continuous curl failure

[kwazineutin]

I am trying to use Reader.ept to access data via a URL.

The URL works fine; I can visualize the number of points in the polygon. However, when I try to use a pipeline to access the points and generate a surface, I get a curl failure. I am using PDAL in a Jupyter Notebook using the code below.

################################################ for index, row in tiles_mec.iterrows():

id = row ['tile_id']

resolution = 1
useful_classes = "Classification[0:6],Classification[17:17],Classification[9:9],Classification[10:10],Classification[11:11],Classification[2:2]"

#fetch all points and create full dsm
reader = pdl.Reader.ept(url, 
                     resolution=resolution, 
                     polygon = row['geometry'].wkt)
rng = pdl.Filter.range(limits=useful_classes)
z = pdl.Writer.gdal(out_fdsm + id +"_rfdsm_r1.tif", 
                resolution=resolution, 
                dimension="Z", 
                data_type="float32", 
                output_type="mean",
                nodata=-9999)

pipeline = reader | rng | z
pipeline.execute()

################################################# I keep getting curl failure errors, as shown in the image. Could the size of the polygon be a factor? Thanks for addressing this.

[hobu]

It is a bad message from PDAL. I think we suppressed it in PDAL 2.6.0.

What's happening is a timeout is reached and then PDAL is retrying again (and succeeding).

[kwazineutin]

I have been running the code for several hours and have less than 10% of the tiles. Is there a more practical way for creating DSM for large areas without downloading the points? I'd be happy to explore that. Thanks.

[hobu]

I have been running the code for several hours and have less than 10% of the tiles

You will have to make requests in smaller chunks. PDAL does not parallelize operations for you. We are available to hire to help with designing such processing approaches.