SEO, Weiterleitungen und strukturierte Daten (Arbeitspaket E)
Migration 20260912000800 ergänzt page.robots, page.og_media_id, website.settings und die Tabelle redirect. Alles Maschinenlesbare wird aus dem veröffentlichten Bestand abgeleitet, nichts separat gepflegt: die Sitemap listet nur veröffentlichte, indexierbare Seiten samt Beitragsübersicht und Begriffsarchiven, und dieselbe Einstellung setzt auch das noindex-Meta - beide Aussagen können nicht auseinanderlaufen. Strukturierte Daten entstehen nur aus gepflegten Werten: ohne Angaben wird kein leerer Organization-Block ausgegeben. JSON-LD ist ein Datenblock, kein Skript, daher bleibt script-src 'none' unberührt. Weiterleitungen greifen vor dem 404, aber nach der Seitensuche: eine vorhandene Seite gewinnt immer, sonst könnte ein alter Eintrag eine neue Seite unerreichbar machen. Ketten werden abgewiesen. robots.txt nimmt zusätzliche Zeilen nur nach festem Muster an, damit ein Eintrag die Datei nicht umdeuten kann. llms.txt ist standardmäßig aus. Geprüft: 128 PHP-Tests / 507 Assertions unter PHP 8.2.12 und 8.4.22, 14 JavaScript-Modelltests, Lint/Format/Container. Dazu eine reine Sichtprüfung im Browser auf einer lokalen Demoseite - weiterhin keine axe- oder Lighthouse-Messung. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
kantarciandClaude Opus 5 committed
1 parent
d43bd91069
commit
a40c2bb27e
30 files changed
+1012
-38
No files matched your search
@@ -15,11 +15,13 @@ final class DocumentValidator
|
||||
public const V7_TYPES = ['tabs', 'tabitem'];
|
||||
/** Schema v8: a form block references a form by its address; the definition lives in the database. */
|
||||
public const V8_TYPES = ['form'];
|
||||
/** Schema v9: the breadcrumb trail, built from the page hierarchy rather than from content. */
|
||||
public const V9_TYPES = ['breadcrumb'];
|
||||
|
||||
public function validate(array $document): void
|
||||
{
|
||||
$schemaVersion = $document['schemaVersion'] ?? null;
|
||||
$this->require(in_array($schemaVersion, [1, 2, 3, 4, 5, 6, 7, 8], true), 'Unsupported document schema.');
|
||||
$this->require(in_array($schemaVersion, [1, 2, 3, 4, 5, 6, 7, 8, 9], true), 'Unsupported document schema.');
|
||||
$this->text($document['title'] ?? null, 160);
|
||||
$this->text($document['description'] ?? null, 320);
|
||||
$this->require(is_array($document['blocks'] ?? null) && array_is_list($document['blocks']), 'Blocks must be a list.');
|
||||
@@ -44,7 +46,7 @@ final class DocumentValidator
|
||||
// page: a deactivated plugin must never make stored content unreadable. Nothing from
|
||||
// props is rendered for such a block, so unvalidated values never reach the output.
|
||||
$plugin = $schemaVersion >= 4 && is_string($type) && 1 === preg_match('/^[a-z][a-z0-9]{1,30}\.[a-z][a-z0-9-]{1,40}$/D', $type);
|
||||
$this->require(in_array($type, ['section', 'row', 'column', 'heading', 'text', 'image', 'button'], true) || ($schemaVersion >= 3 && 'gallery' === $type) || ($schemaVersion >= 5 && 'postlist' === $type) || ($schemaVersion >= 6 && in_array($type, self::V6_TYPES, true)) || ($schemaVersion >= 7 && in_array($type, self::V7_TYPES, true)) || ($schemaVersion >= 8 && in_array($type, self::V8_TYPES, true)) || $plugin, 'Unknown block type. Preserve source until its plugin is available.');
|
||||
$this->require(in_array($type, ['section', 'row', 'column', 'heading', 'text', 'image', 'button'], true) || ($schemaVersion >= 3 && 'gallery' === $type) || ($schemaVersion >= 5 && 'postlist' === $type) || ($schemaVersion >= 6 && in_array($type, self::V6_TYPES, true)) || ($schemaVersion >= 7 && in_array($type, self::V7_TYPES, true)) || ($schemaVersion >= 8 && in_array($type, self::V8_TYPES, true)) || ($schemaVersion >= 9 && in_array($type, self::V9_TYPES, true)) || $plugin, 'Unknown block type. Preserve source until its plugin is available.');
|
||||
// An accordion entry only ever renders inside an accordion; at the top level or anywhere
|
||||
// else it would produce a details element without its group.
