feat: 重构导入逻辑

2026-05-11 00:21:36 +08:00 · 2025-12-01 23:18:27 +08:00
parent 9bbc34e76b
commit 3ef37f16be
9 changed files with 389 additions and 311 deletions
@@ -78,7 +78,8 @@ function createDatabase(sessionId: string): Database.Database {
    CREATE TABLE IF NOT EXISTS member (
      id INTEGER PRIMARY KEY AUTOINCREMENT,
      platform_id TEXT NOT NULL UNIQUE,
-      name TEXT NOT NULL
+      name TEXT NOT NULL,
+      nickname TEXT
    );

    CREATE TABLE IF NOT EXISTS member_name_history (
@@ -148,7 +149,7 @@ export function importData(parseResult: ParseResult): string {
      )

      const insertMember = db.prepare(`
-        INSERT OR IGNORE INTO member (platform_id, name) VALUES (?, ?)
+        INSERT OR IGNORE INTO member (platform_id, name, nickname) VALUES (?, ?, ?)
      `)
      const getMemberId = db.prepare(`
        SELECT id FROM member WHERE platform_id = ?
@@ -157,7 +158,7 @@ export function importData(parseResult: ParseResult): string {
      const memberIdMap = new Map<string, number>()

      for (const member of parseResult.members) {
-        insertMember.run(member.platformId, member.name)
+        insertMember.run(member.platformId, member.name, member.nickname || null)
        const row = getMemberId.get(member.platformId) as { id: number }
        memberIdMap.set(member.platformId, row.id)
      }
@@ -8,7 +8,7 @@ import type { FormatModule } from '../types'
 // 导入所有格式模块
 import chatlab from './chatlab'
 import shuakamiQqExporterV4 from './shuakami-qq-exporter-v4'
-import shuakamiQqExporterLegacy from './shuakami-qq-exporter-legacy'
+import qqNativeTxt from './qq-native-txt'

 /**
 * 所有支持的格式模块（按优先级排序）
@@ -16,8 +16,8 @@ import shuakamiQqExporterLegacy from './shuakami-qq-exporter-legacy'
 export const formats: FormatModule[] = [
  chatlab, // 优先级 1
  shuakamiQqExporterV4, // 优先级 10 - shuakami/qq-chat-exporter V4
-  shuakamiQqExporterLegacy, // 优先级 20 - shuakami/qq-chat-exporter Legacy
+  qqNativeTxt, // 优先级 30 - QQ 官方导出 TXT
 ]

