Update app.py

app.py

@@ -49,14 +49,12 @@ MAIN_SYSTEM_PROMPT = {
 
 WELCOME_PROMPT = {
     "role": "system",
-    "content": """Create a warm welcome message that:
+    "content": """Create a brief welcome message that:
         1. Introduces you as Sam, the English tutor
-        2. Asks for the user's name
-        3. Asks about their background (country, job/studies)
-        4. Inquires about their English learning goals
+        2. Asks for the user's name only
         
         Return in JSON format with key 'greeting'.
-        Keep it friendly and encouraging."""
+        Keep it short, friendly and natural - maximum 2 sentences."""
 }
 
 class EnglishTutor:
@@ -129,14 +127,17 @@ class EnglishTutor:
     def _format_response(self, response_dict):
         formatted = response_dict["response"]
         
-        if response_dict.get("corrections"):
-            formatted += f"\n\n✍️ Corrections: {response_dict['corrections']}"
-        
-        if response_dict.get("vocabulary"):
-            formatted += f"\n\n📚 Vocabulary: {response_dict['vocabulary']}"
-        
         if response_dict.get("encouragement"):
-            formatted += f"\n\n🌟 {response_dict['encouragement']}"
+            encouragement = response_dict["encouragement"]
+            # Make encouragement more impressive
+            if "doing" in encouragement.lower() and "great" in encouragement.lower():
+                formatted += f"\n\n🌟 {encouragement}"
+            elif "wonderful" in encouragement.lower() or "excellent" in encouragement.lower():
+                formatted += f"\n\n✨ {encouragement}"
+            elif "good" in encouragement.lower() or "well" in encouragement.lower():
+                formatted += f"\n\n🎯 {encouragement}"
+            else:
+                formatted += f"\n\n💫 {encouragement}"
         
         return formatted
 
@@ -178,32 +179,66 @@ def initialize_chat():
     try:
         welcome = tutor.get_welcome_message()
         welcome_audio = text_to_speech(welcome)
-        return [(None, welcome)], welcome_audio, ""
+        return [(None, welcome)], welcome_audio, f"🎤 Sam: {welcome}", ""
     except Exception as e:
         print(f"Error initializing chat: {str(e)}")
-        return [(None, "Hi! I'm Sam, your English tutor. What's your name?")], None, ""
+        return [(None, "Hi! I'm Sam, your English tutor. What's your name?")], None, "🎤 Sam: Hi! I'm Sam, your English tutor. What's your name?", ""
 
-def process_audio(audio, history, transcript):
+def process_audio(audio, history, transcript, corrections):
     try:
         if audio is None:
-            return history, None, transcript
+            return history, None, transcript, corrections
 
         user_message = convert_audio_to_text(audio)
         if not user_message:
-            return history, None, transcript
+            return history, None, transcript, corrections
 
-        bot_response = tutor.get_bot_response(user_message)
-        audio_response = text_to_speech(bot_response)
+        bot_response_json = tutor.get_bot_response(user_message)
+        
+        # Extract the JSON response for corrections
+        try:
+            response_dict = json.loads(bot_response_json) if isinstance(bot_response_json, str) else bot_response_json
+        except:
+            response_dict = {"response": bot_response_json}
+        
+        # Format response for TTS (without corrections and vocabulary)
+        tts_response = response_dict.get("response", bot_response_json)
+        if response_dict.get("encouragement"):
+            tts_response += f" {response_dict['encouragement']}"
+        
+        # Generate audio from clean response
+        audio_response = text_to_speech(tts_response)
+        
+        # Format full response for chat display
+        formatted_response = tutor._format_response(response_dict) if hasattr(tutor, '_format_response') else bot_response_json
 
         history = history or []
-        history.append((user_message, bot_response))
+        history.append((user_message, formatted_response))
+        
+        new_transcript = transcript + f"\n\n🎤 You: {user_message}\n🤖 Sam: {tts_response}"
         
-        new_transcript = transcript + f"\n\n🎤 You said: {user_message}\n🤖 Sam replied: {bot_response}"
+        # Update corrections display
+        new_corrections = corrections
+        if response_dict.get("corrections") or response_dict.get("vocabulary"):
+            correction_text = ""
+            if response_dict.get("corrections"):
+                correction_text += f"✍️ Grammar: {response_dict['corrections']}\n\n"
+            if response_dict.get("vocabulary"):
+                vocab = response_dict['vocabulary']
+                if isinstance(vocab, dict):
+                    vocab_text = "\n".join([f"• '{k}' → '{v}'" for k, v in vocab.items()])
+                else:
+                    vocab_text = str(vocab)
+                correction_text += f"📚 Vocabulary:\n{vocab_text}"
+            new_corrections = corrections + f"\n\n{correction_text}" if corrections else correction_text
         
