Spaces:
Building
Building
Update websocket_handler.py
Browse files- websocket_handler.py +6 -9
websocket_handler.py
CHANGED
@@ -491,7 +491,7 @@ async def handle_audio_chunk(websocket: WebSocket, session: RealtimeSession, mes
|
|
491 |
# Decode for processing
|
492 |
decoded_audio = base64.b64decode(audio_data)
|
493 |
|
494 |
-
# Check silence
|
495 |
silence_duration = session.silence_detector.update(decoded_audio)
|
496 |
|
497 |
# Stream to STT if available
|
@@ -510,7 +510,7 @@ async def handle_audio_chunk(websocket: WebSocket, session: RealtimeSession, mes
|
|
510 |
return
|
511 |
|
512 |
try:
|
513 |
-
# Chunk counter
|
514 |
if not hasattr(session, 'chunk_counter'):
|
515 |
session.chunk_counter = 0
|
516 |
session.chunk_counter += 1
|
@@ -522,7 +522,7 @@ async def handle_audio_chunk(websocket: WebSocket, session: RealtimeSession, mes
|
|
522 |
|
523 |
# STT'ye gönder ve sonuçları bekle
|
524 |
async for result in session.stt_manager.stream_audio(decoded_audio):
|
525 |
-
# SADECE FINAL RESULT'LARI
|
526 |
if result.is_final:
|
527 |
log_info(f"✅ FINAL TRANSCRIPTION: '{result.text}'", session_id=session.session.session_id)
|
528 |
|
@@ -549,14 +549,11 @@ async def handle_audio_chunk(websocket: WebSocket, session: RealtimeSession, mes
|
|
549 |
# Process user input
|
550 |
await process_user_input(websocket, session)
|
551 |
|
552 |
-
#
|
553 |
await session.reset_for_new_utterance()
|
554 |
-
return
|
555 |
|
556 |
-
# Interim result'ları
|
557 |
-
elif result.text.strip() and not hasattr(session, 'speech_started'):
|
558 |
-
session.speech_started = True
|
559 |
-
log_info(f"🎤 User started speaking", session_id=session.session.session_id)
|
560 |
|
561 |
except Exception as e:
|
562 |
log_error(f"❌ STT streaming error", error=str(e), traceback=traceback.format_exc(), session_id=session.session.session_id)
|
|
|
491 |
# Decode for processing
|
492 |
decoded_audio = base64.b64decode(audio_data)
|
493 |
|
494 |
+
# Check silence
|
495 |
silence_duration = session.silence_detector.update(decoded_audio)
|
496 |
|
497 |
# Stream to STT if available
|
|
|
510 |
return
|
511 |
|
512 |
try:
|
513 |
+
# Chunk counter
|
514 |
if not hasattr(session, 'chunk_counter'):
|
515 |
session.chunk_counter = 0
|
516 |
session.chunk_counter += 1
|
|
|
522 |
|
523 |
# STT'ye gönder ve sonuçları bekle
|
524 |
async for result in session.stt_manager.stream_audio(decoded_audio):
|
525 |
+
# SADECE FINAL RESULT'LARI İŞLE
|
526 |
if result.is_final:
|
527 |
log_info(f"✅ FINAL TRANSCRIPTION: '{result.text}'", session_id=session.session.session_id)
|
528 |
|
|
|
549 |
# Process user input
|
550 |
await process_user_input(websocket, session)
|
551 |
|
552 |
+
# Reset for new utterance
|
553 |
await session.reset_for_new_utterance()
|
554 |
+
return
|
555 |
|
556 |
+
# Interim result'ları artık göndermiyoruz ve loglama yapmıyoruz
|
|
|
|
|
|
|
557 |
|
558 |
except Exception as e:
|
559 |
log_error(f"❌ STT streaming error", error=str(e), traceback=traceback.format_exc(), session_id=session.session.session_id)
|