def callback(data):
global GLOBAL_FileNamePublisher
dataParts = data.data.split("|")
if dataParts[0] != "TTS":
return
ttsProvider = TTSMemory().GetString("TTSProvider")
usePygame = TTSMemory().GetBoolean("UsePygame")
FileLogger().Info("TTS, callback(), Provider: {0}".format(ttsProvider))
try:
if usePygame and data == "TRIGGER_STOP_AUDIO":
SoundMixer().Stop()
return
if (ttsProvider.lower() == "google"):
data = Google().Speak(dataParts[1])
if (ttsProvider.lower() == "microsoft"):
data = Microsoft().Speak(dataParts[1])
if (ttsProvider.lower() == "ivona"):
data = Ivona().Speak(dataParts[1])
if (ttsProvider.lower() == "watson"):
data = Watson().Speak(dataParts[1])
try:
audio = MP3(data)
delay = Config().GetInt("TextToSpeech", "IntermediateAudioDelay")
TTSMemory().Set("TTS.Until",
(rospy.Time.now().to_sec() +
int(round(audio.info.length)) + delay))
except Exception as e:
FileLogger().Warn(
"TTS, callback() - Error on getting audio duration: {0}".
format(e))
if usePygame:
SoundMixer().Play(data)
else:
audioPlayer = Config().Get("TextToSpeech",
"AudioPlayer") + " '{0}'"
os.system(audioPlayer.format(data))
FileLogger().Info("TTS, callback(), Play Audio: {0}".format(data))
GLOBAL_FileNamePublisher.publish("TTS|{0}".format(data))
user = User().LoadObject()
if (user.GetName() is not None):
user.UpdateSpokenTo()
user.Update()
except Exception as e:
FileLogger().Error(
"TTS, callback(), Error on processing TTS data: {0}".format(e))
def GetName(self):
if self.Formal and self.LastName:
if self.Gender.lower() == "female":
nameWrapper = Config().Get("DEFAULT",
"FormalFormOfAddressFemale")
else:
nameWrapper = Config().Get("DEFAULT",
"FormalFormOfAddressMale")
return nameWrapper.format(self.LastName)
elif not self.Name and self.FirstName:
return self.FirstName
elif self.Name:
return self.Name
return None
class Watson():
__metaclass__ = Singleton
def __init__(self):
self.CHUNK = 1024
self.BUF_MAX_SIZE = self.CHUNK * 10
self.q = Queue(maxsize=int(round(self.BUF_MAX_SIZE / self.CHUNK)))
self.audio_source = AudioSource(self.q, True, True)
self.FORMAT = pyaudio.paInt16
self.CHANNELS = 1
self.RATE = 44100
self.__apikey_stt = Config().Get("SpeechToText", "WatsonSTTAPIKey")
self.__url_stt = Config().Get("SpeechToText", "WatsonSTTUrl")
self.__apikey_tts = Config().Get("TextToSpeech", "WatsonTTSAPIKey")
self.__url_tts = Config().Get("TextToSpeech", "WatsonTTSUrl")
self.__voiceName = Config().Get("TextToSpeech", "WatsonVoiceName")
self.__language_2letter_cc = Config().Get("SpeechToText",
"CountryCode2Letter")
self.__language_4letter_cc = Config().Get("SpeechToText",
"CountryCode4Letter")
self.__audioPlayer = Config().Get("TextToSpeech",
"AudioPlayer") + " '{0}'"
self.text_to_speech = TextToSpeechV1(url=self.__url_tts,
iam_apikey=self.__apikey_tts)
self.text_to_speech.set_default_headers(
{'x-watson-learning-opt-out': "true"})
self.speech_to_text = SpeechToTextV1(url=self.__url_stt,
iam_apikey=self.__apikey_stt)
self.speech_to_text.set_default_headers(
{'x-watson-learning-opt-out': "true"})
self.audio = pyaudio.PyAudio()
# open stream using callback
self.stream = self.audio.open(format=self.FORMAT,
channels=self.CHANNELS,
rate=self.RATE,
input=True,
frames_per_buffer=self.CHUNK,
stream_callback=self.pyaudio_callback,
start=False)
try:
rospy.init_node('STT_watson_node', anonymous=True)
except:
FileLogger().Info('already initialized')
def Speak(self, audioString, playAudio=False):
if (len(audioString) == 0):
return
tmpAudioFile = os.path.join(Global.EmeraldPath, "Data", "TTS", ("Watson_" + \
self.__language_2letter_cc + "_" + \
self.CleanString(audioString) + ".mp3"))
if not os.path.isfile(tmpAudioFile):
with open(join(dirname(__file__), tmpAudioFile),
'wb') as audio_file:
response = self.text_to_speech.synthesize(
audioString, accept='audio/mp3',
voice=self.__voiceName).get_result()
