Improve model and transform_func interaction with Axis limits #3864
Conversation
Compile Times benchmarkNote, that these numbers may fluctuate on the CI servers, so take them with a grain of salt. All benchmark results are based on the mean time and negative percent mean faster than the base branch. Note, that GLMakie + WGLMakie run on an emulated GPU, so the runtime benchmark is much slower. Results are from running: using_time = @ctime using Backend
# Compile time
create_time = @ctime fig = scatter(1:4; color=1:4, colormap=:turbo, markersize=20, visible=true)
display_time = @ctime Makie.colorbuffer(display(fig))
# Runtime
create_time = @benchmark fig = scatter(1:4; color=1:4, colormap=:turbo, markersize=20, visible=true)
display_time = @benchmark Makie.colorbuffer(fig)
|
| ax, plot_top = heatmap(fig[1, 1], randn(10, 10), colormap = [:transparent]) | ||
| # Plot some recipes at the level below the contour | ||
| scatterlineplot_1 = scatterlines!(plot_top, 1:10, 1:10; linewidth = 20, markersize = 20, color = :red) | ||
| scatterlineplot_2 = scatterlines!(plot_top, 1:10, 1:10; linewidth = 20, markersize = 30, color = :blue) | ||
| scatterlineplot_1 = scatterlines!(ax, 1:10, 1:10; linewidth = 20, markersize = 20, color = :red) | ||
| scatterlineplot_2 = scatterlines!(ax, 1:10, 1:10; linewidth = 20, markersize = 30, color = :blue) |
There was a problem hiding this comment.
This test failed because of what I'd say is undefined behavior - plotting to a primitive plot.
Specifically because data_limits has a specific overload for Heatmap, thus using heatmap limits, while boundingbox attempts to recognize primitives by their lack of child plots. Since heatmap has children here, boundingbox uses scatter + lines for limits.
Plotting to a primitive also results in other weirdness - the heatmap doesn't show and the axis uses tight limits.
There was a problem hiding this comment.
Oh, plotting to a primitive outside a recipe?! That's something we shouldn't do in any case ;) Didn't realize we actually do this in the tests!
|
We should merge #3856 first, rebase and let tests run here before merging |
Description
This is an attempt at resolving issues with model transformations (
scale!,translate!) in Axis with or without tranform_func by relying onboundingboxinstead ofdata_limitsand undoing the transform_func afterwards. My gut feeling is that this will result in domain errors for some transformations in some situations...This should fix #3562 and similar issues.
Based on #3856 since there are some overlapping issues (hexbin typo, model application with NaN)
Type of change
Checklist