Desired features (30.08.2024)

[Tara-Lakshmipathy]

Currently desired features (30.08.2024)

• Input handling (only after user has finished typing)
• Positions and arrangement of nodes (which is saved somwhere so that everything doesn't rearrange when the notebook is closed)
• Adding/deleting nodes reflected in underlying workflow
• Duplicate nodes with automatically generated unique IDs
• Hashing of output
• Tab system improvements - adding and closing tabs, connection to treeview for each tab

[liamhuber]

• Hashing of output

Hey Tara, can you elaborate on this a bit? The rest are very clearly GUI issues to me, but it sounds like this might possibly require changes of the back-end

[Tara-Lakshmipathy]

• Hashing of output

Hey Tara, can you elaborate on this a bit? The rest are very clearly GUI issues to me, but it sounds like this might possibly require changes of the back-end

Hey Liam, as it stands, the GUI has two buttons on each node: a button to view the source code and another to run the node (which also displays the output). If we have a node with a pyiron job (e.g. LAMMPS), then there's not much of an issue since the node would just reload the existing job unless specified otherwise by the user. But if we have a node that has a large-ish analysis that is not currently included in pyiron, for e.g. some scipy opitmization, then it would be ideal if pressing "run" just displayed the already performed analysis instead of re-running it (unless specified otherwise by the user).

[liamhuber]

Ah. I think I'd rather call this "caching" than "hashing". Good news is that the underlying `pyiron_workflow infrastructure already does this by default -- as long as the inputs are the same as the last time the node was run. If the inputs aren't the same I guess you do want to actually rerun it anyhow.

[Tara-Lakshmipathy]

We still need to sort out the versions of the pyiron_wokflow module, pyiron_base, pyiron_atomistic etc. I think we are still using a slightly older version of workflow which does not have caching/hashing due to compatibility issues.
I'll try to look into it this week!

[liamhuber]

pyiron_workflow-0.10.0 no longer depends on base or atomistics, so hopefully this will prove relatively straightforward for you!