def init_sound(self):
# Configure audio source and sink.
audio_device = None
if input_audio_file:
self._audio_source = audio_helpers.WaveSource(
open(input_audio_file, 'rb'),
sample_rate=audio_sample_rate,
sample_width=audio_sample_width)
else:
self._audio_source = audio_device = (
audio_device or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
if output_audio_file:
self._audio_sink = audio_helpers.WaveSink(
open(output_audio_file, 'wb'),
sample_rate=audio_sample_rate,
sample_width=audio_sample_width)
else:
self._audio_sink = audio_device = (
audio_device or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
# Create conversation stream with the given audio source and sink.
self._conversation_stream = audio_helpers.ConversationStream(
source=self._audio_source,
sink=self._audio_sink,
iter_size=audio_iter_size,
)
def main(api_endpoint, credentials, verbose, input_audio_file,
output_audio_file, audio_sample_rate, audio_sample_width,
audio_iter_size, audio_block_size, audio_flush_size, grpc_deadline,
*args, **kwargs):
"""Samples for the Google Assistant API.
Examples:
Run the sample with microphone input and speaker output:
$ python -m googlesamples.assistant
Run the sample with file input and speaker output:
$ python -m googlesamples.assistant -i <input file>
Run the sample with file input and output:
$ python -m googlesamples.assistant -i <input file> -o <output file>
"""
# Setup logging.
logging.basicConfig(level=logging.DEBUG if verbose else logging.INFO)
# Load credentials.
try:
creds = auth_helpers.load_credentials(
credentials, scopes=[common_settings.ASSISTANT_OAUTH_SCOPE])
except Exception as e:
logging.error('Error loading credentials: %s', e)
logging.error('Run auth_helpers to initialize new OAuth2 credentials.')
return
# Create an authorized gRPC channel.
grpc_channel = auth_helpers.create_grpc_channel(
api_endpoint,
creds,
ssl_credentials_file=kwargs.get('ssl_credentials_for_testing'),
grpc_channel_options=kwargs.get('grpc_channel_option'))
logging.info('Connecting to %s', api_endpoint)
# Configure audio source and sink.
audio_device = None
if input_audio_file:
audio_source = audio_helpers.WaveSource(
open(input_audio_file, 'rb'),
sample_rate=audio_sample_rate,
sample_width=audio_sample_width)
else:
audio_source = audio_device = (audio_device
or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
if output_audio_file:
audio_sink = audio_helpers.WaveSink(open(output_audio_file, 'wb'),
sample_rate=audio_sample_rate,
sample_width=audio_sample_width)
else:
audio_sink = audio_device = (audio_device
or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
# Create conversation stream with the given audio source and sink.
conversation_stream = audio_helpers.ConversationStream(
source=audio_source,
sink=audio_sink,
iter_size=audio_iter_size,
sample_width=audio_sample_width,
)
with SampleAssistant(conversation_stream, grpc_channel,
grpc_deadline) as assistant:
# If file arguments are supplied:
# exit after the first turn of the conversation.
if input_audio_file or output_audio_file:
assistant.converse()
return
mic = Microphone()
continue_conversation = False
while True:
if continue_conversation or mic.wakeup('respeaker'):
continue_conversation = assistant.converse()
def main(*args, **kwargs):
# init ROS
rospy.init_node('assistant_robin_server')
# get parameters
api_endpoint = rospy.get_param(
'~api_endpoint', 'embeddedassistant.googleapis.com'
) # Address of Google Assistant API service.
credentials = rospy.get_param(
'~credentials',
os.path.join(click.get_app_dir(common_settings.ASSISTANT_APP_NAME),
common_settings.ASSISTANT_CREDENTIALS_FILENAME)
) # Path to read OAuth2 credentials.
verbose = rospy.get_param('~verbose', False) # Verbose logging.
audio_sample_rate = rospy.get_param(
'~audio_sample_rate', common_settings.DEFAULT_AUDIO_SAMPLE_RATE
) # Audio sample rate in hertz.
audio_sample_width = rospy.get_param(
'~audio_sample_width', common_settings.DEFAULT_AUDIO_SAMPLE_WIDTH
) # Audio sample width in bytes.
audio_iter_size = rospy.get_param(
'~audio_iter_size', common_settings.DEFAULT_AUDIO_ITER_SIZE
) # Size of each read during audio stream iteration in bytes.
audio_block_size = rospy.get_param(
'~audio_block_size', common_settings.DEFAULT_AUDIO_DEVICE_BLOCK_SIZE
) # Block size in bytes for each audio device read and write operation.
audio_flush_size = rospy.get_param(
'~audio_flush_size', common_settings.DEFAULT_AUDIO_DEVICE_FLUSH_SIZE
) # Size of silence data in bytes written during flush operation.
grpc_deadline = rospy.get_param(
'~grpc_deadline',
common_settings.DEFAULT_GRPC_DEADLINE) # gRPC deadline in seconds.
