Merge branch 'dev/hz_v1.0.5' into 'main'

realtime

See merge request !19

openai/ws/params.go

@@ -70,12 +70,14 @@ const (
 	ClientAudioCommit            Event = "input_audio_buffer.commit"
 	ClientAudioClear             Event = "input_audio_buffer.clear"
 
-	ResultTypeNone          LLMResultType = 0
-	ResultTypeText          LLMResultType = 1
-	ResultTypeTextDone      LLMResultType = 2
-	ResultTypeAudio         LLMResultType = 3
-	ResultTypeAudioDone     LLMResultType = 4
-	ResultTypeAudioText     LLMResultType = 5
-	ResultTypeAudioTextDone LLMResultType = 6
-	ResultTypeError         LLMResultType = 7
+	ResultTypeNone                   LLMResultType = 0
+	ResultTypeText                   LLMResultType = 1
+	ResultTypeTextDone               LLMResultType = 2
+	ResultTypeAudio                  LLMResultType = 3
+	ResultTypeAudioDone              LLMResultType = 4
+	ResultTypeAudioText              LLMResultType = 5
+	ResultTypeAudioTextDone          LLMResultType = 6
+	ResultTypeError                  LLMResultType = 7
+	ResultTypeAudioBufferSpeechStart LLMResultType = 20
+	ResultTypeAudioBufferCommit      LLMResultType = 30
 )

openai/ws/wsclient.go

@@ -218,10 +218,12 @@ func (c *WSClient) ResolveMessage(response Response, msg []byte) (result LLMResu
 		fmt.Println(responseType, "新项目创建完成...")
 	case ServerAudioCommit:
 		fmt.Println(responseType, "提交音频...")
+		result = LLMResult{Type: ResultTypeAudioBufferCommit}
 	case ServerAudioClear:
 		fmt.Println(responseType, "清除音频...")
 	case ServerAudioSpeechStart:
 		fmt.Println(responseType, "缓冲区收到音频...")
+		result = LLMResult{Type: ResultTypeAudioBufferSpeechStart}
 	case ServerAudioSpeechStop:
 		fmt.Println(responseType, "缓冲区停止接收音频...")
 	case ServerAudioDone: