Goals:

• Currently, DocArray only supports point clouds for 3d data representation.
• We want to add functions to additionally support 3d data representation of meshes, which include more information about a 3d object than a point cloud, and seem to be more commonly used. Meshes basically consist of a vertices tensor and a faces tensor, which I plan to load to load to chunks, i.e.:

from docarray import Document

doc = Document(uri='some/uri')
doc.load_uri_to_vertices_and_faces_tensors()

# resulting in:
# doc = Document(
#     uri='some/uri', 
#     chunks=[
#         Document(name='vertices', tensor=sometensor), 
#         Document(name='faces', tensor=sometensor)
#     ]
# )

• add load_uri_to_vertices_and_faces_chunk_tensors + tests
• load vertices+faces to point cloud
• check and update documentation, if required. See guide