/chatbot_widget_angular

Antwerp UI Chatbot Widget UI (Angular)

Primary LanguageTypeScriptMIT LicenseMIT

Chatbot Smart Widget UI (Angular)

With the chatbot smart widget you can simply implement a chatbot in any page.

You will also need the BFF package in order to get the chatbot smart widget to work: http://github.com/TriangleJuice/chatbot_service_nodejs

Chatbot screenshot

There is a demo app, see below for instructions on running it.

How to use

Installing

> npm install @acpaas-ui-widgets/ngx-chatbot

Import the component in your module:

@NgModule({
  imports: [
    ...,
    ChatbotModule
  ],
  ...
})

In the index.html, include the core branding stylesheet:

<link rel="stylesheet" href="https://cdn.antwerpen.be/core_branding_scss/5.0.1/main.min.css">

For projects that are still using Angular 5, we are maintaining a v1 branch, which will still receive bug fixes if needed.

> npm install @acpaas-ui-widgets/ngx-chatbot@"<2.0.0"

In your template

<aui-chatbot
  url="http://localhost:3000/api/bff"
  session="123456789"
  title="My great chatbot widget"
  placeholder="Type your message here…"
  [pinned]="false"
  [delay]="200"
  [height]="400">
</aui-chatbot>

Supported attributes

url

string BFF URL.

session

string Required session ID for the chatbot engine, where each chat conversation is linked to its chat session. It's important to use a unique id, since all session data, like answers already given, are stored in the chatbot engine.

title

string Title above the chat window.

pinned

boolean Whether the chatbot is inline or pinned to the bottom of the application.

placeholder

string Placeholder string in the chat input field.

delay

number Delay between multiple messages received from the chatbot engine.

height

number Height of the chatbot in pixels. Can only be used when pinned="false".

width

number Width of the chatbot in pixels. Can only be used when pinned="true". Will not be smaller than 18rem.

avatar

string Avatar image URL. Default is the special provided Antwerp icon.

aria

ChatbotMessageAriaLabels Accessibility strings used throughout the chatbot UI:

Default:

{
  chatbot: 'Chatbot',
  close: 'Chatbot minimaliseren',
  avatar: 'Chatbot:',
  user: 'Jij:',
  message: 'Te verzenden bericht',
  send: 'Bericht verzenden',
  toggle: 'Een vraag stellen',
}

Events

actionStarted

Triggers when the user clicks a button of a message with type action. This type of message is injected by the BFF by injecting it into the response. The actionStarted output event requires a defined action and pauses the workings of the chatbot widget, waiting for a method call (completeAction) that has the same defined action. It allows the developer to handle some things outside the chatbot (e.g. a payment) before returning to the chatbot conversation.

How does it work?
<aui-chatbot
  #myChatbotWithAction
  [url]="'http://localhost:3000/api/chatbot'"
  [session]="session1"
  (actionStarted)="triggerMe($event)">
</aui-chatbot>
  1. Your BFF sends a message of type action with one or more actions defined.
{
  actions: [
    {
      action: 'someAction',
      text: 'Click me',
    }
  ]
}
  1. When the user clicks this button, the defined actionStarted output event will fire and the chatbot is disabled.
public triggerMe(event) {
  // Do something with `event`
}
  1. $event can now be handled in the parent controller and whenever finished, the chatbot widget's completeAction method can be triggered to re-enable the chatbot conversation. The result will be sent as a hidden message to the chatbot engine.

Important! The method property of completeAction has to contain the exact same action value that was first defined when disabling the chatbot, otherwise the chatbot will just ignore the method call.

const result = {
  action: 'someAction',
  message: 'success',
}
this.myChatbotWithAction.completeAction(result);

Run the demo app

> npm install
> npm start

Browse to localhost:4200

In order to use the chatbot widget demo app, you will also need to have started the corresponding back-end service.

Contributing

We welcome your bug reports and pull requests.

Please see our contribution guide.

Support

Jasper Van Proeyen (jasper.vanproeyen@digipolis.be)

License

This project is published under the MIT license.