def voice_handler(self, bot, update):
""" Voice recognition """
file = self.bot.getFile(update.message.voice.file_id)
print("file_id: " + str(update.message.voice.file_id))
text = speech_to_text(file)
print(text)
update.message.text = text
return self.plain_text_manager(bot, update)
def detect_voice():
f = open('./file.ogg', 'wb')
f.write(request.data)
f.close()
converted = speech.speech_to_text('file.ogg')
resp = requests.post(url='http://127.0.0.1:5000/send_message',
data={"message": converted})
resp = resp.json()
resp["human_mes"] = converted
return jsonify(resp)
def run():
try:
system.play("listen") #воспроизведение предупреждения о начале записи
system.record("record") #запись голосовой команды
text = speech.speech_to_text("record") #преобразование голосовой записи в текстовую строку
answer = chat.get_answer(text) #получение ответа из преобразованной фразы
if answer != "":
speech.text_to_speech(answer, "temp") #синтез ответа в голосовую запись
system.convert("temp") #конвертация ответа в необходимый формат
system.play("temp") #воспроизведение ответа через динамики
except Exception:
system.play("error") #обработка ошибки и воспроизведение сообщения об ошибке
def record(self, time_limit=2):
# Play start recording
speech.play_file('siri_start.mp3')
self.query = speech.speech_to_text(time_limit=time_limit)
# If error
if self.query == 'speech error':
self.nao.talk({'text': "I didn't quite get that, please repeat"})
self.record()
if self.query == 'connection error':
self.nao.talk({
'text':
"I'm sorry, the connection is not available right now"
})
sys.exit()
# Play stop recording
speech.play_file('siri_end.mp3')
def index():
language = 'en-US'
speak_prompt(langToprompt[language], language)
record_speech(filename='help.wav')
convert_to_flac(filename='help.wav')
texts = speech_to_text('help.flac', lang=language)
is_emergency = True
score = None
if texts:
is_emergency, score = detect_help_intent(texts[0])
if is_emergency:
# print('Emergency text sent')
msg = generate_message(
grandma['name'], grandma['time_since_event'], grandma['address'])
send_message(msg, grandma['emergency_number'])
make_call(msg, grandma['emergency_number'])
# TODO: Send Conversation through the socket to front end
send_conversation('Are you Okay?', texts[0], True)
return jsonify({'resp': texts, 'is_emergency': is_emergency, 'Sentiment score': score})
def run(audio_path: str = 'test.wav',
min_length: int = 1000,
db_threshold: int = -32,
speech_api: str = 'sphinx') -> list:
text = speech_to_text(audio_path,
min_length=min_length,
db_threshold=db_threshold,
speech_api=speech_api)
print(f'Speech to text: {" ".join(t for t, a in text)}\n')
for text_audio in text:
results = dict()
sentence, chunk = text_audio
index = chunk.split('_')[1].split('.')[0]
with open(chunk, 'rb') as chunk_bin:
results['audio_bin'] = Binary(chunk_bin.read())
results.update(parse(sentence, index=index))
results['created'] = datetime.now()
db.texts.insert_one(results)
def process_message(event, user_id, channel_id, text):
global USER_LANGUAGE_MAPPINGS
user_name = get_name_from_user_id(user_id)
logger.info(f"Sender of the message is {user_name}")
# When a new user joins
if event.get("subtype") == 'channel_join':
if not channel_id in USER_LANGUAGE_MAPPINGS:
USER_LANGUAGE_MAPPINGS[channel_id] = {}
USER_LANGUAGE_MAPPINGS[channel_id].update({user_id: 'en'})
with open(f'{Path(__file__).parent}/user_language_mapping.json',
'w') as fp:
json.dump(USER_LANGUAGE_MAPPINGS, fp, sort_keys=True, indent=4)
try:
slack_web_client.chat_postEphemeral(
user=user_id,
attachments=language_selecter["attachments"],
channel=channel_id,
text=f"Hello {user_name}")
return
except SlackApiError:
logger.info("Failed to send message", exc_info=True)
# ENDS HERE
if 'files' in event:
files = event["files"]
audio_url = files[0].get("url_private_download")
user_preferred_language = USER_LANGUAGE_MAPPINGS.get(channel_id).get(
user_id)
file_name = get_audio_file(audio_url, user_id)
logger.info(f"File name of audio received is :{file_name}")
text = speech.speech_to_text(file_name, user_preferred_language)
logger.info(f"Text from audio {file_name} is : {text}")
if text:
text = text.replace("'", "\\'")
source_language = translate_api.detect_language(GOOGLE_TRANSLATE_API_KEY,
text)
logger.info(
f"Detected language of the message is {get_language_from_iso_code(source_language)}"
)
# When a message is posted, translate it for all the users in the channel
for target_user_id, target_language in USER_LANGUAGE_MAPPINGS.get(
channel_id).items():
if target_user_id != user_id:
logger.info(
f'Target user is {user_name} and target language is {get_language_from_iso_code(target_language)} '
)
start = time.time()
translated_text = get_translated_message(text, source_language,
target_language)
logger.info(
f'Translated text in {get_language_from_iso_code(target_language)} is {translated_text}. Completed in {time.time()-start} seconds'
)
try:
slack_web_client.chat_postEphemeral(
user=target_user_id,
channel=channel_id,
text=f"*{user_name}* :\n{translated_text} ")
except SlackApiError:
logger.error("Failed to send message", exc_info=True)
import functions
import yolopy
import speech
import cv2
import os
import detect
import datetime
os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = "dfkey.json"
labelsPath = "yolo/coco.names"
weightsPath = "yolo/yolov3.weights"
configPath = "yolo/yolov3.cfg"
args = {"threshold": 0.3, "confidence": 0.5}
project_id = "blindbot-4f356"
#project_id = "blindbot-286ed"
engine = speech.speech_to_text()
model = yolopy.yolo(labelsPath, weightsPath, configPath)
listening = False
intent = None
while True:
cam = cv2.VideoCapture(1)
if not listening:
resp = engine.recognize_speech_from_mic()
print(resp)
if (resp != None):
intent, text = detect.detect_intent_texts(project_id, 0, [resp],
'en')
if (intent == 'Jyoti' and resp != None):
listening = True