Well it appears from the command you are only doing a rasa init which would just create a basic bot, train the files and then exit so this seems normal to me.
I’m confused and it seems I’m not the only confused newcomer. I want to setup a bot and interact with it from both sides: as a consumer and as a bot admin.
So I guess I should remove the init command when creating the container, and then connect to the container and execute init by hand (?). I run docker run -d -v $dir:/app rasa/rasa and after a few seconds it exits automatically. When I try to restart it (docker start rasacontainer) it exits again.
I think you might want to just look at our docs on using Rasa X and it might shed a lot of light on all of this for you and make it a lot easier to get started.
Yes I went through Getting Started and it says “Once you have built an assistant that can handle the most important happy path stories, you deploy Rasa X to a server and use it to improve your assistant.”. That’s actually what I’m trying to do (1st step: build a basic hello world assistant) and I’m not able to play with it yet. I’m still far away from having a working assistant effectively handling any happy path at all.
I doubt how a frontend would shed light, since I expect it to solve for me (and thereby hide from me) what I want to see. Nevetheless I’ve had a first fast look at the Rasa X doc just to check if it can help me, but while I think it is good to learn how to use and train a bot once available, it indeed seems to hide from me how to build or set it up.
My trouble is at using the docker image, I don’t yet have trouble using RASA itself.
Rasa X isn’t just a frontend, by installing it, you get all the required pieces for all of this to work. Also using the install.sh automated method for docker is pretty straight forward. That is why I was asking if you had looked at those yet.
Also one cool feature after having the rasa init basic bot setup, you can use Rasa X to interact with it, share it out, and annotate and resolve conversations. I’m sure you are aware of this since it sounds like you have read the docs on this already.
Or do you want to just stick right now with the Rasa Open Source platform and just it in a docker setup? If so we have that documented here at Running Rasa with Docker and we can just work through that issue you are seeing.
As far as the original question basically you would run something like this next to interact with the assistant, the first command you ran just basically creates an initial test bot and exits:
The problem is that each docker run execution creates a new container. So a tweeking dialog would create a number of containers (one with each tweak). We newcomers expect to be able to evolve the created container without creating new ones on each interaction (See this doubt example). Perhaps we’re wrong and there’s a good reason to create them all. If so, an explanaiton would be very welcome.
