Dag data difficult to read

[samwaseda]

This is just not very appealing. Can we use a representation like in pyiron_workflow (or DataContainer) to have a better overview?

[liamhuber]

This is just how dataclasses display. I'm not super inclined to overload __repr__, since I like people being able to look see the dataclass the way they usually see the dataclass, but I'm totally on board with adding a method for a pretty representation NodeData.view() or something, or maybe even setting a custom __str__ method.

Is there something in particular you want to see? A particular layout?

[samwaseda]

NodeData.view() would definitely be a possibility. I like the way pyiron_workflow shows the data.

[liamhuber]

Yep, I'm on board. If you're willing to wait long enough, I will get to it eventually; if you wind up having a clear vision and less patience feel free to go into flowrep.retrospective.datastructures and add .view() methods to NodeData and its decedents.

[liamhuber]

Note that the edges already have a .serialize() method which represents them as f"{node}.{port}". I like .view() better, so it would be good to rename that (then keep .serialize() as an alias)

[samwaseda]

ok I'll try to find time. Could you tell me in the meantime which file I should look into in pyiron_workflow?

[liamhuber]

I'm not sure, I don't know what part of it you like. If you're talking about print(wf.nodes) for pyiron_workflow._wfms.api.Workflow or similar, then the relevant printout is here:

https://github.com/pyiron/pyiron_workflow/blob/03be1231146959390d88fb3fc4d75a724a2e0928/pyiron_workflow/_wfms/lexical.py#L139-L143

[samwaseda]

ok the one that I was looking for was from IPython.display import JSON, but ChatGPT just told me a few other possibilities so I look into them as well