def get_tts(self, sentence, wav_file):
""" Generate (remotely) and play mimic2 WAV audio
Args:
sentence (str): Phrase to synthesize to audio with mimic2
wav_file (str): Location to write audio output
"""
# Use the phonetic_spelling mechanism from the TTS base class
if self.phonetic_spelling:
for word in re.findall(r"[\w']+", sentence):
if word.lower() in self.spellings:
sentence = sentence.replace(word,
self.spellings[word.lower()])
chunks = _sentence_chunker(sentence)
LOG.debug("Generating Mimic2 TSS for: "+str(chunks))
try:
for _, req in enumerate(self._requests(chunks)):
results = req.result().json()
audio = base64.b64decode(results['audio_base64'])
vis = results['visimes']
with open(wav_file, 'wb') as f:
f.write(audio)
except (ReadTimeout, ConnectionError, ConnectTimeout, HTTPError):
raise RemoteTTSTimeoutException(
"Mimic 2 server request timed out. Falling back to mimic")
return (wav_file, vis)
def get_tts(self, sentence, wav_file):
"""request and play mimic2 wav audio
Args:
sentence (str): sentence to synthesize from mimic2
ident (optional): Defaults to None.
"""
sentence = self._normalized_numbers(sentence)
# Use the phonetic_spelling mechanism from the TTS base class
if self.phonetic_spelling:
for word in re.findall(r"[\w']+", sentence):
if word.lower() in self.spellings:
sentence = sentence.replace(word,
self.spellings[word.lower()])
chunks = sentence_chunker(sentence, self.chunk_size)
try:
for idx, req in enumerate(self._requests(chunks)):
results = req.result().json()
audio = base64.b64decode(results['audio_base64'])
vis = results['visimes']
with open(wav_file, 'wb') as f:
f.write(audio)
except (ReadTimeout, ConnectionError, ConnectTimeout, HTTPError):
raise RemoteTTSTimeoutException(
"Mimic 2 remote server request timedout. falling back to mimic"
)
return (wav_file, vis)
def get_tts(self, sentence, wav_file):
""" Generate (remotely) and play mimic2 WAV audio
Args:
sentence (str): Phrase to synthesize to audio with mimic2
wav_file (str): Location to write audio output
"""
LOG.debug("Generating Mimic2 TSS for: " + str(sentence))
try:
req = self._requests(sentence)
results = req.result().json()
audio = base64.b64decode(results['audio_base64'])
vis = results['visimes']
with open(wav_file, 'wb') as f:
f.write(audio)
except (ReadTimeout, ConnectionError, ConnectTimeout, HTTPError):
raise RemoteTTSTimeoutException(
"Mimic 2 server request timed out. Falling back to mimic")
return (wav_file, vis)
def execute(self, sentence, ident=None):
"""request and play mimic2 wav audio
Args:
sentence (str): sentence to synthesize from mimic2
ident (optional): Defaults to None.
"""
create_signal("isSpeaking")
sentence = self._normalized_numbers(sentence)
# Use the phonetic_spelling mechanism from the TTS base class
if self.phonetic_spelling:
for word in re.findall(r"[\w']+", sentence):
if word.lower() in self.spellings:
sentence = sentence.replace(word,
self.spellings[word.lower()])
chunks = sentence_chunker(sentence, self.chunk_size)
try:
for idx, req in enumerate(self._requests(chunks)):
results = req.result().json()
audio = base64.b64decode(results['audio_base64'])
vis = self.visime(results['visimes'])
key = str(hashlib.md5(
chunks[idx].encode('utf-8', 'ignore')).hexdigest())
wav_file = os.path.join(
get_cache_directory("tts"),
key + '.' + self.audio_ext
)
with open(wav_file, 'wb') as f:
f.write(audio)
self.queue.put((self.audio_ext, wav_file, vis, ident))
except (ReadTimeout, ConnectionError, ConnectTimeout, HTTPError):
raise RemoteTTSTimeoutException(
"Mimic 2 remote server request timedout. falling back to mimic"
)