# Load credentials.
try:
creds = auth_helpers.load_credentials(
credentials, scopes=[common_settings.ASSISTANT_OAUTH_SCOPE])
except Exception as e:
rospy.logerr('Error loading credentials: %s', e)
rospy.logerr('Run auth_helpers to initialize new OAuth2 credentials.')
return
# Create an authorized gRPC channel.
grpc_channel = auth_helpers.create_grpc_channel(
api_endpoint,
creds,
ssl_credentials_file=kwargs.get('ssl_credentials_for_testing'),
grpc_channel_options=kwargs.get('grpc_channel_option'))
rospy.logdebug('Connecting to %s', api_endpoint)
# Configure audio source and sink.
audio_device = None
audio_source = audio_device = (audio_device
or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
audio_sink = audio_device = (audio_device
or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
# Create conversation stream with the given audio source and sink.
conversation_stream = audio_helpers.ConversationStream(
source=audio_source,
sink=audio_sink,
iter_size=audio_iter_size,
sample_width=audio_sample_width,
)
with SampleAssistant(conversation_stream, grpc_channel,
grpc_deadline) as assistant:
global ass
ass = assistant
s = rospy.Service('activate', Activate, handle_activate)
rospy.loginfo("Assistant ready.")
rospy.spin()
def main(api_endpoint, credentials, verbose, input_audio_file,
output_audio_file, audio_sample_rate, audio_sample_width,
audio_iter_size, audio_block_size, audio_flush_size, grpc_deadline,
once, *args, **kwargs):
"""Samples for the Google Assistant API.
Examples:
Run the sample with microphone input and speaker output:
$ python -m googlesamples.assistant
Run the sample with file input and speaker output:
$ python -m googlesamples.assistant -i <input file>
Run the sample with file input and output:
$ python -m googlesamples.assistant -i <input file> -o <output file>
"""
# Setup logging.
logging.basicConfig(level=logging.DEBUG if verbose else logging.INFO)
# Load credentials.
try:
creds = auth_helpers.load_credentials(
credentials, scopes=[common_settings.ASSISTANT_OAUTH_SCOPE])
except Exception as e:
logging.error('Error loading credentials: %s', e)
logging.error('Run auth_helpers to initialize new OAuth2 credentials.')
return
# Create an authorized gRPC channel.
grpc_channel = auth_helpers.create_grpc_channel(
api_endpoint,
creds,
ssl_credentials_file=kwargs.get('ssl_credentials_for_testing'),
grpc_channel_options=kwargs.get('grpc_channel_option'))
logging.info('Connecting to %s', api_endpoint)
# Configure audio source and sink.
audio_device = None
if input_audio_file:
audio_source = audio_helpers.WaveSource(
open(input_audio_file, 'rb'),
sample_rate=audio_sample_rate,
sample_width=audio_sample_width)
else:
audio_source = audio_device = (audio_device
or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
if output_audio_file:
audio_sink = audio_helpers.WaveSink(open(output_audio_file, 'wb'),
sample_rate=audio_sample_rate,
sample_width=audio_sample_width)
else:
audio_sink = audio_device = (audio_device
or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
# Create conversation stream with the given audio source and sink.
conversation_stream = audio_helpers.ConversationStream(
source=audio_source,
sink=audio_sink,
iter_size=audio_iter_size,
sample_width=audio_sample_width,
)
with SampleAssistant(conversation_stream, grpc_channel,
grpc_deadline) as assistant:
# If file arguments are supplied:
# exit after the first turn of the conversation.
if input_audio_file or output_audio_file:
assistant.converse()
return
# If no file arguments supplied:
# keep recording voice requests using the microphone
# and playing back assistant response using the speaker.
# When the once flag is set, don't wait for a trigger. Otherwise, wait.
wait_for_user_trigger = not once
while True:
if wait_for_user_trigger:
click.pause(info='Press Enter to send a new request...')
continue_conversation = assistant.converse()
# wait for user trigger if there is no follow-up turn in
# the conversation.
wait_for_user_trigger = not continue_conversation
# If we only want one conversation, break.
if once and (not continue_conversation):
break
def main(api_endpoint, credentials, verbose, input_audio_file,
output_audio_file, audio_sample_rate, audio_sample_width,
audio_iter_size, audio_block_size, audio_flush_size, grpc_deadline,
*args, **kwargs):
"""Samples for the Google Assistant API.
