def main():
src = Source(rate=16000, channels=8)
ch1 = ChannelPicker(channels=8, pick=1)
ns = NS(rate=16000, channels=1)
kws = KWS()
doa = DOA(rate=16000)
src.link(ch1)
ch1.link(ns)
ns.link(kws)
src.link(doa)
def on_detected(keyword):
print('detected {} at direction {}'.format(keyword,
doa.get_direction()))
kws.set_callback(on_detected)
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.recursive_stop()
def main():
src = Source(rate=16000, channels=4)
ch0 = ChannelPicker(channels=src.channels, pick=0)
kws = KWS(model='snowboy', sensitivity=0.6, verbose=True)
doa = DOA(rate=16000)
src.link(ch0)
ch0.link(kws)
src.link(doa)
def on_detected(keyword):
print('detected {} at direction {}'.format(keyword,
doa.get_direction()))
kws.set_callback(on_detected)
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.recursive_stop()
# wait a second to allow other threads to exit
time.sleep(1)
def main():
src = Source(rate=16000, channels=4, frames_size=320)
ch0 = ChannelPicker(channels=4, pick=0)
kws = KWS(model='snowboy', sensitivity=0.6, verbose=True)
doa = DOA(rate=16000)
src.link(ch0)
ch0.link(kws)
src.link(doa)
def on_detected(keyword):
position = doa.get_direction()
pixels.wakeup(position)
print('detected {} at direction {}'.format(keyword, position))
kws.set_callback(on_detected)
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.recursive_stop()
def main():
logging.basicConfig(level=logging.DEBUG)
src = Source(rate=16000, channels=8)
ch1 = ChannelPicker(channels=8, pick=1)
ns = NS(rate=16000, channels=1)
kws = KWS(model='alexa')
doa = DOA(rate=16000)
alexa = Alexa()
src.link(ch1)
ch1.link(ns)
ns.link(kws)
kws.link(alexa)
src.link(doa)
def on_detected(keyword):
logging.info('detected {} at direction {}'.format(
keyword, doa.get_direction()))
alexa.listen()
kws.set_callback(on_detected)
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.recursive_stop()
def main():
if len(sys.argv) < 2:
print('Usage: {} audio.wav')
sys.exit(1)
src = Source(sys.argv[1])
ch0 = ChannelPicker(channels=src.channels, pick=0)
kws = KWS()
doa = DOA(rate=16000)
src.link(ch0)
ch0.link(kws)
src.link(doa)
def on_detected(keyword):
print('detected {} at direction {}'.format(keyword, doa.get_direction()))
kws.set_callback(on_detected)
src.recursive_start()
while src.is_active():
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.recursive_stop()
# wait a second to allow other threads to exit
time.sleep(1)
def main():
src = Source(rate=16000, channels=4, frames_size=320)
ch1 = ChannelPicker(channels=4, pick=1)
kws = KWS()
doa = DOA(rate=16000)
src.link(ch1)
ch1.link(kws)
src.link(doa)
def on_detected(keyword):
position = doa.get_direction()
pixels.wakeup(position)
print('detected {} at direction {}'.format(keyword, position))
kws.set_callback(on_detected)
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
print('STOP')
break
src.recursive_stop()
def main():
ard = serial.Serial('/dev/ttyACM0', 9600)
src = Source(rate=16000, channels=8)
ch1 = ChannelPicker(channels=8, pick=1)
kws = KWS()
doa = DOA(rate=16000)
time.sleep(1)
src.link(ch1)
ch1.link(kws)
src.link(doa)
while True:
def on_detected(keyword):
direction = doa.get_direction()
print('detected {} at direction {}'.format(keyword, ))
if (direction >= 0) and (direction <= 180):
ard.write(struct.pack('>B', direction))
kws.set_callback(on_detected)
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.recursive_stop()
def main():
src = Source(rate=16000, channels=4, frames_size=320)
ch1 = ChannelPicker(channels=4, pick=1)