|
||||
$this->require('accordionitem' !== $type || 'accordion' === $parentType, 'An accordion entry belongs into an accordion.');
|
||||
|
||||
@@ -143,7 +143,7 @@ final class PageRepository
|
||||
public function findPublished(string $path, string $locale = 'de'): ?array
|
||||
{
|
||||
$page = $this->db->fetchAssociative(
|
||||
'SELECT p.id, p.slug, p.path, p.type, p.locale, p.author, p.excerpt, p.published_at, r.document, r.id AS revision_id, w.name AS website_name FROM page p JOIN revision r ON r.id = p.published_revision_id AND r.page_id = p.id JOIN website w ON w.id = p.website_id WHERE p.path = ? AND p.locale = ? AND p.website_id = ?',
|
||||
'SELECT p.id, p.slug, p.path, p.type, p.locale, p.author, p.excerpt, p.published_at, p.robots, p.og_media_id, r.document, r.id AS revision_id, w.name AS website_name FROM page p JOIN revision r ON r.id = p.published_revision_id AND r.page_id = p.id JOIN website w ON w.id = p.website_id WHERE p.path = ? AND p.locale = ? AND p.website_id = ?',
|
||||
[$path, $locale, 'default'],
|
||||
);
|
||||
if (false === $page) {
|
||||
@@ -157,7 +157,7 @@ final class PageRepository
|
||||
|
||||
public function listForEditor(): array
|
||||
{
|
||||
return $this->db->fetchAllAssociative('SELECT p.id, p.slug, p.path, p.parent_id, p.type, p.locale, p.version, p.published_revision_id, p.draft_revision_id, r.document FROM page p LEFT JOIN revision r ON r.id = p.draft_revision_id AND r.page_id = p.id WHERE p.website_id = ? ORDER BY p.path', ['default']);
|
||||
return $this->db->fetchAllAssociative('SELECT p.id, p.slug, p.path, p.parent_id, p.type, p.locale, p.version, p.published_revision_id, p.draft_revision_id, p.excerpt, p.author, p.published_at, p.featured_media_id, p.robots, p.og_media_id, r.document FROM page p LEFT JOIN revision r ON r.id = p.draft_revision_id AND r.page_id = p.id WHERE p.website_id = ? ORDER BY p.path', ['default']);
|
||||
}
|
||||
|
||||
public function findForEditor(string $id): ?array
|
||||
|
||||
@@ -40,18 +40,27 @@ final class PostRepository
|
||||
throw new InvalidPage('Das Datum braucht die Form JJJJ-MM-TT, wahlweise mit Uhrzeit.');
|
||||
}
|
||||
$mediaId = trim((string) ($values['featuredMediaId'] ?? ''));
|
||||
if ('' !== $mediaId && (1 !== preg_match('/^[a-f0-9]{32}$/D', $mediaId) || null === $this->media->find($mediaId))) {
|
||||
throw new InvalidPage('Das gewählte Beitragsbild gibt es nicht.');
|
||||
$ogMediaId = trim((string) ($values['ogMediaId'] ?? ''));
|
||||
foreach (['Beitragsbild' => $mediaId, 'Vorschaubild' => $ogMediaId] as $label => $candidate) {
|
||||
if ('' !== $candidate && (1 !== preg_match('/^[a-f0-9]{32}$/D', $candidate) || null === $this->media->find($candidate))) {
|
||||
throw new InvalidPage(sprintf('Das gewählte %s gibt es nicht.', $label));
|
||||
}
|
||||
}
|
||||
$robots = (string) ($values['robots'] ?? 'index');
|
||||
if (!in_array($robots, ['index', 'noindex'], true)) {
|
||||
throw new InvalidPage('Für die Indexierung sind nur „index“ und „noindex“ vorgesehen.');
|
||||
}
|
||||
|
||||
$this->db->transactional(function () use ($pageId, $excerpt, $author, $date, $mediaId, $categories, $tags, $expectedVersion): void {
|
||||
