From 2049bc8c6e79d4fc9b27b5440a73965149da28db Mon Sep 17 00:00:00 2001
From: john <cynell@139.com>
Date: Mon, 21 Sep 2026 09:31:54 +0800
Subject: [PATCH] fix(news): count lead cards in degradation check, add draft
 dry-run script

Trying to push a WeChat draft locally surfaced two problems.

isNewsMorningContentDegraded matched cards with `class="card(?:\s|")`, which
never matches news002's `class="lead-card"`. Headline cards were therefore
invisible to the check, so a healthy news002 page could be flagged as degraded
and regenerated. Count cards through the shared extractor instead, which
handles card, lead-card and highlight-box alike.

Add scripts/dryrun-news-morning-draft.mjs to run a page through the dedup gate,
degradation check, inline conversion and footer layout without calling the
WeChat API, since local IPs are not in the official account whitelist
(errcode 40164) and a real push would write into the production draft box.

Co-authored-by: Cursor <cursoragent@cursor.com>
---
 scripts/dryrun-news-morning-draft.mjs | 85 +++++++++++++++++++++++++++
 wechat-news-morning-dedup.mjs         |  7 ++-
 wechat-news-morning-dedup.test.mjs    | 13 ++++
 wechat-news-morning-draft.mjs         |  7 ++-
 wechat-news-morning-draft.test.mjs    | 13 ++++
 5 files changed, 121 insertions(+), 4 deletions(-)
 create mode 100644 scripts/dryrun-news-morning-draft.mjs

