Update app.py

app.py

@@ -123,7 +123,7 @@ def retrieve_documents(query):
 
 # Function to convert tuple format to message format
 def convert_to_message_format(chat_history):
-    """Convert from [(user, bot)] format to [{"role": "user", "content": user}, {"role": "assistant", "content": bot}] format"""
+    #Convert from [(user, bot)] format to [{"role": "user", "content": user}, {"role": "assistant", "content": bot}] format"""
     message_format = []
     for user_msg, bot_msg in chat_history:
         message_format.append({"role": "user", "content": user_msg})
@@ -132,7 +132,7 @@ def convert_to_message_format(chat_history):
 
 # Function to convert message format to tuple format for processing
 def convert_to_tuple_format(chat_history):
-    """Convert from message format back to tuple format for processing"""
+    #Convert from message format back to tuple format for processing"""
     tuple_format = []
     for i in range(0, len(chat_history), 2):
         if i+1 < len(chat_history):
@@ -278,57 +278,62 @@ def process_document(file):
 """
 # Real-time Whisper setup - cache the model
 #@gr.cache_resource
-def load_realtime_whisper():
-    """Load optimized Whisper model for real-time transcription"""
-    device = "cuda" if torch.cuda.is_available() else "cpu"
-    torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
+#def load_realtime_whisper():
+#    """Load optimized Whisper model for real-time transcription"""
+#    device = "cuda" if torch.cuda.is_available() else "cpu"
+#    torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
     
-    # Use tiny model for real-time speed
-    realtime_transcriber = pipeline(
-        "automatic-speech-recognition",
-        model="openai/whisper-tiny.en",
-        device=device,
-        torch_dtype=torch_dtype,
-    )
+#    # Use tiny model for real-time speed
+#    realtime_transcriber = pipeline(
+#        "automatic-speech-recognition",
+#        model="openai/whisper-tiny.en",
+#        device=device,
+#        torch_dtype=torch_dtype,
+#    )
     
-    return realtime_transcriber
+#    return realtime_transcriber
+
+
+
+
+
 
 # Load model at startup
-realtime_transcriber = load_realtime_whisper()
+#realtime_transcriber = load_realtime_whisper()
 
-def transcribe_audio(audio):
-    """Real-time optimized transcription"""
-    if audio is None:
-        return ""
+#def transcribe_audio(audio):
+#    """Real-time optimized transcription"""
+#    if audio is None:
+#        return ""
     
-    sr, y = audio
+#    sr, y = audio
     
     # Quick preprocessing
-    if y.ndim > 1:
-        y = y.mean(axis=1)  # Convert to mono
-    
-    y = y.astype(np.float32)
-    max_val = np.max(np.abs(y))
-    if max_val > 0:
-        y = y / max_val
+#    if y.ndim > 1:
+#        y = y.mean(axis=1)  # Convert to mono
     
-    try:
-        # Use real-time transcriber with optimized settings
-        result = realtime_transcriber(
-            {"sampling_rate": sr, "raw": y},
-            generate_kwargs={
-                "language": "english",
-                "task": "transcribe",
-                "temperature": 0.0,  # More deterministic
-                "no_repeat_ngram_size": 2,  # Reduce repetitions
-            }
-        )
-        return result["text"]
-    except Exception as e:
-        print(f"Transcription error: {e}")
-        return "Could not transcribe audio. Please try again."
+#    y = y.astype(np.float32)
+#    max_val = np.max(np.abs(y))
+#    if max_val > 0:
+#        y = y / max_val
+#    
+#    try:
+#        # Use real-time transcriber with optimized settings
+#        result = realtime_transcriber(
+#            {"sampling_rate": sr, "raw": y},
+#            generate_kwargs={
+#                "language": "english",
+#                "task": "transcribe",
+#                "temperature": 0.0,  # More deterministic
+#                "no_repeat_ngram_size": 2,  # Reduce repetitions
+#            }
+#        )
+#        return result["text"]
+#    except Exception as e:
+#        print(f"Transcription error: {e}")
+#        return "Could not transcribe audio. Please try again."
+
 
-"""
 
 
 #Common Issue 1: Audio Format Problems