wechaty-puppet-matrix 0.0.52 → 0.0.54

package/src/matrix/service/request.ts

@@ -1109,6 +1109,17 @@ class Client extends EventEmitter {
     return res.data
   }
 
+  /**
+   * 清空后端配置的代理
+   */
+  public async clearProxy (): Promise<void> {
+    const res = await this.postData({
+      path: '/client/set_proxy',
+      data: { proxy: '' },
+    })
+    log.info(PRE, 'clearProxy() res: %s', JSON.stringify(res))
+  }
+
   /**
    * 恢复实例
    */
@@ -1124,7 +1135,7 @@ class Client extends EventEmitter {
       },
     })
     this.restoreTimes++
-    log.info(PRE, 'restore() res: %s', res)
+    log.info(PRE, 'restore() res: %s', JSON.stringify(res))
   }
 
   /**
@@ -1310,9 +1321,15 @@ class Client extends EventEmitter {
   /**
    * 获取当前bot信息
    */
-  public async getSelfInfo (): Promise<ContactPayload | boolean> {
+  public async getSelfInfo (proxyCleared = false): Promise<ContactPayload | boolean> {
     try {
       const res = await this.postData({ path: '/user/get_profile', data: {} })
+      // 后端配置的代理不可用时，清空代理(restore 会带上 proxy: '')后重试一次
+      if (!proxyCleared && typeof res?.errmsg === 'string' && res.errmsg.includes('Could not connect to proxy')) {
+        log.warn(PRE, 'proxy unavailable, clear proxy and retry: %s', res.errmsg)
+        await this.clearProxy()
+        return await this.getSelfInfo(true)
+      }
       if (res.errcode === -102) {
         log.info(PRE, 'offline login:', res?.errmsg)
         return false
@@ -2004,6 +2021,80 @@ class Client extends EventEmitter {
     return ''
   }
 
+  /**
+   * 语音转文字: 1.生成 voice_id -> 2.检查 -> 3.上传语音 -> 4.获取结果
+   * 文档: /msg/new_trans_voice_id, /msg/check_voice_trans, /msg/upload_voice_trans, /msg/get_voice_trans
+   */
+  public async newVoiceTransId (): Promise<any> {
+    return await this.postData({ path: '/msg/new_trans_voice_id', data: {} })
+  }
+
+  public async checkVoiceTrans (voiceId: string, msgId: string, length: number): Promise<any> {
+    return await this.postData({ path: '/msg/check_voice_trans', data: { voice_id: voiceId, msg_id: msgId, length } })
+  }
+
+  public async uploadVoiceTrans (voiceId: string, msgId: string, length: number): Promise<any> {
+    return await this.postData({ path: '/msg/upload_voice_trans', data: { voice_id: voiceId, msg_id: msgId, length } })
+  }
+
+  public async getVoiceTrans (voiceId: string): Promise<any> {
+    return await this.postData({ path: '/msg/get_voice_trans', data: { voice_id: voiceId } })
+  }
+
+  /**
+   * 语音转文字，转写完成(endFlag === 1)返回文字，失败或超时返回空字符串
+   */
+  public async voiceToText (msgId: string, length: number, timeoutMs = 20000): Promise<string> {
+    // timeSlice.buffer 很长且无需关注，打印时去掉
+    const brief = (res: any) => JSON.stringify(res, (k, v) => (k === 'buffer' ? undefined : v))
+    try {
+      const idRes = await this.newVoiceTransId()
+      const voiceId: string = idRes?.voice_id
+      if (!voiceId) {
+        log.warn(PRE, 'voiceToText: no voice_id: %s', JSON.stringify(idRes))
+        return ''
+      }
+      // length 必须是语音数据字节数(不是时长)。check 返回 uploadCtx 时需要 upload，
+      // 直到 uploadCtx 为空(后端已拿到全部数据)
+      let checkRes: any
+      for (let round = 0; round < 20; round++) {
+        checkRes = await this.checkVoiceTrans(voiceId, msgId, length)
+        log.verbose(PRE, 'voiceToText check#%d(length=%d): %s', round, length, brief(checkRes))
+        if (checkRes?.baseResponse?.ret !== 0) {
+          log.warn(PRE, 'voiceToText: check failed: %s', brief(checkRes))
+          return ''
+        }
+        const ctx = checkRes?.uploadCtx
+        if (!ctx?.dataLen) {
+          break
+        }
+        const uploadRes = await this.uploadVoiceTrans(voiceId, msgId, length)
+        log.verbose(PRE, 'voiceToText upload#%d: %s', round, brief(uploadRes))
+        if (uploadRes?.baseResponse?.ret !== 0) {
+          log.warn(PRE, 'voiceToText: upload failed: %s', brief(uploadRes))
+          return ''
+        }
+      }
+      const interval = (checkRes?.queryCtx?.interval || 2) * 1000
+      const deadline = Date.now() + timeoutMs
+      while (Date.now() < deadline) {
+        await new Promise(resolve => setTimeout(resolve, interval))
+        const res = await this.getVoiceTrans(voiceId)
+        if (res?.baseResponse?.ret !== 0) {
+          log.warn(PRE, 'voiceToText: get result failed: %s', brief(res))
+          return ''
+        }
+        if (res?.transRes?.endFlag === 1) {
+          return res.transRes.result || ''
+        }
+      }
+      log.warn(PRE, 'voiceToText: timeout for msg %s', msgId)
+    } catch (e) {
+      log.warn(PRE, 'voiceToText: error: %s', e)
+    }
+    return ''
+  }
+
   /**
    * 下载视频
    * @param imageInfo

