add github copilot manual

[ShaivanBhagat]

manual to use copilot via local vsc using ssh for vscodium in python deskptop workbench.

[jelletreep]

@ShaivanBhagat , thanks for this PR! Looks good. I would need to try this out myself to give more specific feedback. Some general comments:

• we should probably put a disclaimer that we do not endorse any of the coding assistants, using a coding assistant is the responsibility of the user.
• We can maybe make it tool agnostic (using a coding assistant in combination with Research cloud), and primarily explain the concept of connecting to a workspace via ssh.
• I assume it would be possible to run something on the workspace from the local VS Code instance? In that case it matters less which workspace you are connecting to, although it might be helpful to have a workspace that has python env managers pre-installed.

[ShaivanBhagat]

Hi @jelletreep @dometto I have updated the files. Now there is a guide for connecting local vscode to any workspace and to use vscode remotely.
There is a second manual that guides (briefly) how to use copilot in vscode.
I have tried to include disclaimers for ai use from what i could find from UU's website

[jelletreep]

Hi @ShaivanBhagat, I am now testing the tutorial, here my thoughts about the Connecting to Research Cloud from VS Code part:

The part about setting up the SSH connection is very clear I think, well done!

I didn't realize that you first need to transfer files to the workspace. This should be clarified at the start with a conceptual explanation, maybe in a first paragraph that also explains why this setup might be useful (mainly using your own VS code extensions). It should be emphasized that the workshop should be paused whenever not working on it or not running any code, as well as to transfer any updated scripts and output data back to laptop or storage. I think this is important because with this way of working the risk is that people might forget about the remote workspace (that costs credits, has an expiry data, and data loss when a workspace is deleted).
Maybe examples can be added (using the src-python-example?)

The extensions part is still a bit of an unknown for me. I see that some extensions that I have locally, I can just continue to use them. Some other local extensions are grayed out with a button to install them on the workspace, I assume because they need to be installed on the workspace to be able to run them?
Any newly installed extensions are automatically only installed on the workspace it seems.
If you understand this better, could you try to clarify this?

Some structural things I wasn't clear about earlier:
I think we can add two manuals under the "Other" section on the main manuals page:

• Connecting to Research Cloud from VS Code
• Using an AI powered coding assistant on Surf Research Cloud (probably better to remove any preference for e.g. github copilot)

The FAQ section can be removed from the vscode-to-researchcloud page (I was confused with the Glossary, we don't actually have a FAQ page, however this is something to consider as a separate issue)

[jelletreep]

@ShaivanBhagat , I will start with the copilot manual soon

[ShaivanBhagat]

I have updated the documentation for vscode - ssh - workspace

• explaining what it does now and why would it be useful or relevant
• also included an example workflow for working with python env managers
• tried to explain extension use

For the ai assistant, i have now made it general (agnostic tool). Ideally any extension that is supported on local VSCode should work on the workspace