Dem gridded_data shifted coordinates #1682
Conversation
|
Here the PR for salem fmaussion/salem#234 |
oggm/tests/test_prepro.py
Outdated
| assert np.allclose(gridded_topo.data, gtiff_ds.data) | ||
|
|
||
| # compare coordinates of topo.tif with dem.tif | ||
| demtiff_ds = salem.open_xr_dataset(gdir.get_filepath('dem')) |
There was a problem hiding this comment.
Can we try to avoid using salem here, but use rioxarray instead? It might be possible to use rioxarray to open but still use salem's accessor.
There was a problem hiding this comment.
Yes, I can use rioxarray to open the tiff files and compare the coordinates, I will change this. But I can not use salem's accessor because their is no variable .pyproj_srs added to the dataset when opening with rioxarray.
There was a problem hiding this comment.
OK yes I hoped that the coords would be enough, but indeed the proj info is missing - no worries, that's enough as is for now.
|
Ok, now the tiff-files are opened with rioxarray and this PR is now independent of the bug in salem and can be merged if all tests pass. |
|
It seems that for the minimal test |
|
Sorry for the mess @pat-schmitt - just ignore and keep it simple, as long as it worked locally we should be good to go |
|
Yes, it also worked with the other containers, so it is ready to merge. However, we should add |
|
Yes sorry I was not clear: adding rioxarray to minimal defies the purpose of the minimal environment (which is to have no GIS library in it). To this test should rather be moved to tests_prepro, or altered/commented for simplicity. |
|
I used now |
|
Thanks @pat-schmitt ! And sorry for the confusion |
This fixes a coordinate shift between
dem.tifand thegridded_data.nc.However, there seems to be a bug in salem when opening a .tif-file using
ds = salem.open_xr_dataset(tiff_path)and callingds.salem.gridafterwards. I will open a PR for salem and link it here.Due to this bug, one test is currently not working and I commented it out. After salem is fixed the test can be uncommented again (and maybe adapted due to floating point problems...).
Closes #1679
whats-new.rst