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'))
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, )
def __init__(self, console_manager, settings_): self.console_manager = console_manager self.settings = settings_ self.response_creator = ResponseCreator() self.skill_analyzer = SkillAnalyzer( weight_measure=TfidfVectorizer, similarity_measure=cosine_similarity, args=self.settings.SKILL_ANALYZER.get('args'), sensitivity=self.settings.SKILL_ANALYZER.get('sensitivity'), )
class Processor: 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, ) def run(self): start_up() while True: self.skill_controller.wake_up_check() if self.skill_controller.is_assistant_enabled: # Check if the assistant is waked up self._process() def _process(self): self.skill_controller.get_transcript() self.skill_controller.get_skills() if self.skill_controller.to_execute: response = self.response_creator.create_positive_response( self.skill_controller.latest_voice_transcript) else: response = self.response_creator.create_negative_response( self.skill_controller.latest_voice_transcript) self.output_engine.assistant_response(response) self.skill_controller.execute()
class Processor: 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 run(self): start_up() keyboard.add_hotkey(GENERAL_SETTINGS['wake_up_hotkey'], self._process) keyboard.wait() def _process(self): print('Assistant has woken up') self.skill_controller.get_transcript() self.skill_controller.get_skills() if self.skill_controller.to_execute: response = self.response_creator.create_positive_response( self.skill_controller.latest_voice_transcript) else: response = self.response_creator.create_negative_response( self.skill_controller.latest_voice_transcript) self.output_engine.assistant_response(response) self.skill_controller.execute()
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 __init__(self, settings_, db): self.settings = settings_ self.db = db self.input_engine = engines.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 db.get_documents(collection='general_settings')[0]['input_mode'] == InputMode.VOICE.value \ else engines.TTTEngine() self.output_engine = engines.TTSEngine() if db.get_documents(collection='general_settings')[0]['response_in_speech'] \ else engines.TTTEngine() self.response_creator = ResponseCreator() self.skill_analyzer = SkillAnalyzer( weight_measure=TfidfVectorizer, similarity_measure=cosine_similarity, args=self.settings.SKILL_ANALYZER.get('args'), sensitivity=self.settings.SKILL_ANALYZER.get('sensitivity'), db=self.db)
class Processor: def __init__(self, console_manager, settings_): self.console_manager = console_manager self.settings = settings_ self.response_creator = ResponseCreator() self.skill_analyzer = SkillAnalyzer( weight_measure=TfidfVectorizer, similarity_measure=cosine_similarity, args=self.settings.SKILL_ANALYZER.get('args'), sensitivity=self.settings.SKILL_ANALYZER.get('sensitivity'), ) def run(self): """ Assistant starting point. - STEP 1: Get user input based on the input mode (voice or text) - STEP 2: Matches the input with a skill - STEP 3: Create a response - STEP 4: Execute matched skill - STEP 5: Insert user transcript and response in history collection (in MongoDB) """ # STEP 1 transcript = jarvis.input_engine.recognize_input() # STEP 2 skill_to_execute = self._extract_skill(transcript) # STEP 3 response = self.response_creator.create_positive_response(transcript) if skill_to_execute \ else self.response_creator.create_negative_response(transcript) jarvis.output_engine.assistant_response(response) # STEP 4 self._execute_skill(skill_to_execute) # STEP 5 record = { 'user_transcript': transcript, 'response': response if response else '--', 'executed_skill': skill_to_execute if skill_to_execute else '--' } db.insert_many_documents('history', [record]) def _extract_skill(self, transcript): skill = self.skill_analyzer.extract(transcript) if skill: return {'voice_transcript': transcript, 'skill': skill} @staticmethod def _execute_skill(skill): if skill: ActivationSkills.enable_assistant() try: logging.debug('Executing skill {0}'.format( skill.get('skill').get('name'))) skill_func_name = skill.get('skill').get('func') skill_func = skill_objects[skill_func_name] skill_func(**skill) except Exception as e: logging.debug( "Error with the execution of skill with message {0}". format(e))