package/src/puppet-matrix.ts

@@ -7,6 +7,7 @@ import {
 } from 'file-box'
 import type { ContactPayload, MessagePayload } from './engine-schema.js'
 import type { FileBoxMetadataMessage } from './matrix/types.js'
+import { WechatMessageType } from './matrix/types.js'
 import Client from './matrix/service/request.js'
 import { getUnixTime } from 'date-fns'
 
@@ -15,7 +16,8 @@ import path from 'path'
 // 参考
 import { CacheManager, RoomMemberMap } from './matrix/cache-manager.js'
 import { isIMContactId, isRoomId } from './matrix/utils/is-type.js'
-import { AppMessagePayload, parseAppmsgMessagePayload } from './matrix/messages/message-appmsg.js'
+import { AppMessagePayload, AppMessageType, parseAppmsgMessagePayload } from './matrix/messages/message-appmsg.js'
+import { parseRecordItem, RecordItem, RecordItemType } from './matrix/messages/message-chat-history.js'
 import { parseMiniProgramMessagePayload } from './matrix/messages/message-miniprogram.js'
 import { parseEmotionMessagePayload } from './matrix/messages/message-emotion.js'
 import { ImageMessagePayload, parseImageMessagePayload } from './matrix/messages/message-image.js'
@@ -858,6 +860,94 @@ class PuppetMatrix extends PUPPET.Puppet {
     return locationPayload
   }
 
