blog/blog-admin/src/routes/rss/tools.ts

// /api/favicon —— 自动发现并返回站点 favicon
// /api/update —— 接收抓取结果写入 KV（scheduled 已内置，此端点保留兼容外部上报）
// /api/proxy —— RSS 抓取代理（EdgeOne 版迁过来；注意本 Worker 在境外，等价于原 EdgeOne）
// 迁自 edgeone/functions/api/{favicon,update,proxy}.js

import type { Env } from '../../types';
import { respond, requireAuth } from '../../lib/rss/util';
import { notifyAdmin } from '../../lib/admin-notify';
import { checkProxyHealth, notifyProxyDown } from '../../lib/rss/proxy-health';
import { kvGetJson, kvPutJson } from '../../lib/rss/util';

const CACHE_TTL = 604800;

function resolveUrl(href: string, baseUrl: string): string | null {
  if (href.startsWith('http://') || href.startsWith('https://')) return href;
  try {
    return new URL(href, baseUrl).href;
  } catch {
    return null;
  }
}

async function discoverFaviconUrl(siteUrl: string): Promise<string | null> {
  if (!/^https?:\/\//i.test(siteUrl)) siteUrl = 'https://' + siteUrl;

  try {
    const ctrl = new AbortController();
    const t = setTimeout(() => ctrl.abort(), 12000);
    const res = await fetch(siteUrl, {
      signal: ctrl.signal,
      redirect: 'follow',
      headers: {
        'User-Agent': 'Mozilla/5.0 (compatible; RSSBot/1.0)',
        Accept: 'text/html,application/xhtml+xml,*/*',
      },
    });
    clearTimeout(t);
    const html = await res.text();

    const iconPatterns = [
      /<link[^>]+rel=["'](?:shortcut )?icon["'][^>]+href=["']([^"']+)["']/i,
      /<link[^>]+href=["']([^"']+)["'][^>]+rel=["'](?:shortcut )?icon["']/i,
      /<link[^>]+rel=["']apple-touch-icon["'][^>]+href=["']([^"']+)["']/i,
      /<link[^>]+href=["']([^"']+)["'][^>]+rel=["']apple-touch-icon["']/i,
    ];
    for (const pattern of iconPatterns) {
      const match = html.match(pattern);
      if (match) return resolveUrl(match[1].trim(), siteUrl);
    }

    const urlObj = new URL(siteUrl);
    const defaultFavicon = `${urlObj.protocol}//${urlObj.host}/favicon.ico`;
    try {
      const ctrl2 = new AbortController();
      const t2 = setTimeout(() => ctrl2.abort(), 8000);
      const check = await fetch(defaultFavicon, { method: 'HEAD', signal: ctrl2.signal, redirect: 'follow' });
      clearTimeout(t2);
      const ct = (check.headers.get('content-type') || '').toLowerCase();
      // 必须确认返回的真是图片：Typecho/WordPress 等会把 404 页面以 200 + text/html 返回
      if (check.ok && (ct.startsWith('image/') || ct.includes('octet-stream'))) return defaultFavicon;
    } catch {
      /* 继续走兜底 */
    }
    // favicon.ico 不通或返回的不是图片 → 兜底第三方 favicon 服务（国内可达），
    // 避免友圈卡片大面积落回灰色默认头像
    return `https://api.iowen.cn/favicon/${urlObj.host}.png`;
  } catch {
    try {
      const urlObj = new URL(siteUrl);
      // 同上：抓取失败兜底第三方服务
      return `https://api.iowen.cn/favicon/${urlObj.host}.png`;
    } catch {
      return null;
    }
  }
}