diff --git a/scripts/dryrun-news-morning-draft.mjs b/scripts/dryrun-news-morning-draft.mjs
new file mode 100644
index 0000000..0506c0c
--- /dev/null
+++ b/scripts/dryrun-news-morning-draft.mjs
@@ -0,0 +1,85 @@
+#!/usr/bin/env node
+/**
+ * 本地 dry-run：把一张早报 HTML 走完「去重闸门 → 微信 inline 转换 → 统一页脚」全链路，
+ * 输出最终草稿正文，但不调用任何微信接口、不落 DB。
+ *
+ * 用于在本机 IP 不在公众号白名单时验证 news001 / news002 的草稿产出。
+ *
+ * 用法：
+ *   node scripts/dryrun-news-morning-draft.mjs <页面.html> [输出目录]
+ */
+import fs from 'node:fs';
+import path from 'node:path';
+import {
+  dedupeNewsMorningHtml,
+  formatNewsMorningDedupeSummary,
+} from '../wechat-news-morning-dedup.mjs';
+import {
+  convertNewsPageHtmlToWechatArticle,
+  isDailyNewsFormat,
+  isNewsMorningContentDegraded,
+} from '../wechat-news-morning-draft.mjs';
+
+const MAX_WECHAT_CONTENT_CHARS = 20000;
+
+const input = process.argv[2];
+const outDir = process.argv[3] || path.dirname(input ?? '.');
+if (!input || !fs.existsSync(input)) {
+  console.error('用法：node scripts/dryrun-news-morning-draft.mjs <页面.html> [输出目录]');
+  process.exit(1);
+}
+
+const publicUrl = 'https://m.tkmind.cn/MindSpace/dry-run/public/daily-news-dryrun.html';
+const raw = fs.readFileSync(input, 'utf8');
+
+console.log(`输入：${input}（${raw.length} 字符）`);
+
+// 1. 去重闸门
+const deduped = dedupeNewsMorningHtml(raw);
+console.log(`\n[1] 去重闸门：${formatNewsMorningDedupeSummary(deduped)}`);
+if (deduped.removedSectionIds.length) {
+  console.log(`    移除空栏目：${deduped.removedSectionIds.join(', ')}`);
+}
+
+// 2. 降级检测
+const degraded = isNewsMorningContentDegraded(deduped.html);
+console.log(`\n[2] 降级检测：${degraded ? '❌ 判定为降级页（卡片不足/检索失败）' : '✅ 正常'}`);
+
+// 3. 识别版式 + 转换为微信草稿
+console.log(`\n[3] 版式识别：${isDailyNewsFormat(deduped.html) ? '✅ daily-news' : '❌ 非早报版式'}`);
+
+const article = convertNewsPageHtmlToWechatArticle(
+  deduped.html,
+  { publicUrl, qrcodeImageUrl: 'https://mmbiz.qpic.cn/dryrun-qrcode.png' },
+);
+
+console.log('\n[4] 草稿产出：');
+console.log(`    标题：${article.title}`);
+console.log(`    摘要：${String(article.digest ?? '').slice(0, 60)}`);
+console.log(`    正文长度：${article.content.length} / ${MAX_WECHAT_CONTENT_CHARS}`
+  + (article.content.length > MAX_WECHAT_CONTENT_CHARS ? '  ❌ 超限' : '  ✅ 未超限'));
+
+const checks = [
+  // 微信不支持 <style>，早报路径必须全部内联。
+  ['无 <style> 块（已内联）', !/<style/i.test(article.content)],
+  ['正文使用内联 style', (article.content.match(/style="/g) ?? []).length > 20],
+  ['含「阅读原文」引导', /阅读原文/u.test(article.content)],
+  ['含关注二维码区', /<img[^>]+dryrun-qrcode/i.test(article.content)],
+  ['无 <script> 残留', !/<script/i.test(article.content)],
+  ['无 MindSpace 平台水印', !/data-mindspace-page-tag/i.test(article.content)],
+];
+console.log('\n[5] 合规检查：');
+for (const [label, ok] of checks) console.log(`    ${ok ? '✅' : '❌'} ${label}`);
+
+const base = path.basename(input, '.html');
+const outHtml = path.join(outDir, `${base}.wechat-draft.html`);
+fs.writeFileSync(outHtml, `<!DOCTYPE html><html><head><meta charset="utf-8">
+<meta name="viewport" content="width=device-width,initial-scale=1">
+<title>${article.title}（微信草稿 dry-run）</title></head>${article.content}</html>`);
+console.log(`\n草稿正文已写入：${outHtml}`);
+
+const outDeduped = path.join(outDir, `${base}.deduped.html`);
+if (deduped.changed) {
+  fs.writeFileSync(outDeduped, deduped.html);
+  console.log(`去重后页面已写入：${outDeduped}`);
+}
diff --git a/wechat-news-morning-dedup.mjs b/wechat-news-morning-dedup.mjs
index d66a9c2..a24851d 100644
--- a/wechat-news-morning-dedup.mjs
+++ b/wechat-news-morning-dedup.mjs
@@ -17,7 +17,7 @@ import {
 // 必须按 class token 精确匹配：`\bcard\b` 会误命中 weather-card / knowledge-card 等。
 const CARD_OPEN_RE = /<(div|a)\b[^>]*\bclass="([^"]*)"[^>]*>/gi;
 const SECTION_OPEN_RE = /<div\b[^>]*class="[^"]*\bsection\b[^"]*"[^>]*id="([^"]+)"[^>]*>/gi;
-const CARD_CLASS_TOKENS = new Set(['card', 'lead-card']);
+const CARD_CLASS_TOKENS = new Set(['card', 'lead-card', 'highlight-box']);
 
 function hasCardClass(classAttr) {
   return String(classAttr).split(/\s+/).some((token) => CARD_CLASS_TOKENS.has(token));
@@ -138,6 +138,11 @@ export function extractNewsMorningCards(html) {
   return cards;
 }
 