kws = KWS(model='dad.pmdl')
doa = DOA(rate=16000)
src.link(ch1)
ch1.link(kws)
src.link(doa)
def on_detected(keyword):
position = doa.get_direction()
pixels.wakeup(position)
print('detected {} at direction {}'.format(keyword, position))
kws.set_callback(on_detected)
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.recursive_stop()
def main():
logging.basicConfig(level=logging.DEBUG)
src = Source(rate=16000, channels=8)
ch1 = ChannelPicker(channels=8, pick=1)
ns = NS(rate=16000, channels=1)
kws = KWS(model='alexa')
doa = DOA(rate=16000)
src.link(ch1)
ch1.link(ns)
ns.link(kws)
src.link(doa)
def on_detected(keyword):
direction = doa.get_direction()
logging.info('detected {} at direction {}'.format(keyword, direction))
pixels.wakeup(direction)
kws.on_detected = on_detected
src.recursive_start()
while True:
try:
time.sleep(0.05)
#direction = doa.get_direction()
#logging.info('direction {}'.format(direction))
#pixels.wakeup(direction)
except KeyboardInterrupt:
break
src.recursive_stop()
def main():
## RESPEAKER_RATE = 16000
## RESPEAKER_CHANNELS = 1
## RESPEAKER_WIDTH = 2
## CHUNK = 1024
## RECORD_SECONDS = 2
src = Source(rate=16000, channels=4, frames_size=320)
ch1 = ChannelPicker(channels=4, pick=1)
kws = KWS()
doa = DOA(rate=16000)
src.link(ch1)
ch1.link(kws)
src.link(doa)
def on_detected(keyword):
position = doa.get_direction()
pixels.wakeup(position)
print('detected {} at direction {}'.format(keyword, position))
src.stream.start()
print("* recording")
frames = []
for i in range(0, int(RESPEAKER_RATE / CHUNK * RECORD_SECONDS)):
data = stream.read(CHUNK)
frames.append(data)
print("* done recording")
src.stream.stop()
print("start to send to baidu")
# audio_data should be raw_data
text = baidu.server_api(generator_list(frames))
if text:
try:
text = json.loads(text)
for t in text['result']:
print(t)
except KeyError:
print("get nothing")
else:
print("get nothing")
kws.set_callback(on_detected)
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.recursive_stop()
def main():
src = Source(rate=16000, channels=4)
ch1 = ChannelPicker(channels=4, pick=1)
kws = KWS()
doa = DOA(rate=16000)
src.link(ch1)
ch1.link(kws)
src.link(doa)
outf = "./log/test.wav"
fsall = FileSink(outf, rate=16000, channels=4)
src.link(fsall)
#ch1.link(fs)
for i in range(4):
ch = ChannelPicker(channels=4, pick=i)
src.link(ch)
outf = "./log/test"+str(i)+".wav"
fs = FileSink(outf, rate=16000, channels=1)
ch.link(fs)
def on_detected(keyword):
print('detected {} at direction {}'.format(keyword, doa.get_direction()))
kws.set_callback(on_detected)
src.recursive_start()
sec = 0
while sec < 5:
try:
time.sleep(1)
sec = sec + 1
except KeyboardInterrupt:
break
src.recursive_stop()
def t3_bot():
src = Source(rate=16000, channels=4, frames_size=320)
ch1 = ChannelPicker(channels=4, pick=1)
kws = KWS()
doa = DOA(rate=16000)
src.link(ch1)
ch1.link(kws)
src.link(doa)
pixels.listen()
pwm.setPWM(0, 0, 370)
pwm.setPWM(1, 0, 640)
counter = 0
def on_detected(keyword):
position = doa.get_direction()
pixels.wakeup(position)
print('detected {} at direction {}'.format(keyword, position))
if position >= 30 and position <= 180:
pwm.setPWM(0, 0, 175)
pwm.setPWM(1, 0, 500)
elif position > 180 and position <= 330:
pwm.setPWM(0, 0, 560)
pwm.setPWM(1, 0, 500)
elif position > 330 or position < 30:
pwm.setPWM(0, 0, 370)
pwm.setPWM(1, 0, 6200)
else:
pwm.setPWM(0, 0, 370)
pwm.setPWM(1, 0, 640)
#talkassist.os.system("espeak 'may i help you'")
print("How may I help you?")
print("call google assistant here, delete this line.")
