 'xlsx', 'pptx'], true)) {
return $ext;
+129 -94
View File
@@ -10,119 +10,148 @@ class PdfIndexService
{
private string $projectDir;
// pro PHP-Process genau 1x resetten
private bool $didReset = false;
// pro Crawl-Durchlauf: doppelte Verarbeitung vermeiden
private array $seenThisCrawl = [];
public function __construct(ParameterBagInterface $params)
{
$this->projectDir = rtrim((string) $params->get('kernel.project_dir'), '/');
fwrite(STDERR, "[Meili PDF DEBUG] projectDir={$this->projectDir}\n");
}
/**
* Wird aus dem Listener beim ersten Hook-Call pro Crawl aufgerufen.
*/
public function resetTableOnce(): void
{
if ($this->didReset) {
fwrite(STDERR, "[Meili PDF DEBUG] resetTableOnce(): already reset\n");
return;
}
fwrite(STDERR, "[Meili PDF DEBUG] resetTableOnce(): TRUNCATE tl_search_pdf\n");
$this->didReset = true;
$this->seenThisCrawl = [];
try {
Database::getInstance()->execute('TRUNCATE tl_search_pdf');
} catch (\Throwable $e) {
error_log('[ContaoMeilisearch] PDF reset failed: ' . $e->getMessage());
fwrite(STDERR, "[Meili PDF DEBUG] TRUNCATE failed: {$e->getMessage()}\n");
}
}
/**
* @param array<int,array{url:string,linkText:?string}> $pdfLinks
*/
public function handlePdfLinks(array $pdfLinks): void
{
fwrite(
STDERR,
"[Meili PDF DEBUG] handlePdfLinks(): count=" . count($pdfLinks) . "\n"
);
foreach ($pdfLinks as $row) {
$url = (string) ($row['url'] ?? '');
$linkText = $row['linkText'] ?? null;
fwrite(STDERR, "\n[Meili PDF DEBUG] URL={$url}\n");
if ($url === '') {
fwrite(STDERR, "[Meili PDF DEBUG] → empty URL, skip\n");
continue;
}
try {
// innerhalb des Crawls gleiche URL nicht mehrfach parsen
$seenKey = md5($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 Title
// 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
);
} catch (\Throwable $e) {
error_log(
'[ContaoMeilisearch] PDF indexing failed for "' . $url . '": ' . $e->getMessage()
);
$seenKey = md5($url);
if (isset($this->seenThisCrawl[$seenKey])) {
fwrite(STDERR, "[Meili PDF DEBUG] → already processed, skip\n");
continue;
}
$this->seenThisCrawl[$seenKey] = true;
$normalizedPath = $this->normalizePdfUrl($url);
fwrite(
STDERR,
"[Meili PDF DEBUG] normalizePdfUrl() → "
. ($normalizedPath ?? 'NULL')
. "\n"
);
if ($normalizedPath === null) {
fwrite(STDERR, "[Meili PDF DEBUG] → normalization failed, skip\n");
continue;
}
$absolutePath = $this->getAbsolutePath($normalizedPath);
fwrite(STDERR, "[Meili PDF DEBUG] absolutePath={$absolutePath}\n");
if (!is_file($absolutePath)) {
fwrite(STDERR, "[Meili PDF DEBUG] → file does NOT exist\n");
continue;
}
fwrite(STDERR, "[Meili PDF DEBUG] → file exists\n");
$mtime = (int) (filemtime($absolutePath) ?: 0);
$checksum = md5($normalizedPath . '|' . $mtime);
fwrite(
STDERR,
"[Meili PDF DEBUG] mtime={$mtime} checksum={$checksum}\n"
);
$pdfMetaTitle = $this->readPdfMetaTitle($absolutePath);
fwrite(
STDERR,
"[Meili PDF DEBUG] metaTitle="
. ($pdfMetaTitle ?: 'NULL')
. "\n"
);
$title = $linkText ?: ($pdfMetaTitle ?: basename($absolutePath));
fwrite(STDERR, "[Meili PDF DEBUG] final title={$title}\n");
$text = $this->parsePdf($absolutePath);
fwrite(
STDERR,
"[Meili PDF DEBUG] parsed text length=" . strlen($text) . "\n"
);
if ($text === '') {