audio_file.write(response.content)
if (playAudio):
os.system(self.__audioPlayer.format(tmpAudioFile))
return tmpAudioFile
def Listen(self):
self.stream.start_stream()
try:
while True:
recognize_thread = Thread(
target=self.recognize_using_weboscket, args=())
recognize_thread.start()
recognize_thread.join()
except KeyboardInterrupt:
# stop recording
self.audio_source.completed_recording()
self.stream.stop_stream()
self.stream.close()
self.audio.terminate()
def CleanString(self, string):
data = re.sub(r'\W+', '', string)
return (data[:75] + '_TRIMMED') if len(data) > 75 else data
def recognize_using_weboscket(self, *args):
mycallback = MyRecognizeCallback()
self.speech_to_text.recognize_using_websocket(
audio=self.audio_source,
content_type='audio/l16; rate=44100',
recognize_callback=mycallback,
interim_results=True,
model='{0}_BroadbandModel'.format(self.__language_4letter_cc),
smart_formatting=True)
def pyaudio_callback(self, in_data, frame_count, time_info, status):
try:
self.q.put(in_data)
except Full:
pass
return (None, pyaudio.paContinue)
class Ivona(object):
__metaclass__ = Singleton
__language_2letter_cc = 'de'
__language_4letter_cc = 'de-DE'
__audioPlayer = "afplay '{0}'"
__voiceGender = 'Male'
__voiceName = 'Hans'
__accessKey = None
__secretKey = None
__speechRate = None
__sentenceBreak = None
__paragraphBreak = None
__session = None
__host = None
__region = None
__region_options = {
'us-east': 'us-east-1',
'us-west': 'us-west-2',
'eu-west': 'eu-west-1',
}
def __setRegion(self, region_name):
self.__region = self.__region_options.get(region_name, 'eu-west-1')
self.__host = 'tts.{}.ivonacloud.com'.format(self.__region)
__codec = None
__codec_options = {
'mp3': 'mp3',
'ogg': 'ogg',
'mp4': 'mp4',
}
def __setCodec(self, codec_name):
self.__codec = self.__codec_options.get(codec_name, 'mp3')
def __init__(self):
self.__language_2letter_cc = Config().Get("TextToSpeech",
"CountryCode2Letter")
self.__language_4letter_cc = Config().Get("TextToSpeech",
"CountryCode4Letter")
self.__audioPlayer = Config().Get("TextToSpeech",
"AudioPlayer") + " '{0}'"
self.__voiceGender = Config().Get("TextToSpeech", "IvonaVoiceGender")
self.__voiceName = Config().Get("TextToSpeech", "IvonaVoiceName")
self.__accessKey = Config().Get("TextToSpeech", "IvonaAccessKey")
self.__secretKey = Config().Get("TextToSpeech", "IvonaSecretKey")
self.__speechRate = 'medium' # x-slow - slow - medium - fast - x-fast
self.__sentenceBreak = 400
self.__paragraphBreak = 650
self.__setRegion('eu-west')
self.__setCodec('mp3')
def Speak(self, audioString, playAudio=False):
if (len(audioString) == 0):
return
tmpAudioFile = os.path.join(Global.EmeraldPath, "Data", "TTS", ("Ivona_" + \
self.__language_2letter_cc + "_" + \
self.CleanString(audioString) + ".mp3"))
if not os.path.isfile(tmpAudioFile):
with open(tmpAudioFile, "wb") as f:
r = self._send_amazon_auth_packet_v4(
'POST', 'tts', 'application/json', '/CreateSpeech', '',
self._generate_payload(audioString), self.__region,
self.__host)
if r.content.startswith(b'{'):
raise Exception('Error fetching voice: {}'.format(
r.content))
else:
f.write(r.content)
if (playAudio):
os.system(self.__audioPlayer.format(tmpAudioFile))
return tmpAudioFile
def GetVoices(self):
"""Returns all the possible voices
"""
r = self._send_amazon_auth_packet_v4('POST', 'tts', 'application/json',
'/ListVoices', '', '',
self.__region, self.__host)
return r.json()
def CleanString(self, string):
data = re.sub(r'\W+', '', string)
return (data[:75] + '_TRIMMED') if len(data) > 75 else data
def _generate_payload(self, text_to_speak):
return json.dumps({
'Input': {
"Type": "application/ssml+xml",
'Data': text_to_speak
},
'OutputFormat': {
'Codec': self.__codec.upper()
},
'Parameters': {
'Rate': self.__speechRate,
'SentenceBreak': self.__sentenceBreak,
'ParagraphBreak': self.__paragraphBreak
},
'Voice': {