kws.set_callback(on_detected)
src.recursive_start()
while True:
try:
time.sleep(1)
counter += 1
print("counter is at " + str(counter))
except KeyboardInterrupt:
break
src.recursive_stop()
def main():
import time
import datetime
from voice_engine.source import Source
from voice_engine.channel_picker import ChannelPicker
from voice_engine.file_sink import FileSink
src = Source(channels=2, rate=48000, frames_size=48000)
chx = ChannelPicker(channels=src.channels, pick=0)
thd = THD(1000, src.rate)
filename = '2.94dB1KHz.' + datetime.datetime.now().strftime("%Y%m%d.%H:%M:%S") + '.wav'
sink = FileSink(filename, channels=src.channels, rate=src.rate)
src.link(sink)
src.pipeline(chx, thd)
src.pipeline_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.pipeline_stop()
def main():
src = Source(rate=16000, channels=8)
ch1 = ChannelPicker(channels=8, pick=1)
kws = KWS()
filename = '4.kws.' + datetime.datetime.now().strftime(
"%Y%m%d.%H:%M:%S") + '.wav'
sink = FileSink(filename, channels=src.channels, rate=src.rate)
src.link(sink)
src.pipeline(ch1, kws)
def on_detected(keyword):
print('detected {}'.format(keyword))
kws.set_callback(on_detected)
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.recursive_stop()
def main():
import time
from voice_engine.channel_picker import ChannelPicker
from voice_engine.kws import KWS
from voice_engine.source import Source
src = Source(channels=2)
ch0 = ChannelPicker(channels=src.channels, pick=0)
kws = KWS(model='snowboy', sensitivity=0.7)
bing = Bing(BING_KEY)
src.pipeline(ch0, kws, bing)
def on_detected(keyword):
print('detected {}'.format(keyword))
bing.listen()
kws.set_callback(on_detected)
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.recursive_stop()
def main():
if len(sys.argv) < 2:
print('Usage: {} audio.wav')
sys.exit(1)
src = Source(sys.argv[1])
ch0 = ChannelPicker(channels=src.channels, pick=0)
ns = NS(rate=src.rate, channels=1)
kws = KWS()
src.pipeline(ch0, kws)
# src.pipeline(ch0, ns, kws)
def on_detected(keyword):
print('detected {}'.format(keyword))
kws.set_callback(on_detected)
src.pipeline_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.pipeline_stop()
# wait a second to allow other threads to exit
time.sleep(1)
def main():
src = Source('/tmp/ec.output', rate=16000, channels=2)
ch0 = ChannelPicker(channels=src.channels, pick=1)
ns = NS()
kws = KWS(model='alexa', sensitivity=0.7)
alexa = Alexa()
src.pipeline(ch0, ns, kws, alexa)
def on_detected(keyword):
print('detected {}'.format(keyword))
alexa.listen()
kws.set_callback(on_detected)
alexa.state_listener.on_listening = pixels.listen
alexa.state_listener.on_thinking = pixels.think
alexa.state_listener.on_speaking = pixels.speak
alexa.state_listener.on_finished = pixels.off
src.pipeline_start()
is_quit = []
def signal_handler(sig, frame):
is_quit.append(True)
print('quit')
signal.signal(signal.SIGINT, signal_handler)
while src.is_active() and not is_quit:
time.sleep(1)
src.pipeline_stop()
def main():
src = Source(rate=16000, frames_size=320, channels=8)
ch0 = ChannelPicker(channels=src.channels, pick=0)
kws = KWS(model='snowboy', sensitivity=0.5)
doa = DOA(rate=src.rate, chunks=20)
src.pipeline(ch0, kws)
src.link(doa)
def on_detected(keyword):
direction = doa.get_direction()
print('detected {} at direction {}'.format(keyword, direction))
pixel_ring.wakeup(direction)
time.sleep(2)
pixel_ring.off()
kws.on_detected = on_detected
is_quit = []
def signal_handler(sig, frame):
