Update app.py

app.py

@@ -31,7 +31,7 @@ def transcribe(audio):
     global current_speaker
     if audio is None:
         logging.info("No audio input received")
-        return "No audio input received"
+        return json.dumps({"error": "No audio input received"})
     
     try:
         logging.info(f"Audio input received: {type(audio)}")
@@ -61,10 +61,10 @@ def transcribe(audio):
                 logging.info(f"Transcription result: {text}")
             except Exception as e:
                 logging.error(f"Error during transcription: {e}", exc_info=True)
-                return f"Transcription error: {str(e)}"
+                return json.dumps({"error": f"Transcription error: {str(e)}"})
         else:
             logging.error("Transcriber not initialized")
-            return "Transcriber not initialized"
+            return json.dumps({"error": "Transcriber not initialized"})
 
         # 创建结构化数据
         if text:
@@ -84,7 +84,7 @@ def transcribe(audio):
         return formatted_conversation
     except Exception as e:
         logging.error(f"Error in transcribe function: {e}", exc_info=True)
-        return f"Error: {str(e)}"
+        return json.dumps({"error": f"Error: {str(e)}"})
 
 def switch_speaker():
     global current_speaker
@@ -132,7 +132,7 @@ with gr.Blocks() as demo:
     gr.Markdown("点击麦克风图标开始录音，说话后会自动进行语音识别。支持中文识别。")
     
     with gr.Row():
-        audio_input = gr.Audio(sources=["microphone"], type="numpy")
+        audio_input = gr.Audio(sources=["microphone"], type="numpy", sample_rate=16000)
         speaker_button = gr.Button("切换说话者")
     
     speaker_label = gr.Label("当前说话者：患者")