/**
 * favicon 图片响应的统一缓存头。
 *
 * ★ 为什么要显式加 s-maxage / CDN-Cache-Control：
 *   `Cache-Control: max-age=86400` 只让**浏览器**缓存，Cloudflare 边缘默认
 *   **不缓存 Worker 的响应**（实测响应里没有 cf-cache-status）。于是每个访客、
 *   每次清缓存、每换设备，都会真正进 Worker → 第一条语句就是 KV.get → KV 读量被放大。
 *
 *   加上 s-maxage（共享缓存）后，CF 边缘可缓存 24h：边缘命中就完全不进 Worker，
 *   KV 读取 ≈ 0，同时首包更快。浏览器侧仍保留 max-age。
 *   只对**成功拿到图片**的响应使用；失败/兜底响应不缓存（避免把错误固化一天）。
 */
function imageCacheHeaders(contentType: string): Record<string, string> {
  return {
    'Content-Type': contentType,
    'Cache-Control': 'public, max-age=86400, s-maxage=86400',
    'CDN-Cache-Control': 'max-age=86400',
    'Cloudflare-CDN-Cache-Control': 'max-age=86400',
    'Access-Control-Allow-Origin': '*',
  };
}

export async function favicon(request: Request, env: Env): Promise<Response> {
  const url = new URL(request.url);
  const siteUrl = url.searchParams.get('url');
  if (!siteUrl) return respond({ error: 'url parameter required' }, 400);

  const bodyKey = 'favicon:body:v2:' + siteUrl;

  // 一级缓存：图片本体直接存 KV（base64）——命中后零外站回源，响应最快
  const cached = await env.RSS_KV.get(bodyKey);
  if (cached) {
    const [meta, b64] = cached.split('|');
    const body = Uint8Array.from(atob(b64), (ch) => ch.charCodeAt(0));
    return new Response(body, { status: 200, headers: imageCacheHeaders(meta) });
  }

  // 失败短缓存：抓取失败的站点 1 小时内不再反复打外站
  const missKey = 'favicon:miss:' + siteUrl;
  if (await env.RSS_KV.get(missKey)) {
    return letterSvg(siteUrl);
  }

  if (url.searchParams.get('meta')) {
    const u = await discoverFaviconUrl(siteUrl);
    return respond({ url: u });
  }

  // 发现 + 下载图片本体（兼容读旧版 URL 字符串缓存：存量友链的发现结果都在旧 key 里）
  let faviconUrl: string | null = await env.RSS_KV.get('favicon:' + siteUrl).catch(() => null);
  if (!faviconUrl) {
    try {
      faviconUrl = await discoverFaviconUrl(siteUrl);
      if (faviconUrl) await env.RSS_KV.put('favicon:' + siteUrl, faviconUrl, { expirationTtl: CACHE_TTL });
    } catch {
      faviconUrl = null;
    }
  }

  if (faviconUrl) {
    try {
      const ctrl = new AbortController();
      const t = setTimeout(() => ctrl.abort(), 10000);
      const imgRes = await fetch(faviconUrl, {
        signal: ctrl.signal,
        redirect: 'follow',
        headers: {
          'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/126.0.0.0 Safari/537.36',
          Accept: 'image/avif,image/webp,image/apng,image/*,*/*;q=0.8',
        },
      });
      clearTimeout(t);
      if (imgRes.ok) {
        const rawCt = (imgRes.headers.get('content-type') || '').toLowerCase();
        // 只接受真正的图片（含 octet-stream 的 .ico）；HTML/文本一律视为抓取失败
        if (!rawCt.startsWith('image/') && !rawCt.includes('octet-stream')) {
          throw new Error('favicon not an image: ' + rawCt);
        }
        const contentType = rawCt || 'image/x-icon';
        const buf = await imgRes.arrayBuffer();
        // 只缓存小于 300KB 的图（KV 单值上限 25MB，异常大图不占空间）
        if (buf.byteLength <= 307200) {
          let bin = '';
          const bytes = new Uint8Array(buf);
          const chunk = 0x8000;
          for (let i = 0; i < bytes.length; i += chunk) {
            bin += String.fromCharCode.apply(null, Array.from(bytes.subarray(i, i + chunk)) as unknown as number[]);
          }
          await env.RSS_KV.put(bodyKey, `${contentType}|${btoa(bin)}`, { expirationTtl: CACHE_TTL });
        }
        return new Response(buf, { status: 200, headers: imageCacheHeaders(contentType) });
      }
    } catch {
      /* 落到第三方兜底 */
    }
  }

