def process_pending_commands(self):
"""Process all pending commands.
Available commands:
stop() - stop processing and exit the process
flush() - flush input buffers.
Now it only flushes the input connection.
It is not able flush data already send to the sound card.
Return True if the process should terminate.
"""
if self.commands.poll():
command = self.commands.recv()
if self.cfg['AudioIO']['debug']:
self.cfg['Logging']['system_logger'].debug(command)
if isinstance(command, Command):
if command.parsed['__name__'] == 'stop':
# Discard all data in play buffer.
while self.audio_play.poll():
self.audio_play.recv()
return True
if command.parsed['__name__'] == 'flush':
# Discard all data in play buffer.
while self.audio_play.poll():
self.audio_play.recv()
return False
if command.parsed['__name__'] == 'flush_out':
self.audio_to_send = ""
while self.audio_play.poll():
self.audio_play.recv()
msg = AlexToClient()
msg.type = AlexToClient.FLUSH_OUT_AUDIO
msg.priority = -1
msg.seq = self.curr_seq
self.send_to_client(msg.SerializeToString())
elif isinstance(command, ASRHyp):
hyp = command.hyp
asr_hyp = hyp.get_best()
msg = AlexToClient()
msg.type = AlexToClient.ASR_RESULT
msg.asr_result = unicode(asr_hyp).lower()
msg.priority = -1
msg.seq = self.curr_seq
self.send_to_client(msg.SerializeToString())
elif isinstance(command, TTSText):
txt = command.text
msg = AlexToClient()
msg.type = AlexToClient.SYSTEM_PROMPT
msg.priority = -1
msg.seq = self.curr_seq
msg.system_prompt = unicode(txt)
self.send_to_client(msg.SerializeToString())
self.process_command(command)
return False
def read_write_audio(self):
"""Send some of the available data to the output.
It should be a non-blocking operation.
Therefore:
1) do not send more then play_buffer_frames
2) send only if stream.get_write_available() is more then the frame size
"""
if self.audio_play.poll():
data_play = self.audio_play.recv()
if isinstance(data_play, Frame):
buffer = data_play.payload
self.audio_to_send += buffer
self.n_played_frames += len(buffer) / 2
elif isinstance(data_play, Command):
if data_play.parsed['__name__'] == 'utterance_start':
msg = AlexToClient()
msg.seq = self.curr_seq
msg.type = AlexToClient.SPEECH_BEGIN
msg.utterance_id = self.utterance_id
self.send_to_client(msg.SerializeToString())
if data_play.parsed['__name__'] == 'utterance_end':
msg = AlexToClient()
msg.seq = self.curr_seq
msg.type = AlexToClient.SPEECH_END
msg.utterance_id = self.utterance_id
self.send_to_client(msg.SerializeToString())
self.utterance_id += 1
self.process_command(data_play)
while len(self.audio_to_send) > WSIO.AUDIO_FRAMES_PER_MSG:
buffer = self.audio_to_send[:WSIO.AUDIO_FRAMES_PER_MSG]
self.audio_to_send = self.audio_to_send[WSIO.AUDIO_FRAMES_PER_MSG:]
msg = AlexToClient()
msg.type = AlexToClient.SPEECH
msg.speech = buffer
msg.seq = self.curr_seq
self.send_to_client(msg.SerializeToString())
