Hi,
In RASA there is the possibility to read the intension of a table or file, because I want to create a CRUD for the intensions and responses?
Can you give more info on what you’re trying to achieve?
You want to store your NLU in a database instead of YAML files?
I would like to store the questions and answers in a database. Thinking about the scenario: me as a developer, my client (person who bought the project) who should have their users.
Create a solution where the customer would store the questions and answers in a database. When a user sends a question via chatbot, through an action the question would be checked in the table and the registered answer would be returned.
The approach you are trying to use is rule based dialogue management. In such case, you dont need rasa nlu at all. You can write a python script which fetches the corresponding answer based on some keyword spotting/ string matching of questions.
Or if rasa nlu is required, train intent classes for each question, with examples being several variants of the same question the user may ask. For example -
intent -ask_name
Examples -what is your name ?
can you tell your name ?
your name please etc
then map the intent identified to replies in the table. Use rasa nlu only server to fetch the intent in custom python script.
1 Like
You should either create a script that converts the database to valid NLU, domain, and stories/rules files, or write a Custom Importer.
hi sandeeppandey456,
yes in the approach I intend to have both situations. I don’t know how to start (fire) this python script without using NLU.
Can someone help me?
I have already given you the answer here.
Basically, you can’t give any input to Rasa and expect it to understand. You should adapt to the software you are using, not the opposite.
You should find a way to transform your data into Rasa format. Did you read the links I have mentioned?
Hey bro, could you solve it?,