 // 按名称导出，方便单独使用
-export { chatlab, shuakamiQqExporterV4, shuakamiQqExporterLegacy }
+export { chatlab, shuakamiQqExporterV4, qqNativeTxt }
@@ -0,0 +1,278 @@
+/**
+ * QQ 官方导出 TXT 格式解析器
+ * 适配 QQ 群聊旧版导出功能
+ *
+ * 格式特征：
+ * - 文件头：消息记录（此消息记录为文本格式，不支持重新导入）
+ * - 群名：消息对象:xxx
+ * - 消息格式：时间 昵称(QQ号) 或 时间 昵称<邮箱>
+ * - 内容在消息头下一行，可能跨多行
+ *
+ * 字段映射：
+ * - name: platformId（用户ID）
+ * - nickname: 群昵称
+ * - senderName: 群昵称（用于昵称历史追踪）
+ */
+
+import * as fs from 'fs'
+import * as readline from 'readline'
+import { ChatPlatform, ChatType, MessageType } from '../../../../src/types/chat'
+import type {
+  FormatFeature,
+  FormatModule,
+  Parser,
+  ParseOptions,
+  ParseEvent,
+  ParsedMeta,
+  ParsedMember,
+  ParsedMessage,
+} from '../types'
+import { getFileSize, createProgress } from '../utils'
+
+// ==================== 特征定义 ====================
+
+export const feature: FormatFeature = {
+  id: 'qq-native-txt',
+  name: 'QQ 官方导出 (TXT)',
+  platform: ChatPlatform.QQ,
+  priority: 30,
+  extensions: ['.txt'],
+  signatures: {
+    head: [/消息记录（此消息记录为文本格式/, /消息对象:/],
+  },
+}
+
+// ==================== 消息头正则 ====================
+
+// 匹配：2019-07-16 18:15:05 夜喵大人🐱(642163903)
+// 或：2019-07-16 18:15:11 铛🔔<ppbaozi@gmail.com>
+const MESSAGE_HEADER_REGEX = /^(\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}) (.+?)(?:\(([^)]+)\)|<([^>]+)>)$/
+
+// 匹配群名：消息对象:杭州FE
+const GROUP_NAME_REGEX = /^消息对象:(.+)$/
+
+// ==================== 消息类型判断 ====================
+
+function detectMessageType(content: string): MessageType {
+  const trimmed = content.trim()
+  if (trimmed === '[图片]') return MessageType.IMAGE
+  if (trimmed === '[表情]') return MessageType.EMOJI
+  if (trimmed === '[语音]') return MessageType.VOICE
+  if (trimmed === '[视频]') return MessageType.VIDEO
+  if (trimmed === '[文件]') return MessageType.FILE
+  if (trimmed.startsWith('[') && trimmed.endsWith(']')) return MessageType.OTHER
+  return MessageType.TEXT
+}
+
+// ==================== 时间解析 ====================
+
+/**
+ * 解析本地时间字符串为秒级时间戳
+ * @param timeStr 格式：2019-07-16 18:15:05
+ */
+function parseLocalTime(timeStr: string): number {
+  // 直接用 Date 解析，会按本地时区处理
+  const date = new Date(timeStr.replace(' ', 'T'))
+  return Math.floor(date.getTime() / 1000)
+}
+
+// ==================== 昵称清理 ====================
+
+/**
+ * 清理昵称中的前缀污染
+ * 例如：【管理员】张三 -> 张三
+ */
+function cleanNickname(nickname: string): string {
+  // 移除开头的【xxx】前缀（可能有多个）
+  return nickname.replace(/^(【[^】]*】\s*)+/, '').trim()
+}
+
+// ==================== 成员信息 ====================
+
+interface MemberInfo {
+  platformId: string
+  nickname: string // 群昵称
+}
+
+// 用于记录用户最近的有效昵称（昵称 != ID 的情况）
+const lastValidNickname = new Map<string, string>()
+
+// ==================== 解析器实现 ====================
+
+async function* parseTxt(options: ParseOptions): AsyncGenerator<ParseEvent, void, unknown> {
+  const { filePath, batchSize = 5000, onProgress } = options
+
+  const totalBytes = getFileSize(filePath)
+  let bytesRead = 0
+  let messagesProcessed = 0
+
+  // 发送初始进度
+  const initialProgress = createProgress('parsing', 0, totalBytes, 0, '开始解析...')
+  yield { type: 'progress', data: initialProgress }
+  onProgress?.(initialProgress)
+
+  // 收集数据
+  let groupName = '未知群聊'
+  const memberMap = new Map<string, MemberInfo>()
+  const messages: ParsedMessage[] = []
+
+  // 当前正在解析的消息
+  let currentMessage: {
+    timestamp: number
+    platformId: string
+    nickname: string
+    contentLines: string[]
+  } | null = null
+
+  // 保存当前消息
+  const saveCurrentMessage = () => {
+    if (currentMessage) {
+      const content = currentMessage.contentLines.join('\n').trim()
+      const type = detectMessageType(content)
+
+      messages.push({
+        senderPlatformId: currentMessage.platformId,
+        senderName: currentMessage.nickname, // 用于昵称历史追踪
