Docs: how to build aiidalab on top of existing AiiDA installation. about aiidalab HOT 3 OPEN

To clarify, this means to simply install aiidalab into an existing (Python / Jupyter) environment and run it without needing to run any of the docker containers.

from aiidalab.

@ramirezfranciscof @mbercx Could you perhaps share your specific use case where use of the AiiDAlab container would be inconvenient?

The thing is: I don't think we have the manpower to make the AiiDAlab work in arbitrary environments.
Besides some core python packages like aiidalab and aiidalab-widgets-base, the AiiDAlab docker stack installs quite a range of auxiliary software, much of which the AiiDAlab apps may depend on.

If there is something simple that could be done to support your use cases, we can look into it, but I think the canonical distribution of the AiiDAlab will remain the docker image.

from aiidalab.

Hey @ltalirz ! Yes, so in principle my idea was to leverage the aiidalab platform graphical interface existing tools to develop some tools for general AiiDA management (so, make easier the launch of workflows, controlling the runs, etc). Thus for this I need to launch AiiDAlab in a way that it gets access to an existing AiiDA profile instance that is already installed in the system. It could also work if there was a way to launch AiiDAlab in a container that can access an external AiiDA core (not sure if you already support communication through the REST API, like the materials cloud do, or how does this work internally) as long as the container can also have the tools for development (like git or vscode access).

Anyways, now with the new react interface I'm no longer quite sure which would be the best tool for this or how you guys view the different use cases for each.

from aiidalab.