myagent-ai 1.15.93 → 1.15.95

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "myagent-ai",
-  "version": "1.15.93",
+  "version": "1.15.95",
   "description": "本地桌面端执行型AI助手 - Open Interpreter 风格 | Local Desktop Execution-Oriented AI Assistant",
   "main": "main.py",
   "bin": {

package/requirements.txt

@@ -52,6 +52,7 @@ edge-tts>=6.1.0
 # 语音识别 (本地 STT，默认启用)
 # ============================================================
 faster-whisper>=1.0.0
+pydub>=0.25.1
 
 # ============================================================
 # Anthropic Claude (可选)

package/web/ui/chat/chat_main.js

@@ -4433,13 +4433,14 @@ var VoiceInput = {
     }
   },
 
-  /** Start recording（使用 MediaRecorder） */
+  /** Start recording（使用 AudioContext 直接录制 WAV，无需后端 ffmpeg） */
   startRecording: async function() {
     if (this.isRecording) return;
 
     this.isRecording = true;
-    this.audioChunks = [];
     this.rawText = '';
+    // PCM 样本缓冲区（Float32）
+    this._pcmSamples = [];
 
     var btn = document.getElementById('voiceRecordBtn');
     if (btn) btn.classList.add('recording');
@@ -4468,41 +4469,24 @@ var VoiceInput = {
         }
       });
 
-      // 创建 MediaRecorder（优先使用 WAV 格式，回退到 WEBM）
-      var mimeType = 'audio/webm;codecs=opus';
-      if (typeof MediaRecorder !== 'undefined' && MediaRecorder.isTypeSupported) {
-        if (MediaRecorder.isTypeSupported('audio/webm;codecs=opus')) {
-          mimeType = 'audio/webm;codecs=opus';
-        } else if (MediaRecorder.isTypeSupported('audio/webm')) {
-          mimeType = 'audio/webm';
-        } else if (MediaRecorder.isTypeSupported('audio/ogg;codecs=opus')) {
-          mimeType = 'audio/ogg;codecs=opus';
-        }
-      }
-
-      this.mediaRecorder = new MediaRecorder(this._audioStream, { mimeType: mimeType });
+      // [v1.15.94] 使用 AudioContext + ScriptProcessorNode 直接录制 16kHz 单声道 PCM
+      // 最终编码为 WAV，后端无需 pydub/ffmpeg
+      var AudioCtx = window.AudioContext || window.webkitAudioContext;
+      this._audioContext = new AudioCtx({ sampleRate: 16000 });
+      var source = this._audioContext.createMediaStreamSource(this._audioStream);
+      // bufferSize 4096 @ 16kHz ≈ 256ms per chunk
+      this._scriptProcessor = this._audioContext.createScriptProcessor(4096, 1, 1);
       var self = this;
 
-      this.mediaRecorder.ondataavailable = function(e) {
-        if (e.data && e.data.size > 0) {
-          self.audioChunks.push(e.data);
-        }
-      };
-
-      this.mediaRecorder.onstop = function() {
-        self._processAudio();
+      this._scriptProcessor.onaudioprocess = function(e) {
+        if (!self.isRecording) return;
+        var inputData = e.inputBuffer.getChannelData(0);
+        self._pcmSamples.push(new Float32Array(inputData));
       };
 
-      this.mediaRecorder.onerror = function(e) {
-        console.error('MediaRecorder error:', e.error);
-        self.isRecording = false;
-        if (btn) btn.classList.remove('recording');
-        self._showStatus('录音出错，请重试', 'var(--danger)');
-        self._cleanupStream();
-      };
+      source.connect(this._scriptProcessor);
+      this._scriptProcessor.connect(this._audioContext.destination);
 
-      // 开始录音（每100ms收集一次数据）
-      this.mediaRecorder.start(100);
       this._showStatus('正在录音...', 'var(--text3)');
       this._recordingStartTime = Date.now();
 
@@ -4516,7 +4500,7 @@ var VoiceInput = {
 
   /** Stop recording and process audio */
   stopRecording: function() {
-    if (!this.isRecording || !this.mediaRecorder) return;
+    if (!this.isRecording) return;
 
     var btn = document.getElementById('voiceRecordBtn');
     if (btn) btn.classList.remove('recording');
@@ -4526,32 +4510,57 @@ var VoiceInput = {
     if (duration < 500) {
       this._showStatus('录音时间太短，请按住麦克风说话', 'var(--text3)');
       this.isRecording = false;
-      try { this.mediaRecorder.stop(); } catch (e) {}
       this._cleanupStream();
       return;
     }
 
-    this._showStatus('正在识别...', 'var(--text3)');
     this.isRecording = false;
-
-    try {
-      this.mediaRecorder.stop();
-    } catch (e) {}
+    this._cleanupStream();
+    this._showStatus('正在识别...', 'var(--text3)');
+    this._processAudio();
   },
 