+        timestamp: currentMessage.timestamp,
+        type,
+        content: content || null,
+      })
+
+      // 更新成员信息（保留最新昵称）
+      memberMap.set(currentMessage.platformId, {
+        platformId: currentMessage.platformId,
+        nickname: currentMessage.nickname,
+      })
+
+      messagesProcessed++
+    }
+  }
+
+  // 逐行读取文件
+  const fileStream = fs.createReadStream(filePath, { encoding: 'utf-8' })
+  const rl = readline.createInterface({
+    input: fileStream,
+    crlfDelay: Infinity,
+  })
+
+  fileStream.on('data', (chunk: string | Buffer) => {
+    bytesRead += typeof chunk === 'string' ? Buffer.byteLength(chunk) : chunk.length
+  })
+
+  for await (const line of rl) {
+    // 检查群名
+    const groupMatch = line.match(GROUP_NAME_REGEX)
+    if (groupMatch) {
+      groupName = groupMatch[1].trim()
+      continue
+    }
+
+    // 检查消息头
+    const headerMatch = line.match(MESSAGE_HEADER_REGEX)
+    if (headerMatch) {
+      // 保存前一条消息
+      saveCurrentMessage()
+
+      const timeStr = headerMatch[1]
+      const rawNickname = headerMatch[2].trim()
+      let nickname = cleanNickname(rawNickname) // 清理前缀污染
+      const platformId = headerMatch[3] || headerMatch[4] // (id) 或 <email>
+
+      // 如果昵称和 ID 相同，可能是系统故障，使用之前记录的昵称
+      if (nickname === platformId) {
+        const previousNickname = lastValidNickname.get(platformId)
+        if (previousNickname) {
+          nickname = previousNickname
+        }
+        // 如果没有之前的记录，保持使用 ID 作为昵称
+      } else {
+        // 记录有效昵称（昵称 != ID）
+        lastValidNickname.set(platformId, nickname)
+      }
+
+      currentMessage = {
+        timestamp: parseLocalTime(timeStr),
+        platformId,
+        nickname,
+        contentLines: [],
+      }
+
+      // 更新进度
+      if (messagesProcessed % 1000 === 0) {
+        const progress = createProgress(
+          'parsing',
+          bytesRead,
+          totalBytes,
+          messagesProcessed,
+          `已处理 ${messagesProcessed} 条消息...`
+        )
+        onProgress?.(progress)
+      }
+
+      continue
+    }
+
+    // 内容行（追加到当前消息）
+    if (currentMessage) {
+      // 跳过分隔线
+      if (line.startsWith('=====')) continue
+      // 跳过文件头信息
+      if (line.startsWith('消息记录') || line.startsWith('消息分组')) continue
+
+      currentMessage.contentLines.push(line)
+    }
+  }
+
+  // 保存最后一条消息
+  saveCurrentMessage()
+
+  // 发送 meta
+  const meta: ParsedMeta = {
+    name: groupName,
+    platform: ChatPlatform.QQ,
+    type: ChatType.GROUP,
+  }
+  yield { type: 'meta', data: meta }
+
+  // 发送成员（name 使用 platformId，nickname 使用群昵称）
+  const members: ParsedMember[] = Array.from(memberMap.values()).map((m) => ({
+    platformId: m.platformId,
+    name: m.platformId, // name 使用 ID
+    nickname: m.nickname, // nickname 使用群昵称
+  }))
+  yield { type: 'members', data: members }
+
+  // 分批发送消息
+  for (let i = 0; i < messages.length; i += batchSize) {
+    const batch = messages.slice(i, i + batchSize)
+    yield { type: 'messages', data: batch }
+  }
+
+  // 完成
+  const doneProgress = createProgress('done', totalBytes, totalBytes, messagesProcessed, '解析完成')
+  yield { type: 'progress', data: doneProgress }
+  onProgress?.(doneProgress)
+
+  yield {
+    type: 'done',
+    data: { messageCount: messagesProcessed, memberCount: memberMap.size },
+  }
+}
+
+// ==================== 导出解析器 ====================
+
+export const parser_: Parser = {
+  feature,
+  parse: parseTxt,
+}
+
+// ==================== 导出格式模块 ====================
+
+const module_: FormatModule = {
+  feature,
+  parser: parser_,
+  // TXT 格式不需要预处理器
+}
+
+export default module_
+
@@ -1,271 +0,0 @@
-/**
- * shuakami/qq-chat-exporter Legacy 格式解析器
- * 适配项目: https://github.com/shuakami/qq-chat-exporter
- * 版本: V1-V3（早期版本）
- *
- * 特征：
- * - 时间戳使用毫秒数
- * - 没有 metadata.version 或版本号小于 4
- * - sender 中主要使用 uin 字段