-        return history, audio_response, new_transcript
+        return history, audio_response, new_transcript, new_corrections
     except Exception as e:
         print(f"Error in process_audio: {str(e)}")
-        return history, None, transcript
+        return history, None, transcript, corrections
+
+def submit_recording(audio, history, transcript, corrections):
+    return process_audio(audio, history, transcript, corrections)
 
 def clear_chat():
     global tutor
@@ -212,55 +247,75 @@ def clear_chat():
 
 with gr.Blocks(theme=gr.themes.Soft()) as demo:
     gr.Markdown("# 🎓 English Learning Assistant")
-    gr.Markdown("Speak naturally and I'll help you improve your English!")
+    gr.Markdown("🎤 **Record your voice and click submit** - I'll help you improve your English conversation skills!")
     
     with gr.Row():
-        with gr.Column(scale=2):
+        with gr.Column(scale=3):
             chatbot = gr.Chatbot(
-                height=400,
+                height=500,
                 bubble_full_width=False,
-                show_label=False
+                show_label=False,
+                avatar_images=("🎤", "🤖")
             )
             
             with gr.Row():
-                audio_input = gr.Audio(
-                    label="Your Voice",
-                    type="filepath",
-                    show_label=True
+                with gr.Column(scale=2):
+                    audio_input = gr.Audio(
+                        label="🎙️ Record your voice",
+                        type="filepath",
+                        show_label=True
+                    )
+                with gr.Column(scale=1):
+                    submit_btn = gr.Button("📤 Submit Recording", variant="primary", size="lg")
+                with gr.Column(scale=2):
+                    audio_output = gr.Audio(
+                        label="🔊 Sam's response",
+                        type="filepath",
+                        show_label=True,
+                        autoplay=True
+                    )
+        
+        with gr.Column(scale=2):
+            with gr.Box():
+                gr.Markdown("### 📝 Live Transcript")
+                transcript_display = gr.Textbox(
+                    lines=10,
+                    max_lines=10,
+                    show_label=False,
+                    interactive=False,
+                    placeholder="Your conversation will appear here...",
+                    autoscroll=True
                 )
-                audio_output = gr.Audio(
-                    label="Tutor's Voice",
-                    type="filepath",
-                    show_label=True
+            
+            with gr.Box():
+                gr.Markdown("### 📚 Learning Corner")
+                corrections_display = gr.Textbox(
+                    lines=8,
+                    max_lines=8,
+                    show_label=False,
+                    interactive=False,
+                    placeholder="Grammar corrections and vocabulary suggestions will appear here...",
+                    autoscroll=True
                 )
-        
-        with gr.Column(scale=1):
-            gr.Markdown("### 📝 Conversation Transcript")
-            transcript_display = gr.Textbox(
-                lines=15,
-                max_lines=15,
-                show_label=False,
-                interactive=False,
-                placeholder="Your conversation transcript will appear here..."
-            )
     
     with gr.Row():
-        clear_btn = gr.Button("🔄 Clear Chat", variant="secondary")
+        clear_btn = gr.Button("🔄 Start New Conversation", variant="secondary", size="lg")
+        gr.Markdown("💡 **Tip**: Check the Learning Corner for personalized feedback on your English!")
     
-    audio_input.change(
-        process_audio,
-        inputs=[audio_input, chatbot, transcript_display],
-        outputs=[chatbot, audio_output, transcript_display]
+    submit_btn.click(
+        submit_recording,
+        inputs=[audio_input, chatbot, transcript_display, corrections_display],
+        outputs=[chatbot, audio_output, transcript_display, corrections_display]
     )
     
     clear_btn.click(
         clear_chat,
-        outputs=[chatbot, audio_output, transcript_display]
+        outputs=[chatbot, audio_output, transcript_display, corrections_display]
     )
     
     demo.load(
         initialize_chat,
-        outputs=[chatbot, audio_output, transcript_display]
+        outputs=[chatbot, audio_output, transcript_display, corrections_display]
     )
 
 if __name__ == "__main__":