$this->db->transactional(function () use ($pageId, $excerpt, $author, $date, $mediaId, $ogMediaId, $robots, $categories, $tags, $expectedVersion): void {
|
||||
$changed = $this->db->executeStatement(
|
||||
'UPDATE page SET excerpt = ?, author = ?, featured_media_id = ?, published_at = ?, version = version + 1 WHERE id = ? AND version = ?',
|
||||
'UPDATE page SET excerpt = ?, author = ?, featured_media_id = ?, published_at = ?, robots = ?, og_media_id = ?, version = version + 1 WHERE id = ? AND version = ?',
|
||||
[
|
||||
'' === $excerpt ? null : $excerpt,
|
||||
'' === $author ? null : $author,
|
||||
'' === $mediaId ? null : $mediaId,
|
||||
'' === $date ? null : str_replace('T', ' ', 10 === strlen($date) ? $date.' 00:00' : $date).':00',
|
||||
$robots,
|
||||
'' === $ogMediaId ? null : $ogMediaId,
|
||||
$pageId,
|
||||
$expectedVersion,
|
||||
],
|
||||
|
||||
@@ -0,0 +1,91 @@
|
||||
<?php
|
||||
|
||||
namespace App\Content;
|
||||
|
||||
use App\Domain\InvalidPage;
|
||||
use Doctrine\DBAL\Connection;
|
||||
|
||||
/**
|
||||
* Addresses that moved or are deliberately gone. Checked before a page is answered with 404, so a
|
||||
* migrated site keeps every old link working.
|
||||
*/
|
||||
final class RedirectRepository
|
||||
{
|
||||
public const MAX = 2000;
|
||||
|
||||
public function __construct(private Connection $db)
|
||||
{
|
||||
}
|
||||
|
||||
/** @return list<array{id: string, from: string, to: ?string, status: int}> */
|
||||
public function all(): array
|
||||
{
|
||||
return array_map(static fn (array $row): array => [
|
||||
'id' => $row['id'], 'from' => $row['from_path'], 'to' => $row['to_path'], 'status' => (int) $row['status'],
|
||||
], $this->db->fetchAllAssociative('SELECT * FROM redirect WHERE website_id = ? ORDER BY from_path', ['default']));
|
||||
}
|
||||
|
||||
/** @return array{to: ?string, status: int}|null */
|
||||
public function find(string $path): ?array
|
||||
{
|
||||
$row = $this->db->fetchAssociative('SELECT to_path, status FROM redirect WHERE website_id = ? AND from_path = ?', ['default', trim($path, '/')]);
|
||||
|
||||
return false === $row ? null : ['to' => $row['to_path'], 'status' => (int) $row['status']];
|
||||
}
|
||||
|
||||
public function save(string $from, string $to, int $status): void
|
||||
{
|
||||
$from = trim(trim($from), '/');
|
||||
$to = trim(trim($to), '/');
|
||||
if (!in_array($status, [301, 410], true)) {
|
||||
throw new InvalidPage('Erlaubt sind 301 für eine verschobene Adresse und 410 für eine entfernte.');
|
||||
}
|
||||
if (1 !== preg_match('#^[a-z0-9]+(?:[-/][a-z0-9]+)*$#D', $from) || strlen($from) > 400) {
|
||||
throw new InvalidPage('Die alte Adresse darf nur Kleinbuchstaben, Ziffern, Bindestriche und Schrägstriche enthalten.');
|
||||
}
|
||||
if (301 === $status) {
|
||||
if (1 !== preg_match('#^[a-z0-9]+(?:[-/][a-z0-9]+)*$#D', $to) || strlen($to) > 400) {
|
||||
throw new InvalidPage('Das Ziel darf nur Kleinbuchstaben, Ziffern, Bindestriche und Schrägstriche enthalten.');
|
||||
}
|
||||
if ($from === $to) {
|
||||
throw new InvalidPage('Eine Adresse kann nicht auf sich selbst weiterleiten.');
|
||||
}
|
||||
// One hop only: a chain would cost an extra round trip and can become a loop.
