[pdal] Entwine/EPT build failures

[Jean-Roc Morreale (ml)]

Hi Andrew, thank you for your reply.

Unfortunatly, our Windows system configuration prevents me to setup a
msvc build toolchain so I will wait for the next release on this point.

About the LAS missing signature, for now entwine build seems to be the
only tool blocked by it as I have been using many pdal tools an others
(lastools, cloudcompare, fme, potreeconverter) in the last past weeks.
My first reaction was to convert the las files to laz using pdal
stating minor_version=4 in the hope it would automagically solves it.

regards,
jean-roc


Le lundi 18 février 2019 à 10:04 -0500, Andrew Bell a écrit :
> The first error (about legacy flags) is due to an issue in the LASzip
> API code.  You can see this issue 
> https://github.com/PDAL/PDAL/issues/2320 for details.  You can build
> the 1.8-maintenance branch of PDAL to eliminate the error.
> 
> The file signature error is what it says.  LAS files have a signature
> at the beginning of the file.  If the signature isn't found, your LAS
> file is corrupt.
> 
> I'm not sure why you'd get a "too many files" open error.  I think
> this is an issue in entwine.
> 
> On Mon, Feb 18, 2019 at 9:49 AM jrmorreale_ml at enoreth.net <
> jrmorreale_ml at enoreth.net> wrote:
> > Hi,
> > 
> > I am trying to build an EPT with entwine 2.0 (conda) on a
> > classified lidar cloud (1220 tiles, 22billion points) but I did not
> > succeed in getting a usable result.
> > 
> > The command used is :
> > entwine build --srs "EPSG:2154" --threads "4" --dataType "laszip"
> > -i input_dir -o output_dir
> > 
> > The errors are
> > * Exception in pool task: writers.las: legacy flags and extended
> > flags are not identical
> > * During input_xx.laz: readers.las: File signature is not 'LASF',
> > is this an LAS/LAZ file?
> > * Unable to open stream for 'input_xx.laz' with error 'Too many
> > open files'
> > 
> > The last error is only shown when trying to work on the 1220 files
> > at once. The attempts have been made on the original terrascan
> > tiles and on pdal translated laz tiles, the generated ept-data
> > output is empty laz files.
> > 
> > pdal info --all of 1 laz tiles
> > https://gist.github.com/Jean-Roc/bb045bd1e01edbadf7f49e97df0d2568
> > 
> > pdal info all on the same tile but on the original las file from
> > terrascan
> > https://gist.github.com/Jean-Roc/50145b3b7a0b2b432f0b2956807eb181
> > 
> > las sample
> > https://dl.lillemetropole.fr/file.php?h=Rc3bb7eeace8c2923b565e2b7ba7f2370
> > 
> > laz sample
> > https://dl.lillemetropole.fr/file.php?h=R9e3f147ea2f928b9f4437697faad3e51
> > 
> > log of entwine build on 5 tiles (run 5)
> > https://gist.github.com/Jean-Roc/ce231d6770fdaafa5bba307d79c0e992
> > 
> > log of entwine build with on 1220 tiles
> > https://gist.github.com/Jean-Roc/29fcffa60886168e73d0e7146e897801
> > 
> > Could you point to the mistake(s) I've made ? My aim is to publish
> > our data online, the latest usgs example is a great incentive.
> > 
> > Regards,
> > Jean-Roc Morreale _______________________________________________
> > pdal mailing list
> > pdal at lists.osgeo.org
> > https://lists.osgeo.org/mailman/listinfo/pdal
> 
>