- *
- * 注意：此解析器仅适配 shuakami/qq-chat-exporter 项目导出的格式，
- * 其他 QQ 聊天记录导出工具可能需要创建独立的解析器。
- */
-
-import * as fs from 'fs'
-import { parser } from 'stream-json'
-import { pick } from 'stream-json/filters/Pick'
-import { streamValues } from 'stream-json/streamers/StreamValues'
-import { chain } from 'stream-chain'
-import { ChatPlatform, ChatType, MessageType } from '../../../../src/types/chat'
-import type {
-  FormatFeature,
-  FormatModule,
-  Parser,
-  ParseOptions,
-  ParseEvent,
-  ParsedMeta,
-  ParsedMember,
-  ParsedMessage,
-} from '../types'
-import { getFileSize, createProgress, readFileHeadBytes, parseTimestamp, isValidYear } from '../utils'
-
-// ==================== 特征定义 ====================
-
-export const feature: FormatFeature = {
-  id: 'shuakami-qq-exporter-legacy',
-  name: 'shuakami/qq-chat-exporter (Legacy)',
-  platform: ChatPlatform.QQ,
-  priority: 20, // 低于 V4
-  extensions: ['.json'],
-  signatures: {
-    head: [/QQChatExporter/, /"chatInfo"/],
-    requiredFields: ['chatInfo', 'messages'],
-  },
-}
-
-// ==================== 消息结构 ====================
-
-interface LegacyMessage {
-  id?: string
-  timestamp: number // 毫秒时间戳
-  sender: {
-    uid?: string
-    uin: string
-    name: string
-  }
-  type?: string
-  system?: boolean
-  recalled?: boolean
-  content: {
-    text: string
-    html?: string
-    resources?: Array<{ type: string }>
-    elements?: Array<{ type: string }>
-  }
-}
-
-// ==================== 消息类型转换 ====================
-
-function convertMessageType(qqType: string | undefined, content: LegacyMessage['content']): MessageType {
-  // 检查资源类型
-  if (content.resources && content.resources.length > 0) {
-    const resourceType = content.resources[0].type
-    switch (resourceType) {
-      case 'image':
-        return MessageType.IMAGE
-      case 'video':
-        return MessageType.VIDEO
-      case 'voice':
-      case 'audio':
-        return MessageType.VOICE
-      case 'file':
-        return MessageType.FILE
-    }
-  }
-
-  // 检查元素类型
-  if (content.elements) {
-    for (const elem of content.elements) {
-      if (elem.type === 'market_face' || elem.type === 'face') {
-        return MessageType.EMOJI
-      }
-    }
-  }
-
-  // 根据 type 字符串判断
-  switch (qqType) {
-    case 'type_1':
-      return MessageType.TEXT
-    case 'type_17':
-      return MessageType.EMOJI
-    case 'type_3':
-      return MessageType.IMAGE
-    case 'type_7':
-      return MessageType.VOICE
-    default:
-      return MessageType.TEXT
-  }
-}
-
-// ==================== 解析器实现 ====================
-
-async function* parseLegacy(options: ParseOptions): AsyncGenerator<ParseEvent, void, unknown> {
-  const { filePath, batchSize = 5000, onProgress } = options
-
-  const totalBytes = getFileSize(filePath)
-  let bytesRead = 0
-  let messagesProcessed = 0
-
-  // 发送初始进度
-  const initialProgress = createProgress('parsing', 0, totalBytes, 0, '开始解析...')
-  yield { type: 'progress', data: initialProgress }
-  onProgress?.(initialProgress)
-
-  // 读取文件头获取 meta 信息
-  const headContent = readFileHeadBytes(filePath, 100000)
-
-  // 解析 chatInfo
-  let chatInfo = { name: '未知群聊', type: 'group' as const }
-  try {
-    const chatInfoMatch = headContent.match(/"chatInfo"\s*:\s*(\{[^}]+\})/)
-    if (chatInfoMatch) {
-      chatInfo = JSON.parse(chatInfoMatch[1])
-    }
-  } catch {
-    // 使用默认值
-  }
-
-  // 发送 meta
-  const meta: ParsedMeta = {
-    name: chatInfo.name,
-    platform: ChatPlatform.QQ,
-    type: chatInfo.type === 'group' ? ChatType.GROUP : ChatType.PRIVATE,
-  }
-  yield { type: 'meta', data: meta }
-
-  // 收集成员和消息
-  const memberMap = new Map<string, ParsedMember>()
-  let messageBatch: ParsedMessage[] = []
-
-  // 流式解析
-  await new Promise<void>((resolve, reject) => {
-    const readStream = fs.createReadStream(filePath, { encoding: 'utf-8' })
