search

NOAA-EMC / wgrib2

Provides functionality for interacting with, reading, writing, and manipulating GRIB2 files.
27 stars 13 forks source link

build just one version of wgrib2 which handles rotated lat/lon #70

Open edwardhartnett opened 9 months ago

edwardhartnett commented 9 months ago

Build just one version of wgrib2 which handles rotated lat/lon.

edwardhartnett commented 3 months ago

@webisu was this done as part of the recent ip changes?

webisu commented 3 months ago

No, Alex needs to fix ip.

On Thu, Jul 4, 2024 at 10:03 AM Edward Hartnett @.***> wrote:

@webisu https://github.com/webisu was this done as part of the recent ip changes?

— Reply to this email directly, view it on GitHub https://github.com/NOAA-EMC/wgrib2/issues/70#issuecomment-2209080686, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEIB7ZUPGHQOXWIUVH5IYTDZKVIY5AVCNFSM6AAAAABKLT2HR6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMBZGA4DANRYGY . You are receiving this because you were mentioned.Message ID: @.***>

edwardhartnett commented 3 months ago

@AlexanderRichert-NOAA do you know the fix that @webisu needs?

AlexanderRichert-NOAA commented 3 months ago

I believe so yes, it looks like another instance of reconciling ip's and ip2's different "ideas" of rotated lat-lon grids. I should have a candidate fix soon.

AlexanderRichert-NOAA commented 2 months ago

The non-E grid is fixed. Right now I'm looking at the E grid code. The ip and ip2 repositories have some different logic from what's in the wgrib2 vendored copy of ip2. @webisu what version of ip2 is used in wgrib2? If I can figure out where/when the difference within ip2 came in between the two approaches for gdt=1 then hopefully it'll help me implement a clean fix that makes everybody happy.

AlexanderRichert-NOAA commented 2 months ago

@webisu can you please take a look at my fork again (https://github.com/NOAA-EMC/NCEPLIBS-ip/pull/241)? I've restricted the alternate rot lat lon scheme to only be used when gdtn=32769, which should hopefully take care of it. I also added an off switch for "unusing" that scheme. It may ultimately not be necessary if it's now differentating between gdtn=1 and gdtn=32769, but I'll leave it for now just to prevent risk of impact to other users. Anyway, please try the updated ip code again and let me know if it works.

edwardhartnett commented 1 month ago

We believe this is working, testing is underway...