taking out type

src/app.py

@@ -175,59 +175,38 @@ def process_speech(audio_path, history):
         if not audio_path:
             return []
             
-        # The audio_path now contains the transcribed text directly from Gradio
-        transcript = audio_path
+        # Extract just the transcribed text if it's a tuple
+        transcript = audio_path[1] if isinstance(audio_path, tuple) else audio_path
             
         # Query the symptom index
         diagnosis_query = f"""
         Given these symptoms: '{transcript}'
-        
-        Identify the most likely ICD-10 diagnoses and key questions to differentiate between them.
-        Focus only on symptoms mentioned and their clinical implications.
+        Identify the most likely ICD-10 diagnoses and key questions.
+        Focus on clinical implications.
         """
         
         response = symptom_index.as_query_engine().query(diagnosis_query)
         
-        # Format response
-        formatted_response = {
-            "diagnoses": [],
-            "confidences": [],
-            "follow_up": str(response)
-        }
-        
         return [
             {"role": "user", "content": transcript},
-            {"role": "assistant", "content": json.dumps(formatted_response)}
+            {"role": "assistant", "content": json.dumps({
+                "diagnoses": [],
+                "confidences": [],
+                "follow_up": str(response)
+            })}
         ]
         
     except Exception as e:
         print(f"Error processing speech: {e}")
         return []
 
-def text_to_speech(text):
-    """Convert text to speech and return audio HTML element."""
-    tts = gTTS(text=text, lang='en')
-    audio_fp = io.BytesIO()
-    tts.write_to_fp(audio_fp)
-    audio_b64 = base64.b64encode(audio_fp.getvalue()).decode()
-    return f'<audio src="data:audio/mp3;base64,{audio_b64}" autoplay></audio>'
-
-def format_response_for_user(response_dict):
-    """Convert JSON response to user-friendly format."""
-    diagnoses = response_dict.get("diagnoses", [])
-    confidences = response_dict.get("confidences", [])
-    follow_up = response_dict.get("follow_up", "")
-    
-    message = ""
-    if diagnoses and confidences:
-        for d, c in zip(diagnoses, confidences):
-            conf_percent = int(c * 100)
-            message += f"Possible diagnosis ({conf_percent}% confidence): {d}\n"
-    
-    if follow_up:
-        message += f"\n{follow_up}"
-    
-    return message
+def update_transcription(audio_path):
+    """Update transcription box with speech recognition results."""
+    if not audio_path:
+        return ""
+    # Extract transcription from audio result
+    transcript = audio_path[1] if isinstance(audio_path, tuple) else audio_path
+    return transcript
 
 # Build enhanced Gradio interface
 with gr.Blocks(
@@ -261,7 +240,6 @@ with gr.Blocks(
             # Moved microphone row above chatbot
             with gr.Row():
                 microphone = gr.Audio(
-                    type="filepath",  # Use filepath to get the audio file path
                     label="Describe your symptoms",
                     streaming=True
                 )
@@ -302,6 +280,21 @@ with gr.Blocks(
     # Event handlers
     clear_btn.click(lambda: None, None, chatbot, queue=False)
     
+    def format_response_for_user(response_dict):
+        """Format the assistant's response dictionary into a user-friendly string."""
+        diagnoses = response_dict.get("diagnoses", [])
+        confidences = response_dict.get("confidences", [])
+        follow_up = response_dict.get("follow_up", "")
+        result = ""
+        if diagnoses:
+            result += "Possible Diagnoses:\n"
+            for i, diag in enumerate(diagnoses):
+                conf = f" ({confidences[i]*100:.1f}%)" if i < len(confidences) else ""
+                result += f"- {diag}{conf}\n"
+        if follow_up:
+            result += f"\nFollow-up: {follow_up}"
+        return result.strip()
+
     def enhanced_process_speech(audio_path, history, api_key=None, model_tier="small", temp=0.7):
         """Handle speech processing and chat formatting."""
         if not audio_path:
@@ -341,6 +334,19 @@ with gr.Blocks(
         show_progress="hidden"
     )
     
+    microphone.stream(  # Add real-time transcription updates
+        fn=update_transcription,
+        inputs=[microphone],
+        outputs=transcript_box,
+        show_progress="hidden"
+    )
+    
+    clear_btn.click(
+        fn=lambda: (None, ""),  # Clear both chat and transcription
+        outputs=[chatbot, transcript_box],
+        queue=False
+    )
+    
     # Add footer with social links
     gr.Markdown("""
     ---