Form Bot

Rasa Open Source
1

I am trying to use the form bot example given in RASA tutorial for my own bot. However, the bot is not able to extract some of the entities. For example, It is working fine for request_car, num_people, car_type etc but after giving a reply for the source, the bot is not responding properly.

Terminal Conversation with the bot

Your input → Hi
Hello! I am car search assistant! How can I help? Your input → need car
sorry, I didn’t understand you, please try input something else Your input → car please
which type of car? Your input → hatchback
how many people? Your input → 3
Pick up point? Your input → Bellandur
Your input →

I have uploaded the action.py, nlu.md, stories.md and domain.yml file here.

actions.py (3.7 KB) nlu.md (5.4 KB) stories.md (2.4 KB) domain.yml (2.2 KB)

Please suggest the changes.

2

@JiteshGaikwad @Juste please share your views.

3

Hi @avnish_002 - are you sure that the entity is being recognized by your NLU model?

You can check this by running rasa shell --debug and checking the output when you talk to the bot. Alternatively, for a place name, you could do a from_text slot mapping for the location and use an API to validate that it’s a valid place name. Since there is no context, just a single word, it’s tricky for the NER model to know that it’s a place name. So that’s my guess of what’s going wrong.