def main():
"""
Do initial checks, clear the console and print the assistant logo.
STEP 1: Clear log file in each assistant fresh start
STEP 2: Checks for internet connectivity
STEP 3: Configuare MongoDB, load skills and settings
STEP 4: Clear console
"""
# STEP 1
with open(ROOT_LOG_CONF['handlers']['file']['filename'], 'r+') as f:
f.truncate(0)
logging.info('Startup checks..')
# STEP 2
if not internet_connectivity_check():
logging.warning('Skills with internet connection will not work :-(')
time.sleep(3)
# STEP 3
configure_MongoDB(db, settings)
# STEP 4
console_manager = ConsoleManager()
logging.info('Application started')
console_manager.console_output()
processor = Processor(settings, db)
while True:
processor.run()
class TTTEngine:
"""
Text To Text Engine (TTT)
"""
def __init__(self):
self.logger = logging
self.console_manager = ConsoleManager()
def recognize_input(self):
self.logger.info("Waiting for user input.")
try:
text_transcript = input('>> ').lower()
while text_transcript == '':
self.logger.info("User didn't said something")
text_transcript = input('>> ').lower()
return text_transcript
except EOFError as e:
logging.error(e)
return ''
def assistant_response(self, message):
"""
Assistant response in voice or/and in text.
:param message: string
"""
try:
if message:
self.console_manager.console_output(message)
except RuntimeError as e:
self.logger.error(
'Error in assistant response with message {0}'.format(e))
def __init__(self):
super().__init__()
self.logger = logging
self.message_queue = queue.Queue(
maxsize=5
) # Maxsize is the size of the queue / capacity of messages
self.stop_speaking = False
self.console_manager = ConsoleManager()
class TTTEngine:
"""
Text To Text Engine (TTT)
"""
def __init__(self):
self.logger = logging
self.console_manager = ConsoleManager()
def recognize_input(self):
"""
Recognize input from console and returns transcript if its not empty string.
"""
try:
text_transcript = input('>> ').lower()
while text_transcript == '':
self.console_manager.console_output(info_log="User didn't said something")
text_transcript = input('>> ').lower()
return text_transcript
except EOFError as e:
self.console_manager.console_output(error_log='Failed to recognize user input with message: {0}'.format(e))
def assistant_response(self, message):
"""
Assistant response in voice or/and in text.
:param message: string
"""
try:
if message:
self.console_manager.console_output(message)
except RuntimeError as e:
self.console_manager.console_output(error_log='Error in assistant response with message: {0}'.format(e))
class STTEngine:
"""
Speech To Text Engine (STT)
Google API Speech recognition settings
SpeechRecognition API : https://pypi.org/project/SpeechRecognition/2.1.3
"""
def __init__(self):
super().__init__()
self.recognizer = sr.Recognizer()
self.microphone = sr.Microphone()
self.console_manager = ConsoleManager()
self.console_manager.console_output(info_log="Microphone configured")
def recognize_input(self, already_activated=False):
"""
Recognize input from mic and returns transcript if activation tag (assistant name) exist
"""
while True:
transcript = self._recognize_speech_from_mic()
if already_activated or self._activation_name_exist(transcript):
return transcript
def _recognize_speech_from_mic(self, ):
"""
Capture the words from the recorded audio (audio stream --> free text).
Transcribe speech from recorded from `microphone`.
"""
with self.microphone as source:
self.recognizer.adjust_for_ambient_noise(source)
audio = self.recognizer.listen(source)
try:
transcript = self.recognizer.recognize_google(audio).lower()
self.console_manager.console_output(info_log='User said: {0}'.format(transcript))
except sr.UnknownValueError:
# speech was unintelligible
transcript = ''
self.console_manager.console_output(info_log='Unable to recognize speech')
except sr.RequestError:
# API was unreachable or unresponsive
transcript = ''
self.console_manager.console_output(error_log='Google API was unreachable')
return transcript
@staticmethod
def _activation_name_exist(transcript):
"""
Identifies the assistant name in the input transcript.