Examples:
Run the sample with microphone input and speaker output:
$ python -m googlesamples.assistant
Run the sample with file input and speaker output:
$ python -m googlesamples.assistant -i <input file>
Run the sample with file input and output:
$ python -m googlesamples.assistant -i <input file> -o <output file>
"""
# Setup logging.
logging.basicConfig(level=logging.DEBUG if verbose else logging.INFO)
# Load credentials.
try:
creds = auth_helpers.load_credentials(
credentials, scopes=[common_settings.ASSISTANT_OAUTH_SCOPE])
except Exception as e:
logging.error('Error loading credentials: %s', e)
logging.error('Run auth_helpers to initialize new OAuth2 credentials.')
return
# Create gRPC channel
grpc_channel = auth_helpers.create_grpc_channel(
api_endpoint,
creds,
ssl_credentials_file=kwargs.get('ssl_credentials_for_testing'),
grpc_channel_options=kwargs.get('grpc_channel_option'))
logging.info('Connecting to %s', api_endpoint)
# Create Google Assistant API gRPC client.
assistant = embedded_assistant_pb2.EmbeddedAssistantStub(grpc_channel)
# Configure audio source and sink.
audio_device = None
if input_audio_file:
audio_source = audio_helpers.WaveSource(
open(input_audio_file, 'rb'),
sample_rate=audio_sample_rate,
sample_width=audio_sample_width)
else:
audio_source = audio_device = (audio_device
or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
if output_audio_file:
audio_sink = audio_helpers.WaveSink(open(output_audio_file, 'wb'),
sample_rate=audio_sample_rate,
sample_width=audio_sample_width)
else:
audio_sink = audio_device = (audio_device
or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
# Create conversation stream with the given audio source and sink.
conversation_stream = audio_helpers.ConversationStream(
source=audio_source,
sink=audio_sink,
iter_size=audio_iter_size,
)
# Interactive by default.
wait_for_user_trigger = True
# If file arguments are supplied, don't wait for user trigger.
if input_audio_file or output_audio_file:
wait_for_user_trigger = False
# Stores an opaque blob provided in ConverseResponse that,
# when provided in a follow-up ConverseRequest,
# gives the Assistant a context marker within the current state
# of the multi-Converse()-RPC "conversation".
# This value, along with MicrophoneMode, supports a more natural
# "conversation" with the Assistant.
conversation_state_bytes = None
# Stores the current volument percentage.
# Note: No volume change is currently implemented in this sample
volume_percentage = 50
while True:
conversation_stream.start_recording()
logging.info('Recording audio request.')
# This generator yields ConverseRequest to send to the gRPC
# Google Assistant API.
def gen_converse_requests():
converse_state = None
if conversation_state_bytes:
logging.debug('Sending converse_state: %s',
conversation_state_bytes)
converse_state = embedded_assistant_pb2.ConverseState(
conversation_state=conversation_state_bytes, )
config = embedded_assistant_pb2.ConverseConfig(
audio_in_config=embedded_assistant_pb2.AudioInConfig(
encoding='LINEAR16',
sample_rate_hertz=int(audio_sample_rate),
),
audio_out_config=embedded_assistant_pb2.AudioOutConfig(
encoding='LINEAR16',
sample_rate_hertz=int(audio_sample_rate),
volume_percentage=volume_percentage,
),
converse_state=converse_state)
# The first ConverseRequest must contain the ConverseConfig
# and no audio data.
yield embedded_assistant_pb2.ConverseRequest(config=config)
for data in conversation_stream:
# Subsequent requests need audio data, but not config.
yield embedded_assistant_pb2.ConverseRequest(audio_in=data)
def iter_converse_requests():
for c in gen_converse_requests():
assistant_helpers.log_converse_request_without_audio(c)
yield c
conversation_stream.start_playback()
# This generator yields ConverseResponse proto messages
# received from the gRPC Google Assistant API.
for resp in assistant.Converse(iter_converse_requests(),
grpc_deadline):
assistant_helpers.log_converse_response_without_audio(resp)
if resp.error.code != code_pb2.OK:
logging.error('server error: %s', resp.error.message)
break
if resp.event_type == END_OF_UTTERANCE:
logging.info('End of audio request detected')
conversation_stream.stop_recording()
if resp.result.spoken_request_text:
logging.info('Transcript of user request: "%s".',
resp.result.spoken_request_text)
logging.info('Playing assistant response.')
