https://docs.qq.com/document/DY01OdndhR3NZaWVO

weixin.js

/**
 * 微信公众号「发表记录」全量抓取脚本
 *
 * 思路:
 *   后端区分 navigation 请求和 fetch 请求,fetch 拿到 SPA 骨架,只有真正
 *   navigate 才拿得到含数据的 HTML。所以用隐藏 iframe 模拟 navigation,
 *   等渲染完从 contentDocument 拿数据,iframe 立即销毁。
 *   主页面的 console / JS 上下文全程存活。
 *
 * 用法:
 *   1. 登录 mp.weixin.qq.com,打开「发表记录」页(任意一页)
 *   2. F12 → Console → 粘贴本脚本 → 回车
 *   3. 等待结束,自动下载 publish_history_*.json
 *
 * 跑完会在 console 底部输出 [诊断] 区块,如有页面失败 / 解析报错 / 未知状态,
 * 直接把诊断区块复制发给 Claude 修复。
 *
 * 关于 contentStart.js 的 NotFoundError:
 *   这是浏览器扩展(油猴/翻译/拦截类)注入到 iframe 的 content script
 *   在 iframe 销毁时处理不当,与本脚本无关,不影响数据抓取,可忽略。
 *
 * 自定义:
 *   dumpPublishHistory({ concurrency: 5, delayMs: 300, saveFormat: 'both' })
 */

(async function dumpPublishHistory(opts = {}) {
  const {
    count          = 10,
    concurrency    = 5,
    delayMs        = 300,
    iframeTimeout  = 20000,
    minStablePolls = 3,
    pollInterval   = 200,
    retryPerPage   = 1,      // 单页失败重试次数
    startPage      = 1,
    endPage        = null,
    saveFormat     = 'json',
  } = opts;

  const KNOWN_STATUSES = new Set([
    '已发表', '已群发',           // 正常(新/旧版)
    '无法查看', '已删除', '违规', // 不可见
    '审核中', '审核失败',          // 审核态
    '发表失败',                    // 失败
  ]);

  const sleep = (ms) => new Promise(r => setTimeout(r, ms));
  const text  = (el) => el ? el.textContent.replace(/\s+/g, ' ').trim() : null;
  const toNum = (s) => {
    if (s == null) return null;
    const t = String(s).replace(/[,¥￥\s]/g, '').trim();
    if (t === '') return null;
    const n = Number(t);
    return isNaN(n) ? null : n;   // 非数字直接 null,不要污染数据
  };

  // ---------- 提取 block 级共享元数据 ----------
  function extractSharedMeta(block) {
    const meta = {};
    meta.time = text(block.querySelector('.weui-desktop-mass__time'));
    if (meta.time) {
      const m = meta.time.match(/(\d{4})年(\d{1,2})月(\d{1,2})日/);
      meta.send_date = m ? `${m[1]}-${m[2].padStart(2,'0')}-${m[3].padStart(2,'0')}` : null;
    } else meta.send_date = null;

    const st = text(block.querySelector('.weui-desktop-mass__status_text'));
    meta.status        = st ? st.replace(/[▼▲↓↑]/g, '').trim() : null;
    meta.status_detail = text(block.querySelector('.sended_status_desc'));

    const notifyWording = text(block.querySelector('.sended_status_wording'));
    meta.notify_text = notifyWording || meta.status_detail;
    const sd = meta.status_detail || '';
    // 新版"已通知N人 失败N人" / 旧版"发送成功N人 发送失败N人"
    const mn1 = sd.match(/(?:已通知|发送成功)\s*(\d+)/);
    const mn2 = sd.match(/(?:发送)?失败\s*(\d+)/);
    meta.notified_count = mn1 ? parseInt(mn1[1]) : null;
    meta.notify_failed  = mn2 ? parseInt(mn2[1]) : null;

    if (notifyWording && notifyWording.includes('已开启通知')) meta.notify_mode = 'broadcast';
    else if (sd.includes('未开启群发')) meta.notify_mode = 'silent';
    else if (meta.status === '已群发' || meta.notified_count > 0) meta.notify_mode = 'broadcast';
    else meta.notify_mode = null;

    meta.is_broadcast = meta.notify_mode === 'broadcast';

    return meta;
  }

  // ---------- 解析单篇文章(scope 通常是 .publish_hover_content) ----------
  function parseArticle(scope, shared, articleIdx) {
    const item = { ...shared, article_idx: articleIdx, is_headline: articleIdx === 1 };

