Romain Hugonnet
Romain Hugonnet
I think the warnings (still 4000 of them in the CI) come mostly from the default parameters in `xdem.spatialstats` that use the deprecated `nmad()` defined right above, like here: https://github.com/GlacioHack/xdem/blob/e1e2065ea52a7db3c8e88045d743aa1f6efd08bb/xdem/spatialstats.py#L82...
Thanks, we just need to add back the original function as explained above, then we can merge this :slightly_smiling_face:.
Agreed! For deprecation warnings: The filters were neither in the API nor in the doc, we can remove without warnings 😉. The OpenCV removal: Yes, we can remove associated filter...
I was about to write my feedback in the issue following previous comments there (https://github.com/GlacioHack/xdem/issues/692#issuecomment-2727557256), but seeing @adehecq's comment I'll follow up here as well. I already detailed a lot...
Thinking a bit more about this, a couple additional points: 1. To bin in Pandas, one always need to combine `pd.groupby()` with `pd.cut()`. In Xarray, they simplified that by adding...
So the question we need to answer to move forward is: What would be missing in the functionalities of such a `Raster.grouped_statistic()` method to justify building a class wrapper on...
@belletva Thanks for wrapping things up! :slightly_smiling_face: It made me realize I need to clarify further: All my points above are not about proposing enhancements, they relate to the core...
I did not fully realize that under "metrics" were many different functions: from a single value metric (mean, STD), to a distribution (PDF) and visualization (hillshade). I think we might...
I've been working on a very similar problems during the past week in https://github.com/GlacioHack/xdem/pull/486, so while I still know the intricacies of the topic, here's how we could add Numba...