Build the index tensor telling us which coedges belong to a given edge size = [ num_edges x 2 ]

Build the index tensor telling which coedges belong to which face. As the number of coedges in a face can vary then we

Try building point grids from the left coedges of each edge

Build the index tensors single sorted faces array back to the original faces in the different solids

Convert the feature tensors for faces, edges and coedges from numpy to pytorch

Read the point grid and LCS data and convert it to pytorch

Save the cache data for the body

Check a specific feature value for the closest face or edge to a given point

Basically we want to check that the mating coedges are set up correctly. It's quite hard to get the mate information out

Check the mapping from coedges to faces is consistent

We want to check that the array of indices in next is actually a permutation. i.e. there is a bijection (1:1) mapping be

Check the part against the face boxes loaded which were found from the Fusion triangle file.

Append all the data from batch_to_cat into acc_batch We need to re-map the indices to do this. Essentially we combine al

Create batches with a fixed size of num_faces_per_batch. Here we try hard to exactly replicate the original algorithm used to

Create a dictionary which we can contatenate with other data to build up the single solid for the batch

We can check for consistency between the mating coedges and the edge information.

We can cross check the in two ways - For every coedge, we know the index of it's face from array f. We can then check