sirmrak/linux-voice-assistant
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update __main__.py

Browse Source 
(Major change) main() function completely refactored into helpers. _init_controllers now handles decoupled initialization. _process_audio loop logic was fixed to correctly initialize wake words on startup.

 mic.recorder() loop is now much more robust against OS buffering issue when the microphone is muted and wakewords were detected and processed after unmute.
This commit is contained in:
imonlinux
2025-11-17 22:40:53 -08:00
committed by GitHub
parent ada83bdfbb
commit fd5cd13f72
1 changed files with 82 additions and 56 deletions
Download Patch File Download Diff File Expand all files Collapse all files
linux_voice_assistant/__main__.py
+78 -52
View File
@@ -134,7 +134,7 @@ async def main() -> None:
finally:
# --- 10. Cleanup ---
_LOGGER.debug("Shutting down...")
state.mic_muted = True # Signal audio thread to stop
state.shutdown = True # Signal audio thread to stop
audio_thread.join()
if hasattr(state, "mqtt_controller") and state.mqtt_controller:
@@ -404,69 +404,95 @@ def _process_audio(state: ServerState, mic, block_size: int):
last_active: Optional[float] = None
try:
_LOGGER.debug("Opening audio input device: %s", mic.name)
with mic.recorder(samplerate=16000, channels=1, blocksize=block_size) as mic_in:
while True:
if state.mic_muted:
break
while True:
# --- THIS IS THE FIX ---
# Wait while muted. This loop will spin until unmuted.
if state.mic_muted:
_LOGGER.debug("Audio thread muted.")
while state.mic_muted:
if state.shutdown:
_LOGGER.debug("Shutdown signal received, stopping audio thread.")
return # Exit thread
time.sleep(0.1)
audio_chunk_array = mic_in.record(block_size).reshape(-1)
audio_chunk = (
(np.clip(audio_chunk_array, -1.0, 1.0) * 32767.0)
.astype("<i2")
.tobytes()
)
# --- Just unmuted ---
_LOGGER.debug("Audio thread unmuted, clearing buffers.")
micro_features.reset()
if oww_features is not None:
oww_features.reset()
last_active = time.monotonic() # Start refractory period
if state.satellite is None:
time.sleep(0.01)
continue
# Re-open microphone to flush OS buffer
_LOGGER.debug("Re-opening microphone to flush OS buffers...")
with mic.recorder(samplerate=16000, channels=1, blocksize=block_size) as mic_in:
mic_in.flush() # Discard any stale data
_LOGGER.debug("Microphone flushed.")
# Fall through to the main recording loop
# --- END FIX ---
# --- THIS IS THE FIX ---
if (not wake_words) or (state.wake_words_changed):
# --- END FIX ---
state.wake_words_changed = False
wake_words = [
ww for ww in state.wake_words.values()
if ww.id in state.active_wake_words
]
has_oww = any(isinstance(ww, OpenWakeWord) for ww in wake_words)
_LOGGER.debug("Opening audio input device: %s", mic.name)
with mic.recorder(samplerate=16000, channels=1, blocksize=block_size) as mic_in:
# --- This is now the main recording loop ---
while not state.mic_muted:
if state.shutdown:
_LOGGER.debug("Shutdown signal received, stopping audio thread.")
return # Exit thread
if has_oww and (oww_features is None):
_LOGGER.debug("Initializing OpenWakeWord features...")
oww_features = OpenWakeWordFeatures.from_builtin()
audio_chunk_array = mic_in.record(block_size).reshape(-1)
audio_chunk = (
(np.clip(audio_chunk_array, -1.0, 1.0) * 32767.0)
.astype("<i2")
.tobytes()
)
try:
micro_inputs.clear(); micro_inputs.extend(micro_features.process_streaming(audio_chunk))
if state.satellite is None:
time.sleep(0.01)
continue
if has_oww:
assert oww_features is not None
oww_inputs.clear(); oww_inputs.extend(oww_features.process_streaming(audio_chunk))
if (not wake_words) or (state.wake_words_changed):
state.wake_words_changed = False
wake_words = [
ww for ww in state.wake_words.values()
if ww.id in state.active_wake_words
]
has_oww = any(isinstance(ww, OpenWakeWord) for ww in wake_words)
for wake_word in wake_words:
activated = False
if isinstance(wake_word, MicroWakeWord):
if any(wake_word.process_streaming(mi) for mi in micro_inputs): activated = True
elif isinstance(wake_word, OpenWakeWord):
if any(p > 0.5 for oi in oww_inputs for p in wake_word.process_streaming(oi)): activated = True
if has_oww and (oww_features is None):
_LOGGER.debug("Initializing OpenWakeWord features...")
oww_features = OpenWakeWordFeatures.from_builtin()
if activated:
now = time.monotonic()
if (last_active is None) or ((now - last_active) > state.refractory_seconds):
state.loop.call_soon_threadsafe(state.satellite.wakeup, wake_word)
last_active = now
try:
micro_inputs.clear(); micro_inputs.extend(micro_features.process_streaming(audio_chunk))
stopped = False
for micro_input in micro_inputs:
if state.stop_word.process_streaming(micro_input):
stopped = True
if has_oww:
assert oww_features is not None
oww_inputs.clear(); oww_inputs.extend(oww_features.process_streaming(audio_chunk))
if stopped and (state.stop_word.id in state.active_wake_words):
state.loop.call_soon_threadsafe(state.satellite.stop)
for wake_word in wake_words:
activated = False
if isinstance(wake_word, MicroWakeWord):
if any(wake_word.process_streaming(mi) for mi in micro_inputs): activated = True
elif isinstance(wake_word, OpenWakeWord):
if any(p > 0.5 for oi in oww_inputs for p in wake_word.process_streaming(oi)): activated = True
state.loop.call_soon_threadsafe(state.satellite.handle_audio, audio_chunk)
if activated:
now = time.monotonic()
if (last_active is None) or ((now - last_active) > state.refractory_seconds):
state.loop.call_soon_threadsafe(state.satellite.wakeup, wake_word)
last_active = now
except Exception:
_LOGGER.exception("Unexpected error handling audio")
stopped = False
for micro_input in micro_inputs:
if state.stop_word.process_streaming(micro_input):
stopped = True
if stopped and (state.stop_word.id in state.active_wake_words):
state.loop.call_soon_threadsafe(state.satellite.stop)
state.loop.call_soon_threadsafe(state.satellite.handle_audio, audio_chunk)
except Exception:
_LOGGER.exception("Unexpected error handling audio")
except Exception as e:
_LOGGER.critical("A soundcard error occurred: %s", e)