Raise ValueError("event to log must be an instance " ValueError: event to log must be an instance of a subclass of Event

Rasa Open Source
1

Please tell in which city, Hubli or Dharwar? hubli {‘intent’: {‘name’: ‘confirm_specialization’, ‘confidence’: 0.40392767206040286}, ‘entities’: [], ‘intent_ranking’: [{‘name’: ‘confirm_specialization’, ‘confidence’: 0.40392767206040286}, {‘name’: ‘confirm_date’, ‘confidence’: 0.37037497733595404}, {‘name’: ‘greet’, ‘confidence’: 0.085348703255119174}, {‘name’: ‘confirm_city’, ‘confidence’: 0.085215702698709567}, {‘name’: ‘goodbye’, ‘confidence’: 0.023905681599622659}, {‘name’: ‘affirm’, ‘confidence’: 0.014307775582506817}, {‘name’: ‘bye’, ‘confidence’: 0.0082719908852013461}, {‘name’: ‘confirm_time’, ‘confidence’: 0.0055622524680621682}, {‘name’: ‘create_apt’, ‘confidence’: 0.0030852441144210963}], ‘text’: ‘hubli’} Traceback (most recent call last): File “train_rasa_nlu_model.py”, line 178, in responses = agent.handle_message(a) File “C:\Users\MS066115\AppData\Local\Continuum\anaconda3\lib\site-packages\rasa_core\agent.py”, line 114, in handle_message UserMessage(text_message, output_channel, sender_id)) File “C:\Users\MS066115\AppData\Local\Continuum\anaconda3\lib\site-packages\rasa_core\processor.py”, line 84, in handle_message self._predict_and_execute_next_action(message, tracker) File “C:\Users\MS066115\AppData\Local\Continuum\anaconda3\lib\site-packages\rasa_core\processor.py”, line 262, in _predict_and_execute_next_action dispatcher) File “C:\Users\MS066115\AppData\Local\Continuum\anaconda3\lib\site-packages\rasa_core\processor.py”, line 311, in _run_action self._log_action_on_tracker(tracker, action.name(), events) File “C:\Users\MS066115\AppData\Local\Continuum\anaconda3\lib\site-packages\rasa_core\processor.py”, line 369, in _log_action_on_tracker tracker.update(e) File “C:\Users\MS066115\AppData\Local\Continuum\anaconda3\lib\site-packages\rasa_core\trackers.py”, line 296, in update raise ValueError("event to log must be an instance " ValueError: event to log must be an instance of a subclass of Event.

(base) C:\Users\MS066115\Documents\chatbot\doctor_appoint>

################ Domain file

actions:

  • utter_greet
  • utter_bye
  • utter_ask_city
  • utter_ask_specialization
  • utter_ask_date
  • utter_ask_time
  • utter_doctor_details
  • actions.SaveCity
  • actions.SaveSpecialization
  • actions.SaveDate
  • actions.SaveTime
  • actions.ListOfDoctors
  • utter_unclear entities:
  • dr
  • date
  • city
  • time
  • specialist intents:
  • create_apt
  • confirm_city
  • confirm_specialization
  • confirm_date
  • confirm_dr
  • confirm_time
  • greet

  • goodbye slots: city: type: text date: type: text specialist: type: text time: type: text templates: utter_ask_city:

    text: “In which city would you like to book appointment Hubli or Dharwar?”

    text: “Please tell in which city, Hubli or Dharwar?” utter_ask_date:

    text: “When do you want to schedule this appointment?”

    text: “Ok, for what day?” utter_ask_specialization:

    text: “Which specialist are you looking for?”

    text: “Please mention your specialist or Specialization?” utter_ask_time:

    text: “please select the time?”

    text: “Ok, for what time?” utter_bye:
    • “text:It’s My pleasure to help you”
    • text: “Thak you” utter_doctor_details:
    • text: “Here i found similar results: Name,Specialization,Address,City,Phone number,Rating” utter_greet:
    • text: “Hello, how can I help you?”
    • text: “Hello, welcome to the Doctor Appointment Bot!” utter_unclear:
    • text: “I am not sure what you are aiming for.Kindly try it again”

####################################################################

code

from future import absolute_import from future import division from future import unicode_literals from rasa_nlu.training_data import load_data from rasa_nlu.model import Trainer from rasa_nlu import config from rasa_core.policies import FallbackPolicy, KerasPolicy, MemoizationPolicy from rasa_core.agent import Agent from rasa_core.actions import Action from rasa_core.events import SlotSet

from rasa_nlu.evaluate import run_evaluation

import requests

from rasa_core.domain import Domain #from rasa_core.trackers import EventVerbosity

import logging logger = logging.getLogger(name)

#import requests #import json

from rasa_core_sdk import Action from rasa_core_sdk.events import SlotSet from rasa_core_sdk.events import UserUtteranceReverted from rasa_core_sdk.events import AllSlotsReset from rasa_core_sdk.events import Restarted

nlu training data sample to be loaded

train_data = load_data(“data/data.json”)

#loading trainer to teach existing flow trainer = Trainer(config.load(“configuration.yml”))

model training

interpreter = trainer.train(train_data)

store model

model_directory = trainer.persist("./models/nlu", fixed_model_name=“doctornlu”)

“”"

here you can perform task of calling third party api if you integrate it for accessing thier data.