    // 标题 / URL
    const titleA    = scope.querySelector('a.weui-desktop-mass-appmsg__title');
    const titleSpan = scope.querySelector('.weui-desktop-mass-appmsg__title span');
    item.title = text(titleSpan) || text(titleA);
    item.url   = titleA ? titleA.getAttribute('href') : null;

    // 缩略图
    const thumb = scope.querySelector('.weui-desktop-mass-appmsg__thumb');
    if (thumb) {
      const m = (thumb.getAttribute('style') || '').match(/url\(["']?(.+?)["']?\)/);
      item.thumb = m ? m[1] : null;
    } else item.thumb = null;

    // 摘要 / 搜索描述片段
    item.digest = text(scope.querySelector('.weui-desktop-mass-appmsg__digest'));
    item.desc   = text(scope.querySelector('.weui-desktop-mass-appmsg__desc'));

    // 标签 + 文章类型
    const tags = Array.from(scope.querySelectorAll('.weui-desktop-key-tag'))
      .map(t => text(t)).filter(Boolean);
    item.tags         = tags;
    item.is_original  = tags.includes('原创');
    item.is_repost    = tags.includes('转载');
    item.is_modified  = tags.includes('已修改');
    item.article_type = item.is_original ? '原创'
                      : item.is_repost   ? '转载'
                                         : '正常';

    // 指标(兼容 __disable 版)
    const getMetric = (cls) => {
      const el = scope.querySelector(`.${cls} .weui-desktop-mass-media__data__inner`)
              || scope.querySelector(`.${cls}__disable .weui-desktop-mass-media__data__inner`);
      return el ? toNum(text(el)) : null;
    };
    item.view      = getMetric('appmsg-view');
    item.like      = getMetric('appmsg-like');
    item.share     = getMetric('appmsg-share');
    item.recommend = getMetric('appmsg-haokan');
    item.comment   = getMetric('appmsg-comment');
    item.underline = getMetric('appmsg-underline');
    item.reward    = getMetric('appmsg-reward');
    item.forward   = getMetric('appmsg-forward');

    // ID(兼容 appmsg_id / appmsgid / id 三种命名,以及 item_idx / idx)
    const linkWithId = scope.querySelector(
      '.appmsg-underline, .appmsg-underline__disable, .appmsg-reward, .appmsg-forward'
    );
    if (linkWithId) {
      const href = linkWithId.getAttribute('href') || '';
      const mId  = href.match(/[?&](?:appmsg_?id|id)=(\d+)/);
      const mIdx = href.match(/[?&](?:item_)?idx=(\d+)/);
      const mT   = href.match(/[?&]send_time=(\d+)/);
      item.appmsg_id     = mId  ? mId[1]  : null;
      item.item_idx      = mIdx ? mIdx[1] : null;
      item.send_time     = mT   ? parseInt(mT[1]) : null;
      item.send_time_iso = mT   ? new Date(parseInt(mT[1]) * 1000).toISOString() : null;
    } else {
      item.appmsg_id = null; item.item_idx = null;
      item.send_time = null; item.send_time_iso = null;
    }

    // 删除信息
    const tipsContent = text(scope.querySelector('.weui-desktop-mass-appmsg__tips_content'));
    if (tipsContent) {
      item.delete_info     = tipsContent;
      const mt = tipsContent.match(/删除时间\s*(.*?)(?=\s+操作人|$)/);
      const mo = tipsContent.match(/操作人\s*(\S+)/);
      item.delete_time     = mt ? mt[1].trim() : null;
      item.delete_operator = mo ? mo[1].trim() : null;
    } else {
      item.delete_info = null; item.delete_time = null; item.delete_operator = null;
    }

    // 视觉删除态
    const mediaContainer = scope.querySelector('.weui-desktop-mass-media');
    item.is_deleted_in_ui = !!(mediaContainer
      && mediaContainer.classList.contains('weui-desktop-mass-media_del'));

    // is_abnormal: 仅在脚本无法完整 cover 时为 true
    // 已知特殊状态(已删除/违规下架/发表失败等)只要识别到了就不算 abnormal
    const hasId = item.url || item.appmsg_id;
    const isKnownStatus = item.status && KNOWN_STATUSES.has(item.status);
    item.is_abnormal = !item.title || !hasId || !isKnownStatus;
    return item;
  }

  // ---------- 解析一个 block(可能含多篇文章) ----------
  function parseBlock(block) {
    const shared = extractSharedMeta(block);
    const articles = block.querySelectorAll('.publish_hover_content');
    if (!articles.length) {
      // 兜底:整个 block 当一篇
      return [parseArticle(block, shared, 1)];
    }
    return Array.from(articles).map((art, i) => parseArticle(art, shared, i + 1));
  }

