Fix claude mistake v6

app.py

@@ -33,11 +33,14 @@ tts_model = TTS(model_name=TTS_MODEL_NAME, progress_bar=False)
 def chat_with_bot(message, history, chat_history_ids=None):
     """
     Chat with the conversational AI model using DialoGPT.
+    CRITICAL FIX: Ensures history is a list of tuples [("user", "bot"), ...]
     Returns: (updated_history, updated_chat_ids, response_text)
     """
+    # Ensure history is correctly initialized as a list of tuples
+    if history is None:
+        history = []
+
     if not message or not message.strip():
-        # Add an empty entry to history to maintain the structure expected by Gradio
-        history.append(("", ""))
         return history, chat_history_ids, ""
 
     try:
@@ -70,13 +73,13 @@ def chat_with_bot(message, history, chat_history_ids=None):
             chat_history_ids[:, bot_input_ids.shape[-1] :][0], skip_special_tokens=True
         )
 
-        # CRITICAL FIX: Append to history in the Gradio Chatbot (list of tuples) format
+        # FIX: Append to history in the required Gradio Chatbot (list of tuples) format
         history.append((message, response))
 
         return history, chat_history_ids, response
 
     except Exception as e:
-        # CRITICAL FIX: Append error to history in the Gradio Chatbot (list of tuples) format
+        # FIX: Append error to history in the required Gradio Chatbot (list of tuples) format
         history.append((message, f"Error: {e}"))
         return history, chat_history_ids, f"Error: {e}"
 
@@ -134,7 +137,6 @@ def speech_to_text_and_chat(audio_file_path, history, chat_history_ids):
         )
 
     # 2. Chatbot
-    # The third returned value, last_response_text, is the pure text response.
     updated_history, updated_chat_ids, last_response_text = chat_with_bot(
         transcribed_text, history, chat_history_ids
     )
@@ -164,30 +166,24 @@ custom_css = """
 }
 """
 
-# CRITICAL FIX: Removed css argument from gr.Blocks()
 with gr.Blocks() as demo:
     gr.Markdown("# 🗣️ Integrated Voice Assistant (CPU Only)")
     gr.Markdown(
         "**NOTE:** This app is running on CPU-only hardware. The full voice flow will be slow due to **Text-to-Speech**."
     )
 
-    # The global chat state can be used if tabs share history, or use local states per tab
     global_chat_state = gr.State(value=None)
 
     with gr.Tabs():
 
-        # --- NEW FULL VOICE CHAT TAB (STT -> CHAT -> TTS) ---
+        # --- FULL VOICE CHAT TAB (STT -> CHAT -> TTS) ---
         with gr.TabItem("🗣️ Voice Assistant"):
             gr.Markdown("## Talk to the AI Assistant")
-            gr.Markdown(
-                "Speak into the microphone. Your speech will be transcribed, sent to the chatbot, and the chatbot's text response will be converted to audio."
-            )
 
-            # States specific to this tab
             voice_chat_history = gr.Chatbot(
                 label="Conversation Log", elem_classes=["chatbot"], value=[]
             )
-            voice_chat_state = gr.State(value=None)  # Chat state IDs for this tab
+            voice_chat_state = gr.State(value=None)
 
             with gr.Row():
                 audio_in = gr.Audio(
@@ -233,7 +229,7 @@ with gr.Blocks() as demo:
                 ],
             )
 
-        # --- EXISTING CHAT -> TTS TAB ---
+        # --- CHAT -> VOICE OUTPUT TAB ---
         with gr.TabItem("💬 Chat → Voice Output"):
             gr.Markdown("## 💬 Chat with Voice Output")
 
@@ -261,6 +257,7 @@ with gr.Blocks() as demo:
             def chat_and_speak(message, history, chat_ids):
                 """Send message to chat and convert response to speech."""
                 # 1. Chatbot
+                # The chat_with_bot returns history in the correct tuple format
                 updated_history, updated_ids, last_response = chat_with_bot(
                     message, history, chat_ids
                 )
@@ -270,7 +267,8 @@ with gr.Blocks() as demo:
 
                 return updated_history, updated_ids, last_response, audio_path, status
 
-            tts_submit_btn.click(
+            # --- Submit listener for the Enter key ---
+            fn_call = tts_msg.submit(
                 fn=chat_and_speak,
                 inputs=[tts_msg, tts_chatbot, tts_chat_state],
                 outputs=[
@@ -282,7 +280,7 @@ with gr.Blocks() as demo:
                 ],
             ).then(lambda: "", None, tts_msg)
 
-            tts_msg.submit(
+            tts_submit_btn.click(
                 fn=chat_and_speak,
                 inputs=[tts_msg, tts_chatbot, tts_chat_state],
                 outputs=[
@@ -306,7 +304,7 @@ with gr.Blocks() as demo:
                 ],
             )
 
-        # --- EXISTING TEXT CHAT ONLY TAB ---
+        # --- TEXT CHAT ONLY TAB ---
         with gr.TabItem("💬 Text Chat Only"):
             gr.Markdown("## Chat with AI Assistant")
 
@@ -323,8 +321,9 @@ with gr.Blocks() as demo:
                 submit_btn = gr.Button("Send", variant="primary")
                 clear_btn = gr.Button("Clear Chat")
 
-            # Use the global state for the text-only chat
+            # Chat functionality (uses global state, already wired for Enter key)
             fn_call = msg.submit(
+                # Use slice [:2] to discard the third output (response text)
                 lambda message, history, chat_state: chat_with_bot(
                     message, history, chat_state
                 )[:2],
@@ -333,6 +332,7 @@ with gr.Blocks() as demo:
             ).then(lambda: "", None, msg)
 
             submit_btn.click(
+                # Use slice [:2] to discard the third output (response text)
                 lambda message, history, chat_state: chat_with_bot(
                     message, history, chat_state
                 )[:2],
@@ -342,7 +342,7 @@ with gr.Blocks() as demo:
 
             clear_btn.click(lambda: ([], None), None, [chatbot, global_chat_state])
 
-        # --- EXISTING STANDALONE TTS TAB ---
+        # --- STANDALONE TTS TAB ---
         with gr.TabItem("🔊 Text-to-Speech Only"):
             gr.Markdown("## 🔊 Text-to-Speech (TTS)")
 
@@ -361,5 +361,4 @@ with gr.Blocks() as demo:
                 outputs=[standalone_audio_output, standalone_tts_status],
             )
 
-# CRITICAL FIX: Passed css argument to demo.launch()
 demo.launch(css=custom_css)