|
||||
$next = $this->find($to);
|
||||
if (null !== $next) {
|
||||
throw new InvalidPage(sprintf('Das Ziel /%s leitet selbst weiter. Zeige direkt auf das Endziel.', $to));
|
||||
}
|
||||
}
|
||||
if (self::MAX <= (int) $this->db->fetchOne('SELECT COUNT(*) FROM redirect WHERE website_id = ?', ['default'])
|
||||
&& null === $this->find($from)) {
|
||||
throw new InvalidPage(sprintf('Mehr als %d Weiterleitungen sind nicht vorgesehen.', self::MAX));
|
||||
}
|
||||
|
||||
$this->db->transactional(function () use ($from, $to, $status): void {
|
||||
$this->db->executeStatement('DELETE FROM redirect WHERE website_id = ? AND from_path = ?', ['default', $from]);
|
||||
$this->db->insert('redirect', [
|
||||
'id' => bin2hex(random_bytes(16)), 'website_id' => 'default',
|
||||
'from_path' => $from, 'to_path' => 301 === $status ? $to : null,
|
||||
'status' => $status, 'created_at' => gmdate('Y-m-d H:i:s'),
|
||||
]);
|
||||
});
|
||||
}
|
||||
|
||||
public function delete(string $id): void
|
||||
{
|
||||
$this->db->executeStatement('DELETE FROM redirect WHERE id = ? AND website_id = ?', [$id, 'default']);
|
||||
}
|
||||
|
||||
/** Records that an address moved, used when a page is renamed. Never overwrites an entry. */
|
||||
public function remember(string $from, string $to): void
|
||||
{
|
||||
if ($from === $to || null !== $this->find($from)) {
|
||||
return;
|
||||
}
|
||||
try {
|
||||
$this->save($from, $to, 301);
|
||||
} catch (InvalidPage) {
|
||||
// A path the rules reject is simply not remembered; renaming must not fail for it.
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,127 @@
|
||||
<?php
|
||||
|
||||
namespace App\Content;
|
||||
|
||||
use App\Domain\InvalidPage;
|
||||
use Doctrine\DBAL\Connection;
|
||||
|
||||
/**
|
||||
* Company data and site wide switches. The structured data the pages emit is built from exactly
|
||||
* these values, so what a search engine reads is what an editor maintains here – never a separate,
|
||||
* invisible copy.
|
||||
*/
|
||||
final class SiteSettings
|
||||
{
|
||||
public const DEFAULTS = [
|
||||
'legalName' => '',
|
||||
'description' => '',
|
||||
'phone' => '',
|
||||
'email' => '',
|
||||
'street' => '',
|
||||
'postalCode' => '',
|
||||
'city' => '',
|
||||
'country' => 'DE',
|
||||
'sameAs' => [],
|
||||
'localBusiness' => false,
|
||||
'llmsTxt' => false,
|
||||
'robotsExtra' => '',
|
||||
];
|
||||
|
||||
public function __construct(private Connection $db)
|
||||
{
|
||||
}
|
||||
|
||||
public function load(): array
|
||||
{
|
||||
$raw = $this->db->fetchOne('SELECT settings FROM website WHERE id = ?', ['default']);
|
||||
$stored = is_string($raw) ? json_decode($raw, true, 512, JSON_THROW_ON_ERROR) : [];
|
||||
|
||||
return [...self::DEFAULTS, ...(is_array($stored) ? $stored : [])];
|
||||
}
|
||||
|
||||
public function save(array $input): void
|
||||
{
|
||||
$values = self::DEFAULTS;
|
||||
foreach (['legalName' => 200, 'description' => 300, 'phone' => 40, 'email' => 120, 'street' => 120, 'postalCode' => 16, 'city' => 120] as $key => $max) {
|
||||
$values[$key] = mb_substr(trim((string) ($input[$key] ?? '')), 0, $max);
|
||||
}
|
||||
if ('' !== $values['email'] && false === filter_var($values['email'], FILTER_VALIDATE_EMAIL)) {
|
||||
throw new InvalidPage('Bitte gib eine gültige E-Mail-Adresse an.');
|
||||
}
|
||||
if ('' !== $values['phone'] && 1 !== preg_match('/^\+?[0-9 ()\/-]{3,32}$/D', $values['phone'])) {
|
||||
