[gdal-dev] Split out GDAL's GRIB2 tables as a separate repository?

[Even Rouault]

Jack,

~ 90% of the credit for those tables belongs to WMO Information 
Management, as they essentially come from 
https://github.com/wmo-im/GRIB2 (under MIT license)

The GDAL version of those tables is a mixture of those WMO tables + NOAA 
NCEP specific tables that extend the WMO ones here and there.

https://github.com/OSGeo/gdal/blob/master/frmts/grib/degrib/README.TXT 
explains the process we apply to do that. The gist is this script 
https://github.com/OSGeo/gdal/blob/master/frmts/grib/degrib/merge_degrib_and_wmo_tables.py

So essentially what you are seeking for is probably just the original 
WMO .csv tables from https://github.com/wmo-im/GRIB2 . But feel free to 
use the ones of GDAL (under MIT license too), or extract the above 
mentionned script and its input files in a dedicated repository. I'm 
happy if someone else regularly does this refresh & merge operation when 
WMO issue a new release and if I can just copy & paste the output

Even

Le 11/10/2024 à 12:55, Jack Kelly via gdal-dev a écrit :
> Dear GDAL devs,
>
> The directory gdal/frmts/grib/data/ contains the most comprehensive
> collection of GRIB2 code tables in a machine-readable format that I'm
> aware of! Thank you so much for creating this resource. It's a huge
> help!
>
> If possible, I'd love to use these CSV GRIB2 tables in two open-source
> GRIB2 projects that I'm collaborating on: gribberish[1] and
> hypergrib[2].
>
> In the short term, would it be acceptable to the GDAL devs for me to
> copy these CSV files into a new git repository (with some Rust code
> that reads the CSV files and provides a simple Rust API to the GRIB
> tables)? I'd include a README that explains that the CSV files came
> from GDAL, and that I can't take any credit for the CSV files, and
> that any tweaks to the CSV files should be submitted to GDAL (to avoid
> the copied CSV files diverging from GDAL's).
>
> Longer term, would it make sense to move GDAL's CSV GRIB2 tables &
> related Python code into a separate OSGeo github repository that GDAL
> and other GRIB2 readers could depend on? (No worries if not! I
> appreciate that you're very busy. I'm just keen to avoid copying the
> CSV files. But, if you're happy for me to copy the CSV files, then
> maybe there's less need to split the CSV files out from GDAL).
>
> Many thanks,
> Jack
>
> P.S. I'm afraid I may not be of much help in splitting the CSV files
> from GDAL because my knowledge of C++ and cmake is very very rusty!
>
> References
> 1. https://github.com/mpiannucci/gribberish
> 2. https://github.com/JackKelly/hypergrib
> _______________________________________________
> gdal-dev mailing list
> gdal-dev at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/gdal-dev

-- 
http://www.spatialys.com
My software is free, but my time generally not.
Mostly a fool, sometimes acting like a professional (or vice versa)