class Processor: 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')) def run(self): self._traped_until_assistant_is_enabled() transcript = self.input_engine.recognize_input() skill_to_execute = self._extract_skill(transcript) response = self.response_creator.create_positive_response(transcript) if skill_to_execute \ else self.response_creator.create_negative_response(transcript) self.output_engine.assistant_response(response) self._execute_skill(skill_to_execute) def _execute_skill(self, skill): if skill: try: skill_method = skill.get('skill').get('skill') logging.debug('Executing skill {0}'.format(skill)) skill_method(**skill) except Exception as e: logging.debug( "Error with the execution of skill with message {0}". format(e)) def _traped_until_assistant_is_enabled(self): if self.settings.GENERAL_SETTINGS.get( 'input_mode') == InputMode.VOICE.value: while not ExecutionState.is_ready_to_execute(): voice_transcript = self.input_engine.recognize_input() transcript_words = voice_transcript.split() enable_tag = set(transcript_words).intersection( CONTROL_SKILLS.get('enable_assistant').get('tags')) if bool(enable_tag): CONTROL_SKILLS.get('enable_assistant').get('skill')() ExecutionState.update() def _extract_skill(self, transcript): skill = self.skill_analyzer.extract(transcript) if skill: return {'voice_transcript': transcript, 'skill': skill}
class Processor: def __init__(self, settings_, db): self.settings = settings_ self.db = db self.input_engine = engines.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 db.get_documents(collection='general_settings')[0]['input_mode'] == InputMode.VOICE.value \ else engines.TTTEngine() self.output_engine = engines.TTSEngine() if db.get_documents(collection='general_settings')[0]['response_in_speech'] \ else engines.TTTEngine() self.response_creator = ResponseCreator() self.skill_analyzer = SkillAnalyzer( weight_measure=TfidfVectorizer, similarity_measure=cosine_similarity, args=self.settings.SKILL_ANALYZER.get('args'), sensitivity=self.settings.SKILL_ANALYZER.get('sensitivity'), db=self.db) def run(self): """ This method is the assistant starting point. - STEP 1: Waiting for enable keyword (ONLY in 'voice' mode) - STEP 2: Retrieve input (voice or text) - STEP 3: Matches the input with a skill - STEP 4: Create a response - STEP 5: Execute matched skill - STEP 6: Insert user transcript and response in history collection (in MongoDB) """ # STEP 1 self._trapped_until_assistant_is_enabled() # STEP 2 transcript = self.input_engine.recognize_input() # STEP 3 skill_to_execute = self._extract_skill(transcript) # STEP 4 response = self.response_creator.create_positive_response(transcript) if skill_to_execute \ else self.response_creator.create_negative_response(transcript) self.output_engine.assistant_response(response) # STEP 5 self._execute_skill(skill_to_execute) # STEP 6 record = { 'user_transcript': transcript, 'response': response if response else '--', 'executed_skill': skill_to_execute if skill_to_execute else '--' } self.db.insert_many_documents('history', [record]) def _trapped_until_assistant_is_enabled(self): """ In voice mode assistant waiting to hear an enable keyword to start, until then is trapped in a loop. """ if self.db.get_documents(collection='general_settings' )[0]['input_mode'] == InputMode.VOICE.value: while not ExecutionState.is_ready_to_execute(): voice_transcript = self.input_engine.recognize_input() transcript_words = voice_transcript.split() enable_skills = self.db.get_documents( 'control_skills', {'name': 'enable_assistant'}) enable_tags = [skill.get('tags') for skill in enable_skills] enable_tag = set(transcript_words).intersection(enable_tags) if bool(enable_tag): skill_name = self.db.get_documents( 'control_skills', { 'name': 'enable_assistant' }).get('skills') skill_object = skill_objects[skill_name] skill_object() ExecutionState.update() def _extract_skill(self, transcript): skill = self.skill_analyzer.extract(transcript) if skill: return {'voice_transcript': transcript, 'skill': skill} @staticmethod def _execute_skill(skill): if skill: try: logging.debug('Executing skill {0}'.format( skill.get('skill').get('name'))) skill_func_name = skill.get('skill').get('func') skill_func = skill_objects[skill_func_name] skill_func(**skill) except Exception as e: logging.debug( "Error with the execution of skill with message {0}". format(e))