  // 兜底链：第三方 favicon 服务代理返回（仍尝试写缓存，下次直接命中）
  let host = siteUrl.replace(/^https?:\/\//, '').split('/')[0];
  const fallback = `https://api.iowen.cn/favicon/${host}.png`;
  try {
    const ctrl3 = new AbortController();
    const t3 = setTimeout(() => ctrl3.abort(), 10000);
    const fRes = await fetch(fallback, { signal: ctrl3.signal, redirect: 'follow' });
    clearTimeout(t3);
    if (fRes.ok) {
      const contentType = fRes.headers.get('content-type') || 'image/png';
      const buf = await fRes.arrayBuffer();
      let bin = '';
      const bytes = new Uint8Array(buf);
      const chunk = 0x8000;
      for (let i = 0; i < bytes.length; i += chunk) {
        bin += String.fromCharCode.apply(null, Array.from(bytes.subarray(i, i + chunk)) as unknown as number[]);
      }
      await env.RSS_KV.put(bodyKey, `${contentType}|${btoa(bin)}`, { expirationTtl: CACHE_TTL });
      return new Response(buf, { status: 200, headers: imageCacheHeaders(contentType) });
    }
  } catch {
    /* 彻底失败 */
  }

  // 记一次失败，1 小时内不再打外站；返回自绘字母 SVG 兜底
  // （零依赖永远可用，比第三方服务和灰色默认头像都体面）
  await env.RSS_KV.put(missKey, '1', { expirationTtl: 3600 });
  return letterSvg(siteUrl);
}

/** 站点首字母占位图（永远 200 的最后兜底） */
function letterSvg(siteUrl: string): Response {
  let host = siteUrl.replace(/^https?:\/\//, '').split('/')[0].replace(/^www\./, '');
  const letter = (host[0] || '?').toUpperCase();
  const svg =
    `<svg xmlns="http://www.w3.org/2000/svg" width="64" height="64">` +
    `<rect width="64" height="64" rx="12" fill="#2f9e63"/>` +
    `<text x="32" y="43" font-family="Arial,sans-serif" font-size="32" font-weight="600" fill="#ffffff" text-anchor="middle">${letter}</text></svg>`;
  return new Response(svg, {
    status: 200,
    headers: {
      // 兜底图也缓存（否则每个访客都要重跑一遍抓取失败流程 = 3 次 KV 读 + 1 次写）
      'Content-Type': 'image/svg+xml',
      'Cache-Control': 'public, max-age=86400, s-maxage=86400',
      'CDN-Cache-Control': 'max-age=86400',
      'Cloudflare-CDN-Cache-Control': 'max-age=86400',
      'Access-Control-Allow-Origin': '*',
    },
  });
}

export async function update(request: Request, env: Env): Promise<Response> {
  if (request.method !== 'POST') return respond({ error: 'POST only' }, 405);
  if (!(await requireAuth(request, env))) return respond({ error: 'unauthorized' }, 401);
  try {
    const data = await request.json();
    await env.RSS_KV.put('latest', JSON.stringify(data), { expirationTtl: 604800 });
    return respond({ ok: true });
  } catch (err) {
    return respond({ error: (err as Error).message }, 500);
  }
}