-  /** 清理音频流 */
-  _cleanupStream: function() {
-    if (this._audioStream) {
-      this._audioStream.getTracks().forEach(function(t) { t.stop(); });
-      this._audioStream = null;
+  /** 清理音频流 */\n  _cleanupStream: function() {\n    // 断开 ScriptProcessor\n    if (this._scriptProcessor) {\n      try { this._scriptProcessor.disconnect(); } catch(e) {}\n      this._scriptProcessor = null;\n    }\n    // 关闭 AudioContext\n    if (this._audioContext && this._audioContext.state !== 'closed') {\n      try { this._audioContext.close(); } catch(e) {}\n      this._audioContext = null;\n    }\n    // 停止麦克风流\n    if (this._audioStream) {\n      this._audioStream.getTracks().forEach(function(t) { t.stop(); });\n      this._audioStream = null;\n    }\n  },
+
+  /** 将 PCM Float32 样本编码为 WAV Blob（16kHz, 单声道, 16bit） */
+  _encodeWav: function(samples, sampleRate) {
+    var numSamples = 0;
+    for (var i = 0; i < samples.length; i++) numSamples += samples[i].length;
+    var buffer = new ArrayBuffer(44 + numSamples * 2);
+    var view = new DataView(buffer);
+    // RIFF header
+    var writeStr = function(offset, str) {
+      for (var i = 0; i < str.length; i++) view.setUint8(offset + i, str.charCodeAt(i));
+    };
+    writeStr(0, 'RIFF');
+    view.setUint32(4, 36 + numSamples * 2, true);
+    writeStr(8, 'WAVE');
+    writeStr(12, 'fmt ');
+    view.setUint32(16, 16, true); // PCM subchunk size
+    view.setUint16(20, 1, true);  // PCM format
+    view.setUint16(22, 1, true);  // mono
+    view.setUint32(24, sampleRate, true);
+    view.setUint32(28, sampleRate * 2, true); // byte rate
+    view.setUint16(32, 2, true);  // block align
+    view.setUint16(34, 16, true); // bits per sample
+    writeStr(36, 'data');
+    view.setUint32(40, numSamples * 2, true);
+    // 写入 PCM 数据（Float32 → Int16）
+    var offset = 44;
+    for (var i = 0; i < samples.length; i++) {
+      for (var j = 0; j < samples[i].length; j++) {
+        var s = Math.max(-1, Math.min(1, samples[i][j]));
+        view.setInt16(offset, s < 0 ? s * 0x8000 : s * 0x7FFF, true);
+        offset += 2;
+      }
     }
+    return new Blob([buffer], { type: 'audio/wav' });
   },
 
-  /** 处理录音数据：发送到后端 STT */
+  /** 处理录音数据：编码为 WAV 并发送到后端 STT */
   _processAudio: async function() {
-    if (this.audioChunks.length === 0) {
+    if (!this._pcmSamples || this._pcmSamples.length === 0) {
       this._showStatus('未检测到语音，请重试', 'var(--text3)');
-      this._cleanupStream();
       return;
     }
 
@@ -4569,12 +4578,14 @@ var VoiceInput = {
     if (previewSend) previewSend.disabled = true;
 
     try {
-      var audioBlob = new Blob(this.audioChunks, { type: this.mediaRecorder ? this.mediaRecorder.mimeType : 'audio/webm' });
+      // [v1.15.94] 在浏览器端直接编码为 WAV，后端无需 pydub/ffmpeg
+      var audioBlob = this._encodeWav(this._pcmSamples, 16000);
+      this._pcmSamples = []; // 释放内存
 
-      // 发送音频到后端 STT 端点
+      // 发送 WAV 音频到后端 STT 端点
       var formData = new FormData();
-      formData.append('audio', audioBlob, 'recording.webm');
-      formData.append('format', 'webm');
+      formData.append('audio', audioBlob, 'recording.wav');
+      formData.append('format', 'wav');
 
       var resp = await fetch('/api/voice-stt', {
         method: 'POST',
@@ -4616,9 +4627,6 @@ var VoiceInput = {
   /** Cancel recording without processing */
   cancelRecording: function() {
     this.isRecording = false;
-    if (this.mediaRecorder && this.mediaRecorder.state !== 'inactive') {
-      try { this.mediaRecorder.abort(); } catch(e) {}
-    }
     this._cleanupStream();
     var btn = document.getElementById('voiceRecordBtn');
     if (btn) btn.classList.remove('recording');
@@ -4627,7 +4635,7 @@ var VoiceInput = {
       statusEl.textContent = '';
     }
     this.rawText = '';
-    this.audioChunks = [];
+    this._pcmSamples = [];
   },
 
   /** Cancel voice preview and return to voice input mode */