def begin_interaction(stream):
global finished, Paused, userID, witResultOverride
flacrecord.calibrate_input_threshold(stream)
googleTTS("Greetings! After the tone, please speak clearly towards my face. Don't forget to say Hello!")
Paused = False
finished = False
while not finished and not rospy.is_shutdown():
while (Paused):
pass
flac_file = listen_for_block_of_speech(stream)
if finished:
break
if not flac_file == []:
if finished:
break
override = witResultOverride
witResultOverride = None
if override:
pub_speech.publish(str(override))
responseString, finished = responseprocess.messageResponse(override, userID, stream)
print responseString
googleTTS(responseString)
else:
hypothesis = stt_google_wav(flac_file)
if hypothesis:
pub_speech.publish(hypothesis)
if hypothesis.lower() == "what does the fox say":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/fox.wav'))
elif hypothesis.lower() == "do you have a license":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/mate.wav'))
elif hypothesis.lower() == "do you work out":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/work.wav'))
elif hypothesis.lower() == "something strange in the neighborhood":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/ghost.wav'))
elif hypothesis.lower() == "attention":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/attention.wav'))
elif hypothesis.lower() == "i'm talking to you":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/talking.wav'))
elif hypothesis.lower() == "i don't want you":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/want.wav'))
elif hypothesis.lower() == "how did we meet":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/met.wav'))
elif hypothesis.lower() == "do you have any popcorn":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/popcorn.wav'))
elif hypothesis.lower() == "how did you get to work today":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/underground.wav'))
elif hypothesis.lower() == "which man are you":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/tetris.wav'))
elif hypothesis.lower() == "what are you":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/blue.wav'))
elif hypothesis.lower() == "the dogs are out":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/dogs.wav'))
elif hypothesis.lower() == "what did she say":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/love.wav'))
elif hypothesis.lower() == "is there a house in new orleans":
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'easter_eggs/orleans.wav'))
else :
witResult = witLookup(hypothesis)
if witResult != []:
responseString, finished = responseprocess.messageResponse(witResult, userID, stream)
else:
responseString = responseprocess.randomNegative()
flacrecord.calibrate_input_threshold(stream)
googleTTS(responseString)
else:
responseString = responseprocess.randomNegative()
flacrecord.calibrate_input_threshold(stream)
googleTTS(responseString)
else:
responseString = responseprocess.randomNegative()
flacrecord.calibrate_input_threshold(stream)
googleTTS(responseString)
def begin_interaction(stream):
global finished, Paused, userID, witResultOverride
flacrecord.calibrate_input_threshold(stream)
googleTTS(
"Greetings! After the tone, please speak clearly towards my face. Don't forget to say Hello!"
)
Paused = False
finished = False
while not finished and not rospy.is_shutdown():
while (Paused):
pass
flac_file = listen_for_block_of_speech(stream)
if finished:
break
if not flac_file == []:
if finished:
break
override = witResultOverride
witResultOverride = None
if override:
pub_speech.publish(str(override))
responseString, finished = responseprocess.messageResponse(
override, userID, stream)
print responseString
googleTTS(responseString)
else:
hypothesis = stt_google_wav(flac_file)
if hypothesis:
pub_speech.publish(hypothesis)
if hypothesis.lower() == "what does the fox say":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/fox.wav'))
elif hypothesis.lower() == "do you have a license":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/mate.wav'))
elif hypothesis.lower() == "do you work out":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/work.wav'))
elif hypothesis.lower(
) == "something strange in the neighborhood":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/ghost.wav'))
elif hypothesis.lower() == "attention":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/attention.wav'))
elif hypothesis.lower() == "i'm talking to you":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/talking.wav'))
elif hypothesis.lower() == "i don't want you":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/want.wav'))
elif hypothesis.lower() == "how did we meet":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/met.wav'))
elif hypothesis.lower() == "do you have any popcorn":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/popcorn.wav'))
elif hypothesis.lower() == "how did you get to work today":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/underground.wav'))
elif hypothesis.lower() == "which man are you":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/tetris.wav'))
elif hypothesis.lower() == "what are you":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/blue.wav'))
elif hypothesis.lower() == "the dogs are out":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/dogs.wav'))
elif hypothesis.lower() == "what did she say":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/love.wav'))
elif hypothesis.lower(
) == "is there a house in new orleans":
play_wav(
os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'easter_eggs/orleans.wav'))
else:
witResult = witLookup(hypothesis)
if witResult != []:
responseString, finished = responseprocess.messageResponse(
witResult, userID, stream)
else:
responseString = responseprocess.randomNegative()
flacrecord.calibrate_input_threshold(stream)
googleTTS(responseString)
else:
responseString = responseprocess.randomNegative()
flacrecord.calibrate_input_threshold(stream)
googleTTS(responseString)
else:
responseString = responseprocess.randomNegative()
flacrecord.calibrate_input_threshold(stream)
googleTTS(responseString)
def listen_for_block_of_speech(stream):
global interaction_cancelled
global recording_override_start
global recording_override_stop
interaction_cancelled = False
recording_override_start = False
recording_override_stop = False
full_recording = []
max_recording_chunks = 8 * globalvariables.CHUNKS_PER_SECOND
face_controller = FaceController()
calibrate_input_threshold()
recording_started = False
beep_start = time.time()
stream.start_stream()
play_wav(
os.path.join(os.path.dirname(os.path.realpath(__file__)),
'raw/soundstart.wav'))
beep_end = time.time()
print "Removing first " + str(
int(globalvariables.SAMPLERATE *
(beep_end - beep_start))) + " samples from recording"
stream.read(int(globalvariables.SAMPLERATE * (beep_end - beep_start)))
prev_visual_talkingness = 0
listening_start_time = time.time()
face_controller.update(message="Speak now...")