  // ---------- URL 构造 ----------
  const curUrl = new URL(location.href);
  const token  = curUrl.searchParams.get('token');
  const lang   = curUrl.searchParams.get('lang') || 'zh_CN';
  if (!token) { console.error('当前 URL 拿不到 token,请确认在发表记录页'); return; }

  const buildUrl = (page) =>
    `https://mp.weixin.qq.com/cgi-bin/appmsgpublish?sub=list&begin=${(page-1)*count}&count=${count}&token=${token}&lang=${lang}`;

  // ---------- iframe 加载并抓数据 ----------
  async function loadPageViaIframe(pageNum) {
    return new Promise((resolve, reject) => {
      const iframe = document.createElement('iframe');
      iframe.style.cssText = 'position:fixed;left:-99999px;top:0;width:1280px;height:800px;border:0;visibility:hidden;pointer-events:none';
      iframe.setAttribute('aria-hidden', 'true');
      // sandbox 不含 allow-modals,禁用 alert/confirm/prompt 等弹窗
      // 否则页面内 alert 会阻塞主线程,所有 worker 一起卡死
      iframe.setAttribute('sandbox', 'allow-same-origin allow-scripts');

      let cleaned = false;
      let pollTimer = null;
      let lastCount = -1;
      let stableCount = 0;
      let polls = 0;

      const cleanup = () => {
        if (cleaned) return;
        cleaned = true;
        if (pollTimer) clearInterval(pollTimer);
        clearTimeout(timeoutTimer);
        try { iframe.src = 'about:blank'; } catch {}
        try { iframe.remove(); } catch {}
      };

      const timeoutTimer = setTimeout(() => {
        cleanup();
        reject(new Error(`page ${pageNum} 整体超时`));
      }, iframeTimeout);

      iframe.addEventListener('load', () => {
        // 兜底:覆盖 contentWindow 上的弹窗 API,以防 sandbox 在某些情况下没生效
        try {
          const cw = iframe.contentWindow;
          if (cw) {
            cw.alert   = function(){};
            cw.confirm = function(){ return true; };
            cw.prompt  = function(){ return null; };
          }
        } catch (e) {}

        pollTimer = setInterval(() => {
          polls++;
          let blocks = null;
          try {
            const doc = iframe.contentDocument;
            if (doc) {
              const root = doc.querySelector('.publish_record_history');
              if (root) blocks = root.querySelectorAll('.weui-desktop-block');
            }
          } catch (e) {
            cleanup(); reject(e); return;
          }

          const n = blocks ? blocks.length : 0;
          if (n > 0) {
            if (n === lastCount) {
              stableCount++;
              if (stableCount >= minStablePolls) {
                const items = [];
                for (const b of blocks) {
                  try {
                    items.push(...parseBlock(b));
                  } catch (e) {
                    items.push({
                      _parse_error: e.message,
                      _html_snippet: (b.outerHTML || '').slice(0, 800),
                      page: pageNum
                    });
                  }
                }
                cleanup();
                resolve(items);
                return;
              }
            } else {
              lastCount = n;
              stableCount = 0;
            }
          }

          if (polls * pollInterval > iframeTimeout - 1000) {
            cleanup();
            reject(new Error(`page ${pageNum} 渲染等不到数据(超时)`));
          }
        }, pollInterval);
      });

      iframe.addEventListener('error', (e) => {
        cleanup();
        reject(new Error(`iframe error: ${e.message || 'unknown'}`));
      });

      iframe.src = buildUrl(pageNum);
      document.body.appendChild(iframe);
    });
  }

  // 带重试的封装
  async function loadPageWithRetry(pageNum, workerId) {
    let lastErr = null;
    for (let attempt = 0; attempt <= retryPerPage; attempt++) {
      try {
        return await loadPageViaIframe(pageNum);
      } catch (e) {
        lastErr = e;
        if (attempt < retryPerPage) {
          console.warn(`  ⟳ ${pageNum} [w${workerId}] 第 ${attempt+1} 次失败,重试: ${e.message}`);
          await sleep(1000);
        }
      }
    }
    throw lastErr;
  }

