for_loop_pipeline-hobu.py

conda activate myenv
conda env list
import pdal
import glob
import numbpy as np


import logging
logger = logging.getLogger()
logger.setLevel(logging.DEBUG)

files =  glob.glob('F:/PDAL/Batch_Practice/Las_Tiles/*.las')
pipeline  = """
    {
    "pipeline": [
        {
            "type":"readers.las",
            "filename":"%(filename)s"
        },
        {
            "type": "filters.smrf"
        },
        {
            "type":"filters.range",
            "limits":"Classification[2:2]"
        },
        {
            "gdaldriver":"GTiff",
            "output_type":"idw",
            "resolution" :"1.0",
            "type": "writers.gdal"
        }
    ]
}
"""


def run(args, stdin=None, return_json = False):
    import json
    import subprocess
    logger.debug(args)

    p = subprocess.Popen(' '.join(args),
                         stdin = subprocess.PIPE,
                         stdout = subprocess.PIPE,
                         stderr = subprocess.PIPE, shell=True,
                         encoding='utf8')
    ret = p.communicate(input=stdin)
    if p.returncode != 0:
        error = ret[1]
        logger.debug(error)
        sys.exit(0)
        return False

    if return_json:
        logger.debug(ret[0])
        return json.loads(ret[0])
    return True

for f in files:
    # put the filename in a dictionary to allow convenient
    # substitution without requiring order-specific string
    # interpolation
    d = {"filename":f}

    # This is one way to do it, but that means all of the
    # points are read into the python proces. If it's a huge file, that's
    # not so great.
    ll = pdal.Pipeline(pipeline % d)
    ll.execute()


    # an alternative is to kick another process off and run the
    # pipeline using 'pdal pipeline'. The convenient thing about this
    # is you can use the multiprocessing module to create a processing
    # queue to attack things in parallel.
    rargs = ['pdal', 'pipeline', '--stdin','--debug', '--metadata', 'STDOUT']
    results = run(rargs, pipeline % d, return_json=True)

    # results will now contain the "pipeline metadata" that
    # was produced by the 'pdal pipeline' call.

    logger.debug('pipeline results', results)