+  /**
+   * 合并转发消息明细
+   */
+  override async messageChatHistory (messageId: string): Promise<PUPPET.payloads.ChatHistory[]> {
+    const rawPayload = await this.messageRawPayload(messageId)
+    const appPayload = await parseAppmsgMessagePayload(rawPayload.msg)
+    if (appPayload.type !== AppMessageType.ChatHistory) {
+      throw new Error('Can not get chat history from non chat history payload')
+    }
+    log.verbose(PRE, `messageChatHistory recorditem: ${appPayload.recorditem}`)
+    if (!appPayload.recorditem) {
+      return []
+    }
+    const info = await parseRecordItem(appPayload.recorditem)
+    return Promise.all(info.items.map(item => this._recordItemToChatHistory(item)))
+  }
+
+  private async _recordItemToChatHistory (item: RecordItem): Promise<PUPPET.payloads.ChatHistory> {
+    const base = {
+      avatar: item.sourceHeadUrl ? FileBox.fromUrl(item.sourceHeadUrl) : FileBox.fromBuffer(Buffer.alloc(0), 'avatar.png'),
+      corpName: '',
+      senderName: item.sourceName,
+      // sourcetime 形如 "2024-12-30 12:00:00"，解析失败则为 0
+      time: Math.floor((Date.parse(item.sourceTime.replace(/-/g, '/')) || 0) / 1000),
+    }
+    const unknown = { ...base, message: undefined as never, title: item.title, type: PUPPET.types.Message.Unknown } as PUPPET.payloads.ChatHistory
+    const media = async (kind: 'image' | 'video' | 'file' | 'emoticon'): Promise<FileBoxInterface | undefined> => {
+      if (!item.cdnDataUrl) {
+        return undefined
+      }
+      const params = { aes_key: item.cdnDataKey, file_id: item.cdnDataUrl, file_name: item.fileName || item.title, file_size: item.dataSize }
+      const url = kind === 'image' || kind === 'emoticon'
+        ? await this._client?.downloadImage(params as any)
+        : kind === 'video' ? await this._client?.downloadVideo(params as any) : await this._client?.downloadFile(params as any)
+      return url ? FileBox.fromUrl(url, item.fileName ? { name: item.fileName } : undefined) : undefined
+    }
+
+    try {
+      switch (item.type) {
+        case RecordItemType.Text:
+          return { ...base, message: item.desc, type: PUPPET.types.Message.Text }
+        case RecordItemType.Image: {
+          const box = await media('image')
+          return box ? { ...base, message: box, type: PUPPET.types.Message.Image } : unknown
+        }
+        case RecordItemType.Video: {
+          const box = await media('video')
+          return box ? { ...base, message: box, type: PUPPET.types.Message.Video } : unknown
+        }
+        case RecordItemType.File: {
+          const box = await media('file')
+          return box ? { ...base, message: box, type: PUPPET.types.Message.Attachment } : unknown
+        }
+        case RecordItemType.Emoticon: {
+          const box = await media('emoticon')
+          return box ? { ...base, message: box, type: PUPPET.types.Message.Emoticon } : unknown
+        }
+        case RecordItemType.Link:
+          if (!item.link) return unknown
+          return {
+            ...base,
+            message: { description: item.link.desc, thumbnailUrl: item.link.thumbUrl, title: item.link.title, url: item.link.url },
+            type: PUPPET.types.Message.Url,
+          }
+        case RecordItemType.Location:
+          if (!item.location) return unknown
+          return {
+            ...base,
+            message: { accuracy: 15, address: item.location.label, latitude: item.location.lat, longitude: item.location.lng, name: item.location.poiname },
+            type: PUPPET.types.Message.Location,
+          }
+        case RecordItemType.ChatHistory:
+          return {
+            ...base,
+            message: await Promise.all((item.children || []).map(child => this._recordItemToChatHistory(child))),
+            title: item.title,
+            type: PUPPET.types.Message.ChatHistory,
+          }
+        default:
+          log.verbose(PRE, `unsupported chat history item type: ${item.type}`)
+          return unknown
+      }
+    } catch (e) {
+      log.warn(PRE, `parse chat history item failed, type: ${item.type}, error: ${e}`)
+      return unknown
+    }
+  }
+
   /****************************************************************************
    * send message
    ***************************************************************************/
@@ -1206,6 +1296,17 @@ class PuppetMatrix extends PUPPET.Puppet {
     return ret
   }
 
+  /**
+   * 缓存里没有该消息时才写入(用于被引用消息的兜底)
+   */
+  async cacheMessageIfAbsent (payload: MessagePayload): Promise<void> {
+    if (await this._cacheMgr!.hasMessage(payload.id)) {
+      return
+    }
+    log.info(PRE, `cache miss for refer message ${payload.id}, rebuild from refermsg`)
+    await this._cacheMgr!.setMessage(payload.id, payload)
+  }
+
   /**
    * 群数据格式化为Wechaty 支持类型
    * @param rawPayload
@@ -1593,6 +1694,31 @@ class PuppetMatrix extends PUPPET.Puppet {
     }
   }
 
