Add functionalities to DocArray to handle 3D data, like a mesh consisting of vertices and faces.
Currently 3d data handling in DocArray only includes loading point clouds from files like .obj. Point clouds are only a part of 3d data representation. Additionally meshes consisting of vertices and faces are being used and contain more information about a 3d object than point clouds.

Inparticular, we want to add methods to:

• load mesh (consisting of vertices tensor and faces tensor) from file into Document chunks
• convert a mesh to point cloud