plantain
plantain
add_matrix_NaNs doesn't work as a workaround since 0.2.0 >regridder.A = scipy.sparse.coo_matrix(M) >AttributeError: can't set attribute Does anyone have a new workaround?
Thanks for the suggestion, I will try that. It will be quite an overhead in my use cas though as I have very large grids. I wonder is the current...
How do we restart the kernel with the xESMF API? Or should that not leak memory?
That doesn't seem to behave as I expected, it still seems the regridder is never free'd. ``` Line # Mem usage Increment Line Contents ================================================ 4 65.508 MiB 65.508 MiB...
Seeing the same behaviour...
In the absence of a fix here, some documentation would be great - spent a long time trying to debug why my DDS weren't transitioning.
It happens on HEAD, and has been happening for quite a long time. Flags are: cmake -DCMAKE_BUILD_TYPE=Release -DBUILD_TESTING=OFF -DJPEGXL_BUNDLE_GFLAGS=YES Build log below: ``` Submodule 'third_party/testdata' (https://github.com/libjxl/testdata) registered for path 'testdata'...
I'm suspicious that it's the same problem as here: https://github.com/twitter/vireo/issues/8 Something about ARM systems / modern ARM compilers / modern glibc is defining boolean differently to x86.
This is hitting us as well. We store space-projected imagery and access that with rio-tiler to cut WebMercator tiles. 6.2.4+ gives us terrible picture quality. Reproduction case: Data: http://static.skysight.io/rio-bug/orig.tiff ```...
This is great. We're definitely going to use this.