[ncl-install] Issue with ploting shapefile

Wed Jan 1 12:18:04 MST 2025

Ehsan,

1.  You do not need to upload that large shapefile.  Instead, just go to
the shapefile directory, and post the output from this command into github
#176 <https://github.com/NCAR/ncl/issues/176>.  Be sure to show the exact
shape file name that you are using.  (I do not need to see the full
directory path.)  Use the *cut* command to suppress the unnecessary huge
list of individual numerical coordinates.  Reminder, you might need to have
an active conda environment to have access to the *ogrinfo* command.

      *ogrinfo -al FILENAME.shp | cut -c 1-80*

2.  I do not have much experience with conda.  However, I recommend that
you stick with the current stable NCL environment, and do not try to go
back to a historical version.  It would seem that rewinding can be
problematical.  The current stable version is well managed by the conda
team, and includes significant fixes and updates.  If you have new problems
with the current stable version, there is on line advice about how to clean
or uninstall and reload a conda environment.

3.  You can make copies of the original and experimental versions of
*gsn_code.ncl* in separate private directories.  Then you can quickly swap
them in and out of any NCL installation without going through any conda
procedures.

On Wed, Jan 1, 2025 at 5:18 AM Ehsan Taghizadeh <ehsantaghizadeh at yahoo.com>
wrote:

> Dear Dave,
>
> Thank you again for your follow-up.
>
> I tried to upload shapefiles to GitHub, but I encountered the 25 MB file
> size limitation. My shapefiles are 37 MB. Could you advise me on how to
> attach these files in one message?
>
> Additionally, after replacing the gsn_code 2024 version with the previous
> one, every time I remove the ncl_stable environment and try to reinstall
> NCL via conda, I encounter the following errors, and the gsn_code 2024
> version is replaced:
>
> *Downloading and Extracting Packages:*
>
> *Preparing transaction: done*
> *Verifying transaction: \*
> *SafetyError: The package for ncl located at
> /home/taghizade/miniconda3/pkgs/ncl-6.6.2-h7cb714c_54*
> *appears to be corrupted. The path 'lib/ncarg/nclscripts/csm/gsn_code.ncl'*
> *has an incorrect size.*
> *reported size: 504377 bytes*
> *actual size: 509857 bytes*
>
> *ClobberError: This transaction has incompatible packages due to a shared
> path.*
> *packages: conda-forge/linux-64::hdfeos2-2.20-h3e53b52_1004,
> conda-forge/linux-64::hdfeos5-5.1.16-h51d0b48_17*
> *path: 'include/cfortHdf.h'*
>
> *ClobberError: This transaction has incompatible packages due to a shared
> path.*
> *packages: conda-forge/linux-64::hdfeos2-2.20-h3e53b52_1004,
> conda-forge/linux-64::hdfeos5-5.1.16-h51d0b48_17*
> *path: 'include/ease.h'*
>
> *ClobberError: This transaction has incompatible packages due to a shared
> path.*
> *packages: conda-forge/linux-64::hdfeos2-2.20-h3e53b52_1004,
> conda-forge/linux-64::hdfeos5-5.1.16-h51d0b48_17*
> *path: 'include/isin.h'*
>
> *ClobberError: This transaction has incompatible packages due to a shared
> path.*
> *packages: conda-forge/linux-64::hdfeos2-2.20-h3e53b52_1004,
> conda-forge/linux-64::proj-9.4.1-h54d7996_1*
> *path: 'include/proj.h'*
>
> *ClobberError: This transaction has incompatible packages due to a shared
> path.*
> *packages: conda-forge/linux-64::hdfeos2-2.20-h3e53b52_1004,
> conda-forge/linux-64::hdfeos5-5.1.16-h51d0b48_17*
> *path: 'lib/libGctp.a'*
>
> *done*
> *Executing transaction: /*
> *To enable CUDA support, UCX requires the CUDA Runtime library
> (libcudart).*
> *The library can be installed with the appropriate command below:*
>
> ** For CUDA 11, run:    conda install cudatoolkit cuda-version=11*
> ** For CUDA 12, run:    conda install cuda-cudart cuda-version=12*
>
> Does anyone have any suggestions on how to install NCL via conda with the
> default gsn_code version (2019)?
>
>
> *-----------------------------------------------------------------------------*
> *Best regards,*
> *Ehsan*
>
>
> On Tuesday, December 31, 2024 at 06:45:46 AM PST, Dave Allured - NOAA
> Affiliate <dave.allured at noaa.gov> wrote:
>
>
> Ehsan, thank you for testing.  It looks like my changes for "gsn_code.ncl"
> need further debugging.  It is a good sign that "englandshp" worked with
> the new code, because that demonstrated mostly working infrastructure for
> X/Y swapping.
>
> Let us move the debugging conversation off of the mailing lists.  Please
> reply on Github issue #176 <https://github.com/NCAR/ncl/issues/176>.  I
> need the following details:
>
> 1.  Full printed output of those diagnostic messages when trying to plot
> "gadm41_CAN_shp" with the new gsn_code.ncl.
>
> 2.  Attach a zip file with the complete file set for the shapefile
> "gadm41_CAN_shp".
>
>
> On Tue, Dec 31, 2024 at 7:13 AM Ehsan Taghizadeh <
> ehsantaghizadeh at yahoo.com> wrote:
>
> Dear Dave,
> Thank you for your helpful reply.
>
> First, I should mention the GDAL version on our server:
>
> *$ ogrinfo –version*
> *GDAL 3.9.2, released 2024/08/13*
>
> Additionally, the NCL version installed by Conda on our server appears to
> be "h7cb714c_54". I tried to find the versions h8482486_9, h9ce15c7_1, or
> h7290063_6 mentioned as working well in issue *#56*
> <https://github.com/cedadev/ceda-jaspy-envs/issues/56>, but I couldn't
> locate them in the *build details from conda-forge*
> <https://anaconda.org/conda-forge/ncl/files>.
>
> It seems some shapefiles work well with the default NCL version from
> Conda. For instance, gadm41_CAN_shp
> <https://geodata.ucdavis.edu/gadm/gadm4.1/shp/gadm41_CAN_shp.zip> falls
> into this category. However, some shapefiles do not work with the default
> NCL in Conda. For example, "englandshp" (attached file) does not work well
> with the default NCL in Conda.
>
> When I replaced $NCARG_ROOT/lib/ncarg/nclscripts/csm/gsn_code.ncl with the
> one mentioned in issue #176 <https://github.com/NCAR/ncl/issues/176>, the
> situation changed. The "englandshp" worked well with the new gsn_code, but
> for gadm41_CAN_shp
> <https://geodata.ucdavis.edu/gadm/gadm4.1/shp/gadm41_CAN_shp.zip>, the
> following error was displayed:
>
> *(0)     shapefile_order_lat_lon: Error: Cannot parse.*
> *(0)       Unexpected output from 'ogrinfo -al $shapefile'.*
> *...*
> *(0)       Assume order of coordinates = lon/lat.*
> *(0)*
> *(0)     Diagnostics 1 for 'ogrinfo':*
> *GDAL 3.9.2, released 2024/08/13*
> *(0)*
> *(0)     Diagnostics 2 for 'ogrinfo':*
> *...*
> *(0)     Diagnostics 3 for 'ogrinfo':*
> *(0)     Parse results = False missing False False False False*
> *(0)     Number of lines in result = 1*
> *(0)     missing*
> *(0)     ----------------------------*
> *(0)       Assume order of coordinates = lon/lat.*
> *(0)     --- Working on time 2019020106 ---*
> *(0)     shapefile_order_lat_lon: Error: Cannot parse.*
> *(0)       Unexpected output from 'ogrinfo -al $shapefile'.*
>
> However, when I tried to run the simple NCL script from the NCL website to
> plot them, nothing was displayed for both shapefiles with the new gsn_code.
>
> load "./shapefile_utils.ncl"
> sname = "your_shapefile.shp"
> print_shapefile_info(sname)
> plot_shapefile(sname)
>
> For the "englandshp", nothing was displayed. For the CAN shapefile,
> nothing was displayed, and the following error occurred:
>
> *(0)     Diagnostics 3 for 'ogrinfo':*
> *(0)     Parse results = False missing False False False False*
> *(0)     Number of lines in result = 1*
> *(0)     missing*
> *(0)     ----------------------------*
> *(0)       Assume order of coordinates = lon/lat.*
>
> In summary, it seems that for some shape files, we should use the default
> gsn_code.ncl, while for others, we should use your gsn_code.ncl. This is my
> understanding. I would appreciate your thoughts on this.
>
> *-----------------------------------------------------------------------------*
> *Best regards,*
> *Ehsan Taghizade*
>
>
> On Sunday, December 29, 2024 at 02:17:40 PM PST, Dave Allured - NOAA
> Affiliate <dave.allured at noaa.gov> wrote:
>
>
> Ehsan, this is a known issue with recent NCL versions.  The lat and lon
> coordinates are swapped in some shape files, causing missing images, or
> mirror images along a 45 degree fold line.  That is what I see in your
> plots.
>
> This is discussed in Github issues #151
> <https://github.com/NCAR/ncl/issues/151> and #176.  In #176 I show a
> diagnostic method to determine swapped coordinates.  I also provide a
> modified NCL library file that should render plots correctly for any shape
> file.  Please give these suggestions a try, and let us know the results.
>
>
> On Wed, Dec 25, 2024 at 7:23 AM Ehsan Taghizadeh via ncl-install <
> ncl-install at mailman.ucar.edu> wrote:
>
> Hello everyone,
>
> I am trying to add a shapefile to my plot, but the location of the
> shapefile is incorrect as it appears to be rotated 90 degrees clockwise.
> Previously, I had an issue with NCL installed via Conda. I attempted to
> install NCL using a precompiled version, but encountered the following
> error: “ncl: error while loading shared libraries: libssl.so.10: cannot
> open shared object file: No such file or directory.” Since I do not have
> root access, I cannot install libssl.so.10.
>
> Attached are the incorrect and correct output maps. The desired shapefile
> is shown in tan. The correct version is the output from running NCL on
> another Linux version (Rocky), which I do not have access to. The code I
> used to plot these maps is quite long and may not be helpful for this
> issue, but I can provide a shorter version if necessary.
>
> As I am unsure if this issue is directly related to the NCL installation,
> I am sending this to both emails. Apologies for any inconvenience caused.
>
> The output of the command "conda list -n ncl_stable" is attached (
> condalist.txt). Some additional information is provided below:
> $ conda --version
>    *conda 24.11.2*
> $ which ncl
>    *~/miniconda3/envs/ncl_stable/bin/ncl*
> $ ncl -V
>    *6.6.2*
> $ env  | grep NCARG
>    *NCARG_ROOT=/home/taghizade/miniconda3/envs/ncl_stable*
>
> Any help would be appreciated. Happy holidays!
>
> *-----------------------------------------------------------------------------*
> *Best regards,*
> *Ehsan Taghizade*
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.ucar.edu/pipermail/ncl-install/attachments/20250101/4c8bf9cc/attachment-0001.htm>