Feat: Use cloudpickle to run remote functions in jupyter environments

[dleviminzi]

Resolve BE-2218

This PR changes the strategy used for running remote functions in notebooks. Instead of creating a temporary file built by stitching together executed cells, we use cloudpickle to capture the required function + deps. This resolves issues with re-executed cells and is generally more reliable.

This PR also changes the way that the default python environment is selected for Image. It tries to see if the active python version from the local environment is a valid remote python version. If it is, it will use that version by default. I can remove this change or limit it to ipython environments only. This is nice when working in jupyter notebooks because cloudpickle dumps are dependent on the python version. If you create a dump in 3.12 and try to load it in 3.10, things will not work.

TODO:

• Test whether different protocols improve compatibility for cloudpickle python version mixing
• Alert users when they try to run different local/remote python versions in jupyter environments
• Explore ways to detect function content changes to avoid re-syncing when the cell has been re-executed but no code has changed