  // ---------- 总页数 ----------
  function getTotalPagesFromHostPage() {
    const nums = Array.from(document.querySelectorAll('.weui-desktop-pagination__num'))
      .filter(n => !n.classList.contains('weui-desktop-pagination__ellipsis'));
    let maxNum = 1;
    nums.forEach(n => { const v = parseInt(text(n)); if (!isNaN(v) && v > maxNum) maxNum = v; });
    if (maxNum > 1) return maxNum;

    const buttons = document.querySelectorAll('button, span, div');
    for (const b of buttons) {
      const t = b.textContent || '';
      const m = t.match(/全部\s*(\d+)/);
      if (m) return Math.ceil(parseInt(m[1]) / count);
    }
    return null;
  }

  // ---------- 主流程 ----------
  console.log('%c[发表记录]', 'color:#0066cc;font-weight:bold');

  let totalPages = endPage || getTotalPagesFromHostPage();

  if (!totalPages) {
    console.log('  主页面拿不到总页数,iframe 探测中...');
    try {
      const firstItems = await loadPageWithRetry(startPage, 0);
      console.log(`  第一页 ${firstItems.length} 条`);
      totalPages = 1;
    } catch (e) {
      console.error('  连第一页都加载失败:', e.message);
      return;
    }
  }

  console.log(`  总页数 ${totalPages}, 并发 ${concurrency}`);

  const allPages = [];
  for (let p = startPage; p <= totalPages; p++) allPages.push(p);

  const results = [];
  const errors  = [];
  let cursor = 0;
  let completed = 0;

  async function worker(workerId) {
    while (cursor < allPages.length) {
      const p = allPages[cursor++];
      try {
        const items = await loadPageWithRetry(p, workerId);
        items.forEach(i => i.page = p);
        results.push(...items);
        completed++;
        const abn = items.filter(i => i.is_abnormal).length;
        console.log(`  ✓ ${p}/${totalPages} [w${workerId}] — ${items.length}条 (异常 ${abn}) — 已完成 ${completed}/${allPages.length} 累计 ${results.length}`);
      } catch (e) {
        completed++;
        console.error(`  ✗ ${p}/${totalPages} [w${workerId}] (重试后仍失败):`, e.message);
        errors.push({ page: p, error: e.message });
      }
      await sleep(delayMs);
    }
  }

  const t0 = Date.now();
  await Promise.all(Array.from({ length: concurrency }, (_, i) => worker(i + 1)));
  const elapsed = ((Date.now() - t0) / 1000).toFixed(1);

  results.sort((a, b) => {
    if (a.page !== b.page) return a.page - b.page;
    if (a.article_idx !== b.article_idx) return (a.article_idx || 0) - (b.article_idx || 0);
    return (b.send_time || 0) - (a.send_time || 0);
  });

  // ---------- send_date 回填 ----------
  // Pass 1: 从 send_time_iso 派生 (来自 underline/reward URL 的 unix timestamp,可靠)
  // Pass 2: 上下文 + delete_time 推断 (主要处理已删除/违规下架文章)
  const parseDelDate = s => {
    if (!s) return null;
    const m = s.match(/(\d{4})年(\d{2})月(\d{2})日/);
    return m ? `${m[1]}-${m[2]}-${m[3]}` : null;
  };
  let bf1 = 0, bfDel = 0, bfNeighbor = 0, bfFail = 0;
  // Pass 1
  for (const x of results) {
    if (!x.send_date && x.send_time_iso) {
      x.send_date = x.send_time_iso.slice(0, 10);
      x.send_date_inferred = 'send_time';
      bf1++;
    }
  }
  // Pass 2 (数组已按时间倒序: page asc 对应最新在前)
  for (let i = 0; i < results.length; i++) {
    const x = results[i];
    if (x.send_date) continue;
    let newer = null, older = null;
    for (let j = i - 1; j >= 0; j--) if (results[j].send_date) { newer = results[j].send_date; break; }
    for (let j = i + 1; j < results.length; j++) if (results[j].send_date) { older = results[j].send_date; break; }
    const delDate = parseDelDate(x.delete_time);
    if (delDate && (older === null || delDate >= older) && (newer === null || delDate <= newer)) {
      x.send_date = delDate;
      x.send_date_inferred = 'delete_time';
      bfDel++;
    } else if (older) {
      x.send_date = older;
      x.send_date_inferred = 'older_neighbor';
      bfNeighbor++;
    } else if (newer) {
      x.send_date = newer;
      x.send_date_inferred = 'newer_neighbor';
      bfNeighbor++;
    } else {
      bfFail++;
    }
  }
  if (bf1 + bfDel + bfNeighbor + bfFail > 0) {
    console.log(`  send_date 回填: send_time=${bf1}, delete_time=${bfDel}, 邻居=${bfNeighbor}, 仍缺=${bfFail}`);
  }

