NebulaCloud/lib/sitemap-utils.ts

import { locales, defaultLocale } from './i18n';
import { getSitemapStats, generateSitemap } from './sitemap-generator';

/**
 * 站点地图验证工具
 */
export class SitemapValidator {
    private baseUrl: string;

    constructor(baseUrl: string) {
        this.baseUrl = baseUrl;
    }

    /**
     * 验证所有 URL 是否可访问
     */
    async validateUrls(): Promise<{
        valid: string[];
        invalid: string[];
        errors: Array<{ url: string; error: string }>;
    }> {
        const sitemap = generateSitemap(this.baseUrl);
        const valid: string[] = [];
        const invalid: string[] = [];
        const errors: Array<{ url: string; error: string }> = [];

        for (const entry of sitemap) {
            try {
                const response = await fetch(entry.url, { method: 'HEAD' });
                if (response.ok) {
                    valid.push(entry.url);
                } else {
                    invalid.push(entry.url);
                    errors.push({
                        url: entry.url,
                        error: `HTTP ${response.status}: ${response.statusText}`,
                    });
                }
            } catch (error) {
                invalid.push(entry.url);
                errors.push({
                    url: entry.url,
                    error: error instanceof Error ? error.message : 'Unknown error',
                });
            }
        }

        return { valid, invalid, errors };
    }

    /**
     * 检查重复的 URL
     */
    checkDuplicates(): string[] {
        const sitemap = generateSitemap(this.baseUrl);
        const urls = sitemap.map((entry) => entry.url);
        const duplicates: string[] = [];
        const seen = new Set<string>();

        for (const url of urls) {
            if (seen.has(url)) {
                duplicates.push(url);
            } else {
                seen.add(url);
            }
        }

        return duplicates;
    }

    /**
     * 生成站点地图报告
     */
    generateReport(): {
        stats: ReturnType<typeof getSitemapStats>;
        duplicates: string[];
        recommendations: string[];
    } {
        const stats = getSitemapStats(this.baseUrl);
        const duplicates = this.checkDuplicates();
        const recommendations: string[] = [];

        // 生成建议
        if (duplicates.length > 0) {
            recommendations.push(`发现 ${duplicates.length} 个重复的 URL，请检查站点地图生成逻辑`);
        }

        if (stats.urlCount > 50000) {
            recommendations.push('站点地图包含超过 50,000 个 URL，建议分割为多个站点地图文件');
        }

        if (stats.blogPages === 0) {
            recommendations.push('未发现博客页面，请确认博客内容是否正确配置');
        }

        return {
            stats,
            duplicates,
            recommendations,
        };
    }
}

/**
 * 站点地图提交工具
 */
export class SitemapSubmitter {
    private baseUrl: string;

    constructor(baseUrl: string) {
        this.baseUrl = baseUrl;
    }

    /**
     * 生成搜索引擎提交 URL
     */
    getSubmissionUrls(): Record<string, string> {
        const sitemapUrl = `${this.baseUrl}/sitemap.xml`;

        return {
            google: `https://www.google.com/ping?sitemap=${encodeURIComponent(sitemapUrl)}`,
            bing: `https://www.bing.com/ping?sitemap=${encodeURIComponent(sitemapUrl)}`,
            yandex: `https://webmaster.yandex.com/ping?sitemap=${encodeURIComponent(sitemapUrl)}`,
            baidu: `https://ping.baidu.com/ping/RPC2?sitemap=${encodeURIComponent(sitemapUrl)}`,
        };
    }

    /**
     * 自动提交站点地图到搜索引擎
     */
    async submitToSearchEngines(): Promise<{
        success: string[];
        failed: Array<{ engine: string; error: string }>;
    }> {
        const urls = this.getSubmissionUrls();
        const success: string[] = [];
        const failed: Array<{ engine: string; error: string }> = [];

        for (const [engine, url] of Object.entries(urls)) {
            try {
                const response = await fetch(url, { method: 'GET' });
                if (response.ok) {
                    success.push(engine);
                } else {
                    failed.push({
                        engine,
                        error: `HTTP ${response.status}: ${response.statusText}`,
                    });
                }
            } catch (error) {
                failed.push({
                    engine,
                    error: error instanceof Error ? error.message : 'Unknown error',
                });
            }
        }

        return { success, failed };
    }
}

/**
 * 生成站点地图索引文件
 */
export function generateSitemapIndex(baseUrl: string): string {
    const lastmod = new Date().toISOString();

    return `<?xml version="1.0" encoding="UTF-8"?>
<sitemapindex xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
  <sitemap>
    <loc>${baseUrl}/sitemap.xml</loc>
    <lastmod>${lastmod}</lastmod>
  </sitemap>
</sitemapindex>`;
}

