How to get the predicted intent with confidence using agent rasa sdk chatbot

Question

I am loading the trained rasa models manually by using this

agent = Agent.load(
                model,
                action_endpoint=EndpointConfig(ACTION_ENDPOINT)
            )

And i am predicting the result like this

botResponse = await agent.handle_text(query)

but this just returns the response as text, but i need the confidence and intent name as well

I tried the handle_message but still it does not give confidence.

Answer 1

you can retrieve this information from the tracker_store instance of the Agent. To do so, firstly make sure that you pass a sender id while calling agent.handle_text(query, sender_id="some sender id"). Then retrieve the tracker with:

current_tracker = agent.get_or_create_tracker(sender_id="some sender id")

Once you have the tracker, you can retrieve the NLU parsed data of the last sent message with :

user_event = tracker.get_last_event_for(UserUttered)
if user_event:
    nlu_parse_data = user_event.parse_data

nlu_parse_data should look something like this:

"text": "Hi MoodBot.",
        "parse_data": {
          "intent": {
            "id": 3068390702409455462,
            "name": "greet",
            "confidence": 0.9968197345733643
          },
          "entities": [],
          "text": "Hi MoodBot.",
          "message_id": "47efa155fc234abea554242883f0a74e",
          "metadata": {},
          "intent_ranking": [
            {
              "id": 3068390702409455462,
              "name": "greet",
              "confidence": 0.9968197345733643
            },
            {
              "id": -7997748339392136471,
              "name": "bot_challenge",
              "confidence": 0.0019184695556759834
            },
            {
              "id": -3856210704443307570,
              "name": "mood_unhappy",
              "confidence": 0.0010514792520552874
            },

Answer 2

I used two different apis of rasa to achieve the same

To get the intent and confidence of the query parse_message_using_nlu_interpreter and to get the response handle_text

queryResponseList = await agent.handle_text(query)
intentInfo = await agent.parse_message_using_nlu_interpreter(query)
intent = Intent(**{
            "name": intentInfo["intent"]["name"],
            "confidence": intentInfo["intent"]["confidence"]
        })