is_quit.append(True)
print('quit')
signal.signal(signal.SIGINT, signal_handler)
src.pipeline_start()
while not is_quit:
time.sleep(1)
src.pipeline_stop()
def main():
logging.basicConfig(level=logging.WARNING)
# logging.basicConfig(level=logging.WARNING)
logging.getLogger('hpack.hpack').setLevel(logging.INFO)
# logging.getLogger('voice_engine.netT').setLevel(logging.DEBUG)
# src = Source(rate=16000, frames_size=320, channels=6)
src = Source(rate=16000,
frames_size=128,
channels=6,
device_name='ReSpeaker 4 Mic Array')
ch0 = ChannelPicker(channels=src.channels, pick=0)
nett = NetT()
src.pipeline(ch0, nett)
is_quit = []
def signal_handler(sig, frame):
is_quit.append(True)
print('quit')
signal.signal(signal.SIGINT, signal_handler)
src.pipeline_start()
while not is_quit:
time.sleep(1)
src.pipeline_stop()
def main():
src = Source(rate=16000, channels=2)
ch1 = ChannelPicker(channels=2, pick=1)
ns = NS(rate=16000, channels=1)
kws = KWS(model='snowboy', sensitivity=0.6, verbose=True)
src.pipeline(ch1, ns, kws)
def on_detected(keyword):
print('detected {}'.format(keyword))
kws.set_callback(on_detected)
src.pipeline_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.pipeline_stop()
# wait a second to allow other threads to exit
time.sleep(1)
def __init__(self):
self.init_pixelring()
self.source = Source(rate=160000, frames_size=320, channels=8)
self.ch0 = ChannelPicker(channels=self.source.channels, pick=0)
self.kws = KWS(sensitivity=0.7)
self.doa = DOA(rate=self.source.rate, chunks=20)
self.kws.on_detected = self.kws_on_detected
self.cmd_pub = rospy.Publisher('/respeaker/cmd', String, queue_size=1)
def __init__(self, parent=None):
super(MainWindow, self).__init__(parent)
self.setWindowTitle('Audio Measuring Experiment')
self.resize(800, 500)
self.cwidget = QWidget()
self.setCentralWidget(self.cwidget)
layout = QHBoxLayout()
layout.setContentsMargins(0, 0, 0, 0)
self.cwidget.setLayout(layout)
self.plotwidget = pg.PlotWidget()
layout.addWidget(self.plotwidget)
self.plotwidget.setTitle(title="press 'space' to freeze/resume")
self.plot = self.plotwidget.plot()
self.plot.setPen((0, 255, 0))
self.ff = pg.TextItem('fundamental frequency', anchor=(0, 0))
self.plotwidget.addItem(self.ff)
self.ff.setPos(100, 0)
self.arrow = pg.ArrowItem(pos=(100, 0), angle=-45)
self.plotwidget.addItem(self.arrow)
plotitem = self.plotwidget.getPlotItem()
# plotitem.hideAxis('left')
# plotitem.hideAxis('bottom')
self.freeze = False
self.timer = pg.QtCore.QTimer()
self.timer.timeout.connect(self.update)
self.timer.start(50)
self.src = Source(rate=48000, frames_size=4800)
chx = ChannelPicker(channels=self.src.channels, pick=0)
thd = THD(1000, self.src.rate)
filename = '2.94dB1KHz.' + datetime.datetime.now().strftime(
"%Y%m%d.%H:%M:%S") + '.wav'
sink = FileSink(filename,
channels=self.src.channels,
rate=self.src.rate)
def on_data(d):
global data
data = d
thd.on_data = on_data
self.src.link(sink)
self.src.pipeline(thd)
self.src.pipeline_start()
def main():
ser = serial.Serial("/dev/ttyACM0", 9600)
src = Source(rate=16000, channels=4, frames_size=320)
ch1 = ChannelPicker(channels=4, pick=1)
kws = KWS(model='dad.pmdl')
doa = DOA(rate=16000)
src.link(ch1)
ch1.link(kws)
src.link(doa)
def on_detected(keyword):
position = doa.get_direction()
pixels.wakeup(position)
print('detected {} at direction {}'.format(keyword, position))
if position <= 90:
ser.write(b"1000\n")
elif position > 90 and position <= 180:
