Why input-/output_port and not just inputs/outputs in NodeData?

[samwaseda]

In NodeRecipe we have inputs and outputs. Why are they called input-/output_ports in NodeData?

[liamhuber]

NodeRecipe.outputs: list[str]; NodeData.output_ports: dict[str, OutputPort].

Just to remind people that the thing they're touching is relatively complex in comparison to the recipe. E.g., they need to remember to .value to actually get the value out.

Later I added a @property pyiron_workflow.Run.outputs: dict[str, object] that is again pretty simple, built to map {label: data.output_ports[label].value} and intuitively get people back the output. I don't have any objection to pulling that property down onto the NodeData.

(Ok, actually I am cheating here, those str are probably all flowrep.schemas.Labels, and the first dict is a OutputDataPorts: MutableMapping[base_models.Label, OutputDataPort], and the second is a pyiron_snippets.dotdict.DotDict, but everything above is truthy.)