Add Meshes to the pdal API

[runette]

Adds the ability to access face data from meshes in the Python API:

• Adds a new method to pdal.Pipeline called get_meshes() that returns a list of ndarray in the same format as get_arrays(). Each ndarray represents one mesh from one PointViewand is a (1, size) array of Triangles, where each triangle is a tuple of (index of corner a, index of corner b, index of corner c) - the index referring to the points in PointView array.
• Adds the details to libpdalpython.pyx to implement the above,
• Adds a method to the cpp Pipeline class called getMeshes() that returns a vector of PyMesh
• Adds a cpp class PyMesh very much a copy of the PyArray class, in this case the class creates the mesh ndarray.

All of the above is working on my builds and does not introduce any new dependencies. However, I also wanted to create a way to create a useable Mesh object easily.

NOTE - where there is no mesh attached to the PointView, get_meshes will return an ndarray with 0 rows to ensure that the correspondence between arrays and meshes can be retained.

One good way of doing that is using the Meshio package - which allows the Mesh object to be re-used by a wide range of other packages.

So - I have also included the following :

• Adds a new method to pdal.Pipeline called get_meshio(idx: int) -> meshio.Mesh. This uses get_meshes and get_arrays to create a Mesh object. At the moment it does not add the non-geometry data (a ToDo). This method requires meshio and if it is not installed will fail with an informative error.

Usage

The simplest usage would look like the following:

import pdal

...
pl = pdal.Pipeline(pipeline)
pl.execute()

mesh = pl.get_meshio(0)
mesh.write('test.obj')

NOTE that since we are getting back a vector of meshes - we can do something that was not easy before and was actually the use-case that I needed to solve:

USE-CASE : Take a LiDAR map, create a mesh from the ground points, split into tiles and store the tiles in PostGIS.

(example using 1.2-with-color.las and not doing the ground classification for clarity)

import pdal
import json
import psycopg2
import io

pipe = [
     '.../python/test/data/1.2-with-color.las',
     {"type":  "filters.splitter", "length": 1000}, 
     {"type":  "filters.delaunay"}
]

pl = pdal.Pipeline(json.dumps(pipe))
pl.execute()

conn = psycopg(%CONNNECTION_STRING%)
buffer = io.StringIO

for idx in range(len(pl.meshes)):
    m =  pl.get_meshio(idx)
    if m:
        m.write(buffer,  file_format = "wkt")
        with conn.cursor() as curr:
           curr.execute(
               "INSERT INTO %table-name% (mesh) VALUES (ST_GeomFromEWKT(%(ewkt)s)", 
               { "ewkt": buffer.getvalue()}
           )

conn.commit()
conn.close()
buffer.close()

This creates 24 tile meshes in PostGIS

[runette]

Can you verify that it doesn't leak PDAL objects?

I am not aware of any leaks and tried to follow the existing Array structure as closely as possible - but as usual I don't know what I don't know!

[hobu]

Please add a test that does filters.delaunay on https://github.com/PDAL/python/blob/master/test/data/1.2-with-color.las and tests that you get what you expect. With that, I think we're good to merge.

[runette]

Please add a test that does filters.delaunay on https://github.com/PDAL/python/blob/master/test/data/1.2-with-color.las and tests that you get what you expect. With that, I think we're good to merge.

Will do - probably tomorrow

[runette]

@hobu Added tests and documentation

[hobu]

2.4.0 pushed https://pypi.org/project/PDAL/