Rasa intent server with entities extraction

Rasa is just AI bot which does all the hard work. Integration once you have Rasa running is very simple.

With this workflow by extending your training data you could do a

A bot which you can ask for a stock price
A bot which you ask for a next match for the games

We will need few things

Running Rasa service. https://rasa.com/docs/rasa/installation
Rest API configuration in Live Helper Chat
Bot configuration in LHC

For more information please read

https://medium.com/better-programming/chatbots-and-whats-new-in-rasa-2-0-a51c61ca3c33
https://github.com/RasaHQ their main github repository where you can find also few more sample bots
https://towardsdatascience.com/a-beginners-guide-to-rasa-nlu-for-intent-classification-and-named-entity-recognition-a4f0f76b2a96

Install instructions for docker version

git clone https://github.com/LiveHelperChat/rasa-intent-entities.git && cd rasa-intent-entities

Now you can edit data/nlu.yml and write your model data in this file. As example this file contains few examples about user requesting specific stock price

Build docker image

docker-compose build

Run one time

docker-compose up

Run as a service

docker-compose up -d

You can try out Rasa rest API using curl commands

curl -i http://localhost:5005

# Send demo request
curl localhost:5005/model/parse -d '{"text":"how much does the apple cost?"}'

Example of JSON response.

{
  "text": "how much does the apple cost?",
  "intent": {
    "id": -5100266420140976000,
    "name": "stock_price",
    "confidence": 0.9999998807907104
  },
  "entities": [
    {
      "entity": "stock",
      "start": 18,
      "end": 23,
      "confidence_entity": 0.999945878982544,
      "value": "apple",
      "extractor": "DIETClassifier"
    }
  ],
  "intent_ranking": [
    {
      "id": -5100266420140976000,
      "name": "stock_price",
      "confidence": 0.9999998807907104
    },
    {
      "id": -3041718864255617000,
      "name": "faq",
      "confidence": 1.3706980439565086e-7
    }
  ],
  "response_selector": {
    "all_retrieval_intents": [],
    "out_of_scope": {
      "response": {
        "id": null,
        "response_templates": null,
        "confidence": 0,
        "intent_response_key": null,
        "template_name": "utter_None"
      },
      "ranking": []
    },
    "faq": {
      "response": {
        "id": -2384809603512026000,
        "response_templates": [
          {
            "text": "faq/ask_location"
          }
        ],
        "confidence": 0.9957873821258545,
        "intent_response_key": "faq/ask_location",
        "template_name": "utter_faq/ask_location"
      },
      "ranking": [
        {
          "id": -2384809603512026000,
          "confidence": 0.9957873821258545,
          "intent_response_key": "faq/ask_location"
        },
        {
          "id": -7675572599941173000,
          "confidence": 0.004212635103613138,
          "intent_response_key": "faq/ask_gender"
        }
      ]
    },
    "chitchat": {
      "response": {
        "id": null,
        "response_templates": null,
        "confidence": 0,
        "intent_response_key": null,
        "template_name": "utter_None"
      },
      "ranking": []
    }
  }
}

As you see we have intent stock_price and extracted entity stock. Now the fun part to implement that in Rest API configuration.

Configuring Rest API in Live Helper Chat

Create a new Rest API by navigating to

System configuration > Live help configuration > Rest API Calls

Just create a new. Configuration looks like this

We set body request as JSON and set content.

We also set Outpout parsing

Now just save.

Bot configuration in Live Helper Chat

For bot configuration we only need four triggers

Default it has checked Default, Default for unknown message
Inten parser searches for a messages with intent
nlu_fallback - this message we will send if Rasa did not returned anything or returned not what we expected.
stock_price - this trigger will be executed once extracted entity was found.

Default trigger configuration

Inten parser - searches our bot for an action

It's possible just to execute directly response what you want to do with stock price. So there is many ways you can do same way. Our way has advantage if you define let say goodbye intent in Rasa you just need define keyword to search for goodbye.