while True:
if interaction_cancelled:
print "-----------Recording cancelled-----------"
full_recording = []
break
try:
chunk_data = stream.read(globalvariables.SAMPLES_PER_CHUNK)
except IOError, e:
if e.args[1] == pyaudio.paInputOverflowed:
chunk_data = '\x00' * globalvariables.SAMPLES_PER_CHUNK
else:
raise
full_recording.append(chunk_data)
print "Visual talkingness: " + str(visual_talkingness)
if recording_started:
if not recording_override_start:
if visual_talkingness <= visual_talkingness_passive_avg:
print "-----------Stopped recording due to stopped speaking-----------"
break
else:
if visual_talkingness > visual_talkingness_passive_max or recording_override_start:
print("starting to record")
face_controller.update(message="I'm listening...")
recording_started = True
else:
if len(full_recording) > globalvariables.PRE_CHUNKS:
full_recording.pop(0)
if time.time() - listening_start_time > 5:
print "--------Stopped recording as nothing heard--------"
full_recording = []
break
if recording_override_stop:
print "-----------Stopped recording due to override-----------"
break
if len(full_recording) >= max_recording_chunks:
print "-----------Stopped recording due to max recording length-----------"
break
print "--------Stopped recording as nothing heard--------"
full_recording = []
break
if recording_override_stop:
print "-----------Stopped recording due to override-----------"
break
if len(full_recording) >= max_recording_chunks:
print "-----------Stopped recording due to max recording length-----------"
break
stream.stop_stream()
face_controller.update(message="Hang on a moment...")
play_wav(
os.path.join(os.path.dirname(os.path.realpath(__file__)),
'raw/soundstop.wav'))
if len(full_recording):
wav_filename = save_speech(full_recording, p)
flac_filename = convert_wav_to_flac(wav_filename)
return flac_filename
else:
return ""
def save_speech(data, p):
filename = os.path.join(os.path.dirname(os.path.realpath(__file__)),
'output_' + str(int(time.time())))
# write data to WAVE file
data = ''.join(data)
def listen_for_block_of_speech(stream):
global interaction_cancelled
global recording_override_start
global recording_override_stop
interaction_cancelled = False
recording_override_start = False
recording_override_stop = False
full_recording = []
max_recording_chunks = 8 * globalvariables.CHUNKS_PER_SECOND
face_controller = FaceController()
calibrate_input_threshold()
recording_started = False
beep_start = time.time()
stream.start_stream()
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'raw/soundstart.wav'))
beep_end = time.time()
print "Removing first " + str(int(globalvariables.SAMPLERATE*(beep_end - beep_start))) + " samples from recording"
stream.read(int(globalvariables.SAMPLERATE*(beep_end - beep_start)))
prev_visual_talkingness = 0
listening_start_time = time.time()
face_controller.update(message="Speak now...")
while True:
if interaction_cancelled:
print "-----------Recording cancelled-----------"
full_recording = []
break
try:
chunk_data = stream.read(globalvariables.SAMPLES_PER_CHUNK)
except IOError, e:
if e.args[1] == pyaudio.paInputOverflowed:
chunk_data = '\x00'*globalvariables.SAMPLES_PER_CHUNK
else:
raise
full_recording.append(chunk_data)
print "Visual talkingness: " + str(visual_talkingness)
if recording_started:
if not recording_override_start:
if visual_talkingness <= visual_talkingness_passive_avg:
print "-----------Stopped recording due to stopped speaking-----------"
break
else:
if visual_talkingness > visual_talkingness_passive_max or recording_override_start:
print("starting to record")
face_controller.update(message="I'm listening...")
recording_started = True
else:
if len(full_recording) > globalvariables.PRE_CHUNKS:
full_recording.pop(0)
if time.time() - listening_start_time > 5:
print "--------Stopped recording as nothing heard--------"
full_recording = []
break
if recording_override_stop:
print "-----------Stopped recording due to override-----------"
break
if len(full_recording) >= max_recording_chunks:
print "-----------Stopped recording due to max recording length-----------"
break
if time.time() - listening_start_time > 5:
print "--------Stopped recording as nothing heard--------"
full_recording = []
break
if recording_override_stop:
print "-----------Stopped recording due to override-----------"
break
if len(full_recording) >= max_recording_chunks:
print "-----------Stopped recording due to max recording length-----------"
break
stream.stop_stream()
face_controller.update(message="Hang on a moment...")
play_wav(os.path.join(os.path.dirname(os.path.realpath(__file__)), 'raw/soundstop.wav'))
if len(full_recording):
wav_filename = save_speech(full_recording,p)
flac_filename = convert_wav_to_flac(wav_filename)
return flac_filename
else:
return ""
def save_speech(data, p):
filename = os.path.join(os.path.dirname(os.path.realpath(__file__)),'output_'+str(int(time.time())))
# write data to WAVE file
data = ''.join(data)
wf = wave.open(filename+'.wav', 'wb')
wf.setnchannels(1)