"""
if transcript:
transcript_words = transcript.split()
return bool(set(transcript_words).intersection([jarvis.assistant_name]))
else:
return False
def __init__(self, settings_):
self.settings = settings_
self.input_engine = STTEngine(
pause_threshold=self.settings.SPEECH_RECOGNITION.get(
'pause_threshold'),
energy_theshold=self.settings.SPEECH_RECOGNITION.get(
'energy_threshold'),
ambient_duration=self.settings.SPEECH_RECOGNITION.get(
'ambient_duration'),
dynamic_energy_threshold=self.settings.SPEECH_RECOGNITION.get(
'dynamic_energy_threshold'),
sr=sr) if self.settings.GENERAL_SETTINGS.get(
'input_mode') == InputMode.VOICE.value else TTTEngine()
self.console_manager = ConsoleManager(
log_settings=self.settings.ROOT_LOG_CONF, )
self.output_engine = TTSEngine(
console_manager=self.console_manager,
speech_response_enabled=self.settings.GENERAL_SETTINGS.get(
'response_in_speech'))
self.response_creator = ResponseCreator()
self.skill_analyzer = SkillAnalyzer(
weight_measure=TfidfVectorizer,
similarity_measure=cosine_similarity,
args=self.settings.SKILL_ANALYZER.get('args'),
skills_=SKILLS,
sensitivity=self.settings.SKILL_ANALYZER.get('sensitivity'))
class AssistantSkill:
"""
This class is the parent of all skill classes.
"""
first_activation = True
console_manager = ConsoleManager()
engine = engines.TTSEngine() if GENERAL_SETTINGS.get(
'response_in_speech') else engines.TTTEngine()
@classmethod
def response(cls, text):
cls.engine.assistant_response(text)
@classmethod
def extract_tags(cls, voice_transcript, tags):
"""
This method identifies the tags from the user transcript for a specific skill.
e.x
Let's that the user says "hi jarvis!".
The skill analyzer will match it with enable_assistant skill which has tags 'hi, hello ..'
This method will identify the that the enabled word was the 'hi' not the hello.
:param voice_transcript: string
:param tags: string
:return: set
"""
try:
transcript_words = voice_transcript.split()
tags = tags.split(',')
return set(transcript_words).intersection(tags)
except Exception as e:
logging.error("Failed to extract tags with message {0}".format(e))
return set()
def __init__(self):
self.input_engine = STTEngine(
pause_threshold=SPEECH_RECOGNITION['pause_threshold'],
energy_theshold=SPEECH_RECOGNITION['energy_threshold'],
ambient_duration=SPEECH_RECOGNITION['ambient_duration'],
dynamic_energy_threshold=SPEECH_RECOGNITION[
'dynamic_energy_threshold'],
sr=sr) if GENERAL_SETTINGS['user_voice_input'] else TTTEngine()
self.console_manager = ConsoleManager(log_settings=ROOT_LOG_CONF, )
self.output_engine = TTSEngine(
console_manager=self.console_manager,
speech_response_enabled=GENERAL_SETTINGS['response_in_speech'])
self.response_creator = ResponseCreator()
self.skill_analyzer = SkillAnalyzer(
weight_measure=TfidfVectorizer,
similarity_measure=cosine_similarity,
args=ANALYZER['args'],
skills_=SKILLS,
sensitivity=ANALYZER['sensitivity'])
self.skill_controller = SkillController(
settings_=GENERAL_SETTINGS,
input_engine=self.input_engine,
analyzer=self.skill_analyzer,
control_skills=CONTROL_SKILLS,
)
class AssistantSkill:
"""
This class is the parent of all skill classes.
"""
first_activation = True
console_manager = ConsoleManager()
@classmethod
def console(cls, text):
"""
Prints the text only in console and update the console info.
"""
cls.console_manager.console_output(text)
@classmethod
def response(cls, text):
"""
The mode of the response depends on the output engine:
- TTT Engine: The response is only in text
- TTS Engine: The response is in voice and text
"""
jarvis.output_engine.assistant_response(text)
@classmethod
def extract_tags(cls, voice_transcript, tags):
"""
This method identifies the tags from the user transcript for a specific skill.
e.x
Let's that the user says "hi jarvis!".
The skill analyzer will match it with enable_assistant skill which has tags 'hi, hello ..'
This method will identify the that the enabled word was the 'hi' not the hello.