'Name': self.__voiceName,
'Language': self.__language_4letter_cc,
'Gender': self.__voiceGender
}
})
def _send_amazon_auth_packet_v4(self, method, service, content_type,
canonical_uri, canonical_querystring,
request_parameters, region, host):
"""Send a packet to a given amazon server using Amazon's signature Version 4,
Returns the resulting response object
"""
algorithm = 'AWS4-HMAC-SHA256'
signed_headers = 'content-type;host;x-amz-content-sha256;x-amz-date'
# Create date for headers and the credential string
t = datetime.datetime.utcnow()
amazon_date = t.strftime('%Y%m%dT%H%M%SZ')
date_stamp = t.strftime('%Y%m%d')
# Step 1: Create canonical request
payload_hash = self._sha_hash(request_parameters)
canonical_headers = 'content-type:{}\n'.format(content_type)
canonical_headers += 'host:{}\n'.format(host)
canonical_headers += 'x-amz-content-sha256:{}\n'.format(payload_hash)
canonical_headers += 'x-amz-date:{}\n'.format(amazon_date)
canonical_request = '\n'.join([
method, canonical_uri, canonical_querystring, canonical_headers,
signed_headers, payload_hash
])
# Step 2: Create the string to sign
credential_scope = '{}/{}/{}/aws4_request'.format(
date_stamp, region, service)
string_to_sign = '\n'.join([
algorithm, amazon_date, credential_scope,
self._sha_hash(canonical_request)
])
# Step 3: Calculate the signature
signing_key = self._get_signature_key(self.__secretKey, date_stamp,
region, service)
signature = hmac.new(signing_key, string_to_sign.encode('utf-8'),
hashlib.sha256).hexdigest()
# Step 4: Create the signed packet
endpoint = 'https://{}{}'.format(host, canonical_uri)
authorization_header = '{} Credential={}/{}, ' +\
'SignedHeaders={}, Signature={}'
authorization_header = authorization_header.format(
algorithm, self.__accessKey, credential_scope, signed_headers,
signature)
headers = {
'Host': host,
'Content-type': content_type,
'X-Amz-Date': amazon_date,
'Authorization': authorization_header,
'x-amz-content-sha256': payload_hash,
'Content-Length': str(len(request_parameters))
}
# Send the packet and return the response
# Use requests.Session() for HTTP keep-alive
if self.__session is None:
self.__session = requests.Session()
return self.__session.post(endpoint,
data=request_parameters,
headers=headers)
def _sha_hash(self, to_hash):
return hashlib.sha256(to_hash.encode('utf-8')).hexdigest()
def _sign(self, key, msg):
return hmac.new(key, msg.encode('utf-8'), hashlib.sha256).digest()
def _get_signature_key(self, key, date_stamp, region_name, service_name):
k_date = self._sign(('AWS4{}'.format(key)).encode('utf-8'), date_stamp)
k_region = self._sign(k_date, region_name)
k_service = self._sign(k_region, service_name)
k_signing = self._sign(k_service, 'aws4_request')
return k_signing
class Google(object):
__metaclass__ = Singleton
__language_2letter_cc = 'de'
__language_4letter_cc = 'de-DE'
__audioPlayer = "afplay '{0}'"
__apiKey = None
__asyncInit = False
__asyncResultList = []
def __init__(self):
self.__language_2letter_cc = Config().Get("TextToSpeech",
"CountryCode2Letter")
self.__language_4letter_cc = Config().Get("TextToSpeech",
"CountryCode4Letter")
self.__audioPlayer = Config().Get("TextToSpeech",
"AudioPlayer") + " '{0}'"
self.__asyncInit = False
self.__apiKey = Config().Get("TextToSpeech", "GoogleAPIKey")
if (len(self.__apiKey) == 0):
self.__apiKey = None
self.__microphoneID = None
microphoneName = Config().Get("SpeechToText", "Microphone")
for i, microphone_name in enumerate(
sr.Microphone().list_microphone_names()):
if microphone_name == microphoneName:
self.__microphoneID = i
if self.__microphoneID is None:
FileLogger().Error(
"Google Line 38: No microphone found - skip listen initialisation"
)
return
self.__recognizer = sr.Recognizer()
#Represents the minimum length of silence (in seconds) that will register as the
#end of a phrase. Can be changed.
