feat: 添加语音模块支持，优化服务启动方式

api/.env.example

@@ -0,0 +1,2 @@
+MINIMAX_API_KEY=your_api_key_here
+MINIMAX_GROUP_ID=your_group_id_here

api/app.ts

@@ -23,8 +23,12 @@ import { apiModules } from './modules/index.js'
 import { validateModuleConsistency } from './infra/moduleValidator.js'
 import path from 'path'
 import fs from 'fs'
+import { fileURLToPath } from 'url'
 
-dotenv.config()
+const __filename = fileURLToPath(import.meta.url)
+const __dirname = path.dirname(__filename)
+
+dotenv.config({ path: path.resolve(__dirname, './.env') })
 
 const app: express.Application = express()
 export const container = new ServiceContainer()

api/config/index.ts

@@ -38,6 +38,18 @@ export const config = {
   get isDev(): boolean {
     return !this.isElectron && !this.isVercel
   },
+
+  get minimaxApiKey(): string | undefined {
+    return process.env.MINIMAX_API_KEY
+  },
+
+  get minimaxGroupId(): string | undefined {
+    return process.env.MINIMAX_GROUP_ID
+  },
+
+  get openaiApiKey(): string | undefined {
+    return process.env.OPENAI_API_KEY
+  },
 }
 
 export const PATHS = {

api/modules/index.ts

@@ -63,6 +63,13 @@ async function getStaticModules(): Promise<ApiModule[]> {
     console.warn('[ModuleLoader] Failed to load opencode module:', e)
   }
 
+  try {
+    const { createVoiceModule } = await import('./voice/index.js')
+    modules.push(createVoiceModule())
+  } catch (e) {
+    console.warn('[ModuleLoader] Failed to load voice module:', e)
+  }
+
   modules.sort((a, b) => {
     const orderA = a.metadata.order ?? 0
     const orderB = b.metadata.order ?? 0
@@ -82,3 +89,4 @@ export * from './document-parser/index.js'
 export * from './ai/index.js'
 export * from './remote/index.js'
 export * from './opencode/index.js'
+export * from './voice/index.js'

api/modules/voice/index.ts

@@ -0,0 +1,25 @@
+import type { Router } from 'express'
+import type { ServiceContainer } from '../../infra/container.js'
+import { createApiModule } from '../../infra/createModule.js'
+import { voiceModule, VOICE_MODULE } from '../../../shared/modules/voice/index.js'
+import { createVoiceRoutes } from './routes.js'
+
+export * from './routes.js'
+
+export const createVoiceModule = () => {
+  return createApiModule(
+    {
+      ...voiceModule,
+      basePath: '/voice',
+      version: '1.0.0',
+      order: 100,
+    },
+    {
+      routes: (_container: ServiceContainer): Router => {
+        return createVoiceRoutes()
+      },
+    }
+  )
+}
+
+export default createVoiceModule

