[Draft] support PointCloud pickle

[chunibyo-wly]

Type

• Bug fix (non-breaking change which fixes an issue): Fixes #
• New feature (non-breaking change which adds functionality). Resolves #
• Breaking change (fix or feature that would cause existing functionality to not work as expected) Resolves #

Motivation and Context

Currently, the Non-tensor version of Geometry Class can not work with multiprocessing very well, as these classes miss the implementation of py::pickle.

May I ask if it is the proper implementation to support pickle? If so, I will continue to finish the remaining classes (lineset, triangle mesh, ...).

[update-docs]

Thanks for submitting this pull request! The maintainers of this repository would appreciate if you could update the CHANGELOG.md based on your changes.

[chunibyo-wly]

The minimal test case is

import pickle

import numpy as np
import open3d as o3d

pcd = o3d.io.read_point_cloud(o3d.data.KnotMesh().path)
pcd.estimate_normals()
pcd.estimate_covariances()
print(pcd.has_covariances(), pcd.has_normals())
data = pickle.dumps(pcd)
pcd_loaded = pickle.loads(data)


print(np.all(np.asarray(pcd.points) == np.asarray(pcd_loaded.points)))
print(np.all(np.asarray(pcd.normals) == np.asarray(pcd_loaded.normals)))
print(np.all(np.asarray(pcd.colors) == np.asarray(pcd_loaded.colors)))
print(np.all(np.asarray(pcd.covariances) == np.asarray(pcd_loaded.covariances)))

And the stdout is

True True
True
True
True
True