Closed rti closed 9 months ago
The parts of the code that I can read make sense for why those edits were made. But I cannot comment further.
I would like to ask though: how do we generate frontend/dist
to begin with?
I hoped that it would be built for us during the docker build
. Now I assume that we have to enter the frontend
directory, build the VueJS manually, and then build the docker.
Is that correct? If so
Dockerfile
without diving into docker-compose?frontend
director" language to the README.md?Best, ew
Unrelated to the above:
Do we agree to rename the director ./excellent-articles
to ./json_input
?
If so, could we please change that in this branch-PR as well?
@exowanderer Thanks for the review. Very good catches.
I added the frontend build step to the docker container, so it is automatic now.
Do we agree to rename the director ./excellent-articles to ./json_input?
I am fine with that. I merged the other branch from #9 to main, merged main into this one, so this is fixed too.
@exowanderer Thanks for the review. Very good catches.
I added the frontend build step to the docker container, so it is automatic now.
It is always helpful to have a noob onboard. I have no idea how to build VueJS, so I knew it was not in the Dockerfile.
PS: New T-shirt for Pregnant friends: "Noob On Board"
In the andrew-frontend
branch, I am still unable to build the Docker because of the send_message
in haystack.telemetry
issue #6, but that is not related to my Change request above.
Merges in the work of @andrewtavis from https://github.com/andrewtavis/gbnc
Closes #3
This approach is different from #4 as the frontend is still served via the unicorn fastapi server. There is no further containerization going on here.
I think #4 is the right approach on the long run. This PRs solution is very simple, keeping the deployment easier for now (one container instead of multiple). For real world production deployments, we most probably will still need #4.