if len(resp.audio_out.audio_data) > 0:
conversation_stream.write(resp.audio_out.audio_data)
if resp.result.spoken_response_text:
logging.info(
'Transcript of TTS response '
'(only populated from IFTTT): "%s".',
resp.result.spoken_response_text)
if resp.result.conversation_state:
conversation_state_bytes = resp.result.conversation_state
if resp.result.volume_percentage != 0:
volume_percentage = resp.result.volume_percentage
logging.info('Volume should be set to %s%%', volume_percentage)
if resp.result.microphone_mode == DIALOG_FOLLOW_ON:
wait_for_user_trigger = False
logging.info('Expecting follow-on query from user.')
elif resp.result.microphone_mode == CLOSE_MICROPHONE:
wait_for_user_trigger = True
logging.info('Finished playing assistant response.')
conversation_stream.stop_playback()
# If file arguments are supplied, end the conversation.
if input_audio_file or output_audio_file:
break
if wait_for_user_trigger:
break
conversation_stream.close()
def begin_play(self):
"""Samples for the Google Assistant API.
Examples:
Run the sample with microphone input and speaker output:
$ python -m googlesamples.assistant
Run the sample with file input and speaker output:
$ python -m googlesamples.assistant -i <input file>
Run the sample with file input and output:
$ python -m googlesamples.assistant -i <input file> -o <output file>
"""
ue.log('Initializing Google Samples API.')
# Setup logging.
logging.basicConfig(level=logging.DEBUG if verbose else logging.INFO)
# Load credentials.
try:
creds = auth_helpers.load_credentials(
credentials, scopes=[common_settings.ASSISTANT_OAUTH_SCOPE])
except Exception as e:
#logging.error('Error loading credentials: %s', e)
#logging.error('Run auth_helpers to initialize new OAuth2 credentials.')
ue.log_error('Error loading credentials: %s', e)
ue.log_error(
'Run auth_helpers to initialize new OAuth2 credentials.')
return
# Create an authorized gRPC channel.
grpc_channel = auth_helpers.create_grpc_channel(
api_endpoint,
creds,
ssl_credentials_file=kwargs.get('ssl_credentials_for_testing'),
grpc_channel_options=kwargs.get('grpc_channel_option'))
ue.log('Connecting to %s', api_endpoint)
# Configure audio source and sink.
audio_device = None
if input_audio_file:
audio_source = audio_helpers.WaveSource(
open(input_audio_file, 'rb'),
sample_rate=audio_sample_rate,
sample_width=audio_sample_width)
else:
audio_source = audio_device = (audio_device
or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
if output_audio_file:
audio_sink = audio_helpers.WaveSink(
open(output_audio_file, 'wb'),
sample_rate=audio_sample_rate,
sample_width=audio_sample_width)
else:
audio_sink = audio_device = (audio_device
or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
# Create conversation stream with the given audio source and sink.
conversation_stream = audio_helpers.ConversationStream(
source=audio_source,
sink=audio_sink,
iter_size=audio_iter_size,
sample_width=audio_sample_width,
)
def begin_play(self):
"""Samples for the Google Assistant API.
Examples:
Run the sample with microphone input and speaker output:
$ python -m googlesamples.assistant
Run the sample with file input and speaker output:
$ python -m googlesamples.assistant -i <input file>
Run the sample with file input and output:
$ python -m googlesamples.assistant -i <input file> -o <output file>
"""
ue.log('Initializing Google Samples API.')
# Initialize defaults
credentials = os.path.join(
sys.path[0], common_settings.ASSISTANT_CREDENTIALS_FILENAME)
# Load credentials.
try:
creds = auth_helpers.load_credentials(
credentials, scopes=[common_settings.ASSISTANT_OAUTH_SCOPE])
except Exception:
# Maybe we didn't load the credentials yet?
# This could happen on first run
creds = auth_helpers.credentials_flow_interactive(
credentials, common_settings.ASSISTANT_OAUTH_SCOPE)
auth_helpers.save_credentials(credentials, creds)
try:
creds = auth_helpers.load_credentials(
credentials,
scopes=[common_settings.ASSISTANT_OAUTH_SCOPE])
except Exception as e:
ue.log_error('Error loading credentials: ' + str(e))
ue.log_error(
'Run auth_helpers to initialize new OAuth2 credentials.')
return
ue.log('Begin play done!')
