Mapping or parity between WG UUIDs and aiida-core UUIDs

[GeigerJ2]

Currently, the UUIDs obtained from WG (e.g., via WG.to_dict() for the parent WG, as well as the sub-tasks in the nested dictionary), and the UUIDs of the aiida-core objects, e.g., the WorkChainNodes and CalcJobNodes shown, for example, through verdi process list and verdi process show differ.

I think this can cause a lot of confusion (as it just did for me), so we should find a way to solve this. Either call the keys in the WG dictionary representation different (wg_uiid or sthg similar), overwrite the WG ones when the WG process finishes with the corresponding aiida-core UUIDs (maybe this is difficult), or assign the WG UUIDs to the aiida-core ProcessNodes when they are created during the execution of the workflow.

To be thought about.

[superstar54]

In principle, when creating a WorkGraph, we don't need to assign a UUID to it. We assign the UUID only when saving it into the database. This means we can just use the UUID of the AiiDA process node as soon as we get the process node in the run() and submit() methods.

For the tasks, uuid is not needed, because inside WorkGraph, we always use the task's name as a reference. Also, in some cases (e.g., in a while loop), one task corresponds to multiple process nodes.