#Smaller values result in the recognition completing more quickly, but might result
#in slower speakers being cut off.
self.__recognizer.pause_threshold = 0.5
self.__recognizer.operation_timeout = 3
self.__microphone = sr.Microphone(device_index=self.__microphoneID)
with self.__microphone as source:
self.__recognizer.dynamic_energy_threshold = True
self.__recognizer.adjust_for_ambient_noise(source)
def Speak(self, audioString, playAudio=False):
if (len(audioString) == 0):
return
tmpAudioFile = os.path.join(Global.EmeraldPath, "Data", "TTS", ("Google_" + \
self.__language_2letter_cc + "_" + \
self.CleanString(audioString) + ".mp3"))
if not os.path.isfile(tmpAudioFile):
tts = gTTS(text=audioString, lang=self.__language_2letter_cc)
tts.save(tmpAudioFile)
if (playAudio):
os.system(self.__audioPlayer.format(tmpAudioFile))
return tmpAudioFile
def AsyncCallback(self, recognizer, audio):
if self.__microphoneID is None:
raise Exception("Google: No microphone found - Exit")
data = ""
try:
data = self.__recognizer.recognize_google(
audio,
key=self.__apiKey,
language=self.__language_4letter_cc,
show_all=False)
except sr.UnknownValueError as e:
FileLogger().Warn(
"Google Line 83: Google Speech Recognition could not understand audio: {0}"
.format(e))
except sr.RequestError as e:
FileLogger().Warn(
"Google Line 85: Could not request results from Google Speech Recognition service: {0}"
.format(e))
except Exception as e:
FileLogger().Warn(
"Google Line 87: Error on executing Google Speech Recognition service: {0}"
.format(e))
if (len(data) > 0):
self.__asyncResultList.append(data)
def ListenAsync(self):
if self.__microphoneID is None:
raise Exception("Google: No microphone found - Exit")
if not self.__asyncInit:
self.__recognizer.listen_in_background(self.__microphone,
self.AsyncCallback)
self.__asyncInit = True
if (len(self.__asyncResultList) > 0):
return self.__asyncResultList.pop(0)
return ""
def Listen(self):
if self.__microphoneID is None:
raise Exception("Google: No microphone found - Exit")
with self.__microphone as source:
self.__audio = self.__recognizer.listen(source)
data = ""
try:
data = self.__recognizer.recognize_google(
self.__audio,
key=self.__apiKey,
language=self.__language_4letter_cc,
show_all=False)
except sr.UnknownValueError as e:
FileLogger().Warn(
"Google Line 75: Google Speech Recognition could not understand audio: {0}"
.format(e))
except sr.RequestError as e:
FileLogger().Warn(
"Google Line 77: Could not request results from Google Speech Recognition service: {0}"
.format(e))
except Exception as e:
FileLogger().Warn(
"Google Line 81: Error on executing Google Speech Recognition service: {0}"
.format(e))
return data
def CleanString(self, string):
data = re.sub(r'\W+', '', string)
return (data[:75] + '_TRIMMED') if len(data) > 75 else data
def GetAvailiabeMicrophones(self):
return sr.Microphone().list_microphone_names()
class Microsoft(object):
__metaclass__ = Singleton
__language_2letter_cc = 'de'
__language_4letter_cc = 'de-DE'
__audioPlayer = "afplay '{0}'"
__voiceGender = 'Male'
__voiceName = 'Microsoft Server Speech Text to Speech Voice (de-DE, Stefan, Apollo)'
__apiKey = None
__accesstoken = None
__ssmlTemplate = """<speak version='1.0' xml:lang='{0}'>
<voice xml:lang='{0}' xml:gender='{1}' name='{2}'>
{3}
</voice>
</speak>"""
def __init__(self):