/**
 * POST /api/notify-mail —— 通用提醒邮件（给自己发一封）
 * body: { subject, text, dedupeKey?, minGapHours? }
 * 鉴权：requireAuth；收件人固定 MAIL_ADMIN，不能指定别人。
 */
export async function notifyMail(request: Request, env: Env): Promise<Response> {
  if (request.method !== 'POST') return respond({ error: 'POST only' }, 405);
  if (!(await requireAuth(request, env))) return respond({ error: 'unauthorized' }, 401);

  let body: any;
  try {
    body = await request.json();
  } catch {
    return respond({ error: 'invalid json' }, 400);
  }

  const result = await notifyAdmin(env, {
    subject: String(body?.subject || '提醒'),
    text: String(body?.text || ''),
    dedupeKey: body?.dedupeKey,
    minGapHours: Number(body?.minGapHours) || 0,
  });
  if (result.error) return respond({ error: result.error }, 400);
  return respond(result);
}

/**
 * GET /api/proxy-health —— 国内代理（scfapi.usj.cc）体检
 * ?notify=1 时若不健康还会给站长发一封提醒邮件（48h 节流）。
 */
export async function proxyHealth(request: Request, env: Env): Promise<Response> {
  if (!(await requireAuth(request, env))) return respond({ error: 'unauthorized' }, 401);
  const health = await checkProxyHealth(env);
  let mailed: unknown = null;
  if (!health.ok && new URL(request.url).searchParams.get('notify') === '1') {
    mailed = await notifyProxyDown(env, health);
  }
  return respond({ ...health, mailed });
}

export async function proxy(request: Request, env: Env): Promise<Response> {
  if (request.method === 'OPTIONS') {
    return new Response(null, { status: 204, headers: { 'Access-Control-Allow-Origin': '*', 'Access-Control-Allow-Methods': 'POST, OPTIONS', 'Access-Control-Allow-Headers': 'Content-Type' } });
  }
  if (request.method !== 'POST') return respond({ error: 'POST only' }, 405);
  if (!(await requireAuth(request, env))) return respond({ error: 'unauthorized' }, 401);

  let body: any;
  try {
    body = await request.json();
  } catch {
    return respond({ error: 'invalid json' }, 400);
  }
  const { url, timeout = 15000 } = body;
  if (!url) return respond({ error: 'url required' }, 400);

  const controller = new AbortController();
  const timer = setTimeout(() => controller.abort(), timeout);
  try {
    const targetRes = await fetch(url, {
      signal: controller.signal,
      redirect: 'follow',
      headers: {
        'User-Agent': 'Mozilla/5.0 (compatible; RSSBot/1.0)',
        Accept: 'text/html,application/xhtml+xml,application/xml,application/json;q=0.9,*/*;q=0.8',
      },
    });
    clearTimeout(timer);
    const text = await targetRes.text();
    const ct = targetRes.headers.get('content-type') || '';
    return respond({ ok: true, status: targetRes.status, contentType: ct, body: text });
  } catch (err) {
    return respond({ ok: false, error: (err as Error).message }, 502);
  }
}

/**
 * POST /api/cron/run?offset=0&limit=35 —— 手动触发一批抓取（token 鉴权）。
 * 免费版 subrequest 限制下用于补跑/首灌数据。
 */
export async function cronRun(request: Request, env: Env): Promise<Response> {
  if (!(await requireAuth(request, env))) return respond({ error: 'unauthorized' }, 401);
  const url = new URL(request.url);
  const offset = Math.max(parseInt(url.searchParams.get('offset') || '0', 10) || 0, 0);
  const limit = Math.max(parseInt(url.searchParams.get('limit') || '35', 10) || 0, 0);
  // dry=1：只跑抓取、不写 KV、不发通知 —— 用于安全地测耗时/成功率（不会污染线上数据）
  const dry = url.searchParams.get('dry') === '1';
  // rotate=1：从 KV 游标继续抓（模拟定时任务的行为）
  const rotate = url.searchParams.get('rotate') === '1';
  // 抓取主要是 IO 等待（不计 CPU），HTTP 触发也没有时长上限，35 个源串行能跑完
  const { runCron } = await import('../../lib/rss/cron');
  const stats = await runCron(env, { offset, limit, dry, rotate });
  return respond({ ok: true, offset, limit, stats });
}

