def test_16khz_file():
audio_file = AudioFile(io.BytesIO(b'\x52\x49\x46\x46\x22\xe2\x01\x00\x57\x41\x56\x45\x66\x6d\x74\x20\x10\x00'
b'\x00\x00\x01\x00\x01\x00\x80\x3e\x00\x00\x00\x7d\x00\x00\x02\x00\x10\x00'
b'\x64\x61\x74\x61\xfe\xe1\x01\x00\x74\x65\x73\x74'))
assert audio_file.header.encoding == PCM_SIGNED_INT
assert audio_file.header.bits_per_sample == 16
assert audio_file.header.sample_rate == 16000
assert audio_file.header.channels == 1
assert audio_file.header.big_endian is False
assert audio_file.read(4) == b'test'
def test_48khz_big_endian_file():
audio_file = AudioFile(io.BytesIO(b'\x52\x49\x46\x58\x00\x0a\x00\x32\x57\x41\x56\x45\x66\x6d\x74\x20\x00\x00'
b'\x00\x12\x00\x03\x00\x01\x00\x00\xbb\x80\x00\x02\xee\x00\x00\x04\x00\x20'
b'\x00\x00\x66\x61\x63\x74\x00\x00\x00\x04\x00\x02\x62\x01\x64\x61\x74\x61'
b'\x00\x09\x88\x04\x74\x65\x73\x74'))
assert audio_file.header.encoding == PCM_FLOAT
assert audio_file.header.bits_per_sample == 32
assert audio_file.header.sample_rate == 48000
assert audio_file.header.channels == 1
assert audio_file.header.big_endian is True
assert audio_file.read(4) == b'test'
def test_16khz_file():
audio_file = AudioFile(
io.BytesIO(
b'\x52\x49\x46\x46\x22\xe2\x01\x00\x57\x41\x56\x45\x66\x6d\x74\x20\x10\x00'
b'\x00\x00\x01\x00\x01\x00\x80\x3e\x00\x00\x00\x7d\x00\x00\x02\x00\x10\x00'
b'\x64\x61\x74\x61\xfe\xe1\x01\x00'))
assert audio_file.header.bits_per_sample == 16
assert audio_file.header.sample_rate == 16000
def test_raw_audio():
audio_file = AudioFile(
io.BytesIO(
b'\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00'
b'\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00'
b'\x00\x00\x00\x00\x00\x00\x00\x00'))
assert audio_file.header.bits_per_sample == 16
assert audio_file.header.sample_rate == 16000
def test_48khz_file():
audio_file = AudioFile(
io.BytesIO(
b'\x52\x49\x46\x46\x32\x00\x0a\x00\x57\x41\x56\x45\x66\x6d\x74\x20\x12\x00'
b'\x00\x00\x03\x00\x01\x00\x80\xbb\x00\x00\x00\xee\x02\x00\x04\x00\x20\x00'
b'\x00\x00\x66\x61\x63\x74\x04\x00'))
assert audio_file.header.bits_per_sample == 32
assert audio_file.header.sample_rate == 48000
def __init__(self, audio_file=None, **kwargs):
Device.__init__(self, **kwargs)
self.time_between_chunks = kwargs.get('time_between_chunks', 0.08)
self._queue = queue.Queue()
self._last_chunk_time = datetime.datetime.utcfromtimestamp(0)
self.audio_file = AudioFile(audio_file)
if self.audio_file.header is not None:
self.encoding = self.audio_file.header.encoding
self.sample_rate = self.audio_file.header.sample_rate
self.bits_per_sample = self.audio_file.header.bits_per_sample
self.channels = self.audio_file.header.channels
self.big_endian = self.audio_file.header.big_endian
else:
self.encoding = None
self.sample_rate = None
self.bits_per_sample = None
self.channels = None
self.big_endian = None
class FileDevice(Device):
def __init__(self, audio_file=None, **kwargs):
Device.__init__(self)
self.time_between_chunks = kwargs.get('time_between_chunks', 0.08)
self._queue = queue.Queue()
self._last_chunk_time = datetime.datetime.utcfromtimestamp(0)
self.audio_file = AudioFile(audio_file)
def stream(self, client, recording_stopper):
self.start_recording()
recording_stopper.started()
query = client.stream_audio(self.generate_frames(), notification_handler=recording_stopper.stop_recording,
audio_type=self.audio_type())
recording_stopper.stop_recording(None)
return query
def start_recording(self):
self._queue.queue.clear()
self.audio_to_frames()
def stop_recording(self):
self._queue.queue.clear()
def is_recording(self):
return not (self._queue.empty())
def generate_frames(self):
while not self._queue.empty():
data = self._queue.get_nowait()
if data:
now = datetime.datetime.utcnow()