  const totalAbn = results.filter(i => i.is_abnormal).length;
  const byType = results.reduce((acc, r) => {
    const k = r.article_type || '_未知';
    acc[k] = (acc[k] || 0) + 1; return acc;
  }, {});

  console.log(`%c[完成] ${results.length} 篇,异常态 ${totalAbn} 篇,失败页 ${errors.length},耗时 ${elapsed}s`,
              'color:#00aa00;font-weight:bold');
  console.log(`  类型分布:`, byType);

  const ids = results.map(r => r.appmsg_id).filter(Boolean);
  if (ids.length !== new Set(ids).size) {
    // 注意:同一 appmsg_id 的头条+次条共用 ID,所以重复可能是正常的,看 (appmsg_id, item_idx) 才唯一
    const pairs = results.map(r => r.appmsg_id && `${r.appmsg_id}_${r.item_idx || r.article_idx}`).filter(Boolean);
    if (pairs.length !== new Set(pairs).size) {
      console.warn(`%c⚠ (appmsg_id, idx) 仍重复: ${pairs.length} → ${new Set(pairs).size}`,
                   'color:orange;font-weight:bold');
    }
  }

  // ---------- 诊断报告 ----------
  const parseErrors    = results.filter(r => r._parse_error);
  const unknownStatuses = [...new Set(
    results.filter(r => r.status && !KNOWN_STATUSES.has(r.status)).map(r => r.status)
  )];
  const hasDiag = errors.length || parseErrors.length || unknownStatuses.length;

  if (hasDiag) {
    console.group(`%c[诊断] 复制以下内容发给 Claude 修复 ↓`, 'color:#cc6600;font-weight:bold;font-size:13px');

    if (errors.length) {
      console.group(`页面级失败 ${errors.length} 个`);
      errors.forEach(e => console.log(`page ${e.page}: ${e.error}`));
      console.groupEnd();
    }

    if (parseErrors.length) {
      console.group(`解析失败 ${parseErrors.length} 条`);
      parseErrors.forEach((e, i) => {
        console.log(`── #${i+1} page=${e.page} ── ${e._parse_error}`);
        console.log(e._html_snippet);
      });
      console.groupEnd();
    }

    if (unknownStatuses.length) {
      console.group(`未知状态文本 ${unknownStatuses.length} 种(可能需要扩展 KNOWN_STATUSES)`);
      unknownStatuses.forEach(s => {
        const sample = results.find(r => r.status === s);
        console.log(`"${s}" — 样本: ${sample.title || '(无标题)'} (page ${sample.page})`);
      });
      console.groupEnd();
    }

    console.groupEnd();
  } else {
    console.log('%c[诊断] 无异常 ✓', 'color:#00aa00');
  }

  window.__publish_dump   = results;
  window.__publish_errors = errors;
  window.__publish_diag   = { parseErrors, errors, unknownStatuses };
  console.log('结果: window.__publish_dump  失败页: window.__publish_errors  诊断: window.__publish_diag');
  if (errors.length) {
    console.log('补抓失败页: dumpPublishHistory({ startPage: N, endPage: N })');
  }

  // ---------- 下载 ----------
  const stamp = new Date().toISOString().slice(0, 19).replace(/[T:]/g, '-');
  function downloadBlob(content, filename, mime) {
    const blob = new Blob([content], { type: mime });
    const url  = URL.createObjectURL(blob);
    const a    = document.createElement('a');
    a.href = url; a.download = filename;
    document.body.appendChild(a); a.click(); a.remove();
    URL.revokeObjectURL(url);
  }
  function toCSV(rows) {
    if (!rows.length) return '';
    const cols = Array.from(new Set(rows.flatMap(r => Object.keys(r))));
    const esc = (v) => {
      if (v == null) return '';
      const s = (typeof v === 'string') ? v
              : (Array.isArray(v) ? v.join('|') : JSON.stringify(v));
      return /[",\n\r]/.test(s) ? '"' + s.replace(/"/g, '""') + '"' : s;
    };
    return [cols.join(','), ...rows.map(r => cols.map(c => esc(r[c])).join(','))].join('\n');
  }

  if (saveFormat === 'json' || saveFormat === 'both')
    downloadBlob(JSON.stringify(results, null, 2),
                 `publish_history_${stamp}.json`, 'application/json');
  if (saveFormat === 'csv' || saveFormat === 'both')
    downloadBlob('\ufeff' + toCSV(results),
                 `publish_history_${stamp}.csv`, 'text/csv;charset=utf-8');

  return results;
})();