dimension of input coordinates and voronoi patch are not the same

[wjlei1990]

Hi Tyler,

I just downloaded your code and run the ipynb example(in the documentation).

I found you have 1000 pionts as input to the voronoi_instance. However, the output only contains 999 patch. There is one patch missing. Could you explain to me why is that?

[tylerjereddy]

Ah yes, that was a bug I think. Most of our development efforts are focused on the spherical Voronoi code integration into scipy (pull request: scipy/scipy#5232). We'd be greatful if you could take a look at our scipy fork (https://github.com/tylerjereddy/scipy) and try the scipy.spatial.SphericalVoronoi code there--it should hopefully be working quite well!

I think it was @niknow who managed to patch the issue you describe (for SphericalVoronoi in scipy fork).

[wjlei1990]

I just checked out your scipy fork. It looks much cleaner :)

However, I also need this method:

dictionary_voronoi_polygon_surface_areas=voronoi_instance.voronoi_region_surface_areas_spherical_surface()

But I didn't see it in the scipy code.

Any idea?

By the way, I compared the old code with the scipy one. The area method should be pretty easy to added in...But I think it might be safer for you guys to added in.

[tylerjereddy]

The surface area code was removed because we felt it was just too much new stuff to add to scipy in a single pull request, and because calculating the surface area of a spherical polygon is a bit of a separate functionality. I will likely open a pull request to scipy soon for including that functionality somewhere else in the package.

For now, I guess my suggestion is to use the scipy spherical Voronoi code and then use the surface area calculation function(s) from the 'old' code / my repo. I hope that is not too annoying, and we hope to improve the situation soon, but I'm sure you can understand it takes quite a while to get these things done correctly for scipy standards, etc.

Do keep us in the loop if you notice anything else that might need fixing, and if you'd like to help even better!