Update app.py

app.py

@@ -19,8 +19,6 @@ from twilio.rest import Client
 import logging
 import whisper
 import speech_recognition as sr
-from audiorecorder import audiorecorder
-import numpy as np
 #model = whisper.load_model("base")
 
 
@@ -817,20 +815,7 @@ def get_ice_servers():
 
     return token.ice_servers
 
-# Function to generate question audio and save to a valid path
-async def generate_question_audio(question, voice="en-IE-EmilyNeural"):
-    clean_question = question.strip().replace("\n", " ")
-    tts = edge_tts.Communicate(text=clean_question, voice=voice)
-
-    
-    with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3", dir=os.getcwd()) as tmp_file:
-        output_path = tmp_file.name
-        await tts.save(output_path)
-    return output_path
 
-@st.cache_data(show_spinner=False)
-def get_cached_question_audio(question_text):
-    return asyncio.run(generate_question_audio(question_text))
 
 # === Main QA Interface ===
 if st.session_state["generated_questions"]:
@@ -839,36 +824,17 @@ if st.session_state["generated_questions"]:
         question = st.session_state["generated_questions"][idx].lstrip("1234567890. ").strip()
 
         # Phase 0: Play audio first and wait 5s before countdown
-        """
         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": ""
+                "recorded_text": ""
             })
-        """
-        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",
-            })
-        
-
-        #st.markdown(f"**Q{idx + 1}:** {question}")
-        #st.audio(st.session_state["question_audio_file"], format="audio/mp3")
 
         st.markdown(f"**Q{idx + 1}:** {question}")
-        audio_file = st.session_state["question_audio_file"]
-        if os.path.exists(audio_file) and os.path.getsize(audio_file) > 0:
-            st.audio(audio_file, format="audio/mp3")
-        else:
-            st.warning("⚠️ Could not load question audio. Please retry.")
-
-        
+        st.audio(st.session_state["question_audio_file"], format="audio/mp3")
 
         now = time.time()
         elapsed = now - st.session_state.get("question_start_time", 0)
@@ -915,78 +881,7 @@ if st.session_state["generated_questions"]:
             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}")
-                audio = audiorecorder("🎙️ Start Recording", "⏹️ Stop Recording")
-                if len(audio) > 0 and "response_file" not in st.session_state:
-                    if np.max(audio) > 0.01:
-                        wav_path = f"response_{idx}.wav"
-                        with open(wav_path, "wb") as f:
-                            f.write(audio.tobytes())
-                        st.session_state["response_file"] = wav_path
-                        st.session_state["record_phase"] = "listening"
-                        st.success("✅ Audio recorded. Please confirm to proceed.")
-                        st.audio(wav_path, format="audio/wav")
-                        st.rerun()
-                    else:
-                        st.warning("⚠️ Silence detected. Please try again.")
-                        st.rerun()
-
-                else:
-                    time.sleep(1)
-                    st.rerun()
-
-            else:
-                if not st.session_state.get("response_file"):
-                    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()
-        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"):
-                recognizer = sr.Recognizer()
-                try:
-                    with sr.AudioFile(st.session_state["response_file"]) 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["answers"].append({
-                    "question": question,
-                    "response_file": st.session_state["response_file"],
-                    "response": transcript
-                })
-
-                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()
-                        
-                """
+                audio_value = st.audio_input("🎤 Tap to record — then stop when done", key=f"audio_{idx}")
                 if audio_value and "response_file" not in st.session_state:
                     wav_path = f"response_{idx}.wav"
                     with open(wav_path, "wb") as f:
@@ -1009,7 +904,7 @@ 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..."):
@@ -1021,7 +916,7 @@ 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,
@@ -1096,7 +991,7 @@ if st.session_state["generated_questions"]:
                     evaluate_answers()
                     st.session_state["show_summary"] = True
                 st.rerun()
-                """
+
 
 # === Summary Display ===