ser.write(b"0100\n")
elif position > 180 and position <= 270:
ser.write(b"0010\n")
else:
ser.write(b"0001\n")
vibrate_time = 0.5
time.sleep(vibrate_time)
ser.write(b"0000\n")
kws.set_callback(on_detected)
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.recursive_stop()
ser.close()
def main():
logging.basicConfig(level=logging.DEBUG)
src = Source(rate=16000, channels=4)
ch1 = ChannelPicker(channels=4, pick=1)
ns = NS(rate=16000, channels=1)
kws = KWS(model='snowboy')
doa = DOA(rate=16000)
alexa = Alexa()
alexa.state_listener.on_listening = pixels.listen
alexa.state_listener.on_thinking = pixels.think
alexa.state_listener.on_speaking = pixels.speak
alexa.state_listener.on_finished = pixels.off
def on_detected(keyword):
direction = doa.get_direction()
logging.info('detected {} at direction {}'.format(keyword, direction))
pixels.wakeup(direction)
alexa.listen()
kws.on_detected = on_detected
src.link(ch1)
ch1.link(ns)
ns.link(kws)
kws.link(alexa)
src.link(doa)
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.recursive_stop()
def main():
logging.basicConfig(level=logging.DEBUG)
src = Source(rate=16000, channels=4)
ch1 = ChannelPicker(channels=4, pick=1)
ns = NS(rate=16000, channels=1)
kws = KWS(model='snowboy')
doa = DOA(rate=16000)
alexa = Alexa()
alexa.state_listener.on_listening = pixels.listen
alexa.state_listener.on_thinking = pixels.think
alexa.state_listener.on_speaking = pixels.speak
alexa.state_listener.on_finished = pixels.off
def on_detected(keyword):
direction = doa.get_direction()
logging.info('detected {} at direction {}'.format(keyword, direction))
pixels.wakeup(direction)
alexa.listen()
kws.on_detected = on_detected
src.link(ch1)
ch1.link(ns)
ns.link(kws)
kws.link(alexa)
src.link(doa)
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.recursive_stop()
def main():
src = Source(rate=16000, frames_size=320, channels=8)
# ds = DelaySum(channels=8, frames_size=src.frames_size, max_offset=max_offset)
ch0 = ChannelPicker(channels=src.channels, pick=0)
kws = KWS()
# doa = DOA(rate=16000, chunks=20)
alexa = Alexa()
# alexa.state_listener.on_listening = pixel_ring.listen
# alexa.state_listener.on_thinking = pixel_ring.think
# alexa.state_listener.on_speaking = pixel_ring.speak
# alexa.state_listener.on_finished = pixel_ring.off
src.link(ch0)
ch0.link(kws)
kws.link(alexa)
# src.link(doa)
def on_detected(keyword):
# direction = doa.get_direction()
alexa.listen()
# pixel_ring.wakeup((direction + 0) % 360)
# print('detected {} at direction {}'.format(keyword, direction))
print('detected keyword')
kws.on_detected = on_detected
src.recursive_start()
while True:
try:
time.sleep(1)
except KeyboardInterrupt:
print('quit')
break
src.recursive_stop()
def main():
src = Source(rate=16000, frames_size=160, channels=2)
ch0 = ChannelPicker(channels=src.channels, pick=0)
ns = NS(rate=src.rate, channels=1)
kws = KWS()
doa = DOA(rate=16000, chunks=50)
alexa = Alexa()
alexa.state_listener.on_listening = pixel_ring.listen
alexa.state_listener.on_thinking = pixel_ring.think
alexa.state_listener.on_speaking = pixel_ring.speak
alexa.state_listener.on_finished = pixel_ring.off
# data flow between elements
# ---------------------------
# src -> ns -> kws -> alexa
# \
# doa
src.pipeline(ch0, ns, kws, alexa)
src.link(doa)
def on_detected(keyword):
direction = doa.get_direction()
print('detected {} at direction {}'.format(keyword, direction))
alexa.listen()
pixel_ring.wakeup(direction)
