mhm-ufz / basinex Goto Github PK

Basinex seems to be not compatible with netcdf4 1.6.

Quick fix for now is applied in these two commits commit 1c5dd35 & commit a17c4a6

Basinex has some issue on EVE. Running the example case throws the error messages as shown in the screenshot below.

Tagging @MuellerSeb and @StephanThober

p.s. The problem persists while testing @lotruheawea 's fork as well.

When I init a new instance of a geoarray, e.g. by its wrapper calls full_like, a new geoarray is created. BUT some attributes are copied over from the like-array, which makes sense but not for _fobj. This is done in geoarray/core.py, line 281, fobj is passed on. But in the write section of geoarray/gdalio.py line 160, _fobj is taken by default for creating the output GDAL bands from the new geoarray.

This means that all the previous data manipulation is pointless as the original array's _fobj is taken.

This should be changed, likely by removing _fobj from the argument list in geoarray/core.py.

(It took my like 2 hours to figure that out. I also assume that the current behavior leads to bugs.)

BasinEx fails at some entries for me. Funnily these entries pass in older version of Basin Extractor when used by Olda.

The error message that I get is as shown in the screenshot below.

The point to be noted here is that the facc, fdir are at ULYSSES resolution of 0.1. I am not sure how much that plays a role here in BasinEx failing.

I have placed the failing test case here /work/shresthp/basinex_testcase/. PS> It doesn't fail at all entries.

Tagging @StephanThober for updates.