-
-    readStream.on('data', (chunk: string | Buffer) => {
-      bytesRead += typeof chunk === 'string' ? Buffer.byteLength(chunk) : chunk.length
-    })
-
-    const pipeline = chain([readStream, parser(), pick({ filter: /^messages\.\d+$/ }), streamValues()])
-
-    const processMessage = (msg: LegacyMessage): ParsedMessage | null => {
-      // 获取 platformId
-      const platformId = msg.sender.uin || msg.sender.uid
-      if (!platformId) return null
-
-      // 获取发送者名称
-      const senderName = msg.sender.name || platformId
-
-      // 更新成员
-      memberMap.set(platformId, { platformId, name: senderName })
-
-      // 解析时间戳（毫秒）
-      const timestamp = parseTimestamp(msg.timestamp)
-      if (timestamp === null || !isValidYear(timestamp)) return null
-
-      // 消息类型
-      const type = msg.system ? MessageType.SYSTEM : convertMessageType(msg.type, msg.content)
-
-      // 文本内容
-      let textContent = msg.content?.text || ''
-      if (msg.recalled) {
-        textContent = '[已撤回] ' + textContent
-      }
-
-      return {
-        senderPlatformId: platformId,
-        senderName,
-        timestamp,
-        type,
-        content: textContent || null,
-      }
-    }
-
-    // 用于收集批次的临时数组
-    const batchCollector: ParsedMessage[] = []
-
-    pipeline.on('data', ({ value }: { value: LegacyMessage }) => {
-      const parsed = processMessage(value)
-      if (parsed) {
-        batchCollector.push(parsed)
-        messagesProcessed++
-
-        // 达到批次大小
-        if (batchCollector.length >= batchSize) {
-          messageBatch.push(...batchCollector)
-          batchCollector.length = 0
-
-          const progress = createProgress(
-            'parsing',
-            bytesRead,
-            totalBytes,
-            messagesProcessed,
-            `已处理 ${messagesProcessed} 条消息...`
-          )
-          onProgress?.(progress)
-        }
-      }
-    })
-
-    pipeline.on('end', () => {
-      // 收集剩余消息
-      if (batchCollector.length > 0) {
-        messageBatch.push(...batchCollector)
-      }
-      resolve()
-    })
-
-    pipeline.on('error', reject)
-  })
-
-  // 发送成员
-  yield { type: 'members', data: Array.from(memberMap.values()) }
-
-  // 分批发送消息
-  for (let i = 0; i < messageBatch.length; i += batchSize) {
-    const batch = messageBatch.slice(i, i + batchSize)
-    yield { type: 'messages', data: batch }
-  }
-
-  // 完成
-  const doneProgress = createProgress('done', totalBytes, totalBytes, messagesProcessed, '解析完成')
-  yield { type: 'progress', data: doneProgress }
-  onProgress?.(doneProgress)
-
-  yield {
-    type: 'done',
-    data: { messageCount: messagesProcessed, memberCount: memberMap.size },
-  }
-}
-
-// ==================== 导出解析器 ====================
-
-export const parser_: Parser = {
-  feature,
-  parse: parseLegacy,
-}
-
-// ==================== 导出预处理器 ====================
-
-import { qqPreprocessor } from './shuakami-qq-preprocessor'
-export const preprocessor = qqPreprocessor
-
-// ==================== 导出格式模块 ====================
-
-const module_: FormatModule = {
-  feature,
-  parser: parser_,
-  preprocessor: qqPreprocessor,
-}
-
-export default module_
@@ -1,15 +1,19 @@
 /**
 * shuakami/qq-chat-exporter V4 格式解析器
 * 适配项目: https://github.com/shuakami/qq-chat-exporter
- * 版本: V4.x
+ * 版本: V4.x (2024年12月更新)
 *
 * 特征：
- * - 时间戳使用 ISO 字符串格式（如 "2017-12-30T03:24:36.000Z"）
+ * - 时间戳使用 ISO 字符串格式（如 "2022-10-29T06:42:53.000Z"）
 * - metadata.version 为 "4.x.x"
- * - sender 中有 uid 字段
+ * - rawMessage 中包含 sendNickName（QQ昵称）、sendMemberName（群昵称）
 *
- * 注意：此解析器仅适配 shuakami/qq-chat-exporter 项目导出的格式，
- * 其他 QQ 聊天记录导出工具可能需要创建独立的解析器。
+ * 名字字段说明：
+ * - sendNickName: QQ原始昵称（始终存在）
+ * - sendMemberName: 群昵称（可选，用户未设置时不存在）
+ * - sendRemarkName: 导出者的备注名（不使用）
+ *
+ * 显示名优先级: sendMemberName > sendNickName
 */

