LasPy to PDAL

First post of the new year … after spending way too much time writing a function that I didn’t need at the end :/

A simple and correct function to convert a laspy.LasData object to a structured numpy array compatible with PDAL. This doesn’t seem very fast, my guess is that merge_arrays is doing a lot of copying around that could be optimized away.

import laspy
import numpy as np
from numpy.lib.recfunctions import drop_fields, merge_arrays, unstructured_to_structured


def laspy_to_pdal(las_cloud: laspy.LasData) -> np.ndarray:
    """
    cloud_xyz = unstructured_to_structured(las_cloud.xyz, names=['X', 'Y', 'Z'])
    cloud = drop_fields(las_cloud.points.array, ['X', 'Y', 'Z'])
    cloud = merge_arrays([cloud_xyz, cloud], flatten=True)
    # rename fields to CamelCase
    new_names = [snake2camel(name) for name in cloud.dtype.names]
    cloud.dtype.names = new_names
    return cloud