+/** 统计页面里的新闻卡片数（含 news001 的 .card/.highlight-box 与 news002 的 .lead-card）。 */
+export function countNewsMorningStoryCards(html) {
+  return extractNewsMorningCards(html).filter((card) => card.title).length;
+}
+
 /** 提取 section 区块边界，用于判断某个栏目是否被清空。 */
 export function extractNewsMorningSections(html) {
   const source = String(html ?? '');
diff --git a/wechat-news-morning-dedup.test.mjs b/wechat-news-morning-dedup.test.mjs
index 74c02a9..c72dcce 100644
--- a/wechat-news-morning-dedup.test.mjs
+++ b/wechat-news-morning-dedup.test.mjs
@@ -162,3 +162,16 @@ test('formatNewsMorningDedupeSummary reports reasons and removed sections', () =
   assert.match(summary, /剔除 1 条重复（event-key:1）/);
   assert.match(summary, /移除空栏目 living/);
 });
+
+test('countNewsMorningStoryCards counts news001 and news002 card shapes', async () => {
+  const { countNewsMorningStoryCards } = await import('./wechat-news-morning-dedup.mjs');
+  const html = page([
+    ['headlines', [
+      card({ title: '头条一条', cls: 'lead-card' }),
+      card({ title: '头条两条', cls: 'card highlight-box' }),
+    ]],
+    ['domestic', [card({ title: '国内一条' })]],
+    ['weather', ['<div class="weather-card"><div class="wc">北京</div></div>']],
+  ]);
+  assert.equal(countNewsMorningStoryCards(html), 3, 'lead-card/highlight-box 要计入，weather-card 不计入');
+});
diff --git a/wechat-news-morning-draft.mjs b/wechat-news-morning-draft.mjs
index 3a35d4a..687bd1e 100644
--- a/wechat-news-morning-draft.mjs
+++ b/wechat-news-morning-draft.mjs
@@ -21,6 +21,7 @@ import {
   uploadWechatDraftBrandAssets,
 } from './wechat-draft-article-layout.mjs';
 import { getLocalParts, localDateKey, startOfLocalDay } from './schedule-time.mjs';
+import { countNewsMorningStoryCards } from './wechat-news-morning-dedup.mjs';
 import {
   NEWS001_FRESH_CONTENT_SPEC,
   NEWS001_SEO_SPEC,
@@ -110,9 +111,9 @@ export function isNewsMorningContentDegraded(html, { minStoryCards = 12 } = {})
   if (/关于今日早报的重要说明|未填充任何当日具体新闻条目|未能获取当日可核验新闻/u.test(source)) {
     return true;
   }
-  const cards = (source.match(/class="card(?:\s|")/gi) ?? []).length;
-  const highlights = (source.match(/class="card highlight-box/gi) ?? []).length;
-  return cards + highlights < minStoryCards;
+  // 必须按 class token 统计：news002 的头条卡是 a.lead-card，
+  // 旧写法 `class="card(?:\s|")` 匹配不到，会把正常页误判成降级页。
+  return countNewsMorningStoryCards(source) < minStoryCards;
 }
 
 export function isNewsMorningPageDegraded(page, { minStoryCards = 6 } = {}) {
diff --git a/wechat-news-morning-draft.test.mjs b/wechat-news-morning-draft.test.mjs
index 015816f..58ccb34 100644
--- a/wechat-news-morning-draft.test.mjs
+++ b/wechat-news-morning-draft.test.mjs
@@ -314,3 +314,16 @@ test('internals normalize booleans consistently', () => {
   assert.equal(wechatNewsMorningDraftInternals.normalizeBoolean('1', false), true);
   assert.equal(wechatNewsMorningDraftInternals.normalizeBoolean('off', true), false);
 });
+
+test('isNewsMorningContentDegraded counts news002 lead cards', async () => {
+  const { isNewsMorningContentDegraded } = await import('./wechat-news-morning-draft.mjs');
+  const leads = Array.from({ length: 4 }, (_, i) =>
+    `<a class="lead-card" href="#"><div class="body"><h3>头条${i}</h3><p>正文</p></div></a>`).join('');
+  const cards = Array.from({ length: 4 }, (_, i) =>
+    `<div class="card" data-event-key="k${i}"><div class="body"><h3>条目${i}</h3><p>正文</p></div></div>`).join('');
+  const html = `<body><div class="container">${leads}${cards}</div></body>`;
+
+  // 旧实现用 `class="card(?:\s|")` 统计，匹配不到 lead-card，4 条头条会被漏数。
+  assert.equal(isNewsMorningContentDegraded(html, { minStoryCards: 8 }), false);
+  assert.equal(isNewsMorningContentDegraded(html, { minStoryCards: 9 }), true);
+});