Set up backend structure and list all available drives

[DenisaCG]

Set up the backend structure.

Created managers as an abstraction layer to enable extensibility, namely for the providers. For this purpose, a base manager class JupyterDrivesManager was created, and the corresponding parent class S3Manager. The entrypoints package is also used to list providers for the managers.

A base handler class was also instated, as JupyterDrivesAPIHandler. All other handlers implement their logic on top of it. To make it as simple and extensible as possible, the handlers and endpoint responses are meant to be provider independent.

In base.py, we handle the provider selection based on jupyter_notebook_config.py, through the DrivesConfig class. The user can set a session token, a secret access token and an access key id to enable access to the desired S3 drives.

[github-actions]

👈 Launch a Binder on branch DenisaCG/jupyter-drives/listAvailableDrives

[fcollonval]

Thanks @DenisaCG this looks very promising

[fcollonval]

We may not be as lucky as the pull requests extension about this. But let's see how it goes in the future.

[fcollonval]

Thanks @DenisaCG

I have some minor suggestion, but it looks really great.

[DenisaCG]

Thanks for the review and suggestions @fcollonval!

[fcollonval]

@DenisaCG I have some post mortem comment. If you could take them into account in a follow-up PR it will be great

[fcollonval]

What is the error code if the credentials are invalid? Do we get an error code like 401 that could in the future be used in the frontend to notify the user credentials are invalid?

[DenisaCG]

Thanks for the additional comments, @fcollonval! All of this has been improved in #9 (the credentials errors are still a work in progress, but will be pushed to the same PR).