Rasa X in production (docker) - loading training data from file

I am setting up rasa x on a linux server in docker following the docs. Is is possible to load my existing training data and configurations (data/nlu.md, config.yml, etc.) in rasa x?

Hi @BoyanH, yes, so long as you are in your project directory and it is set up like a normal rasa project (so far with data/nlu.md and config.yml, sounds like yours is fine), doing rasa x in this repository should load your existing data and configurations.

What about if rasa-x is run using docker? how can this behaviour be replicated? thanks


As @aherrera said, in docker (or in production) things seem to be quite different. As soon as one adds the --production flag, none of nlu.md, domain.yml, etc. seem to be loaded. Further one can’t load domain.yml data by any means within the GUI or with HTTP requests, as far as I know, but this is probably off-topic.

I am having the same problem.

Sorry, I was mistaken, if you serve rasa x in docker, you should upload your models and data via the HTTP API. Currently, the domain is populated by the uploaded model, but in a future release there will be endpoints for uploading a domain separately, too.

@erohmensing , I’m running RasaX in docker container and production mode. The RasaX is NOT loading NLU\Stories data from my files automatically when the container starts. We are following the default structure data\nlu.md and data\stories.md.

It requires manual intervention to upload data from RasaX UI. I believe uploading data through HTTP API or RasaX UI will be same and requires manual intervention.

Is there anyway to load data from files automatically while starting the docker container for RasaX? Please note we cannot use git repository to load data. Uploading data manually is not an option for real time production environments.

Hope rasateam can suggest an alternative approach or advise if there is any existing feature that we can leverage.

Thank you.

1 Like