Error in plotting points in DOE

Nodeworks
#1

Here is the error that shows up:
If you continue running, the application may become unstable. Consider saving your work now.
Error: can’t set attribute
File “…/defaultnodes/surrogate_modeling_analysis/design_of_experiments.py”, line 833, in build_doe
self.update_plot()
File “…/defaultnodes/surrogate_modeling_analysis/design_of_experiments.py”, line 1053, in update_plot
self.plot.clearAxesData()
File “…/tools/matplotlib_helper.py”, line 88, in func_wrapper
return func(*args, **kwargs)
File “…/tools/matplotlib_helper.py”, line 1399, in clearAxesData
self._clearAxesData()
File “…/tools/matplotlib_helper.py”, line 2017, in _clearAxesData
self.axes.clearData()
File “…/tools/matplotlib_helper.py”, line 575, in clearData
self._clearMPLData(True)
File “…/tools/matplotlib_helper.py”, line 704, in _clearMPLData
self.ax.lines = []

How should I fix this issue?

About building the DOE Matrix
#2

can you share the *.nc file?

#3

Here is the .nc file link: https://file.io/JZBIL0ldiFOB

#4

Any update? Is there a way to fix this issue?

#5

nodeworks version: 20.2.0
Python version: 3.7.10 (default, Feb 26 2021, 18:47:35)
[GCC 7.3.0]
Qt wrapper: PyQt5
Qt version: 5.9.6
qtpy version: 2.0.1
Numpy version: 1.21.5

#6

I fixed it (and a couple other issues with changes in library calls). Building a point release now.

#7

I really appreciate your update. Is a notification sent once the point release is available in the public domain?

#8

I just pushed a conda package release candidate, version 22.1.0. I am still testing it, need to iron out a couple issues with the pipeline, but you can install it via:

conda  install nodeworks=22.1.0 -c https://mfix.netl.doe.gov/s3/<key>/conda/dist -c conda-forge
#9

Hello onlyjus,
I installed the point release and everything is working well now. I really appreciate your help. Are the issues in the pipeline related to surrogate modelling?

Thank you.

Abhijeet C.

#10

No, building the other packaging options and deploying updated documentation.

Although, the Gaussian Process method in the Response Surface Node is acting a little strange. Seems like Scikit-learn changed something. I am still playing around with that.