Infinite warnings while using mask

[ayushpatnaikgit]

I am getting infinite warnings while using Rasters.mask

using Rasters, Shapefile, DataFrames
raster = Raster("DATA/Harmonized_DN_NTL_2020_simVIIRS.tif")
shp = Shapefile.Table("DATA/world_shapefile/in_primary_countries.shp") |> DataFrame
japan_index = findall( x -> x == "Japan", shp.name)
japan = crop(raster; to = shp.geometry[japan_index])
japan_masked = mask(japan, with = shp.geometry[japan_index]) 

┌ Warning: (Band,) dims were not found in object
└ @ DimensionalData.Dimensions ~/.julia/packages/DimensionalData/yMNeq/src/Dimensions/primitives.jl:648
┌ Warning: (Band,) dims were not found in object
└ @ DimensionalData.Dimensions ~/.julia/packages/DimensionalData/yMNeq/src/Dimensions/primitives.jl:648
┌ Warning: (Band,) dims were not found in object
└ @ DimensionalData.Dimensions ~/.julia/packages/DimensionalData/yMNeq/src/Dimensions/primitives.jl:648
┌ Warning: (Band,) dims were not found in object
└ @ DimensionalData.Dimensions ~/.julia/packages/DimensionalData/yMNeq/src/Dimensions/primitives.jl:648

I am also getting these while using Rasters.zonal

zonal(sum, raster; of=shp.geometry)

The performance also seems bad and variable. On some occasions, it takes 30 seconds and on some, it takes up to an hour. I am unable to replicate that, perhaps it's related to my memory usage.

For replication, the tif file can be obtained here. And the Shapefile can be obtained here.

I've also tried the same with a different tif file.

PS:

I am making a tutorial on using Rasters, and this is something I found while making the Jupyter notebook for it.
https://github.com/xKDR/datascience-tutorials/blob/main/rasters.ipynb

[rafaqz]

I think this is already fixed on master, and the 0.4.5 patch version should be registered in the next half hour.

Good work on the notebook! will this be in a blog post or something similar?

[ayushpatnaikgit]

Is it fixed! The speed is spectacular.

I've been teaching raster analysis at a few places, so I thought I'll make a notebook + video and circulate it.