Other types of mean using volume_average

[shashwat97]

Hi,

I want to calculate the harmonic mean for a tree mesh based on an input tensor mesh. Just like the volume_average function, just a different kind of average.

I am doing it like this:

avg_df = pd.DataFrame([tree_mesh.point2index(fine_mesh.gridCC), df[prop].values]).T
harmonic_means = avg_df.groupby(0)[1].apply(hmean)

The problem is, this is much slower than the volume_average. Is there any way to use the volume_average code to find the harmonic average of the overlapping cells?

Thanks for your assistance!
@jcapriot

[jcapriot]

I would look at how volume_average finds intersecting cells.

[jcapriot]

The relevant code is inside cython at

discretize/discretize/_extensions/tree_ext.pyx

Line 6302 in 90f8482

def _vol_avg_to_tens(self, out_tens_mesh, values=None, output=None):

, so it is compiled.