seconds_since_last = (now - self._last_chunk_time).total_seconds()
if seconds_since_last < self.time_between_chunks:
time.sleep(self.time_between_chunks - seconds_since_last)
self._last_chunk_time = now
yield data
def audio_to_frames(self):
while True:
data = self.audio_file.read(self.chunk_size)
if not data:
break
self._queue.put(data)
def audio_type(self):
return 'audio/pcm;bits={};rate={}'.format(self.audio_file.header.bits_per_sample,
self.audio_file.header.sample_rate)
def test_raw_audio():
audio_file = AudioFile(io.BytesIO(b'\x0f\x0e\x0d\x0c\x0b\x0a\x09\x08\x07\x06\x05\x04\x03\x02\x01\x00\x00\x00'
b'\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00'
b'\x00\x00\x00\x00\x00\x00\x00\x00'))
assert audio_file.header is None
assert audio_file.read(16) == b'\x0f\x0e\x0d\x0c\x0b\x0a\x09\x08\x07\x06\x05\x04\x03\x02\x01\x00'
class FileDevice(Device):
def __init__(self, audio_file=None, **kwargs):
Device.__init__(self, **kwargs)
self.time_between_chunks = kwargs.get('time_between_chunks', 0.08)
self._queue = queue.Queue()
self._last_chunk_time = datetime.datetime.utcfromtimestamp(0)
self.wakeword_detected = False
self.audio_file = AudioFile(audio_file)
if self.audio_file.header is not None:
self.encoding = self.audio_file.header.encoding
self.sample_rate = self.audio_file.header.sample_rate
self.bits_per_sample = self.audio_file.header.bits_per_sample
self.channels = self.audio_file.header.channels
self.big_endian = self.audio_file.header.big_endian
else:
self.encoding = None
self.sample_rate = None
self.bits_per_sample = None
self.channels = None
self.big_endian = None
def stream(self, client, recording_stopper):
self.start_recording()
recording_stopper.started()
query = client.stream_audio(
self.generate_frames(),
notification_handler=recording_stopper.stop_recording,
audio_type=self.audio_type())
recording_stopper.stop_recording(None)
return query
def stream_with_wakeword(self, client, recording_stopper,
wakeword_detector):
self.start_recording()
recording_stopper.started()
self.wakeword_detected = wakeword_detector.stream_audio(
self.generate_frames())
if self.wakeword_detected:
print("Wakeword detected.")
else:
print("No wakeword detected.")
query = client.stream_audio(
self.generate_frames(),
notification_handler=recording_stopper.stop_recording,
audio_type=self.audio_type())
recording_stopper.stop_recording(None)
return query
def test_wakeword(self, recording_stopper, wakeword_detector):
self.start_recording()
recording_stopper.started()
wakeword_indices = wakeword_detector.test_wakeword(
self.generate_frames())
recording_stopper.stop_recording(None)
return wakeword_indices
def start_recording(self):
log.info(
'Sending %s channels at %sHz, %s bits per sample using encoding %s',
self.channels, self.sample_rate, self.bits_per_sample,
self.encoding)
self._queue.queue.clear()
self.audio_to_frames()
def stop_recording(self):
self._queue.queue.clear()
def is_recording(self):
return not (self._queue.empty())
def generate_frames(self):
while not self._queue.empty():
data = self._queue.get_nowait()
if data:
now = datetime.datetime.utcnow()
seconds_since_last = (now -
self._last_chunk_time).total_seconds()
if seconds_since_last < self.time_between_chunks:
time.sleep(self.time_between_chunks - seconds_since_last)
self._last_chunk_time = now
yield data
def audio_to_frames(self):
while True:
data = self.audio_file.read(self.chunk_size)
if not data:
break
self._queue.put(data)
@abc.abstractmethod
def audio_type(self):
pass
def __init__(self, audio_file=None, **kwargs):
Device.__init__(self)
self.time_between_chunks = kwargs.get('time_between_chunks', 0.08)
self._queue = queue.Queue()
self._last_chunk_time = datetime.datetime.utcfromtimestamp(0)
self.audio_file = AudioFile(audio_file)