fwrite(STDERR, "[Meili PDF DEBUG] → empty text, skip\n");
continue;
}
fwrite(STDERR, "[Meili PDF DEBUG] → writing to DB\n");
$this->upsertPdf(
$normalizedPath,
$title,
$text,
$checksum,
$mtime
);
}
}
private function normalizePdfUrl(string $url): ?string
{
// Fall 1: direkter /files/-Pfad
fwrite(STDERR, "[Meili PDF DEBUG] normalizePdfUrl(): {$url}\n");
if (str_starts_with($url, '/files/') && preg_match('~\.pdf(\?.*)?$~i', $url)) {
return preg_replace('~\?.*$~', '', $url);
$r = preg_replace('~\?.*$~', '', $url);
fwrite(STDERR, "[Meili PDF DEBUG] → direct /files path {$r}\n");
return $r;
}
$decoded = html_entity_decode($url);
$parts = parse_url($decoded);
// Fall 2: absolute URL auf gleiche Site
if (
!empty($parts['path'])
&& str_starts_with($parts['path'], '/files/')
&& str_ends_with(strtolower($parts['path']), '.pdf')
) {
fwrite(STDERR, "[Meili PDF DEBUG] → absolute URL path {$parts['path']}\n");
return $parts['path'];
}
// Fall 3: Contao-Download-Link mit ?p=
if (empty($parts['query'])) {
fwrite(STDERR, "[Meili PDF DEBUG] → no query\n");
return null;
}
@@ -130,9 +159,12 @@ class PdfIndexService
if (!empty($query['p'])) {
$p = urldecode((string) $query['p']);
return '/files/' . ltrim($p, '/');
$r = '/files/' . ltrim($p, '/');
fwrite(STDERR, "[Meili PDF DEBUG] → p= normalized {$r}\n");
return $r;
}
fwrite(STDERR, "[Meili PDF DEBUG] → no usable parameter\n");
return null;
}
@@ -141,8 +173,13 @@ class PdfIndexService
return $this->projectDir . '/' . ltrim($relativePath, '/');
}
private function upsertPdf(string $url, string $title, string $text, string $checksum, int $mtime): void
{
private function upsertPdf(
string $url,
string $title,
string $text,
string $checksum,
int $mtime
): void {
try {
Database::getInstance()
->prepare('
@@ -165,9 +202,12 @@ class PdfIndexService
$checksum,
$mtime
);
fwrite(STDERR, "[Meili PDF DEBUG] → DB write OK\n");
} catch (\Throwable $e) {
error_log(
'[ContaoMeilisearch] Failed to write PDF index entry (' . $url . '): ' . $e->getMessage()
fwrite(
STDERR,
"[Meili PDF DEBUG] DB write failed: {$e->getMessage()}\n"
);
}
}
@@ -177,18 +217,39 @@ class PdfIndexService
try {
$parser = new Parser();
$pdf = $parser->parseFile($absolutePath);
$text = $this->cleanPdfContent($pdf->getText());
return mb_substr($text, 0, 20000);
} catch (\Throwable $e) {
error_log(
'[ContaoMeilisearch] Failed to parse PDF "' . $absolutePath . '": ' . $e->getMessage()
fwrite(
STDERR,
"[Meili PDF DEBUG] parsePdf failed: {$e->getMessage()}\n"
);
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[$key])) {
return trim($details[$key]);
}
}
} catch (\Throwable $e) {
fwrite(
STDERR,
"[Meili PDF DEBUG] readPdfMetaTitle failed: {$e->getMessage()}\n"
);
}
return null;
}
private function cleanPdfContent(string $text): string
{
if (class_exists(\Normalizer::class)) {
@@ -198,34 +259,8 @@ class PdfIndexService
$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 = str_replace(["\\'", "", ""], "'", $text);
$text = preg_replace('/\s+/u', ' ', $text);
return trim($text);
}
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 $e) {
error_log(
'[ContaoMeilisearch] Failed to read PDF metadata "' . $absolutePath . '": ' . $e->getMessage()
);
}
return null;
}
}