How to separate JSON output for Chatterbot

Question

I'm creating a simple chatbot with Python and the chatterbot library. I've been able to get something able to work, where the user inputs something and the bot responds based on a .json of patterns and responses.
Every time I execute the bot is able to tell what "tag" the user input and responds accordingly only in this format: "patterns": ["cya", "See you later", "Goodbye", "I am Leaving", "Have a Good day"], But I would like it to remove the quotations, commas, etc and just pick one response at random. I've tried this tokenizing thing? Seems not to be working
Code

## MODULES
from chatterbot import ChatBot #This imports the chatbot
from chatterbot.trainers import ListTrainer #Method used for training the chatbot
import chatterbot #Just in case
import os

## BOT VARIABLES
bot = ChatBot("Test") #This creates the chatbot and names it 'Test'
trainer = ListTrainer(bot) #Creates the trainer
data = "C:/Users/PAVILION/Desktop/ChatBotTest2/INTS/"

for files in os.listdir(data):
    conv = open(data + files, 'r').readlines() #This opens the direrctory full of data
    trainer.train(conv) #This trains the chatbot trainer with data from the directory

## CHATTERBOT OUTPUT
while True:
    message = input("You: ")

    if message.strip() != "Bye":
        response = bot.get_response(message)
        print("ChatBot: ", response)

    if message.strip() == "Bye":
        print("ChatBot: Farewell")
        break

Intentions.json

{"intents": [
    {"tag": "greeting",
     "patterns": ["Hi", "How are you", "Is anyone there?", "Hello", "Good day", "Whats up"],
     "responses": ["Hello!", "Good to see you again!", "Hi there, how can I help?"],
     "context_set": ""
    },
    {"tag": "goodbye",
     "patterns": ["cya", "See you later", "Goodbye", "I am Leaving", "Have a Good day"],
     "responses": ["Sad to see you go :(", "Talk to you later", "Goodbye!"],
     "context_set": ""
    },
    {"tag": "age",
     "patterns": ["how old", "how old is tim", "what is your age", "how old are you", "age?"],
     "responses": ["I am 18 years old!", "18 years young!"],
     "context_set": ""
    },
    {"tag": "name",
     "patterns": ["what is your name", "what should I call you", "whats your name?"],
     "responses": ["You can call me Tim.", "I'm Tim!", "I'm Tim aka Tech With Tim."],
     "context_set": ""
    },
    {"tag": "shop",
     "patterns": ["Id like to buy something", "whats on the menu", "what do you reccommend?", "could i get something to eat"],
     "responses": ["We sell chocolate chip cookies for $2!", "Cookies are on the menu!"],
     "context_set": ""
    },
    {"tag": "hours",
     "patterns": ["when are you guys open", "what are your hours", "hours of operation"],
     "responses": ["We are open 7am-4pm Monday-Friday!"],
     "context_set": ""
    }
]
}

json was gotten from Tech with Tim tutorial btw.

elembie · Accepted Answer · 2019-09-20T05:11:51.657

1

You can load that json into a python object by removing the top level 'intents' (it's not needed) and loading the list into a variable;

intents = [

    {"tag": "greeting",
    "patterns": ["Hi", "How are you", "Is anyone there?", "Hello", "Good day", "Whats up"],
    "responses": ["Hello!", "Good to see you again!", "Hi there, how can I help?"],
    "context_set": ""
    },

    # rest of json pasted here
]

You can then access the dicts in the list by using next and a generator

selected_intent = next((i for i in intents if i['tag'] == 'greeting'), None)

Now selected_intent will be the intent dict from the list if the tag is greeting (in this example), but if no intents have that tag it will return None - so be sure to handle this.

You can now access any parts of the intent dict by their keys e.g. selected_intent['patterns'] - this would return the list of patterns for this tag. So you can do;

import random

patterns = selected_intent['patterns']
return_pattern = patterns[ random.randint(0, len(patterns)-1) ]

print(return_pattern) # output: whats up

This will generate a random int between 0 and the length of the list -1, which is used as an index to extract a random string from the list.

p.s. your if statements can be condensed with an else block because the second if is just the inverse of the first.

if message.strip() != "Bye":
    response = bot.get_response(message)
    print("ChatBot: ", response) 
else:
    print("ChatBot: Farewell")
    break

edited Sep 20 '19 at 05:11

answered Sep 20 '19 at 04:59

elembie

600
2
8

Your solution works, but how do I implement it into the chatbot? – Draxalot2 Sep 22 '19 at 21:36
And in the `i for i in in intents if i["tag"] == "greeting" ` how do I make the tag equal to all the tags instead of just greeting? – Draxalot2 Sep 22 '19 at 23:03
I'm not sure on the details of `chatterbot` - I was under the impression you already knew the `tag` and so needed to respond with a random `pattern`? If you get the `tag` from user input you can just load that into a variable e.g. `tag = input()` and then you can use `tag` in the generator `selected_intent = next((i for i in intents if i['tag'] == tag), None)`. How to implement the chat bot is beyond the scope of a single SO question - you might have to dig around in the docs :) – elembie Sep 23 '19 at 04:20
Yeah sorry, I found a way to implement it into the chatbot. For loading `input()` into a variable and using it like that, it doesn't seem to be working, anytime I try the terminal gets stuck on execution. Is there a way to loop through the file and get all the tags? – Draxalot2 Sep 23 '19 at 16:30
You can get a list of the tags using [list comprehensions](https://medium.com/better-programming/list-comprehension-in-python-8895a785550b) by `tags = [intent['tag'] for intent in intents]` – elembie Sep 23 '19 at 23:05

How to separate JSON output for Chatterbot

1 Answers1