Erik Schytt Mannerfelt

Thanks for taking on the problem so fast! If it helps, here are the messages that show up when running `ksnip` from the terminal: ``` ❯ ksnip Critical: Unknown platform,...

This is most likely due to the `dilate_mask` argument, which we deemed was a preferable behavior over potentially introducing weird artefacts. I'm not saying this isn't an issue, just that...

That's a good idea. A small problem right now (or an unintended feature) is that all metadata are lost after raster recreation. Only the CRS and transform are identical to...

Good idea! I would first prefer to have some kind of synthetic approach which could be parametrized over different nan cases, sizes, and other settings. I know I'm talking a...

What about squeezing automatically then for GeoUtils compatibility? So (1, height, width) automatically becomes (height, width). Also, a dtype check could maybe be done: make sure it's a float32 or...

Nice! I can try to get it running once #87 is fixed (since it changes the Coreg structure quite considerably).

Hey, we're up to 100 issues! That sounds so bad in any non-GitHub context haha but here I think it's a great achievement!

What exactly do you mean with vectorize? Speed up? The np.vectorize decorator won't work as it's using numba, not numpy for the brunt of the work.

It seems like it struggles with many nans. This has to be checked further! EDIT: Also @rhugonnet , you have my permission to say "I told you so!", but it...

That would be an awesome feature!