/**
 * 生成人类可读的站点地图 HTML
 */
export function generateHtmlSitemap(baseUrl: string): string {
    const sitemap = generateSitemap(baseUrl);
    const stats = getSitemapStats(baseUrl);

    // 按页面类型分组
    const groupedPages = sitemap.reduce(
        (acc, entry) => {
            let category = 'other';

            if (entry.url.includes('/blog/') && !entry.url.endsWith('/blog')) {
                category = 'blog-posts';
            } else if (entry.url.endsWith('/blog')) {
                category = 'blog';
            } else if (entry.url.endsWith('/contact')) {
                category = 'contact';
            } else if (entry.url.endsWith('/products')) {
                category = 'products';
            } else if (entry.url === baseUrl || entry.url.match(/\/[a-z-]+$/)) {
                category = 'main';
            }

            if (!acc[category]) acc[category] = [];
            acc[category].push(entry);
            return acc;
        },
        {} as Record<string, typeof sitemap>,
    );

    const categoryNames = {
        main: '主要页面',
        blog: '博客',
        'blog-posts': '博客文章',
        contact: '联系我们',
        products: '产品',
        other: '其他页面',
    };

    let html = `<!DOCTYPE html>
<html lang="zh-CN">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>站点地图 - Eco Life</title>
    <style>
        body { font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif; margin: 40px; line-height: 1.6; }
        .header { border-bottom: 2px solid #10b981; padding-bottom: 20px; margin-bottom: 30px; }
        .stats { background: #f0fdf4; padding: 20px; border-radius: 8px; margin-bottom: 30px; }
        .category { margin-bottom: 30px; }
        .category h2 { color: #10b981; border-bottom: 1px solid #d1d5db; padding-bottom: 10px; }
        .url-list { list-style: none; padding: 0; }
        .url-item { margin: 8px 0; padding: 12px; background: #f9fafb; border-radius: 6px; }
        .url-item a { text-decoration: none; color: #1f2937; font-weight: 500; }
        .url-item a:hover { color: #10b981; }
        .url-meta { font-size: 0.875rem; color: #6b7280; margin-top: 4px; }
        .priority { display: inline-block; padding: 2px 8px; border-radius: 12px; font-size: 0.75rem; }
        .priority-high { background: #fef3c7; color: #92400e; }
        .priority-medium { background: #dbeafe; color: #1e40af; }
        .priority-low { background: #f3f4f6; color: #374151; }
    </style>
</head>
<body>
    <div class="header">
        <h1>站点地图</h1>
        <p>生成时间: ${new Date().toLocaleString('zh-CN')}</p>
    </div>
    
    <div class="stats">
        <h3>统计信息</h3>
        <p><strong>总页面数:</strong> ${stats.urlCount}</p>
        <p><strong>静态页面:</strong> ${stats.staticPages}</p>
        <p><strong>博客文章:</strong> ${stats.blogPages}</p>
        <p><strong>支持语言:</strong> ${stats.languages} (${locales.join(', ')})</p>
    </div>`;

    Object.entries(groupedPages).forEach(([category, pages]) => {
        html += `
    <div class="category">
        <h2>${categoryNames[category as keyof typeof categoryNames] || category}</h2>
        <ul class="url-list">`;

        pages.forEach((page) => {
            const priorityClass =
                page.priority >= 0.8
                    ? 'priority-high'
                    : page.priority >= 0.6
                      ? 'priority-medium'
                      : 'priority-low';

            html += `
            <li class="url-item">
                <a href="${page.url}" target="_blank">${page.url}</a>
                <div class="url-meta">
                    <span class="priority ${priorityClass}">优先级: ${page.priority}</span>
                    <span>更新频率: ${page.changeFrequency}</span>
                    <span>最后修改: ${page.lastModified.toLocaleDateString('zh-CN')}</span>
                </div>
            </li>`;
        });

        html += `
        </ul>
    </div>`;
    });

    html += `
</body>
</html>`;

    return html;
}