Error when Ymodulus chart activated, but 0 events in

[jarogumulec]

In datasets with extreme deformation data (under ridiculously high flow) , when HE-2D-FEM-22 is activated, 0 Ymoduli is calculated for cells (in milder flow cases, but still far above suggested, the event numbers between deformation and Ymodulus charts do not match). Then, when a chart with 0 events (0 Ymoduli calculated) is activated, an error popups.

When such dataset is saved, it cannot be reopened as this popup message blocks dataset opening.

In particular problem for HE-2D-FEM-22 model, 2019 calculates more data even from extreme deformations

Unhandled exception in DCscope version 2.23.0:
Traceback (most recent call last):
  File "dcscope\gui\pipeline_plot.py", line 91, in update_content
    self.update_content_plot(plot_state, slot_states, dslist)
  File "dcscope\gui\pipeline_plot.py", line 211, in update_content_plot
    pp.redraw([ds], [sl], plot_state_scatter, hash_flag)
  File "dcscope\gui\pipeline_plot.py", line 349, in redraw
    sct = add_scatter(plot_item=self,
  File "dcscope\gui\pipeline_plot.py", line 592, in add_scatter
    x, y, kde, idx = plot_cache.get_scatter_data(
  File "dcscope\plot_cache.py", line 63, in get_scatter_data
    kde = rtdc_ds.get_kde_scatter(
  File "dclab\rtdc_dataset\core.py", line 677, in get_kde_scatter
  File "dclab\kde\base.py", line 373, in get_scatter
  File "dclab\kde\methods.py", line 116, in new_kde_method
  File "dclab\cached.py", line 67, in __call__
  File "dclab\kde\methods.py", line 209, in kde_histogram
  File "dclab\kde\methods.py", line 25, in bin_num_doane
  File "dclab\kde\methods.py", line 55, in bin_width_doane
  File "numpy\core\_methods.py", line 41, in _amax
ValueError: zero-size array to reduction operation maximum which has no identity

[paulmueller]

Thanks, I have implemented a fix for the kernel density estimation in dclab 0.67.2 (DC-analysis/dclab@e76f5c7).
This means that this will be fixed with the next release of DCscope.

For the sake of completeness, would you mind exporting an .rtdc file with ~20 events that I can use to reproduce this issue in DCscope? Then I could write a test function to make sure this never happens again in DCscope.
You should be able to attach the .rtdc file to this issue by putting it into a zip archive.

[jarogumulec]

What a great fast debugging! This dataset: after filtering porosity 1-1.05 and area 50-1000 there are 99 events (deformation plot). When Ymodulus is calculated with 2019 model, only 3 events remain, and with 2022 model 0 (hence error appears). I'm of extreme flow conditions out of model optimum, but consider this a demonstration.
14 pc3 2.2uls.zip

[jarogumulec]

Thanks, update 2.24 solves quickview problem with no data datasets, but issue still persists in Plot view

Unhandled exception in DCscope version 2.24.0:
Traceback (most recent call last):
  File "dcscope\gui\widgets\wait_cursor.py", line 31, in wrapper
    ret = func(*args, **kwargs)
  File "dcscope\gui\quick_view\qv_main.py", line 816, in plot
    self.widget_scatter.plot_data(rtdc_ds=self.rtdc_ds,
  File "C:\Program Files (x86)\DCscope\_internal\dcscope\gui\quick_view\qv_scatter.py", line 151, in plot_data
    self.yax, slot_state=self.slot.__getstate__())
AttributeError: 'NoneType' object has no attribute '__getstate__'