roi-theme/Shared/Infrastructure/Scripts/validate-fix-lists.php

<?php
/**
 * Validador de Correcciones - Genera archivos HTML para revisión visual
 *
 * PROPÓSITO: Crear archivos comparativos ANTES/DESPUÉS para validar
 * que la corrección no rompe el contenido.
 *
 * USO: php validate-fix-lists.php
 *
 * GENERA:
 *   /tmp/list-fix-validation/
 *     ├── post_ID_before.html
 *     ├── post_ID_after.html
 *     └── comparison_report.html
 *
 * @package ROI_Theme
 */

error_reporting(E_ALL);
ini_set('display_errors', 1);
ini_set('memory_limit', '256M');

$db_config = [
    'host'     => 'localhost',
    'database' => 'preciosunitarios_seo',
    'username' => 'preciosunitarios_seo',
    'password' => 'ACl%EEFd=V-Yvb??',
    'charset'  => 'utf8mb4'
];

$output_dir = '/tmp/list-fix-validation';
$sample_size = 5;

echo "==============================================\n";
echo "  VALIDADOR DE CORRECCIONES\n";
echo "  Fecha: " . date('Y-m-d H:i:s') . "\n";
echo "==============================================\n\n";

// Crear directorio de salida
if (!is_dir($output_dir)) {
    mkdir($output_dir, 0755, true);
}

// Limpiar archivos anteriores
array_map('unlink', glob("$output_dir/*.html"));

/**
 * Detectar problemas en HTML
 */
function detectIssues(string $html): array {
    $issues = [];
    libxml_use_internal_errors(true);

    $doc = new DOMDocument('1.0', 'UTF-8');
    $wrapped = '<div id="temp-wrapper">' . $html . '</div>';
    $doc->loadHTML('<?xml encoding="UTF-8">' . $wrapped, LIBXML_HTML_NOIMPLIED | LIBXML_HTML_NODEFDTD);
    libxml_clear_errors();

    $validChildren = ['li', 'script', 'template'];

    foreach (['ul', 'ol'] as $listTag) {
        foreach ($doc->getElementsByTagName($listTag) as $list) {
            foreach ($list->childNodes as $child) {
                if ($child->nodeType === XML_ELEMENT_NODE) {
                    $tagName = strtolower($child->nodeName);
                    if (!in_array($tagName, $validChildren)) {
                        $issues[] = "<$listTag> contiene <$tagName>";
                    }
                }
            }
        }
    }

    return $issues;
}

/**
 * Corregir listas mal formadas
 */
function fixMalformedLists(string $html): array {
    $result = ['fixed' => false, 'html' => $html, 'changes' => 0, 'details' => []];

    libxml_use_internal_errors(true);
    $doc = new DOMDocument('1.0', 'UTF-8');
    $wrapped = '<div id="temp-wrapper">' . $html . '</div>';
    $doc->loadHTML('<?xml encoding="UTF-8">' . $wrapped, LIBXML_HTML_NOIMPLIED | LIBXML_HTML_NODEFDTD);
    libxml_clear_errors();

    $lists = [];
    foreach ($doc->getElementsByTagName('ul') as $ul) { $lists[] = $ul; }
    foreach ($doc->getElementsByTagName('ol') as $ol) { $lists[] = $ol; }

    $changes = 0;
    $validChildren = ['li', 'script', 'template'];

    foreach ($lists as $list) {
        $nodesToProcess = [];
        foreach ($list->childNodes as $child) {
            if ($child->nodeType === XML_ELEMENT_NODE) {
                $tagName = strtolower($child->nodeName);
                if (!in_array($tagName, $validChildren)) {
                    $nodesToProcess[] = $child;
                }
            }
        }

        foreach ($nodesToProcess as $node) {
            $tagName = strtolower($node->nodeName);
            $prevLi = null;
            $prev = $node->previousSibling;

