MokoSuiteOpenGraph/source/packages/plg_system_mokoog/src/Helper/SitemapBuilder.php

<?php

/**
 * @package     MokoSuiteOpenGraph
 * @subpackage  plg_system_mokoog
 * @author      Moko Consulting <hello@mokoconsulting.tech>
 * @copyright   Copyright (C) 2026 Moko Consulting. All rights reserved.
 * @license     GNU General Public License version 3 or later; see LICENSE
 */

namespace Joomla\Plugin\System\MokoOG\Helper;

defined('_JEXEC') or die;

use Joomla\CMS\Factory;
use Joomla\CMS\Uri\Uri;

/**
 * XML Sitemap builder.
 *
 * Generates a sitemap.xml containing all published articles, excluding
 * those marked with noindex robots directives in the mokoog_tags table.
 */
class SitemapBuilder
{
    /**
     * Generate sitemap XML content.
     *
     * @param   string  $changefreq  Default change frequency for entries
     *
     * @return  string  Complete sitemap XML
     */
    public static function generate(string $changefreq = 'weekly'): string
    {
        $allowed = ['always', 'hourly', 'daily', 'weekly', 'monthly', 'yearly', 'never'];
        $changefreq = \in_array($changefreq, $allowed, true) ? $changefreq : 'weekly';

        $db = Factory::getContainer()->get(\Joomla\Database\DatabaseInterface::class);

        // Only include content the public (guest, user id 0) can view — never
        // leak registered/special-access articles into the public sitemap.
        $publicLevels = array_map('intval', \Joomla\CMS\Access\Access::getAuthorisedViewLevels(0));

        // Get all published articles
        $query = $db->getQuery(true)
            ->select($db->quoteName(['a.id', 'a.alias', 'a.catid', 'a.modified', 'a.language']))
            ->from($db->quoteName('#__content', 'a'))
            ->where($db->quoteName('a.state') . ' = 1');

        if (!empty($publicLevels)) {
            $query->where($db->quoteName('a.access') . ' IN (' . implode(',', $publicLevels) . ')');
        }

        $db->setQuery($query);
        $articles = $db->loadObjectList();

        // Get noindex articles from mokoog_tags
        $noindexQuery = $db->getQuery(true)
            ->select($db->quoteName('content_id'))
            ->from($db->quoteName('#__mokoog_tags'))
            ->where($db->quoteName('content_type') . ' = ' . $db->quote('com_content'))
            ->where($db->quoteName('robots') . ' LIKE ' . $db->quote('%noindex%'));

        $db->setQuery($noindexQuery);
        $noindexIds = array_map('intval', $db->loadColumn());

        $root = rtrim(Uri::root(), '/');
        $xml  = '<?xml version="1.0" encoding="UTF-8"?>' . "\n";
        $xml .= '<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">' . "\n";

        // Homepage
        $xml .= '  <url>' . "\n";
        $xml .= '    <loc>' . $root . '/</loc>' . "\n";
        $xml .= '    <changefreq>daily</changefreq>' . "\n";
        $xml .= '    <priority>1.0</priority>' . "\n";
        $xml .= '  </url>' . "\n";

        foreach ($articles as $article) {
            // Skip noindexed
            if (\in_array((int) $article->id, $noindexIds, true)) {
                continue;
            }

            $url     = self::articleUrl($article, $root);
            $lastmod = $article->modified && $article->modified !== '0000-00-00 00:00:00'
                ? date('Y-m-d', strtotime($article->modified)) : '';

            $xml .= '  <url>' . "\n";
            $xml .= '    <loc>' . htmlspecialchars($url, ENT_XML1) . '</loc>' . "\n";

            if ($lastmod) {
                $xml .= '    <lastmod>' . $lastmod . '</lastmod>' . "\n";
            }

            $xml .= '    <changefreq>' . $changefreq . '</changefreq>' . "\n";
            $xml .= '    <priority>0.8</priority>' . "\n";
            $xml .= '  </url>' . "\n";
        }

        $xml .= '</urlset>';

        return $xml;
    }

    /**
     * Build the SEF/canonical site URL for an article, with a safe fallback.
     *
     * Routes through the site router so the sitemap matches the canonical URLs
     * the plugin emits. If routing fails (or SEF is off), falls back to the
     * non-SEF index.php URL — never an empty or broken URL.
     *
     * @param   object  $article  Row with id, alias, catid, language
     * @param   string  $root     Site root without trailing slash
     *
     * @return  string  Absolute URL
     */
    private static function articleUrl(object $article, string $root): string
    {
        $fallback = $root . '/index.php?option=com_content&view=article&id=' . (int) $article->id;

        $internal = 'index.php?option=com_content&view=article&id=' . (int) $article->id
            . (!empty($article->alias) ? ':' . $article->alias : '')
            . (!empty($article->catid) ? '&catid=' . (int) $article->catid : '');

        try {
            $routed = \Joomla\CMS\Router\Route::link(
                'site',
                $internal,
                false,
                \Joomla\CMS\Router\Route::TLS_IGNORE,
                true
            );

            if (\is_string($routed) && $routed !== '') {
                return $routed;
            }
        } catch (\Throwable $e) {
            // Fall back to the non-SEF URL below.
        }

        return $fallback;
    }

    /**
     * Write sitemap XML to the site root.
     *
     * @param   string  $xml  The sitemap XML content
     *
     * @return  bool  True on success
     */
    public static function writeToFile(string $xml): bool
    {
        $path = JPATH_ROOT . '/sitemap.xml';
        $tmp  = $path . '.' . uniqid('tmp', true);

        if (file_put_contents($tmp, $xml) === false) {
            return false;
        }

        // Atomic replace so concurrent saves never expose a half-written sitemap.
        if (!@rename($tmp, $path)) {
            @unlink($tmp);

            return false;
        }

        return true;
    }
}