 import * as fs from 'fs'
@@ -36,7 +40,7 @@ export const feature: FormatFeature = {
  id: 'shuakami-qq-exporter-v4',
  name: 'shuakami/qq-chat-exporter V4',
  platform: ChatPlatform.QQ,
-  priority: 10, // 高优先级
+  priority: 10,
  extensions: ['.json'],
  signatures: {
    head: [/QQChatExporter V4/, /"version"\s*:\s*"4\./],
@@ -46,6 +50,14 @@ export const feature: FormatFeature = {

 // ==================== 消息结构 ====================

+interface V4RawMessage {
+  senderUin?: string
+  senderUid?: string
+  sendNickName?: string // QQ原始昵称
+  sendMemberName?: string // 群昵称
+  msgTime?: string // 秒级时间戳字符串
+}
+
 interface V4Message {
  messageId?: string
  timestamp: string // ISO 格式
@@ -63,7 +75,17 @@ interface V4Message {
    raw?: string
    resources?: Array<{ type: string }>
    elements?: Array<{ type: string }>
+    emojis?: Array<{ type: string }>
  }
+  rawMessage?: V4RawMessage
+}
+
+// ==================== 成员信息追踪 ====================
+
+interface MemberInfo {
+  platformId: string
+  displayName: string // 显示名（sendMemberName || sendNickName）
+  nickname: string // QQ原始昵称（sendNickName）
 }

 // ==================== 消息类型转换 ====================
@@ -85,13 +107,9 @@ function convertMessageType(messageType: number | undefined, content: V4Message[
    }
  }

-  // 检查元素类型
-  if (content.elements) {
-    for (const elem of content.elements) {
-      if (elem.type === 'market_face' || elem.type === 'face') {
-        return MessageType.EMOJI
-      }
-    }
+  // 检查 emojis 字段
+  if (content.emojis && content.emojis.length > 0) {
+    return MessageType.EMOJI
  }

  // 根据 messageType 判断
@@ -99,11 +117,13 @@ function convertMessageType(messageType: number | undefined, content: V4Message[
    case 1:
      return MessageType.TEXT
    case 2:
-      return MessageType.IMAGE
+      return MessageType.TEXT // 普通消息
    case 3:
-      return MessageType.VOICE
+      return MessageType.IMAGE
    case 7:
      return MessageType.VIDEO
+    case 9:
+      return MessageType.TEXT // 回复消息
    default:
      return MessageType.TEXT
  }
@@ -146,7 +166,7 @@ async function* parseV4(options: ParseOptions): AsyncGenerator<ParseEvent, void,
  yield { type: 'meta', data: meta }

  // 收集成员和消息
-  const memberMap = new Map<string, ParsedMember>()
+  const memberMap = new Map<string, MemberInfo>()
  let messageBatch: ParsedMessage[] = []

  // 流式解析
@@ -160,15 +180,31 @@ async function* parseV4(options: ParseOptions): AsyncGenerator<ParseEvent, void,
    const pipeline = chain([readStream, parser(), pick({ filter: /^messages\.\d+$/ }), streamValues()])

    const processMessage = (msg: V4Message): ParsedMessage | null => {
-      // 获取 platformId
-      const platformId = msg.sender.uin || msg.sender.uid
+      // 获取 platformId：优先使用 uin（QQ号），fallback 到 uid
+      const platformId = msg.sender.uin || msg.sender.uid || msg.rawMessage?.senderUin || msg.rawMessage?.senderUid
      if (!platformId) return null

-      // 获取发送者名称
-      const senderName = msg.sender.name || platformId
+      // 从 rawMessage 获取名字信息
+      const raw = msg.rawMessage
+      const sendNickName = raw?.sendNickName || msg.sender.name || platformId
+      const sendMemberName = raw?.sendMemberName