kws.on_detected = on_detected
is_quit = []
def signal_handler(sig, frame):
is_quit.append(True)
print('quit')
signal.signal(signal.SIGINT, signal_handler)
src.pipeline_start()
while not is_quit:
time.sleep(1)
src.pipeline_stop()
pixel_ring.off()
# wait a second to allow other threads to exit
time.sleep(1)
def main():
src = Source('/tmp/ec.output', rate=16000, channels=2)
ch0 = ChannelPicker(channels=src.channels, pick=1)
kws = KWS(sensitivity=0.7)
src.pipeline(ch0, kws)
def on_detected(keyword):
print('detected {}'.format(keyword))
kws.set_callback(on_detected)
src.pipeline_start()
while src.is_active():
try:
time.sleep(1)
except KeyboardInterrupt:
break
src.pipeline_stop()
def main():
# src = Source(rate=16000, frames_size=320, channels=6)
src = Source(rate=16000, frames_size=160, channels=6, device_name='hw:1')
ch0 = ChannelPicker(channels=src.channels, pick=0)
kws = KWS(model='snowboy', sensitivity=0.5)
# doa = DOA(rate=src.rate, chunks=20)
alexa = Alexa()
alexa.state_listener.on_listening = pixel_ring.listen
alexa.state_listener.on_thinking = pixel_ring.think
alexa.state_listener.on_speaking = pixel_ring.speak
alexa.state_listener.on_finished = pixel_ring.off
src.pipeline(ch0, kws, alexa)
# src.link(doa)
def on_detected(keyword):
# direction = doa.get_direction()
# print('detected {} at direction {}'.format(keyword, direction))
print('detected {} '.format(keyword))
alexa.listen()
# pixel_ring.wakeup(direction)
kws.on_detected = on_detected
is_quit = []
def signal_handler(sig, frame):
is_quit.append(True)
print('quit')
signal.signal(signal.SIGINT, signal_handler)
src.pipeline_start()
while not is_quit:
time.sleep(1)
src.pipeline_stop()
'''
CTRL + C
'''
def CtrlC(signum, frame):
os.kill(os.getpid(), signal.SIGKILL)
'''
main start ...
'''
if __name__ == "__main__":
try:
src = Source(channels=8)
ch0 = ChannelPicker(channels=src.channels, pick=0)
ns = NS()
bing = Bing(BING_KEY)
src.pipeline(ch0, ns, bing)
bing_api_call()
signal.signal(signal.SIGINT, CtrlC)
signal.signal(signal.SIGTERM, CtrlC)
thread_key = threading.Thread(target=key_hander, args=(1, ))
thread_play = threading.Thread(target=pre_play, args=(1, ))
thread_key.setDaemon(True)
def main():
src = Source(rate=16000, channels=1, frames_size=21000)
ch1 = ChannelPicker(channels=1, pick=1)
# doa = DOA(rate=16000, chunks=3)
model_path = get_model_path()
config = Decoder.default_config()
# config.set_string('-hmm', os.path.join(model_path, 'en-us'))
# config.set_string('-lm', os.path.join(model_path, 'en-us.lm.bin'))
config.set_string('-hmm', os.path.join(model_path, 'en-us'))
config.set_string('-lm', '2823.lm')
config.set_string('-verbose', 'False')
config.set_string('-dict', '2823.dic')
# config.set_string('-dict', os.path.join(model_path, 'cmudict-en-us.dict'))
config.set_string('-kws', 'keyphrase.list')
config.set_string('-logfn', '/dev/null')
# config.set_string('-keyphrase', 'hey there')
# config.set_float('-kws_threshold', 1e-30)
sphinx = Sphinx(config)
src.link(ch1)
# src.link(doa)
ch1.link(sphinx)
graph = XMLParser(graph_file="basic.xml", debug=True).parse()
arduino = serial.Serial('/dev/ttyACM0', 57600)
arduino.timeout = 0.1
# Check currnet state
print("Current State: {}".format(graph.get_current_state().name))
def on_graph_state_change():
print("onStateChange()")
# Runs through state responses
print("\tNew Current State: {}".format(graph.state))
print("\tExecuting responses for nextState...")
