Adding a chatbot interface to Mrl.

This simple html page works to chat with rasa bot server on localhost....

 

<head>
  <link rel="stylesheet" href="https://npm-scalableminds.s3.eu-central-1.amazonaws.com/@scalableminds/chatroom@master/dist/Chatroom.css" />
</head>
<body>
  <div class="chat-container"></div>
 
  <script src="https://npm-scalableminds.s3.eu-central-1.amazonaws.com/@scalableminds/chatroom@master/dist/Chatroom.js"/></script>
  <script type="text/javascript">
    var chatroom = new window.Chatroom({
      host: "http://localhost:5005",
      title: "Chat with Lola",
      container: document.querySelector(".chat-container"),
      welcomeMessage: "Hi, I am Lola. How may I help you?",
      speechRecognition: "en-US",
      voiceLang: "en-US"
    });
    chatroom.openChat();
  </script>
</body>
 
 
What do i need to do to add something similar to MRL, to chat with the bot on localhost?
GroG's picture

Hi Shaun, what your showing

Hi Shaun,
what your showing is a very minimal amount of UI code .. the "real" working code exists here - https://npm-scalableminds.s3.eu-central-1.amazonaws.com/@scalableminds/chatroom@master/dist/Chatroom.js, and as expected its minified ... meaning its incomprehenisble in its current form.

I tried a quick test and loaded your example, but it noWorky for me  ... not really surprised if its a paid or tokenized cloud service.

https://rasa.com/ looked possibly promising ... they say "open source" alot ... good

After digging a little more I saw they had a pypi pip3 install so I tried it on my Ubuntu laptop

ITS A BIG ONE .... TENSORFLOW IS HUGE ... but I knew that ...

dang .. they are downloading the internet ... reminds me of myrobotlab.jar --install :)

So, after it downloaded the internet I started it ... very nice command line initializaion, it trained a model and ran !   WHOOHOO !

The bot is not very smart, but I expected that - it had a few responses.

But the important part is that it WORKED !

Alright ... I saw an HTTP API reference somewhere ...

Here - https://rasa.com/docs/rasa/user-guide/configuring-http-api/

Weird .. its very cloud-centric ... so is that how they make their income ?  Setting up an easy cloud deployment system...  hmmm interesting..

ah .. that was "configuring" the http-api

This is docs on the actual api ?

https://rasa.com/docs/rasa/api/http-api/

WOOHOO .. finally ... 

rasa run

Running locally, getting closer ...

Finally ! ... Success !

curl -XPOST \
    --header 'content-type: application/json' \
    --data '{"message": "Hi Bot"}' \
    http://localhost:5005/webhooks/rest/webhook

 

That appears to be the root of what is needed to be sent from a remote system... Pretty awesome ..
I think we could bang a service out quickly for this.