VITEC-website/packages/vitec/Classes/Command/NormalizeSlugsCommand.php

<?php

declare(strict_types=1);

namespace Evomedien\Vitec\Command;

use Symfony\Component\Console\Attribute\AsCommand;
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 TYPO3\CMS\Core\Core\Bootstrap;
use TYPO3\CMS\Core\Database\ConnectionPool;
use TYPO3\CMS\Core\DataHandling\DataHandler;
use TYPO3\CMS\Core\Utility\GeneralUtility;

/**
 * Normalises every page slug below a root to <parent-slug>/<slugified-title>,
 * recursively and cascading (a fixed parent slug carries into its children).
 * EXT:redirects covers the old URLs automatically.
 *
 * Why: TYPO3's own slug generator KEEPS slashes that live in page titles
 * ("RTCC / RTIC" -> rtcc-/-rtic), which breaks routes and canonicals - the
 * defect hit us three times (passenger page, solution page 467, RTCC).
 * This command repairs a whole subtree in one sweep.
 *
 * --style picks the ampersand convention of the tree:
 *   plain  "A & B" -> a-b      (TYPO3 default, used below /solutions)
 *   and    "A & B" -> a-and-b  (used below /markets since the 10-23 align)
 * Do not mix styles within one tree: the wrong style renames EVERY page.
 *
 *   vendor/bin/typo3 vitec:normalize-slugs --root=/solutions
 *   vendor/bin/typo3 vitec:normalize-slugs --root=/solutions --apply
 */
#[AsCommand(
    name: 'vitec:normalize-slugs',
    description: 'Normalise page slugs below a root to parent-slug/title-slug (dry-run by default)'
)]
final class NormalizeSlugsCommand extends Command
{
    /** Page types that carry no meaningful slug of their own. */
    private const SKIP_DOKTYPES = [199, 254, 255];

    protected function configure(): void
    {
        $this->addOption('root', null, InputOption::VALUE_REQUIRED, 'Slug of the root page, e.g. /solutions');
        $this->addOption('style', null, InputOption::VALUE_REQUIRED, "Ampersand style: 'plain' (A & B -> a-b) or 'and' (-> a-and-b)", 'plain');
        $this->addOption('apply', null, InputOption::VALUE_NONE, 'Write changes (default: dry run)');
    }

    protected function execute(InputInterface $input, OutputInterface $output): int
    {
        Bootstrap::initializeBackendAuthentication();
        $apply = (bool)$input->getOption('apply');
        $rootSlug = rtrim((string)$input->getOption('root'), '/');
        $style = (string)$input->getOption('style') === 'and' ? 'and' : 'plain';
        if ($rootSlug === '') {
            $output->writeln('<error>--root is required (e.g. --root=/solutions)</error>');
            return Command::FAILURE;
        }

        $connection = GeneralUtility::makeInstance(ConnectionPool::class)->getConnectionForTable('pages');
        $root = $connection->fetchAssociative('SELECT uid, slug FROM pages WHERE deleted = 0 AND slug = ?', [$rootSlug]);
        if (!$root) {
            $output->writeln('<error>No page with slug ' . $rootSlug . '</error>');
            return Command::FAILURE;
        }

        $output->writeln(($apply ? 'APPLY' : 'DRY-RUN') . ' · root ' . $rootSlug . ' (uid ' . (int)$root['uid'] . ') · style ' . $style);
        $datamap = [];
        $changes = 0;
        $this->walk($connection, (int)$root['uid'], $rootSlug, $style, $datamap, $changes, $output);

        if ($changes === 0) {
            $output->writeln('All slugs already normalised.');
            return Command::SUCCESS;
        }
        if (!$apply) {
            $output->writeln("\n" . $changes . ' change(s). DRY RUN - nothing written. Add --apply to write.');
            return Command::SUCCESS;
        }
        $dataHandler = GeneralUtility::makeInstance(DataHandler::class);
        $dataHandler->start($datamap, []);
        $dataHandler->process_datamap();
        if ($dataHandler->errorLog !== []) {
            $output->writeln('<error>DataHandler: ' . implode(' | ', $dataHandler->errorLog) . '</error>');
            return Command::FAILURE;
        }
        $output->writeln('<info>' . $changes . ' slug(s) written (redirects follow). Flush the cache: vendor/bin/typo3 cache:flush</info>');
        return Command::SUCCESS;
    }

    /**
     * Depth-first: the parent's EFFECTIVE (possibly just-fixed) slug prefixes
     * the children. Collisions between siblings get a numeric suffix and a
     * warning - that needs an editorial look, never silent overwriting.
     *
     * @param array<string,array<int,array<string,mixed>>> $datamap
     */
    private function walk($connection, int $parentUid, string $parentSlug, string $style, array &$datamap, int &$changes, OutputInterface $output): void
    {
        $taken = [];
        foreach ($connection->fetchAllAssociative(
            'SELECT uid, title, slug, doktype FROM pages WHERE deleted = 0 AND pid = ? AND sys_language_uid IN (0,-1) ORDER BY sorting',
            [$parentUid]
        ) as $child) {
            if (in_array((int)$child['doktype'], self::SKIP_DOKTYPES, true)) {
                $this->walk($connection, (int)$child['uid'], $parentSlug, $style, $datamap, $changes, $output);
                continue;
            }
            $expected = $parentSlug . '/' . $this->slugify((string)$child['title'], $style);
            if (isset($taken[$expected])) {
                $suffixed = $expected . '-' . ((int)$child['uid']);
                $output->writeln('  <comment>! sibling collision on ' . $expected . ' - "' . (string)$child['title'] . '" (uid ' . (int)$child['uid'] . ') gets ' . $suffixed . ', review editorially</comment>');
                $expected = $suffixed;
            }
            $taken[$expected] = true;
            if ((string)$child['slug'] !== $expected) {
                $datamap['pages'][(int)$child['uid']]['slug'] = $expected;
                $this->queueRedirect($datamap, (string)$child['slug'], (int)$child['uid']);
                $changes++;
                $output->writeln(sprintf('  ~ "%s" (uid %d) %s -> %s', (string)$child['title'], (int)$child['uid'], (string)$child['slug'], $expected));
            }
            $this->walk($connection, (int)$child['uid'], $expected, $style, $datamap, $changes, $output);
        }
    }

    /**
     * Queues the 301 for a vacated slug in the same DataHandler run -
     * EXT:redirects' automatic creation does not fire for CLI DataHandler
     * runs (found 2026-10-26: every renamed URL answered 404).
     *
     * @param array<string,array<int|string,array<string,mixed>>> $datamap
     */
    private function queueRedirect(array &$datamap, string $oldSlug, int $pageUid): void
    {
        $oldSlug = rtrim($oldSlug, '/');
        if ($oldSlug === '') {
            return;
        }
        $datamap['sys_redirect']['NEWR' . substr(md5($oldSlug), 0, 10)] = [
            'pid' => 0,
            'source_host' => '*',
            'source_path' => $oldSlug,
            'target' => 't3://page?uid=' . $pageUid,
            'target_statuscode' => 301,
        ];
    }
    private function slugify(string $value, string $style): string
    {
        $value = mb_strtolower(trim($value));
        $value = str_replace('&', $style === 'and' ? ' and ' : ' ', $value);
        return trim((string)preg_replace('/[^a-z0-9]+/', '-', $value), '-');
    }
}