Closed chronis10 closed 2 years ago
The problem here is that, when the GIT action calls the checkout function, it does not include the submodules. There is a quick fix for this, that I've already done (i.e., adding the submodules='true'
) in the git action for building the image, but requires the repository teaching-base to be public.
The long-term solution is to translate teaching-base to a python package to be built locally by the other repositories (i.e., by cloning teaching-base internally and doing pip install . within teaching-base).
I had a quick chat with the coordinator. We plan to make all the repos public soon! This should fix the re-build feature.
@chronis10 try again once the build of the ai-toolkit image completes. Check directly by yourself in the teaching-ai-toolkit workflow.
Hi again, I am sorry for the spamming today, but I am trying to ensure the normal operation of the platform before the take-over. I tried the scenario 5 and returns me the error "from base.node import TEACHINGNode ai_toolkit | ModuleNotFoundError: No module named 'base.node'". Also, I am receiving the same error with the scenario 2 (RL personalization), that module created in the last integration meeting based on the stress module.![image_2022_05_23T10_57_12_084Z](https://user-images.githubusercontent.com/34183674/169817140-4393d51c-602d-467b-81dc-03478135669c.png)