feat: 完善日报图片与公众号消息展示

修复日报图片读取格式与模型不支持时的友好降级
恢复话题关键词回退，确保词云在缺少顶层关键词时仍可展示
支持公众号多文章消息在聊天、日报、检索和 HTML 导出中完整呈现
迁移账号身份缓存并优化首次连接的账号占位信息
补充图片、词云、账号缓存与多文章消息回归测试

tests/unit/bootstrap-cache.test.ts

@@ -1,4 +1,5 @@
-import { mkdtempSync, readFileSync, readdirSync, rmSync, writeFileSync } from 'fs'
+import { mkdtempSync, readFileSync, readdirSync, rmSync, writeFileSync, mkdirSync } from 'fs'
+import { createHash } from 'crypto'
 import { tmpdir } from 'os'
 import { join } from 'path'
 import { afterAll, beforeAll, beforeEach, describe, expect, it, vi } from 'vitest'
@@ -105,4 +106,39 @@ describe('bootstrap cache', () => {
       }).nickname
     ).toBe('濑岛田井卫')
   })
+
+  it('migrates the previous startup cache so account discovery can show identity before unlock', () => {
+    const legacyRoot = '/fixture/legacy-account'
+    const digest = createHash('sha1')
+      .update(`${process.platform}:${legacyRoot}`)
+      .digest('hex')
+      .slice(0, 16)
+    const legacyFile = join(userData, 'cache', 'bootstrap', `${process.platform}-${digest}.json`)
+    mkdirSync(join(userData, 'cache', 'bootstrap'), { recursive: true })
+    writeFileSync(
+      legacyFile,
+      JSON.stringify({
+        version: 1,
+        platform: process.platform,
+        accountRoot: legacyRoot,
+        updatedAt: Date.now(),
+        self: {
+          wxid: 'wxid_legacy',
+          nickname: '缓存昵称',
+          avatar: 'data:image/png;base64,fixture',
+          accountRoot: legacyRoot
+        },
+        contacts: []
+      }),
+      'utf8'
+    )
+
+    clearBootstrapCache()
+    expect(getBootstrapCache(legacyRoot)?.self).toMatchObject({
+      wxid: 'wxid_legacy',
+      nickname: '缓存昵称',
+      avatar: 'data:image/png;base64,fixture'
+    })
+    flushBootstrapCacheWritesSync()
+  })
 })

tests/unit/group-report.test.ts

@@ -0,0 +1,46 @@
+import { describe, expect, it } from 'vitest'
+import { parseGroupDailyReport } from '../../src/renderer/src/utils/group-report'
+import type { GroupReportMetadata } from '../../src/shared/group-report'
+
+const metadata: GroupReportMetadata = {
+  groupName: '测试群',
+  reportDate: '2026-08-06',
+  dateRange: '今日',
+  messageCount: 3,
+  activeUsers: 2,
+  timeSpan: '1 h',
+  generatedAt: '2026/8/6 12:00:00',
+  recordNote: 'fixture',
+  footerNote: '',
+  heroParticipants: [],
+  reportMode: 'full'
+}
+
+const media = {
+  gallery: [],
+  voiceHighlights: [],
+  funBadges: []
+}
+
+describe('group report parsing', () => {
+  it('falls back to topic keywords when the model omits top-level keywords', () => {
+    const report = parseGroupDailyReport(
+      JSON.stringify({
+        topics: [
+          {
+            title: '健身安排',
+            summary: '讨论训练时间和肌酸。',
+            keywords: ['肌酸', '训练']
+          }
+        ]
+      }),
+      [],
+      '',
+      [],
+      metadata,
+      media
+    )
+
+    expect(report.keywords).toEqual(['肌酸', '训练', '健身安排'])
+  })
+})

tests/unit/message-parser.test.ts

@@ -69,6 +69,23 @@ describe('message parser', () => {
     expect(parsed).toMatchObject({ type: 'share', title: '普通分享', typeVal: '5' })
   })
 
+  it('preserves every article in a public-account multi-article message', () => {
+    const parsed = parseMessageContent(
+      `<appmsg><type>5</type><appname>长江日报</appname><mmreader><category count="3"><item><title><![CDATA[女子吃酒席时意外发现]]></title><url><![CDATA[https://mp.weixin.qq.com/a]]></url><cover><![CDATA[https://img.test/a.jpg]]></cover></item><item><title>霍尔木兹海峡开放临时协议</title><digest>国际油价短期走势</digest><url>https://mp.weixin.qq.com/b</url></item><item><title>东野圭吾新作</title><url>https://mp.weixin.qq.com/c</url></item></category></mmreader></appmsg>`,
+      49
+    )
+
+    expect(parsed).toMatchObject({ type: 'share', appname: '长江日报' })
+    if (parsed.type === 'share') {
+      expect(parsed.articles).toHaveLength(3)
+      expect(parsed.articles?.map((article) => article.title)).toEqual([
+        '女子吃酒席时意外发现',
+        '霍尔木兹海峡开放临时协议',
+        '东野圭吾新作'
+      ])
+    }
+  })
+
   it('uses the quoted group member id instead of the chatroom id', () => {
     const parsed = parseMessageContent(
       '<appmsg><type>57</type><title>回复内容</title><refermsg><type>1</type><fromusr>123456789@chatroom</fromusr><chatusr>wxid_fixture_member</chatusr><content>被引用内容</content></refermsg></appmsg>',