            while ($prev) {
                if ($prev->nodeType === XML_ELEMENT_NODE && strtolower($prev->nodeName) === 'li') {
                    $prevLi = $prev;
                    break;
                }
                $prev = $prev->previousSibling;
            }

            if ($prevLi) {
                $prevLi->appendChild($node);
                $result['details'][] = "Movido <$tagName> dentro del <li> anterior";
                $changes++;
            } else {
                $newLi = $doc->createElement('li');
                $list->insertBefore($newLi, $node);
                $newLi->appendChild($node);
                $result['details'][] = "Envuelto <$tagName> en nuevo <li>";
                $changes++;
            }
        }
    }

    if ($changes > 0) {
        $wrapper = $doc->getElementById('temp-wrapper');
        if ($wrapper) {
            $innerHTML = '';
            foreach ($wrapper->childNodes as $child) {
                $innerHTML .= $doc->saveHTML($child);
            }
            $result['html'] = $innerHTML;
            $result['fixed'] = true;
            $result['changes'] = $changes;
        }
    }

    return $result;
}

/**
 * Generar HTML wrapper para visualización
 */
function wrapForVisualization(string $content, string $title, string $status): string {
    $statusColor = $status === 'error' ? '#dc3545' : '#28a745';
    return <<<HTML
<!DOCTYPE html>
<html lang="es">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>$title</title>
    <style>
        body { font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif; margin: 20px; line-height: 1.6; }
        .status { padding: 10px 20px; background: $statusColor; color: white; border-radius: 4px; margin-bottom: 20px; }
        .content { border: 1px solid #ddd; padding: 20px; border-radius: 4px; background: #fafafa; }
        ul, ol { background: #fff3cd; padding: 15px 15px 15px 35px; border-left: 4px solid #ffc107; margin: 10px 0; }
        li { background: #d4edda; padding: 5px 10px; margin: 5px 0; border-left: 3px solid #28a745; }
        h1, h2, h3, h4, h5, h6 { color: #333; }
        p { color: #555; }
    </style>
</head>
<body>
    <div class="status">$status</div>
    <div class="content">
        $content
    </div>
</body>
</html>
HTML;
}

// Conectar a DB
$conn = new mysqli($db_config['host'], $db_config['username'], $db_config['password'], $db_config['database']);
$conn->set_charset($db_config['charset']);

if ($conn->connect_error) {
    die("Error de conexión: " . $conn->connect_error);
}

echo "✓ Conexión establecida\n\n";

// Buscar posts con problemas
$query = "SELECT id, page, html FROM datos_seo_pagina WHERE html IS NOT NULL AND html != '' ORDER BY id LIMIT 500";
$result = $conn->query($query);

$samples = [];
while ($row = $result->fetch_assoc()) {
    $issues = detectIssues($row['html']);
    if (!empty($issues) && count($samples) < $sample_size) {
        $samples[] = $row;
    }
}

echo "Encontrados " . count($samples) . " posts con problemas para validar\n\n";

$comparison_data = [];

foreach ($samples as $idx => $post) {
    $id = $post['id'];
    $url = $post['page'];
    $html_before = $post['html'];

    echo "─────────────────────────────────\n";
    echo "POST $id: $url\n";

    // Detectar problemas antes
    $issues_before = detectIssues($html_before);
    echo "  Problemas ANTES: " . count($issues_before) . "\n";

    // Aplicar corrección
    $fixResult = fixMalformedLists($html_before);
    $html_after = $fixResult['html'];

    // Detectar problemas después
    $issues_after = detectIssues($html_after);
    echo "  Problemas DESPUÉS: " . count($issues_after) . "\n";
    echo "  Cambios aplicados: " . $fixResult['changes'] . "\n";

