Problem with slots

Hi guys, I have a problem with my bot. Rasa version:

Rasa Version      :         3.1.0
Minimum Compatible Version: 3.0.0
Rasa SDK Version  :         3.1.1
Rasa X Version    :         1.1.0
Python Version    :         3.8.13

Now, I have 2 intents:

First intent:

- intent: what_is
  examples: |

    - What is [Conocchia]{"entity":"topic", "value":"conocchia"}?
    - What is a [cryptoporticus]{"entity":"topic", "value":"cryptoporticus"}?
    - Tell me about the [cryptoporticus]{"entity":"topic", "value":"cryptoporticus"} of ancient Capua.
AND OTHER QUOTES

Second intent:

- intent: monument_depictions
  examples: |
    - Are there any drawings of the [cryptoporticus]{"entity":"drawings", "value":"labruzzi_cryptoporticus"} in antiquity?
    - How was the [cryptoporticus]{"entity":"drawings", "value":"labruzzi_cryptoporticus"} of Carlo Labruzzi?
    - Tell me about the [cryptoporticus]{"entity":"drawings", "value":"labruzzi_cryptoporticus"} of Carlo Labruzzi.
AND OTHER QUOTES

The problem is that when I chat with the bot , for example saying it “What is a cryptoporticus” he goes in the right utter_action but the bot wants to save this :

Is the intent 'what_is' correct for 'what is a [cryptoporticus]{"entity": "topic", "value": "labruzzi_cryptoporticus"}?' and are
 all entities labeled correctly?  (Y/n)

and this is not correct because in domain.yml there isn’t “labruzzi_cryptoporticus” in the intent : topic

Can you help me ? Pratically I’d like that when the user writes a “what_is” INTENT with the word cryptoporticus → the bot has to save CRYPTOPORTICUS in the entity : TOPIC but when the user writes a “monument_depictions” INTENT with the word cryptoporticus → the bot has to save LABRUZZI_CRYPTOPORTICUS in the entity : DRAWINGS

Thanks !

hi @luisshinken - I see a couple of issues here. first I believe the training data might not be sufficient and apt. For example i see that you have similar phrases Tell me about the [cryptoporticus] under two different intents - what_is and monument_depiction.

i believe you will need to design your bot a bit differently to accomplish your job or give it a training data which has sufficient differences between these two intents.

for example -

  1. you ca just have one intent called monument_depiction.
  2. have a single entity called topic.
  3. once the entity is detected, you can have a follow question or menu that will let the user choose from - Drawings or What Is option.
  4. When the user selects one the relevant options, you can trigger a specific action to accomplish that option.

thank you so much utsuk, worked for me having 1 entity : Topic and changing some intent phrases!

1 Like

Happy that it solved your problemđź‘Ť