Surfaces don't mask each other

[domist07]

Describe the bug The surfaces don't mask each other instead I have installed the scikit-image version of Leguark as described in #216. I only get the warning "WARNING (theano.tensor.blas): Using NumPy C-API based implementation for BLAS functions." when executing GemPy, but I think thats not relevant for the problem. The warning "UserWarning: Surfaces not computed due to: marching_cubes_lewiner() got an unexpected keyword argument 'mask' " doesn't arrises.

Desktop

• System: Ubuntu 19.10 as Virtual machine on Windows 10
• GemPy version 2.1.1

To Reproduce Today I run these two installation commands:

pip3 install git+git://github.com/Leguark/scikit-image@master
pip3 install gempy

Then I load my model, set the relations, calculate the surfaces and plot it:

[...] # loading data and set relations
gp.compute_model(geo_data, sort_surfaces=False, compute_mesh=True, debug=False)
gv = vista.Vista(geo_data, plotter_type='background', notebook=False, real_time=False)
gv.plot_surfaces()

Expected behavior The blue surface has an onlap relation according to the red surface and should be cut off by it.

Screenshots

[javoha]

Hi @domist07 , is it correct though that you get the "correct" behavior when plotting the lithology block model (e.g. in a cross section?). If I rememebr correctly that is the case and that means we "just"have to fix the marching cubes ... Thanks for pointing this out!

[domist07]

Yes, the lithology block model is correct, but the contact lines aren't cut off too.

[Leguark]

Were you able to compile scikit image in windows? Can you tell me how? I am stuck there at the moment so I can not test it (yes, I feel ashamed 😄 ) In scikit image 0.17 the necessary code comes in the pip installation so I was waiting till then to really support the feature

[domist07]

No that's the reason why I run Ubuntu as a virtual machine. All Python files run on Ubuntu and I've the scikit image version 17-dev0, but that doesn't mask the marching cubes.