api/modules/voice/routes.ts

@@ -0,0 +1,117 @@
+import { Router } from 'express'
+import { config } from '../../config/index.js'
+
+interface AIMLAPISTTCreateResponse {
+  generation_id?: string
+  status?: string
+  error?: {
+    message: string
+  }
+}
+
+interface AIMLAPISTTQueryResponse {
+  status?: string
+  result?: {
+    results?: {
+      channels?: Array<{
+        alternatives?: Array<{
+          transcript?: string
+        }>
+      }>
+    }
+  }
+  error?: {
+    message: string
+  }
+}
+
+export function createVoiceRoutes(): Router {
+  const router = Router()
+
+  router.post('/stt', async (req, res) => {
+    console.log('[Voice] Request received, body keys:', Object.keys(req.body || {}))
+    try {
+      const apiKey = config.minimaxApiKey
+
+      console.log('[Voice] API Key exists:', !!apiKey)
+
+      if (!apiKey) {
+        res.status(500).json({ error: 'MiniMax API key not configured' })
+        return
+      }
+
+      if (!req.body.audio) {
+        res.status(400).json({ error: 'No audio data provided' })
+        return
+      }
+
+      const audioBuffer = Buffer.from(req.body.audio, 'base64')
+      console.log('[Voice] Audio buffer size:', audioBuffer.length)
+
+      const formData = new FormData()
+      const blob = new Blob([audioBuffer], { type: 'audio/webm' })
+      formData.append('file', blob, 'audio.webm')
+      formData.append('model', '#g1_whisper-large')
+
+      console.log('[Voice] Creating STT job via MiniMax...')
+
+      const createResponse = await fetch('https://api.minimax.chat/v1/stt/create', {
+        method: 'POST',
+        headers: {
+          'Authorization': `Bearer ${apiKey}`,
+        },
+        body: formData,
+      })
+
+      console.log('[Voice] Create response status:', createResponse.status)
+      const createData: AIMLAPISTTCreateResponse = await createResponse.json()
+      console.log('[Voice] Create response:', createData)
+
+      if (!createResponse.ok || !createData.generation_id) {
+        console.error('[Voice] Failed to create STT job:', createData.error?.message)
+        res.status(500).json({ error: createData.error?.message || 'Failed to create STT job' })
+        return
+      }
+
+      const jobId = createData.generation_id
+      console.log('[Voice] Job ID:', jobId)
+
+      console.log('[Voice] Polling for result...')
+      let resultText = ''
+      for (let i = 0; i < 30; i++) {
+        await new Promise(resolve => setTimeout(resolve, 1000))
+
+        const queryResponse = await fetch(`https://api.minimax.chat/v1/stt/${jobId}`, {
+          headers: {
+            'Authorization': `Bearer ${apiKey}`,
+          },
+        })
+
+        const queryData: AIMLAPISTTQueryResponse = await queryResponse.json()
+        console.log('[Voice] Query response:', queryData)
+
+        if (queryData.status === 'succeeded') {
+          resultText = queryData.result?.results?.channels?.[0]?.alternatives?.[0]?.transcript || ''
+          break
+        } else if (queryData.status === 'failed') {
+          console.error('[Voice] STT job failed:', queryData.error?.message)
+          res.status(500).json({ error: queryData.error?.message || 'STT processing failed' })
+          return
+        }
+      }
+
+      if (!resultText) {
+        res.status(500).json({ error: 'STT processing timeout' })
+        return
+      }
+
+      console.log('[Voice] Final result:', resultText)
+      res.json({ text: resultText })
+    } catch (error) {
+      console.error('[Voice] STT error:', error)
+      res.status(500).json({ error: 'Failed to process audio' })
+    }
+  })
+
+  return router
+}

electron/services/opencodeService.ts

@@ -150,7 +150,8 @@ class OpenCodeService {
     try {
       log.info('[OpenCodeService] Stopping...');
 
-      const pid = this.process.pid;
+      const pid = this.process?.pid;
+      const processRef = this.process;
       this.process = null;
       this._isRunning = false;
 
@@ -165,11 +166,11 @@ class OpenCodeService {
             resolve({ success: true });
           });
         });
-      } else if (pid) {
-        this.process?.kill('SIGTERM');
+      } else if (pid && processRef) {
+        processRef.kill('SIGTERM');
         await new Promise<void>((resolve) => setTimeout(resolve, 1000));
-        if (this.process && !this.process.killed) {
-          this.process.kill('SIGKILL');
+        if (!processRef.killed) {
+          processRef.kill('SIGKILL');
         }
         this.restartAttempts = 0;
         log.info('[OpenCodeService] Stopped');

package.json

@@ -65,6 +65,7 @@
     "dotenv": "^17.2.1",
     "electron-log": "^5.4.3",
     "express": "^4.21.2",
+    "form-data": "^4.0.5",
     "github-slugger": "^2.0.0",
     "jszip": "^3.10.1",
     "lucide-react": "^0.511.0",

remote/config/default.json

@@ -23,12 +23,10 @@
     "tokenExpiry": 3600
   },
   "frp": {
-    "enabled": true,
-    "frpcPath": "./frp/frpc.exe",
-    "configPath": "./frp/frpc.toml"
+    "enabled": true
   },
   "opencode": {
-    "enabled": true
+    "enabled": false
   },
   "xcopencodeweb": {
     "enabled": true,

remote/src/core/App.js

@@ -118,16 +118,6 @@ class App {
       });
     });
 
-    this.container.register('openCodeService', (c) => {
-      const OpenCodeService = require('../services/opencode/OpenCodeService');
-      const config = c.resolve('config');
-      const opencodeConfig = config.getSection('opencode') || {};
-      return new OpenCodeService({
-        enabled: opencodeConfig.enabled !== false,
-        port: opencodeConfig.port
-      });
-    });
-
     this.container.register('xcOpenCodeWebService', (c) => {
       const XCOpenCodeWebService = require('../services/opencode/XCOpenCodeWebService');
       const config = c.resolve('config');
@@ -212,10 +202,6 @@ class App {
     frpService.start();
     logger.info('FRP service started');
 
-    const openCodeService = this.container.resolve('openCodeService');
-    openCodeService.start();
-    logger.info('OpenCode service started');
-
     const xcOpenCodeWebService = this.container.resolve('xcOpenCodeWebService');
     xcOpenCodeWebService.start();
     logger.info('XCOpenCodeWeb service started');
@@ -498,10 +484,6 @@ class App {
       frpService.stop();
       logger.info('FRP service stopped');
 