    // Guardar archivos HTML
    $file_before = "$output_dir/post_{$id}_BEFORE.html";
    $file_after = "$output_dir/post_{$id}_AFTER.html";

    file_put_contents($file_before, wrapForVisualization(
        $html_before,
        "Post $id - ANTES (con errores)",
        "ANTES: " . count($issues_before) . " problemas de listas"
    ));

    file_put_contents($file_after, wrapForVisualization(
        $html_after,
        "Post $id - DESPUÉS (corregido)",
        "DESPUÉS: " . count($issues_after) . " problemas - " . $fixResult['changes'] . " correcciones aplicadas"
    ));

    echo "  ✓ Archivos generados:\n";
    echo "    - $file_before\n";
    echo "    - $file_after\n";

    // Guardar datos para reporte
    $comparison_data[] = [
        'id' => $id,
        'url' => $url,
        'issues_before' => count($issues_before),
        'issues_after' => count($issues_after),
        'changes' => $fixResult['changes'],
        'file_before' => "post_{$id}_BEFORE.html",
        'file_after' => "post_{$id}_AFTER.html"
    ];
}

// Generar reporte comparativo
$report_html = <<<HTML
<!DOCTYPE html>
<html lang="es">
<head>
    <meta charset="UTF-8">
    <title>Reporte de Validación - Corrección de Listas</title>
    <style>
        body { font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif; margin: 20px; }
        h1 { color: #333; border-bottom: 2px solid #007bff; padding-bottom: 10px; }
        table { width: 100%; border-collapse: collapse; margin: 20px 0; }
        th, td { padding: 12px; text-align: left; border: 1px solid #ddd; }
        th { background: #007bff; color: white; }
        tr:nth-child(even) { background: #f8f9fa; }
        .success { color: #28a745; font-weight: bold; }
        .warning { color: #ffc107; font-weight: bold; }
        .error { color: #dc3545; font-weight: bold; }
        a { color: #007bff; text-decoration: none; }
        a:hover { text-decoration: underline; }
        .instructions { background: #e7f3ff; padding: 15px; border-radius: 4px; margin: 20px 0; }
    </style>
</head>
<body>
    <h1>Reporte de Validación - Corrección de Listas HTML</h1>

    <div class="instructions">
        <strong>Instrucciones:</strong>
        <ol>
            <li>Abre cada par de archivos (ANTES/DESPUÉS) en el navegador</li>
            <li>Verifica que el contenido se muestre correctamente</li>
            <li>Las listas (fondo amarillo) deben contener solo items (fondo verde)</li>
            <li>Si todo se ve bien, la corrección es segura</li>
        </ol>
    </div>

    <table>
        <thead>
            <tr>
                <th>ID</th>
                <th>URL</th>
                <th>Problemas Antes</th>
                <th>Problemas Después</th>
                <th>Cambios</th>
                <th>Archivos</th>
            </tr>
        </thead>
        <tbody>
HTML;

foreach ($comparison_data as $data) {
    $status_class = $data['issues_after'] == 0 ? 'success' : ($data['issues_after'] < $data['issues_before'] ? 'warning' : 'error');

    $report_html .= <<<HTML
            <tr>
                <td>{$data['id']}</td>
                <td><a href="{$data['url']}" target="_blank">{$data['url']}</a></td>
                <td class="error">{$data['issues_before']}</td>
                <td class="$status_class">{$data['issues_after']}</td>
                <td>{$data['changes']}</td>
                <td>
                    <a href="{$data['file_before']}" target="_blank">ANTES</a> |
                    <a href="{$data['file_after']}" target="_blank">DESPUÉS</a>
                </td>
            </tr>
HTML;
}

$report_html .= <<<HTML
        </tbody>
    </table>

    <p><strong>Generado:</strong> {$_SERVER['REQUEST_TIME_FLOAT']}</p>
</body>
</html>
HTML;

$report_file = "$output_dir/comparison_report.html";
file_put_contents($report_file, $report_html);

echo "\n─────────────────────────────────\n";
echo "REPORTE GENERADO:\n";
echo "  $report_file\n\n";
echo "Para revisar, descarga el directorio:\n";
echo "  scp -r VPSContabo:$output_dir ./validation/\n\n";

$conn->close();
echo "✓ Validación completada.\n";