if len(graph.state.get_responses()) > 0:
print('Responses: {}'.format(len(graph.state.get_responses())))
for response in graph.state.get_responses():
print('\tRunning Response {}'.format(response))
# do response action whether it has to do with moving motors, turning led, etc
if response.typ == ResponseType.GO_TO_STATE:
graph.set_current_state(response.value)
elif response.typ == ResponseType.LED:
pixels != null:
if response.value == 'listening':
pixels.think()
elif response.value == 'off':
pixels.off()
elif response.value == 'hello':
pixels.speak()
elif response.value == 'following':
pixels.following()
elif response.value == 'doa':
if mic != null:
pixels.wakeup(mic.direction)
else:
print("Unknown LED value: {} was found.".format(response.value))
elif response.typ == ResponseType.MOTOR_MOVE:
if response.value == 'forward':
arduino.write("d:f;")
elif response.value == 'stop':
arduino.write("d:s;")
elif response.typ == ResponseType.CAMERA_MOVE:
if response.value == 'doa':
if mic != null:
voice_direction = mic.direction
print "voice from " + str(voice_direction)
arduino_command = "m:" + str(voice_direction) + ";"
if voice_direction < 180:
#voice is coming from behind
voice_direction = (voice_direction + 180) % 360
else:
#voice is coming from in front
voice_direction = 90
arduino_command = arduino_command + "c:" + str(voice_direction) + ",120;"
arduino.write(arduino_command)
last_time_motor_moved = simpletime.time()
print("@done@")
elif response.typ == ResponseType.VOICE_RESPONSE:
text = response.value.replace(' ', '_')
#Calls the Espeak TTS Engine to read aloud a Text
call([cmd_beg+cmd_out+text+cmd_end], shell=True)
else:
print("Unused response type: {}.".format(response.typ))
class XYRobot(object):
jike = config.get('jike', False)
src = Source(rate=16000, channels=4, frames_size=320)
ch1 = ChannelPicker(channels=4, pick=1)
#这里需要填写pmdl的绝对路径
#kws = KWS()
#kws = KWS(os.path.join(constants.DATA_PATH ,config.get('kws_file_name','')))
kws = KWS(
os.path.join(constants.DATA_PATH, config.get('kws_file_name', '')))
doa = DOA(rate=16000)
def __init__(self):
print('极客模式:' + str(self.jike))
self.src.link(self.ch1)
self.ch1.link(self.kws)
self.src.link(self.doa)
def on_detected(keyword):
position = self.doa.get_direction()
pixels.wakeup(position)
voice = os.path.join(
constants.DATA_PATH, 'sysvoices',
'sysvoice' + str(random.randint(1, 8)) + '.mp3')
direction = pixels.positionToDirection()
print('detected {} at direction {} is {}'.format(
keyword, position, direction))
#这里是唤醒后进行打招呼的语音播放
pixels.speak(voice)
print(str(keyword) + voice)
record.speech('你在我' + direction)
if self.jike:
print('进入极客模式')
# 创建线程01，不指定参数
thread_01 = Thread(target=instructions.jikeThread,
args=(
"jikeThread01",
record,
pixels,
))
# 启动线程01
thread_01.start()
else:
#开始录音
outputtext = record.record()
#这里是指令，语音识别后返回文字，将文字传入instructionsFunc，进行处理和播放
instructions.instructionsFunc(outputtext, record, pixels)
self.kws.set_callback(on_detected)
self.src.recursive_start()
config.init()
def sigint_handler(self, signum, frame):
record.stream.stop_stream()
record.stream.close()
record.p.terminate()
self.src.recursive_stop()
print 'catched interrupt signal!'
sys.exit(0)
def train(self, w1, w2, w3, m):
'''
传入三个wav文件，生成snowboy的.pmdl模型
'''
def get_wave(fname):
with open(fname, 'rb') as infile:
return base64.b64encode(infile.read()).decode('utf-8')
url = 'https://snowboy.kitt.ai/api/v1/train/'
data = {
"name":
"wukong-robot",
"language":
"zh",
"token":
config.get('snowboy_token', ''),
"voice_samples": [{
"wave": get_wave(w1)
}, {
"wave": get_wave(w2)
}, {
"wave": get_wave(w3)
}]
}
response = requests.post(url, json=data)
print(response.ok)
if response.ok:
with open(m, "wb") as outfile:
outfile.write(response.content)
return 'Snowboy模型已保存至{}'.format(m)
else:
return "Snowboy模型生成失败，原因:{}".format(response.text)