// ============================================================================
// 部署状态（国内中转机同步监控）
// CI 的 build job 在 COS 上传后上报 phase=built；中转机 sync.sh 同步完成后
// 上报 phase=synced。GET 公开可查，对比两者就知道国内线路部署完了没。
// ============================================================================

interface DeployPhase {
  hash: string;
  phase: string;
  at: string;
}

export async function deployStatusGet(_request: Request, env: Env): Promise<Response> {
  const st = await kvGetJson<{ builds: DeployPhase[] }>(env, 'deploy_status', { builds: [] });
  const builds = (st.builds || []).slice(0, 5);
  const built = builds.find((b) => b.phase === 'built');
  const synced = builds.find((b) => b.phase === 'synced');
  let state = 'unknown';
  if (built && synced && synced.hash === built.hash) state = 'synced';
  else if (built && !synced) state = 'syncing';
  else if (synced && !built) state = 'synced'; // 未配置 CI 上报时，以中转机记录为准
  else if (built && synced && synced.hash !== built.hash) state = 'syncing';
  return respond({
    state, // synced = 国内已生效；syncing = 构建完成等待中转机同步
    latest_build: built || null,
    latest_sync: synced || null,
    builds,
  });
}

export async function deployStatusPost(request: Request, env: Env): Promise<Response> {
  if (!(await requireAuth(request, env))) return respond({ error: 'unauthorized' }, 401);
  let body: { hash?: string; phase?: string } = {};
  try {
    body = (await request.json()) as { hash?: string; phase?: string };
  } catch {
    return respond({ error: 'invalid json' }, 400);
  }
  const hash = String(body.hash || '').slice(0, 64);
  const phase = String(body.phase || '').slice(0, 20);
  if (!hash || !phase) return respond({ error: 'hash and phase required' }, 400);

  const st = await kvGetJson<{ builds: DeployPhase[] }>(env, 'deploy_status', { builds: [] });
  st.builds = (st.builds || []).filter((b) => !(b.phase === phase && b.hash === hash));
  st.builds.unshift({ hash, phase, at: new Date().toISOString() });
  st.builds = st.builds.slice(0, 20);
  await kvPutJson(env, 'deploy_status', st);
  return respond({ ok: true });
}


/**
 * POST /api/deploy-notify —— 部署完成邮件（token 鉴权）。
 * 历史用途：与广州中转机 `sync.sh` 配套 —— 构建通知告诉你"构建好了"，
 * 这封邮件告诉你"国内线路真正生效了"。
 * （2026-10-04 起国内线路改由 CNB 国内节点直传又拍云，该轮询链路已停用；
 *   端点保留以兼容旧调用方。）
 */
export async function deployNotify(request: Request, env: Env): Promise<Response> {
  if (!(await requireAuth(request, env))) return respond({ error: 'unauthorized' }, 401);
  let body: { hash?: string } = {};
  try {
    body = (await request.json()) as { hash?: string };
  } catch {
    return respond({ error: 'invalid json' }, 400);
  }
  const hash = String(body.hash || '').slice(0, 12) || '未知';
  const { sendMail } = await import('../../lib/mail');
  const okMail = await sendMail(env, {
    to: 'imql@qq.com',
    subject: `✅ 博客国内线路部署完成 ${hash}`,
    text: `构建产物 ${hash} 已同步到又拍云并刷新多吉云 CDN，国内线路已生效。\n时间：${new Date().toISOString()}`,
    html:
      `<p>构建产物 <code>${hash}</code> 已同步到又拍云并刷新多吉云 CDN，<b>国内线路已生效</b>。</p>` +
      `<p style="color:#8a8e98">时间：${new Date().toISOString()} · artalk-cf deploy monitor</p>`,
  });
  return respond({ ok: okMail });
}