:param voice_transcript: string
:param tags: string
:return: set
"""
try:
transcript_words = voice_transcript.split()
tags = tags.split(',')
return set(transcript_words).intersection(tags)
except Exception as e:
cls.console_manager.console_output(
info_log='Failed to extract tags with message: {0}'.format(e))
return set()
class AssistantSkill:
first_activation = True
console_manager = ConsoleManager(
log_settings=ROOT_LOG_CONF,
)
tts_engine = TTSEngine(
console_manager=console_manager,
speech_response_enabled=GENERAL_SETTINGS['response_in_speech']
)
@classmethod
def response(cls, text):
cls.tts_engine.assistant_response(text)
@classmethod
def _extract_tags(cls, voice_transcript, skill_tags):
transcript_words = voice_transcript.split()
return set(transcript_words).intersection(skill_tags)
def __init__(self):
self.input_engine = SPEECH_ENGINES[SPEECH_RECOGNITION['recognizer']]()
self.console_manager = ConsoleManager(log_settings=ROOT_LOG_CONF, )
self.output_engine = TTSEngine(
console_manager=self.console_manager,
speech_response_enabled=GENERAL_SETTINGS['response_in_speech'])
self.response_creator = ResponseCreator()
self.skill_analyzer = SkillAnalyzer(
weight_measure=TfidfVectorizer,
similarity_measure=cosine_similarity,
args=ANALYZER['args'],
skills_=SKILLS,
sensitivity=ANALYZER['sensitivity'])
self.skill_controller = SkillController(
settings_=GENERAL_SETTINGS,
input_engine=self.input_engine,
analyzer=self.skill_analyzer,
control_skills=CONTROL_SKILLS,
)
def main():
"""
Do initial checks, clear the console and print the assistant logo.
"""
console_manager = ConsoleManager()
console_manager.console_output(info_log='Startup checks..')
internet_connectivity_check()
console_manager.console_output(info_log='Application started')
processor = Processor(console_manager=console_manager, settings_=settings)
while True:
processor.run()
def __init__(self):
self.logger = logging
self.console_manager = ConsoleManager()
# The above copyright notice and this permission notice shall be included in all
# copies or substantial portions of the Software.
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
# SOFTWARE.
import jarvis
from jarvis.core.console_manager import ConsoleManager
console_manager = ConsoleManager()
def validate_digits_input(message, values_range=None):
"""
Checks users input to be only numbers else it will be in infinite loop for a right value.
Extra parameter 'values_range' checks the input to be between a range.
"""
input_number = None
while True:
jarvis.output_engine.assistant_response(message)
user_input = jarvis.input_engine.recognize_input(
already_activated=True)
try:
input_number = int(user_input)
except ValueError:
class TTSEngine(TTS):
def __init__(self):
super().__init__()
self.logger = logging
self.message_queue = queue.Queue(
maxsize=5
) # Maxsize is the size of the queue / capacity of messages
self.stop_speaking = False
self.console_manager = ConsoleManager()
def assistant_response(self, message):
"""
Assistant response in voice.
:param message: string
"""
self._insert_into_message_queue(message)
try:
speech_tread = threading.Thread(target=self._speech_and_console)
speech_tread.start()
except RuntimeError as e:
self.logger.error(
'Error in assistant response thread with message {0}'.format(
e))
def _insert_into_message_queue(self, message):
try:
self.message_queue.put(message)
except Exception as e:
self.logger.error(
"Unable to insert message to queue with error message: {0}".
format(e))
def _speech_and_console(self):
"""
Speech method translate text batches to speech and print them in the console.
:param text: string (e.g 'tell me about google')
"""
try:
while not self.message_queue.empty():
cumulative_batch = ''
message = self.message_queue.get()
if message:
batches = self._create_text_batches(raw_text=message)
for batch in batches:
self.tts_engine.say(batch)
cumulative_batch += batch
self.console_manager.console_output(cumulative_batch)
self.run_engine()
if self.stop_speaking:
self.logger.debug('Speech interruption triggered')
self.stop_speaking = False
break
except Exception as e:
self.logger.error(
"Speech and console error message: {0}".format(e))
@staticmethod
def _create_text_batches(raw_text, number_of_words_per_batch=8):
"""
Splits the user speech message into batches and return a list with the split batches
:param raw_text: string
:param number_of_words_per_batch: int
:return: list
"""
raw_text = raw_text + ' '
list_of_batches = []
total_words = raw_text.count(' ')
letter_id = 0
for split in range(0, int(total_words / number_of_words_per_batch)):
batch = ''
words_count = 0
while words_count < number_of_words_per_batch:
batch += raw_text[letter_id]
if raw_text[letter_id] == ' ':
words_count += 1
letter_id += 1
list_of_batches.append(batch)
if letter_id < len(raw_text): # Add the rest of word in a batch
list_of_batches.append(raw_text[letter_id:])
return list_of_batches
def __init__(self):
super().__init__()
self.recognizer = sr.Recognizer()
self.microphone = sr.Microphone()
self.console_manager = ConsoleManager()
self.console_manager.console_output(info_log="Microphone configured")