class ApiAction(Action): def name(self): return “action_retrieve_image”

def run(self, dispatcher, tracker, domain):
    dispatcher.utter_message("looking for a good img")
    url = "https://picsum.photos/200/300/?random"
    return [SlotSet("img_api_response", url)]
	"""

################################################################################################################ class SaveCity(Action): def name(self): return ‘action_save_city’ def run(self, dispatcher, tracker, domain): city_name = next(tracker.get_latest_entity_values(“city”), None) return [SlotSet(‘city’,city_name)]

class SaveSpecialization(Action): def name(self): return ‘action_save_specialization’

def run(self, dispatcher, tracker, domain):
	specialization_in = next(tracker.get_latest_entity_values("specialist"), None)
	
	return [SlotSet('specialist',specialization_in)]

class SaveDate(Action): def name(self): return ‘action_save_date’

def run(self, dispatcher, tracker, domain):
	dates = next(tracker.get_latest_entity_values("date"), None)
	
	return [SlotSet('date',dates)]

class SaveTime(Action): def name(self): return ‘action_save_time’

def run(self, dispatcher, tracker, domain):
	timing= next(tracker.get_latest_entity_values("time"), None)
	
	return [SlotSet('time',timing)]

#import re import pandas as pd

class ListOfDoctors(Action):

def __init__(self):
	#with open("C:/Users/MS066115/Documents/chatbot/doctor_appoint/data/dlist.csv",'rb') as f:
	#	self.Doctors_data = f.read()
	self.Doctors_data = pd.read_csv("C:/Users/MS066115/Documents/chatbot/doctor_appoint/data/dlist.csv",encoding='ISO-8859-1')

def name(self):
	return 'action_list_doctors'
	
def run(self, dispatcher, tracker, domain):
	city_name=tracker.get_slot('city')
	specialization_in=tracker.get_slot('specialist')
	#dates=tracker.get_slot('date')
	timing=tracker.get_slot('time')
	
	## create objects of the each class and access each the functions 
	#SaveCityObj = SaveCity()
	#SaveSpecializationObj = SaveSpecialization()
	#SaveTimeObj= SaveTime()
	#cityx = SaveCityObj.run()
	#data1= data[(data['time'].str.contains(y) == True) & (data['specialization'].str.contains(spec) == True) & (data['location'].str.contains(loc) == True)]
	filterd_data=self.Doctors_data[(self.Doctors_data['City'].str.contains(city_name)==True) & (self.Doctors_data['Specialization'].str.contains(specialization_in)==True)&(self.Doctors_data['Time'].str.contains(timing)==True)]

	if filterd_data is None: 
		response = "Sorry I couldn't find anything relevant to your query!"
	else:
		res_name=filterd_data['Name'].head(5).to_string(index=False)
		res_specialization=filterd_data['Specialization'].head(5).to_string(index=False)
		res_address=filterd_data['Address'].head(5).to_string(index=False)
		res_city=filterd_data['Location'].head(5).to_string(index=False)
		res_phonenum=filterd_data['Phone number'].head(5).to_string(index=False)
		res_timing=filterd_data['Time'].head(5).to_string(index=False)
		res_rating=filterd_data['Rating'].head(5).to_string(index=False)
		response = "Here's something I found, \n\n res_name:\n{} \t res_specialization: {} \t res_address: {} \t res_city: {} \t res_phonenum: {} \t res_timing{} \t res_rating{}\n".format(res_name, res_specialization, res_address, res_city,res_phonenum,res_timing,res_rating)
	dispatcher.utter_message(response)

"""
	res_name=self.Doctors_data['Name']
	res_specialization=self.Doctors_data['Specialization']
	res_address=self.Doctors_data['Address']
	res_city=self.Doctors_data['City']
	res_phonenum=self.Doctors_data['Phone number']
	res_timing=self.Doctors_data['Time']
	res_rating=self.Doctors_data['Rating']
	response = "Here's something I found, \n\n res_name: {} \n res_specialization: {} \n res_address: {} \n res_city: {} \n res_phonenum: {} 
	\n res_timing{} \n res_rating{}\n".format(res_name, res_specialization, res_address, res_city,res_phonenum,res_timing,res_rating)

else:
	response = "Sorry I couldn't find anything relevant to your query!"
	"""

#############################################################################################################################

this will catch predictions the model isn’t very certain about

there is a threshold for the NLU predictions as well as the action predictions

fallback = FallbackPolicy(fallback_action_name=“utter_unclear”, core_threshold=0.1, # this core threshold limits the fallback_Action you can change it. nlu_threshold=0.15) # nlu threshold help model to choose reply based on identification of intent and only responses above the specified range.

agent = Agent(‘doctor_domain.yml’, policies=[MemoizationPolicy(), KerasPolicy(), fallback])

loading our neatly defined training dialogues

training_data = agent.load_data(‘data/stories.md’)

agent.train(training_data,validation_split=0.1,epochs=800, batch_size=1)

agent.persist(‘models/dialogue’)

agent = Agent.load(‘models/dialogue’, interpreter=model_directory)

print(“Your bot is ready to talk! Type your messages here or send ‘stop’”) while True: a = input() print(interpreter.parse(a)) if a == ‘stop’: break responses = agent.handle_message(a) for response in responses: print(response[“text”])

run_evaluation(“nlu.json”, model_directory)