[BUG] heatmap output shifts relative to axis when using logarithmic xscale and setting xlims on GR

[LordCMonkey]

Details

When plotting with heatmap(x, y, z::matrix) and a logarithmic xscale, setting xlims causes the heatmap output to shift to an incorrect position on the x axis.
The same happens when first plotting using heatmap(x, y, z::matrix) and then plotting 2D data, even without explicitly setting xlims, The 2D data is plotted in the correct location.
This also occurs when using :ln or :log2 and scatter instead of plot.

The expected output can be observed when using plotly or pyplot (for :log10).

Minimal working example:

using Plots

function test_bug()
    x = [10.0^x for x in 0:1]
    A = zeros(2, 2)

    heatmap(x, 1:2, A, xscale=:log10,
            xlims=[10.0^-1, 10.0^2]
    )
    #plot!(x, x -> 1.8, label="")
end

function demonstrate()
    gr()
    gui(test_bug())
    
    plotly()
    gui(test_bug())
    
    pyplot()
    gui(test_bug())
end

When commenting the xlims=.. line and uncommenting the plot! call, the same behaviour occurs.

Backends

This bug occurs on ( insert x below )

Backend	yes	no	untested
gr (default)	x
pythonplot		x
plotlyjs		x
pgfplotsx			x
unicodeplots			x
inspectdr			x
gaston			x

Versions

Plots.jl version: 1.40.2
Backend version: GR: 0.73.3, PlotlyBase 0.8.19 & PlotlyKaleide 2.2.4, PyPlot 2.11.2
Output of versioninfo():

Julia Version 1.10.2
Commit bd47eca2c8 (2024-03-01 10:14 UTC)
Build Info:
Official https://julialang.org/ release
Platform Info:
OS: Windows (x86_64-w64-mingw32)
CPU: 4 × Intel(R) Core(TM) i5-7400 CPU @ 3.00GHz
WORD_SIZE: 64
LIBM: libopenlibm
LLVM: libLLVM-15.0.7 (ORCJIT, skylake)
Threads: 4 default, 0 interactive, 2 GC (on 4 virtual cores)
Environment:
JULIA_NUM_THREADS = auto

[ktitimbo]

This is still an issue

Plots v1.40.4
GR v0.72.8

I adapted an example from here, and I now get

x = collect(LinRange(0.001,4,1001))
y = 1:10
z = exp.(-x) * y'
xlog10 = log10.(x)

plot(
    heatmap(x, y, z, xscale = :log10, title = "Ideal", color=:plasma),
    heatmap(xlog10, y, z, xformatter = xx -> "10^{$(Int(xx))}", title = "Non-ideal", color=:plasma)
)

The "ideal" construction doesn't produce the right output

[pkalaczynski]

You need to transpose() your matrix, this should yield correct result

[angerpointnerd]

Still appears to be an issue (just ran it with Plots v.1.41.2 and Julia 1.10.10).

However, it seems to be an issue with GR itself (?). When I run this code (without Plots.jl) I get the same behavior. I'll make an issue there as well (EDIT: link).

# `add GR` in a temporary environment (installed v0.73.22)

using GR

x = exp10.(-10:1:-1)
y = 0.1:0.1:0.9
z = rand(9, 10)

# looks correct (up to the alignment of the data and the x-axis being slightly off)
heatmap(x, y, z; xlog=true)

# looks the same (limits correspond to the actual extent of the data)
heatmap(x, y, z; xlog=true, xlim=(1e-10, 1e-1))

# data gets "compressed" at the lower end of the x-axis (upper limit greater than actual max x)
heatmap(x, y, z; xlog=true, xlim=(1e-10, 1e0))

# data gets "stretched" beyond the upper end of the x-axis (upper limit smaller than actual max x)
heatmap(x, y, z; xlog=true, xlim=(1e-10, 1e-2))

[angerpointnerd]

By the way, the issue with the stretched/squeezed pixels along the logarithmic direction has been fixed upstream in GR.jl.

Not sure if this solves the original issue, but using gr() the output of OP's example looks correct to me now.