Makie image recipe is putting the data on the wrong lat lon position

[felixcremer]

When I plot a Raster with image that gives a wrong position. I can use heatmap to properly plot the data. The mismatch depends on some properties of the dimensions.
Below is an example script and I will attach the exemplary nc file. I would have expected that the position of the rectangle in the corner would be the same for heatmap and for image.
I could not reproduce the mismatch with a simple reverse ordered Y-Axis.

julia> r = Raster("./image_vs_heatmap_corner.nc")
[ Info: No `name` or `key` keyword provided, using first valid layer with name `:unnamed`
559×561 Raster{Union{Missing, Float32},2} unnamed with dimensions: 
  X Mapped{Float64} Float64[635090.3685878773, 636033.1138663981, …, 1.1601994887240704e6, 1.1607739741281688e6] ForwardOrdered Irregular Points crs: EPSG mappedcrs: EPSG,
  Y Mapped{Float64} Float64[6.306992429753325e6, 6.306049684474804e6, …, 5.779997819060089e6, 5.779202377731337e6] ReverseOrdered Irregular Points crs: EPSG mappedcrs: EPSG
extent: Extent(X = (635090.3685878773, 1.1607739741281688e6), Y = (5.779202377731337e6, 6.306992429753325e6))
missingval: missing
crs: EPSG:4326
mappedcrs: EPSG:4326
parent:
            6.30699e6  6.30605e6  6.30511e6  6.30416e6  …  5.78188e6  5.78094e6  5.78e6  5.7792e6
 6.3509e5   1.0        1.0        1.0        1.0           0.0        0.0        0.0     0.0
 6.36033e5  1.0        1.0        1.0        1.0           0.0        0.0        0.0     0.0
 ⋮                                                      ⋱                                
 1.1602e6   0.0        0.0        0.0        0.0        …  0.0        0.0        0.0     0.0
 1.16077e6  0.0        0.0        0.0        0.0           0.0        0.0        0.0     0.0

julia> image(r, interpolate=false)


julia> heatmap(r, interpolate=false)

julia> figim, axim, im = image(r, interpolate=false)

julia> heatmap(r, interpolate=false, colormap=:greys)

julia> display(GLMakie.Screen(), figim)
GLMakie.Screen(...)

image_vs_heatmap_corner.zip

I had to zip it because of github.

[rafaqz]

Weird that they are different, will have to look into it.

But possibly we can ignore this now as all the lower level plotting will move to
https://github.com/rafaqz/DimensionalData.jl/pull/

And just make sure they match over there

[felixcremer]

This is fixed on Rasters main with DD main.