self.__language_2letter_cc = Config().Get("TextToSpeech",
"CountryCode2Letter")
self.__language_4letter_cc = Config().Get("TextToSpeech",
"CountryCode4Letter")
self.__audioPlayer = Config().Get("TextToSpeech",
"AudioPlayer") + " '{0}'"
self.__voiceGender = Config().Get("TextToSpeech",
"MicrosoftVoiceGender")
self.__voiceName = Config().Get("TextToSpeech", "MicrosoftVoiceName")
self.__apiKey = Config().Get("TextToSpeech", "MicrosoftAPIKey")
params = ""
headers = {"Ocp-Apim-Subscription-Key": self.__apiKey}
__AccessTokenHost = "api.cognitive.microsoft.com"
path = "/sts/v1.0/issueToken"
conn = httplib.HTTPSConnection(__AccessTokenHost)
conn.request("POST", path, params, headers)
response = conn.getresponse()
data = response.read()
conn.close()
self.__accesstoken = data.decode("UTF-8")
self.__microphoneID = None
microphoneName = Config().Get("SpeechToText", "Microphone")
for i, microphone_name in enumerate(
sr.Microphone().list_microphone_names()):
if microphone_name == microphoneName:
self.__microphoneID = i
if self.__microphoneID is None:
FileLogger().Error(
"Microsoft Line 44: No microphone found - skip listen initialisation"
)
return
self.__recognizer = sr.Recognizer()
self.__microphone = sr.Microphone(device_index=self.__microphoneID)
with self.__microphone as source:
self.__recognizer.dynamic_energy_threshold = True
self.__recognizer.adjust_for_ambient_noise(source)
def Speak(self, audioString, playAudio=False):
if (len(audioString) == 0):
return
tmpAudioFile = os.path.join(Global.EmeraldPath, "Data", "TTS", ("Microsoft_" + \
self.__language_2letter_cc + "_" + \
self.CleanString(audioString) + ".mp3"))
if not os.path.isfile(tmpAudioFile):
ssml = self.__ssmlTemplate.format(self.__language_4letter_cc,
self.__voiceGender,
self.__voiceName, audioString)
body = ssml # .encode('utf8')
headers = {
"Content-type": "application/ssml+xml",
"X-Microsoft-OutputFormat": "audio-16khz-32kbitrate-mono-mp3",
"Authorization": "Bearer " + self.__accesstoken,
"X-Search-AppId": "07D3234E49CE426DAA29772419F436CA",
"X-Search-ClientID": "1ECFAE91408841A480F00935DC390960",
"User-Agent": "TTSForPython"
}
# Connect to server to synthesize the wave
conn = httplib.HTTPSConnection("speech.platform.bing.com")
conn.request("POST", "/synthesize", body, headers)
response = conn.getresponse()
data = response.read()
conn.close()
with open(tmpAudioFile, "wb") as f:
f.write(data)
if (playAudio):
os.system(self.__audioPlayer.format(tmpAudioFile))
return tmpAudioFile
def Listen(self):
if self.__microphoneID is None:
raise Exception("Microsoft: No microphone found - Exit")
with self.__microphone as source:
self.__audio = self.__recognizer.listen(source)
data = ""
try:
data = self.__recognizer.recognize_bing(
self.__audio,
key=self.__apiKey,
language=self.__language_4letter_cc,
show_all=False)
except sr.UnknownValueError as e:
FileLogger().Warn(
"Microsoft Line 119: Microsoft Bing Voice Recognition could not understand audio: {0}"
.format(e))
except sr.RequestError as e:
FileLogger().Warn(
"Microsoft Line 121: Could not request results from Microsoft Bing Voice Recognition service: {0}"
.format(e))
return data
def CleanString(self, string):
data = re.sub(r'\W+', '', string)
return (data[:75] + '_TRIMMED') if len(data) > 75 else data
def GetAvailiabeMicrophones(self):
return sr.Microphone().list_microphone_names()