-      // 更新成员
-      memberMap.set(platformId, { platformId, name: senderName })
+      // 显示名优先级：群昵称 > QQ昵称
+      const displayName = sendMemberName || sendNickName
+
+      // 更新成员信息（保留最新的名字）
+      const existingMember = memberMap.get(platformId)
+      if (!existingMember) {
+        memberMap.set(platformId, {
+          platformId,
+          displayName,
+          nickname: sendNickName,
+        })
+      } else {
+        // 更新为最新的名字
+        existingMember.displayName = displayName
+        existingMember.nickname = sendNickName
+      }

      // 解析时间戳
      const timestamp = parseTimestamp(msg.timestamp)
@@ -185,7 +221,7 @@ async function* parseV4(options: ParseOptions): AsyncGenerator<ParseEvent, void,

      return {
        senderPlatformId: platformId,
-        senderName,
+        senderName: displayName, // 用于昵称历史追踪
        timestamp,
        type,
        content: textContent || null,
@@ -229,8 +265,13 @@ async function* parseV4(options: ParseOptions): AsyncGenerator<ParseEvent, void,
    pipeline.on('error', reject)
  })

-  // 发送成员
-  yield { type: 'members', data: Array.from(memberMap.values()) }
+  // 发送成员（包含 nickname）
+  const members: ParsedMember[] = Array.from(memberMap.values()).map((m) => ({
+    platformId: m.platformId,
+    name: m.displayName,
+    nickname: m.nickname,
+  }))
+  yield { type: 'members', data: members }

  // 分批发送消息
  for (let i = 0; i < messageBatch.length; i += batchSize) {
@@ -11,7 +11,7 @@ import * as path from 'path'
 import * as os from 'os'
 import { parser } from 'stream-json'
 import { pick } from 'stream-json/filters/Pick'
-import { streamArray } from 'stream-json/streamers/StreamArray'
+import { streamValues } from 'stream-json/streamers/StreamValues'
 import { chain } from 'stream-chain'
 import type { ParseProgress, Preprocessor } from '../types'
 import { getFileSize, createProgress } from '../utils'
@@ -53,11 +53,19 @@ interface SlimQQMessage {
    text: string
    elements?: Array<{ type: string }>
    resources?: Array<{ type: string }>
+    emojis?: Array<{ type: string }>
  }
  recalled?: boolean
  isRecalled?: boolean
  system?: boolean
  isSystemMessage?: boolean
+  // V4 新增：保留 rawMessage 中的名字字段
+  rawMessage?: {
+    sendNickName?: string
+    sendMemberName?: string
+    senderUin?: string
+    senderUid?: string
+  }
 }

 /**
@@ -66,6 +74,7 @@ interface SlimQQMessage {
 function slimMessage(msg: Record<string, unknown>): SlimQQMessage {
  const sender = msg.sender as { uin?: string; uid?: string; name?: string } | undefined
  const content = msg.content as Record<string, unknown> | undefined
+  const rawMessage = msg.rawMessage as Record<string, unknown> | undefined

  const slimContent: SlimQQMessage['content'] = {
    text: (content?.text as string) || '',
@@ -83,6 +92,12 @@ function slimMessage(msg: Record<string, unknown>): SlimQQMessage {
    }))
  }

+  if (content?.emojis && Array.isArray(content.emojis)) {
+    slimContent.emojis = (content.emojis as Array<{ type: string }>).map((e) => ({
+      type: e.type,
+    }))
+  }
+
  const slimMsg: SlimQQMessage = {
    timestamp: msg.timestamp as number | string,
    sender: { name: sender?.name || '' },
@@ -105,6 +120,15 @@ function slimMessage(msg: Record<string, unknown>): SlimQQMessage {
  if (sender?.uin) slimMsg.sender.uin = sender.uin
  if (sender?.uid) slimMsg.sender.uid = sender.uid

