import time
import ctypes
import io
import wave
import os
from tempfile import NamedTemporaryFile
from ..voice import Voice
from . import _espeak, toUtf8, fromUtf8
def buildDriver(proxy):
return EspeakDriver(proxy)
class EspeakDriver(object):
_moduleInitialized = False
_defaultVoice = ''
def __init__(self, proxy):
if not EspeakDriver._moduleInitialized:
# espeak cannot initialize more than once per process and has
# issues when terminating from python (assert error on close)
# so just keep it alive and init once
rate = _espeak.Initialize(_espeak.AUDIO_OUTPUT_RETRIEVAL, 1000)
if rate == -1:
raise RuntimeError('could not initialize espeak')
EspeakDriver._defaultVoice = 'default'
EspeakDriver._moduleInitialized = True
_espeak.SetSynthCallback(self._onSynth)
# make sure all props reset
self.setProperty('voice', EspeakDriver._defaultVoice)
self.setProperty('rate', 200)
self.setProperty('volume', 1.0)
self._proxy = proxy
self._looping = True
self._stopping = False
self._data_buffer = b''
self._numerise_buffer = []
def numerise(self, data):
self._numerise_buffer.append(data)
return ctypes.c_void_p(len(self._numerise_buffer))
def decode_numeric(self, data):
return self._numerise_buffer[int(data) - 1]
def destroy(self):
_espeak.SetSynthCallback(None)
def say(self, text):
self._proxy.setBusy(True)
self._proxy.notify('started-utterance')
_espeak.Synth(toUtf8(text), flags=_espeak.ENDPAUSE |
_espeak.CHARS_UTF8)
def stop(self):
if _espeak.IsPlaying():
self._stopping = True
def getProperty(self, name):
if name == 'voices':
voices = []
for v in _espeak.ListVoices(None):
kwargs = {}
kwargs['id'] = fromUtf8(v.name)
kwargs['name'] = fromUtf8(v.name)
if v.languages:
kwargs['languages'] = [v.languages]
genders = [None, 'male', 'female']
kwargs['gender'] = genders[v.gender]
kwargs['age'] = v.age or None
voices.append(Voice(**kwargs))
return voices
elif name == 'voice':
voice = _espeak.GetCurrentVoice()
return fromUtf8(voice.contents.name)
elif name == 'rate':
return _espeak.GetParameter(_espeak.RATE)
elif name == 'volume':
return _espeak.GetParameter(_espeak.VOLUME) / 100.0
else:
raise KeyError('unknown property %s' % name)
def setProperty(self, name, value):
if name == 'voice':
if value is None:
return
try:
utf8Value = toUtf8(value)
_espeak.SetVoiceByName(utf8Value)
except ctypes.ArgumentError as e:
raise ValueError(str(e))
elif name == 'rate':
try:
_espeak.SetParameter(_espeak.RATE, value, 0)
except ctypes.ArgumentError as e:
raise ValueError(str(e))
elif name == 'volume':
try:
_espeak.SetParameter(
_espeak.VOLUME, int(round(value * 100, 2)), 0)
except TypeError as e:
raise ValueError(str(e))
else:
raise KeyError('unknown property %s' % name)
def startLoop(self):
first = True
self._stopping = False
self._looping = True
while self._looping:
if first:
# kick the queue
self._proxy.setBusy(False)
first = False
if self._stopping and self._looping:
# have to do the cancel on the main thread, not inside the
# callback else deadlock
_espeak.Cancel()
self._stopping = False
self._proxy.notify('finished-utterance', completed=False)
self._proxy.setBusy(False)
time.sleep(0.01)
def save_to_file(self, text, filename):
code = self.numerise(filename)
_espeak.Synth(toUtf8(text), flags=_espeak.ENDPAUSE |
_espeak.CHARS_UTF8, user_data=code)
def endLoop(self):
self._looping = False
def iterate(self):
self._proxy.setBusy(False)
while 1:
if self._stopping:
# have to do the cancel on the main thread, not inside the
# callback else deadlock
_espeak.Cancel()
self._stopping = False
self._proxy.notify('finished-utterance', completed=False)
self._proxy.setBusy(False)
yield
def _onSynth(self, wav, numsamples, events):
i = 0
while True:
event = events[i]
if event.type == _espeak.EVENT_LIST_TERMINATED:
break
if event.type == _espeak.EVENT_WORD:
self._proxy.notify('started-word',
location=event.text_position - 1,
length=event.length)
elif event.type == _espeak.EVENT_MSG_TERMINATED:
stream = NamedTemporaryFile()
with wave.open(stream, 'wb') as f:
f.setnchannels(1)
f.setsampwidth(2)
f.setframerate(22050.0)
f.writeframes(self._data_buffer)
if event.user_data:
os.system('ffmpeg -y -i {} {} -loglevel quiet'.format(stream.name, self.decode_numeric(event.user_data)))
else:
os.system('aplay {} -q'.format(stream.name)) # -q for quiet
self._data_buffer = b''
self._proxy.notify('finished-utterance', completed=True)
self._proxy.setBusy(False)
i += 1
if numsamples > 0:
self._data_buffer += ctypes.string_at(wav, numsamples *
ctypes.sizeof(ctypes.c_short))
return 0