+  /**
+   * 收到语音消息时转文字，结果放到 message.voiceText，由 audio parser 写入消息 text。
+   * 自己发出的语音不处理；失败或超时保持原样
+   */
+  private async _transcribeVoice (message: MessagePayload): Promise<void> {
+    if (message.msgType !== WechatMessageType.Voice || message.talkerId === this.currentUserId) {
+      return
+    }
+    try {
+      const audio = await parseAudioMessagePayload(message)
+      // length 是语音数据字节数，voicelength 是时长(毫秒)；接口的 length 参数要的是字节数
+      const length = parseInt(audio.length, 10)
+      if (!length) {
+        return
+      }
+      const text = await this._client?.voiceToText(message.id, length)
+      if (text) {
+        message.voiceText = text
+        await this._cacheMgr!.setMessage(message.id, message)
+      }
+    } catch (e) {
+      log.warn(PRE, `transcribe voice failed for ${message.id}: ${e}`)
+    }
+  }
+
   private async _onPushMessage (message: MessagePayload): Promise<void> {
     const messageId = message.id
     log.silly(PRE, `on push original message: ${JSON.stringify(message)}`)
@@ -1600,6 +1726,7 @@ class PuppetMatrix extends PUPPET.Puppet {
       return
     }
     await this._cacheMgr!.setMessage(message.id, message)
+    await this._transcribeVoice(message)
     const event = await parseEvent(this, message)
     log.info('event-------', JSON.stringify(event))
     switch (event.type) {
@@ -1751,13 +1878,23 @@ class PuppetMatrix extends PUPPET.Puppet {
    * @param messagePayload
    * @private
    */
+  /**
+   * 下载失败(后端没返回 url，常见于媒体已过期)时给出明确错误，而不是 FileBox 的 Invalid URL
+   */
+  private _fileBoxFromUrl (url: string | undefined, messageId: string, name?: string): FileBoxInterface {
+    if (!url) {
+      throw new Error(`download media failed for message ${messageId}: backend returned no url (media may have expired)`)
+    }
+    return name ? FileBox.fromUrl(url, { name }) : FileBox.fromUrl(url)
+  }
+
   private async _getMessageImageFileBox (messageId: string, messagePayload: MessagePayload): Promise<FileBoxInterface> {
     const imageInfo: ImageMessagePayload = await parseImageMessagePayload(messagePayload)
     if (!imageInfo.file_id) {
       log.error(`Can not get file for message: ${messageId}`)
     }
     const url = imageInfo.is_ww ? await this._client?.downloadWWMedia(imageInfo, 1) : await this._client?.downloadImage(imageInfo)
-    return FileBox.fromUrl(url)
+    return this._fileBoxFromUrl(url, messageId)
   }
 
   private async _getMessageAudioFileBox (messageId: string, messagePayload: MessagePayload): Promise<FileBoxInterface> {
@@ -1766,7 +1903,7 @@ class PuppetMatrix extends PUPPET.Puppet {
       log.error(`Can not get file for message: ${messageId}`)
     }
     const url = await this._client?.downloadAudio(messageId, audioInfo)
-    return FileBox.fromUrl(url)
+    return this._fileBoxFromUrl(url, messageId)
   }
 
   private async _getMessageVideoFileBox (messageId: string, messagePayload: MessagePayload): Promise<FileBoxInterface> {
@@ -1776,7 +1913,7 @@ class PuppetMatrix extends PUPPET.Puppet {
     }
     const url = videoInfo.is_ww ? await this._client?.downloadWWMedia(videoInfo, 2) : await this._client?.downloadVideo(videoInfo)
     log.info(PRE, '_getMessageVideoFileBox(): %s', url)
-    return FileBox.fromUrl(url)
+    return this._fileBoxFromUrl(url, messageId)
   }
 
   /**
@@ -1805,7 +1942,7 @@ class PuppetMatrix extends PUPPET.Puppet {
       log.error(`Can not get file url: ${JSON.stringify(params)}`)
     }
 
-    return FileBox.fromUrl(url, { name: appPayload.title })
+    return this._fileBoxFromUrl(url, messageId, appPayload.title)
   }
 
   // @ts-ignore