+  // V4 新增：保留 rawMessage 中的关键名字字段
+  if (rawMessage) {
+    slimMsg.rawMessage = {}
+    if (rawMessage.sendNickName) slimMsg.rawMessage.sendNickName = rawMessage.sendNickName as string
+    if (rawMessage.sendMemberName) slimMsg.rawMessage.sendMemberName = rawMessage.sendMemberName as string
+    if (rawMessage.senderUin) slimMsg.rawMessage.senderUin = rawMessage.senderUin as string
+    if (rawMessage.senderUid) slimMsg.rawMessage.senderUid = rawMessage.senderUid as string
+  }
+
  return slimMsg
 }

@@ -173,11 +197,12 @@ async function preprocessQQJson(inputPath: string, onProgress?: (progress: Parse

      const header = { metadata, chatInfo, messages: [] }
      const headerJson = JSON.stringify(header)
-      writeStream.write(headerJson.slice(0, -3) + '\n')
+      // 移除最后的 ]} 保留 [，结果如 {"metadata":...,"chatInfo":...,"messages":[
+      writeStream.write(headerJson.slice(0, -2) + '\n')

      let isFirstMessage = true

-      const pipeline = chain([readStream, parser(), pick({ filter: /^messages\.\d+$/ }), streamArray()])
+      const pipeline = chain([readStream, parser(), pick({ filter: /^messages\.\d+$/ }), streamValues()])

      pipeline.on('data', ({ value }: { value: Record<string, unknown> }) => {
        const slimMsg = slimMessage(value)
@@ -5,6 +5,7 @@

 import { FormatSniffer, createSniffer } from './sniffer'
 import { formats } from './formats'
+import { getFileSize } from './utils'
 import type {
  ParseOptions,
  ParseEvent,
@@ -80,7 +80,8 @@ function createDatabase(sessionId: string): Database.Database {
    CREATE TABLE IF NOT EXISTS member (
      id INTEGER PRIMARY KEY AUTOINCREMENT,
      platform_id TEXT NOT NULL UNIQUE,
-      name TEXT NOT NULL
+      name TEXT NOT NULL,
+      nickname TEXT
    );

    CREATE TABLE IF NOT EXISTS member_name_history (
@@ -165,7 +166,7 @@ export async function streamImport(filePath: string, requestId: string): Promise
    INSERT INTO meta (name, platform, type, imported_at) VALUES (?, ?, ?, ?)
  `)
  const insertMember = db.prepare(`
-    INSERT OR IGNORE INTO member (platform_id, name) VALUES (?, ?)
+    INSERT OR IGNORE INTO member (platform_id, name, nickname) VALUES (?, ?, ?)
  `)
  const getMemberId = db.prepare(`SELECT id FROM member WHERE platform_id = ?`)
  const insertMessage = db.prepare(`
@@ -209,7 +210,7 @@ export async function streamImport(filePath: string, requestId: string): Promise
      onMembers: (members: ParsedMember[]) => {
        console.log(`[StreamImport] 收到 ${members.length} 个成员`)
        for (const member of members) {
-          insertMember.run(member.platformId, member.name)
+          insertMember.run(member.platformId, member.name, member.nickname || null)
          const row = getMemberId.get(member.platformId) as { id: number } | undefined
          if (row) {
            memberIdMap.set(member.platformId, row.id)
@@ -236,7 +237,7 @@ export async function streamImport(filePath: string, requestId: string): Promise
          // 确保成员存在
          if (!memberIdMap.has(msg.senderPlatformId)) {
            const memberName = msg.senderName || msg.senderPlatformId
-            insertMember.run(msg.senderPlatformId, memberName)
+            insertMember.run(msg.senderPlatformId, memberName, null)
            const row = getMemberId.get(msg.senderPlatformId) as { id: number } | undefined
            if (row) {
              memberIdMap.set(msg.senderPlatformId, row.id)
@@ -79,7 +79,8 @@ export interface DbMeta {
 export interface DbMember {
  id: number // 自增ID
  platform_id: string // 平台标识（QQ号等）
-  name: string // 最新昵称
+  name: string // 最新显示昵称（群昵称优先）
+  nickname: string | null // QQ原始昵称
 }

 /**
@@ -100,7 +101,8 @@ export interface DbMessage {
 */
 export interface ParsedMember {
  platformId: string // 平台标识
-  name: string // 昵称
+  name: string // 显示昵称（群昵称优先，否则QQ昵称）
+  nickname?: string // QQ原始昵称（sendNickName）
 }

 /**