Mathias Hauser

Proof of concept code (not nice yet, and not yet fully functional), requires xarray-contrib/cf-xarray#478. ```python import cf_xarray import matplotlib.pyplot as plt import numpy as np import shapely import xarray as...

Thanks for using regionmask and opening your first PR! Unfortunately I don't think its the correct approach for the problem. The offset needs to be as small as possible and...

I thought some more about how to best fix #410. It would be possible to shift the grid points more, but I think it's cleaner to fix the coordinates of...

> I tested the newest PR you did in #488 (using the testing code above), the edge issue at lon=180 persists even I use the latest natural_earth_v5_1_2, although it fixes...

The bug should be fixed in gdal 3.7.2 — can try again - ah now that's wrong - it's not yet in 3.7.2. I have to wait some more, probably...

It now works with gdal 3.7.3 :slightly_smiling_face: (why the hell is this the `slightly_smiling_face`) Should that be accessed via a param or get its own method? ```python def mask_3D( lon_or_obj,...

I think I now prefer `mask_3D_all_touched`.

More upstream issues OSGeo/gdal#8918 -> already fixed (unbelievable) The threshold to include the gridcell is 1e-4 https://github.com/OSGeo/gdal/blob/0d4d8757090bb1a992f9431427c2a8b0e92330a7/alg/llrasterize.cpp#L412

Will probably need to update `git update-index --assume-unchanged ci/requirements/docs.yml docs/conf.py` And move `pip install -e .` to `post_install` - the order seems different.

It is faster (by about 3 min) but currently fails & I don't have the nerves to debug this right now. I would probably need to pin the versions better,...