Update app.py

app.py

@@ -818,6 +818,7 @@ def get_ice_servers():
 
 
 # === Main QA Interface ===
+"""
 if st.session_state["generated_questions"]:
     idx = st.session_state["current_question_index"]
     if idx < len(st.session_state["generated_questions"]):
@@ -904,7 +905,7 @@ if st.session_state["generated_questions"]:
                     st.session_state["record_phase"] = "listening"
                     st.success("✅ Audio uploaded. You may now confirm your answer.")
                     #st.audio(wav_path, format="audio/wav")
-                    """
+                    ""
                     if st.button("⏹️ Confirm & Next"):
                         try:
                             with st.spinner("🧠 Transcribing your answer..."):
@@ -916,7 +917,7 @@ if st.session_state["generated_questions"]:
                         except Exception as e:
                                 st.error(f"❌ Transcription error: {e}")
                                 transcript = "[Transcription error]"
-                        """
+                        ""
                             
                     st.session_state["answers"].append({
                         "question": question,
@@ -991,9 +992,165 @@ if st.session_state["generated_questions"]:
                     evaluate_answers()
                     st.session_state["show_summary"] = True
                 st.rerun()
+"""
 
+if st.session_state.get("generated_questions"):
+    idx = st.session_state.get("current_question_index", 0)
+    if idx < len(st.session_state["generated_questions"]):
+        question = st.session_state["generated_questions"][idx].lstrip("1234567890. ").strip()
 
