Connect Rasa to your chatbot

Use on-premise Rasa NLP in CSML chatbots.

In order to use this App, you will need to have an existing Rasa server with the --enable-api tag.
Please refer to this documentation.

AUTH_TOKEN and JWT_SECRET are both optional environment variables you can use to secure the requests.

Methods

analyze

Predict the intent and entities of the message posted to this endpoint.

do intent_entities = App(
  "rasa",
  method = "analyze"
  params = {
    "text": "Hello",
  }
)

Response:

{
  "success": true,
  "response": {
    "text": "Hello",
    "intent": {
      "id": -1915024040213881600,
      "name": "greet",
      "confidence": 0.9999943375587463
    },
    "entities": [],
    "intent_ranking": [
      {
        "id": "-1915024040213881600",
        "name": "greet",
        "confidence": "0.9999943375587463"
      },
      {
        "id": "-7222733869911494000",
        "name": "affirm'",
        "confidence": "0.0000022538461053045467"
      },
      ...
    ],
    "response_selector": {
      "all_retrieval_intents": [],
      "default": {
        "response": {
          "id": null,
          "response_templates": null,
          "confidence": 0,
          "intent_response_key": null,
          "template_name": "utter_None"
        },
        "ranking": []
      }
    }
  }
}