Ryan May
> Looking back at the relevant code, this should definitely be possible, but I suspect it may not be the wisest approach as a uniform principle (e.g., not sure we...
Code is here: https://github.com/Unidata/MetPy/blob/e9c41bdeba68ac9ec2e3a5f34e947181427c2532/src/metpy/interpolate/geometry.py#L94-L146 I should also note that there's an [open PR to scipy](https://github.com/scipy/scipy/pull/19665) adding support for inverse distance weighting interpolation to PR with custom kernels, so this might...
Can you share the data file somewhere so I can try to reproduce the issue?
In that notebook, what does: ```python import sys sys.path ``` show?
I think the delayed warning isn't ideal but, as you pointed out, is par for the course with Dask, so that's fine.
Thanks for starting the issue. A few thoughts I have right now: > To what extent will MetPy support Dask Arrays? Should it be limited to certain modules? I think...
Just to clarify something I said: Complexity is always an option in terms of MetPy implementation; we just don't want to present any more complexity to the users than is...
Here it's using OPeNDAP, so it should be possible to only download data on demand. However, I'm not really clear on how to engage dask with a single dataset loaded...
@rpmanser Well while we might not be looking to use Dask internally as a performance solution, we definitely do want to support its use for others who have gone to...
@lpilz The problem is that dask doesn't support the following in general: ```python a[a > 2] += 4 ``` I haven't dug in to figure out why it doesn't or...