-      const openCodeService = this.container.resolve('openCodeService');
-      openCodeService.stop();
-      logger.info('OpenCode service stopped');
-
       const xcOpenCodeWebService = this.container.resolve('xcOpenCodeWebService');
       xcOpenCodeWebService.stop();
       logger.info('XCOpenCodeWeb service stopped');

remote/src/services/network/FRPService.js

@@ -64,9 +64,15 @@ class FRPService {
         configPath: runtimeConfigPath 
       });
 
-      this.process = spawn(this.frpcPath, ['-c', runtimeConfigPath], {
+      this.process = spawn('cmd.exe', [
+        '/c',
+        this.frpcPath,
+        '-c',
+        runtimeConfigPath
+      ], {
         stdio: ['ignore', 'pipe', 'pipe'],
-        windowsHide: true
+        windowsHide: true,
+        shell: false
       });
 
       this.isRunning = true;

remote/src/services/opencode/OpenCodeService.js

@@ -11,9 +11,17 @@ class OpenCodeService {
     this.enabled = options.enabled !== false;
     
     try {
-      const result = execSync('where opencode', { encoding: 'utf8', windowsHide: true });
-      const firstLine = result.split('\n')[0].trim();
-      this.opencodePath = firstLine.endsWith('.cmd') ? firstLine : firstLine + '.cmd';
+      const result = execSync('powershell -Command "Get-Command opencode -ErrorAction SilentlyContinue | Select-Object -ExpandProperty Source"', { encoding: 'utf8', windowsHide: true, maxBuffer: 1024 * 1024 });
+      const opencodePath = result.trim();
+      if (opencodePath && opencodePath.length > 0) {
+        let finalPath = opencodePath;
+        if (!finalPath.toLowerCase().endsWith('.cmd') && !finalPath.toLowerCase().endsWith('.exe')) {
+          finalPath = finalPath + '.cmd';
+        }
+        this.opencodePath = finalPath;
+      } else {
+        this.opencodePath = null;
+      }
     } catch (e) {
       this.opencodePath = null;
     }
@@ -49,11 +57,10 @@ class OpenCodeService {
         OPENCODE_SERVER_PASSWORD: password
       };
 
-      const portArg = this.port ? ` --port ${this.port}` : '';
-      this.process = spawn('powershell.exe', [
-        '-NoProfile',
-        '-Command',
-        `& '${this.opencodePath}' serve${portArg}`
+      const portArg = this.port ? ' --port ' + this.port : '';
+      this.process = spawn('cmd.exe', [
+        '/c',
+        'chcp 65001 >nul && ' + this.opencodePath + ' serve' + portArg
       ], {
         stdio: ['ignore', 'pipe', 'pipe'],
         env,
@@ -63,14 +70,14 @@ class OpenCodeService {
       this.isRunning = true;
 
       this.process.stdout.on('data', (data) => {
-        const output = data.toString().trim();
+        const output = data.toString('utf8').trim();
         if (output) {
           logger.info(`[OpenCode] ${output}`);
         }
       });
 
       this.process.stderr.on('data', (data) => {
-        const output = data.toString().trim();
+        const output = data.toString('utf8').trim();
         if (output) {
           logger.error(`[OpenCode] ${output}`);
         }

remote/src/services/opencode/XCOpenCodeWebService.js

@@ -15,7 +15,7 @@ class XCOpenCodeWebService {
 
   getExePath() {
     const exeName = 'XCOpenCodeWeb.exe';
-    const basePath = path.join(__dirname, '../../xcopencodeweb');
+    const basePath = path.join(__dirname, '../../../xcopencodeweb');
     return path.join(basePath, exeName);
   }
 

remote/src/utils/paths.js

@@ -5,7 +5,7 @@ function getBasePath() {
   if (process.pkg) {
     return path.dirname(process.execPath);
   }
-  return path.join(__dirname, '../..');
+  return process.cwd();
 }
 
 function getPublicPath() {

shared/modules/voice/index.ts

@@ -0,0 +1,13 @@
+export const VOICE_MODULE = 'voice' as const
+
+export interface VoiceModule {
+  id: typeof VOICE_MODULE
+  name: '语音'
+  icon: 'mic'
+}
+
+export const voiceModule: VoiceModule = {
+  id: VOICE_MODULE,
+  name: '语音',
+  icon: 'mic',
+}

src/components/chat/icons.tsx

@@ -9,6 +9,7 @@ import {
   Trash,
   PanelLeftClose,
   PanelLeft,
+  Mic,
 } from 'lucide-react'
 
 export {
@@ -22,4 +23,5 @@ export {
   Trash as TrashIcon,
   PanelLeftClose as PanelLeftCloseIcon,
   PanelLeft as PanelLeftOpenIcon,
+  Mic as MicIcon,
 }