throw new InvalidPage('Bitte gib eine gültige Telefonnummer an.');
|
||||
}
|
||||
$country = strtoupper(trim((string) ($input['country'] ?? 'DE')));
|
||||
if (1 !== preg_match('/^[A-Z]{2}$/D', $country)) {
|
||||
throw new InvalidPage('Das Land wird als zweibuchstabiger Code angegeben, zum Beispiel DE.');
|
||||
}
|
||||
$values['country'] = $country;
|
||||
foreach (preg_split('/[\r\n,]+/', (string) ($input['sameAs'] ?? '')) ?: [] as $profile) {
|
||||
$profile = trim($profile);
|
||||
if ('' === $profile) {
|
||||
continue;
|
||||
}
|
||||
if (!str_starts_with($profile, 'https://') || false === filter_var($profile, FILTER_VALIDATE_URL)) {
|
||||
throw new InvalidPage(sprintf('„%s“ ist keine HTTPS-Adresse.', $profile));
|
||||
}
|
||||
$values['sameAs'][] = $profile;
|
||||
}
|
||||
$values['sameAs'] = array_slice(array_unique($values['sameAs']), 0, 20);
|
||||
$values['localBusiness'] = (bool) ($input['localBusiness'] ?? false);
|
||||
if ($values['localBusiness'] && ('' === $values['street'] || '' === $values['city'])) {
|
||||
throw new InvalidPage('Ein lokales Unternehmen braucht mindestens Straße und Ort.');
|
||||
}
|
||||
$values['llmsTxt'] = (bool) ($input['llmsTxt'] ?? false);
|
||||
// Only comment and simple directive lines; nothing that could smuggle in a second policy.
|
||||
$extra = [];
|
||||
foreach (preg_split('/\R/', (string) ($input['robotsExtra'] ?? '')) ?: [] as $line) {
|
||||
$line = trim($line);
|
||||
if ('' !== $line && 1 === preg_match('/^(#.{0,180}|(?:Disallow|Allow|Crawl-delay|User-agent):[ \t][^\s][^\r\n]{0,180})$/D', $line)) {
|
||||
$extra[] = $line;
|
||||
}
|
||||
}
|
||||
$values['robotsExtra'] = implode("\n", array_slice($extra, 0, 30));
|
||||
|
||||
$this->db->executeStatement('UPDATE website SET settings = ? WHERE id = ?', [
|
||||
json_encode($values, JSON_THROW_ON_ERROR | JSON_UNESCAPED_UNICODE),
|
||||
'default',
|
||||
]);
|
||||
}
|
||||
|
||||
/**
|
||||
* Structured data for the organisation, built from the maintained values only. Returns null
|
||||
* while nothing is maintained: an empty Organization block would be noise, not data.
|
||||
*/
|
||||
public function organisation(array $settings, string $origin, string $siteName): ?array
|
||||
{
|
||||
$hasAddress = '' !== $settings['street'] && '' !== $settings['city'];
|
||||
if ('' === $settings['legalName'] && !$hasAddress && '' === $settings['phone']) {
|
||||
return null;
|
||||
}
|
||||
$data = [
|
||||
'@context' => 'https://schema.org',
|
||||
'@type' => $settings['localBusiness'] && $hasAddress ? 'LocalBusiness' : 'Organization',
|
||||
'name' => '' !== $settings['legalName'] ? $settings['legalName'] : $siteName,
|
||||
'url' => $origin.'/',
|
||||
];
|
||||
foreach (['description' => 'description', 'phone' => 'telephone', 'email' => 'email'] as $key => $property) {
|
||||
if ('' !== $settings[$key]) {
|
||||
$data[$property] = $settings[$key];
|
||||
}
|
||||
}
|
||||
if ($hasAddress) {
|
||||
$data['address'] = array_filter([
|
||||
'@type' => 'PostalAddress',
|
||||
'streetAddress' => $settings['street'],
|
||||
'postalCode' => $settings['postalCode'],
|
||||
'addressLocality' => $settings['city'],
|
||||
'addressCountry' => $settings['country'],
|
||||
], static fn (string $value): bool => '' !== $value);
|
||||
}
|
||||
if ([] !== $settings['sameAs']) {
|
||||
$data['sameAs'] = $settings['sameAs'];
|
||||
}
|
||||
|
||||
return $data;
|
||||
}
|
||||
}
|
||||
Reference in new issue
Block a user