-# === Summary Display ===
+        # Phase 0: Generate & play question audio
+        if not st.session_state.get("question_played"):
+            st.session_state["question_audio_file"] = asyncio.run(generate_question_audio(question))
+            st.session_state.update({
+                "question_played": True,
+                "question_start_time": time.time(),
+                "record_phase": "audio_playing",
+                "recorded_text": "",
+                "response_file": None
+            })
+        st.markdown(f"**Q{idx + 1}:** {question}")
+        st.audio(st.session_state["question_audio_file"], format="audio/mp3")
+
+        now = time.time()
+        elapsed = now - st.session_state.get("question_start_time", 0)
+
+        # Phase 1: Audio Playing
+        if st.session_state["record_phase"] == "audio_playing":
+            if elapsed < 5:
+                st.markdown("<h4 class='timer-text'>🔊 Playing question audio... Please listen</h4>", unsafe_allow_html=True)
+                time.sleep(1)
+                st.rerun()
+            else:
+                st.session_state["record_phase"] = "waiting_to_start"
+                st.session_state["question_start_time"] = time.time()
+                st.rerun()
+        # Phase 2: Waiting to Start Recording
+        elif st.session_state["record_phase"] == "waiting_to_start":
+            remaining = 10 - int(elapsed)
+            if remaining > 0:
+                st.markdown(f"<h4 class='timer-text'>⏳ {remaining} seconds to click 'Start Recording'...</h4>", unsafe_allow_html=True)
+                if st.button("🎙️ Start Recording"):
+                    st.session_state.update({
+                        "record_phase": "recording",
+                        "timer_start": time.time(),
+                        "recording_started": True,
+                        "response_file": None
+                    })
+                    st.rerun()
+                time.sleep(1)
+                st.rerun()
+            else:
+                st.markdown("<div style='padding:10px; background:#fff8e1; border-left:5px solid orange;color: #212529;'>⚠️ <strong>No action detected.</strong> Automatically skipping to next question...</div>", unsafe_allow_html=True)
+                st.session_state["answers"].append({"question": question, "response": "[No response]"})
+                st.session_state.update({
+                    "record_phase": "idle",
+                    "question_played": False,
+                    "question_start_time": 0.0,
+                    "current_question_index": idx + 1
+                })
+                if st.session_state["current_question_index"] == len(st.session_state["generated_questions"]):
+                    evaluate_answers()
+                    st.session_state["show_summary"] = True
+                st.rerun()
+        # Phase 3: Recording
+        elif st.session_state["record_phase"] == "recording":
+            remaining = 15 - int(now - st.session_state.get("timer_start", 0))
+            if remaining > 0:
+                st.markdown(f"<h4 class='timer-text'>🎙️ {remaining} seconds to answer...</h4>", unsafe_allow_html=True)
+
+                audio_value = st.audio_input("🎤 Tap to record — then stop when done", key=f"audio_{idx}")
+
+                # Defer processing by saving to temp_audio
+                if audio_value:
+                    st.session_state["temp_audio"] = audio_value
+                    st.rerun()
+
+                # Process the deferred audio input
+                if "temp_audio" in st.session_state and st.session_state.get("response_file") is None:
+                    audio_value = st.session_state.pop("temp_audio")
+                    wav_path = f"response_{idx}.wav"
+                    with open(wav_path, "wb") as f:
+                        f.write(audio_value.getbuffer())
+
+                    recognizer = sr.Recognizer()
+
+                    try:
+                        with sr.AudioFile(wav_path) as source:
+                            audio = recognizer.record(source)
+                            transcript = recognizer.recognize_google(audio)
+                    except sr.UnknownValueError:
+                        transcript = "[Could not understand audio]"
+                    except sr.RequestError:
+                        transcript = "[Google API error]"
+                    except Exception as e:
+                        transcript = f"[Transcription failed: {e}]"
+
+                    st.session_state["response_file"] = wav_path
+                    st.session_state["record_phase"] = "listening"
+                    st.success("✅ Audio uploaded. You may now confirm your answer.")
+
+                    st.session_state["answers"].append({
+                        "question": question,
+                        "response_file": wav_path,
+                        "response": transcript
+                    })
+
+                    if st.session_state["current_question_index"] == len(st.session_state["generated_questions"]):
+                        evaluate_answers()
+                        st.session_state["show_summary"] = True
+
+                    st.rerun()
+
+                    if elapsed > 15 and st.session_state.get("response_file") is None:
+                        st.warning("⚠️ No audio captured. Moving to next question.")
+                        st.session_state["answers"].append({"question": question, "response": "[No response]"})
+                        st.session_state.update({
+                            "record_phase": "idle",
+                            "question_played": False,
+                            "current_question_index": idx + 1
+                        })
+                        if st.session_state["current_question_index"] == len(st.session_state["generated_questions"]):
+                            evaluate_answers()
+                            st.session_state["show_summary"] = True
+                        st.rerun()
+
+            else:
+                st.markdown("<div style='padding:10px; background:#fff3e0; border-left:5px solid orange;'>⚠️ <strong>No response detected.</strong> Moving to next question...</div>", unsafe_allow_html=True)
+                st.session_state["answers"].append({"question": question, "response": "[No response]"})
+                st.session_state.update({
+                    "record_phase": "idle",
+                    "recording_started": False,
+                    "question_played": False,
+                    "question_start_time": 0.0,
+                    "current_question_index": idx + 1
+                })
+                if st.session_state["current_question_index"] == len(st.session_state["generated_questions"]):
+                    evaluate_answers()
+                    st.session_state["show_summary"] = True
+                st.rerun()
+
+        # Phase 4: Listening / Confirming
+        elif st.session_state["record_phase"] == "listening":
+            st.success("🎧 Review your recorded response below:")
+            st.audio(st.session_state["response_file"], format="audio/wav")
+
+            if st.button("⏹️ Confirm & Next"):
+                st.session_state.update({
+                    "record_phase": "idle",
+                    "recording_started": False,
+                    "question_played": False,
+                    "question_start_time": 0.0,
+                    "current_question_index": idx + 1,
+                    "response_file": None,
+                    "audio_waiting": True
+                })
+
+                if st.session_state["current_question_index"] == len(st.session_state["generated_questions"]):
+                    evaluate_answers()
+                    st.session_state["show_summary"] = True
+                st.rerun()
+                    
  
 # === Summary Display ===
 if st.session_state.get("show_summary", False):