Return output when calling run

[liamhuber]

And downstream stuff like update and thus __call__. This was requested by @JNmpi and now makes things really start to feel like regular python.

from pyiron_contrib.workflow import Workflow

@Workflow.wrap_as.single_value_node()
def plus_one(x=0):
    y = x + 1
    return y

wf = Workflow("wf", run_on_updates=True)
# wf.executor = CloudpickleProcessPoolExecutor()
wf.a = plus_one()
wf.b = plus_one(wf.a)

# For workflows we now return the DotDict of the output values
print(wf(a_x=2))
>>> {'b_y': 4}

out = wf(a_x=4)
print(out.b_y)
>>> 6

# And of course it works for Function nodes too, but with just the values
print(wf.a(42))
>>> 43

# The returned value is its own object
# So when we updated the node just there, the workflow output changes
# but `out` stays the same:
print(out.b_y, wf.outputs.to_value_dict())
>>> 6 {'b_y': 44}

# IMO that's desirable, because then you can collect up outputs
# over multiple runs of your workflow and plot them in the notebook
# or whatever

TODO:

• Update docstrings
• Update example notebook

[github-actions]

👈 Launch a binder notebook on branch pyiron/pyiron_contrib/return_on_run

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB