How to use Multiprojectimporter in Deployment?

Rasa Open Source
1

Hi,

I use the multiprojectimporter in my project which works just fine locally. It is also trained via rasa train (so no fancy training argument is required). But when I want to deploy it the training data is not found. Is there anything I can change in my docker-compose.yml file to make sure that it is found?

That is the message I get:

No training data given. Please provide stories and NLU data in order to train a Rasa model using the '--data' argument

And this is my current command:

command: bash -c “rm -rf models/* && rasa train && rasa run --enable-api --cors "*" --debug”

That is my docker-compose.yml:

    volumes:
    - ./actions:/app/actions
    - ./data:/app/data

while I have this in my config.yml:

importers:
- name: MultiProjectImporter
imports:
- projects/Baseline

Many thanks in advance!

2

Hi, can you provide your complete Dockerfile?

3

hi, sorry if I’m a bit off-topic but I’m trying to deploy my rasa in a docker container and the tutorial in the documentation is not rally helpful could you help?

4

Hi @jean-on-hub, please take a look at this thread Deploy my chatbot using docker

1 Like