# Define endpoint
# This might where you can inject custom API.AI behaviors?
api_endpoint = ASSISTANT_API_ENDPOINT
# Create an authorized gRPC channel.
grpc_channel = auth_helpers.create_grpc_channel(api_endpoint, creds)
ue.log('Connecting to ' + str(api_endpoint))
# Set up audio parameters
audio_sample_rate = common_settings.DEFAULT_AUDIO_SAMPLE_RATE
audio_sample_width = common_settings.DEFAULT_AUDIO_SAMPLE_WIDTH
audio_iter_size = common_settings.DEFAULT_AUDIO_ITER_SIZE
audio_block_size = common_settings.DEFAULT_AUDIO_DEVICE_BLOCK_SIZE
audio_flush_size = common_settings.DEFAULT_AUDIO_DEVICE_FLUSH_SIZE
# Configure audio source and sink.
audio_device = None
audio_source = audio_device = (audio_device
or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
audio_sink = audio_device = (audio_device
or audio_helpers.SoundDeviceStream(
sample_rate=audio_sample_rate,
sample_width=audio_sample_width,
block_size=audio_block_size,
flush_size=audio_flush_size))
# Create conversation stream with the given audio source and sink.
conversation_stream = audio_helpers.ConversationStream(
source=audio_source,
sink=audio_sink,
iter_size=audio_iter_size,
sample_width=audio_sample_width,
)
ue.log('Audio device: ' + str(audio_device))
self.assistant = SampleAssistant(conversation_stream, grpc_channel,
common_settings.DEFAULT_GRPC_DEADLINE)
self.assistant.converse()
def assist(self):
# Configure audio source and sink.
self.audio_device = None
self.audio_source = self.audio_device = (
self.audio_device or audio_helpers.SoundDeviceStream(
sample_rate=self.audio_sample_rate,
sample_width=self.audio_sample_width,
block_size=self.audio_block_size,
flush_size=self.audio_flush_size))
self.audio_sink = self.audio_device = (
self.audio_device or audio_helpers.SoundDeviceStream(
sample_rate=self.audio_sample_rate,
sample_width=self.audio_sample_width,
block_size=self.audio_block_size,
flush_size=self.audio_flush_size))
# Create conversation stream with the given audio source and sink.
self.conversation_stream = audio_helpers.ConversationStream(
source=self.audio_source,
sink=self.audio_sink,
iter_size=self.audio_iter_size,
sample_width=self.audio_sample_width)
restart = False
continue_dialog = True
try:
while continue_dialog:
continue_dialog = False
# snowboydecoder.play_audio_file(snowboydecoder.DETECT_DING)
self.conversation_stream.start_recording()
self.logger.info('Recording audio request.')
# This generator yields ConverseResponse proto messages
# received from the gRPC Google Assistant API.
for resp in self.assistant.Converse(
self._iter_converse_requests(), self.grpc_deadline):
assistant_helpers.log_converse_response_without_audio(resp)
if resp.error.code != code_pb2.OK:
self.logger.error('server error: %s',
resp.error.message)
break
if resp.event_type == Assistant.END_OF_UTTERANCE:
self.logger.info('End of audio request detected')
self.conversation_stream.stop_recording()
if resp.result.spoken_request_text:
self.logger.info('Transcript of user request: "%s".',
resp.result.spoken_request_text)
self.logger.info('Playing assistant response.')
if len(resp.audio_out.audio_data) > 0:
self.conversation_stream.write(
resp.audio_out.audio_data)
if resp.result.spoken_response_text:
self.logger.info(
'Transcript of TTS response '
'(only populated from IFTTT): "%s".',
resp.result.spoken_response_text)
if resp.result.conversation_state:
self.conversation_state_bytes = resp.result.conversation_state
if resp.result.volume_percentage != 0:
volume_percentage = resp.result.volume_percentage
self.logger.info('Volume should be set to %s%%',
volume_percentage)
if resp.result.microphone_mode == self.DIALOG_FOLLOW_ON:
continue_dialog = True
self.logger.info(
'Expecting follow-on query from user.')
self.logger.info('Finished playing assistant response.')
self.conversation_stream.stop_playback()
except Exception as e:
self._create_assistant()
self.logger.exception('Skipping because of connection reset')
restart = True
try:
self.conversation_stream.close()
if restart:
self.assist()
except Exception:
self.logger.error('Failed to close conversation_stream.')