Generate training datasets for chatbots in a breeze!

Chatito is a natural language generation (NLG) tool and a domain specific language (DSL) for creating chatbot training datasets, with first class suport for RasaNLU.

Slot filling chatbots are a type of chatbots that link short text sentences to actions and extract the action arguments from the sentence.

Recommended videos to watch:

Slot-Filling in Conversations with Deep Learning

Open-source language understanding for bots by RASA author

Given a sentence, we link an (action/intent) to it and map some of the sentece words to arguments or entities that are useful for the action. e.g.:

Sentence -> Hey Bot turn the lights off

Parsed Intent -> {
    id: "Hey Bot turn the lights off",
    action: "lightChange",
    arg: { switch: "off" }
}

RASA_nlu is a tool for training intent and entity extraction models in a breeze. Chatito generates training datasets compatible with the RASA format.

Test it online at https://rodrigopivi.github.io/Chatito/

• npm i chatito --save
• create a definition file. e.g.: trainClimateBot.chatito with your DSL definitions.
• npx chatito trainClimateBot.chatito
• The training set should be available at trainClimateBot.json

The json file contains all the possible combination sentences. Each traning example is an object that contains sentence, action and arguments.

You can also use it programmatically:

const chatito = require("chatito");
const dataset = chatito.datasetFromString(dslDefinitionString);

A simple example si that given this DSL definition:

%[lightChange]
    lights @[switch]

@[switch]
    ~[on]
    ~[off]

~[on]
    on
    active

~[off]
    off
    inactive

It will generate this dataset:

[
  {
    "text": "lights on",
    "intent": "lightChange",
    "entities": [
      {
        "start": 7,
        "end": 9,
        "value": "on",
        "entity": "switch"
      }
    ]
  },
  {
    "text": "lights active",
    "intent": "lightChange",
    "entities": [
      {
        "start": 7,
        "end": 13,
        "value": "on",
        "entity": "switch"
      }
    ]
  },
  {
    "text": "lights off",
    "intent": "lightChange",
    "entities": [
      {
        "start": 7,
        "end": 10,
        "value": "off",
        "entity": "switch"
      }
    ]
  },
  {
    "text": "lights inactive",
    "intent": "lightChange",
    "entities": [
      {
        "start": 7,
        "end": 15,
        "value": "off",
        "entity": "switch"
      }
    ]
  }
]

A Chatito domain specific language file should contain the '.chatito' extension, and is just a text file the operators definitions.

Important Note: The DSL uses 4 space identation to define the code block scope nesting.

Operators are the way to declare keywords with special behaviors. An operator is a token that starts with an operator symbol (% or @ or ~), followed by an opening squared bracket [, a cammel cased word with no spaces and a closing squared bracket ]. e.g.: %[someAction], @[someArgument], ~[someAlias]

When using operators inside a sentence, operators can be made optional by adding a ? symbol after the closing squared bracket. E.g.: ~[hi]?

Here is the full list of operators:

The action operator is how we link a sentence with an actual bot command. Each action is an entry point for the generator. Actions cannot be nested, but can contain other operators (arguments and aliases). We can think about it, as if each action maps to a function call. E.g.:

%[turnOnLights]
    turn on the lights
    lights on
    its too dark
    please lights on

Then, everytime we receive a sentence that links to turnOnLights action. We can call a function, if the function we want to call needs arguments, we can use the argument operator to define them.

Encapsulates a set of keywords as possible values for an input variable. Argument definitions can only include alias operators or words.

Important note: When an argument only defines an alias, like in this example @city > ~[newYork], the argument value variations will map to the alias id. (e.g.: all new york variations inside city, will return newYork as value).

%[timeAtCity]
    what time is at @[city]

@[city]
    miami
    ~[newYork]
    ~[losAngeles]

~[newYork]
    new york
    ny
    ny city
    new york city

~[losAngeles]
    los angeles
    la
    los angeles city
    la city

A word or list of words that are equivalent. e.g.:

%[greet]
    ~[hi]

~[hi]
    hi
    hello
    hi there
    hey

%[findByCityAndCategory]
    ~[greet]? ~[botName]? ~[please]? ~[find]? ~[restaurants]? ~[nearby] @[city]

~[greet]
    hey
    hi
    hello
    greetings

~[botName]
    Fred
    Fredy
    Frederick

~[please]
    please
    pls

~[find]
    find
    search
    lookup

~[nearby]
    close to
    in the area of
    within
    located at
    nearby

~[restaurants]
    restaurants
    places to eat
    where to eat

~[newYork]
    new york
    ny
    new york city
    ny city

~[sanFrancisco]
    san francisco
    san francisco city

~[atlanta]
    atlanta
    atlanta city

@[city]
    ~[newYork]
    ~[sanFrancisco]
    ~[atlanta]

• nalgene - Similar tool.
• PEG.js - Simple and powerfull parser generator.
• RASA_NLU - Great NLU tool for intent and entity model training.