Hi, I’d like to plot an heatmap using Makie (I don’t care about the backend). The data are stored in a 2D matrix, and the x/y dimensions are actually categorical, represented with a vector of strings. In Plots, I can do something like
julia> program_names = [t.program for t in results]
58-element Vector{String}:
"bench_2mm_core2_1_100_100"
"bench_2mm_orig"
"bench_3mm_core2_1_100_100"
"bench_3mm_orig"
"bench_adi_core2_1_100_100"
"bench_adi_orig"
"bench_atax_core2_1_100_100"
"bench_atax_orig"
"bench_bicg_core2_1_100_100"
"bench_bicg_orig"
"bench_cholesky_core2_1_100_100"
"bench_cholesky_orig"
"bench_covariance_core2_1_100_100"
"bench_covariance_orig"
"bench_deriche_core2_1_100_100"
"bench_deriche_orig"
⋮
"bench_mvt_core2_1_100_100"
"bench_mvt_orig"
"bench_nussinov_core2_1_100_100"
"bench_nussinov_orig"
"bench_seidel_2d_core2_1_100_100"
"bench_seidel_2d_orig"
"bench_symm_core2_1_100_100"
"bench_symm_orig"
"bench_syr2k_core2_1_100_100"
"bench_syr2k_orig"
"bench_syrk_core2_1_100_100"
"bench_syrk_orig"
"bench_trisolv_core2_1_100_100"
"bench_trisolv_orig"
"bench_trmm_core2_1_100_100"
"bench_trmm_orig"
julia> distance_matrix_measures = pairwise_difference(measures_vec)
58×58 Matrix{Quantity{Float64, 𝐋^2 𝐌 𝐓^-3, Unitful.FreeUnits{(W,), 𝐋^2 𝐌 𝐓^-3, nothing}}}:
0.0 W -0.138593 W … -0.000670795 W -0.152928 W
0.138593 W 0.0 W 0.137922 W -0.0143344 W
0.000644075 W -0.137949 W -2.67202e-5 W -0.152284 W
0.153729 W 0.0151362 W 0.153059 W 0.000801762 W
0.000987496 W -0.137606 W 0.000316701 W -0.15194 W
0.000341562 W -0.138252 W … -0.000329233 W -0.152586 W
0.00310015 W -0.135493 W 0.00242935 W -0.149828 W
0.149652 W 0.011059 W 0.148981 W -0.00327543 W
0.00110525 W -0.137488 W 0.000434454 W -0.151822 W
0.132496 W -0.00609756 W 0.131825 W -0.020432 W
0.145352 W 0.00675866 W … 0.144681 W -0.00757578 W
0.139371 W 0.000777863 W 0.1387 W -0.0135566 W
0.00121418 W -0.137379 W 0.000543384 W -0.151714 W
0.151652 W 0.0130586 W 0.150981 W -0.00127589 W
0.118514 W -0.0200795 W 0.117843 W -0.0344139 W
0.151647 W 0.0130542 W … 0.150977 W -0.00128026 W
⋮ ⋱
0.000762225 W -0.137831 W 9.14301e-5 W -0.152165 W
0.149316 W 0.0107225 W 0.148645 W -0.00361189 W
0.000843365 W -0.13775 W 0.00017257 W -0.152084 W
0.13425 W -0.00434351 W … 0.133579 W -0.0186779 W
0.00074294 W -0.13785 W 7.21451e-5 W -0.152185 W
0.158092 W 0.0194988 W 0.157421 W 0.00516438 W
0.000947786 W -0.137645 W 0.000276991 W -0.15198 W
0.132462 W -0.00613168 W 0.131791 W -0.0204661 W
0.00110807 W -0.137485 W … 0.000437276 W -0.15182 W
0.1351 W -0.00349302 W 0.134429 W -0.0178275 W
0.000920184 W -0.137673 W 0.000249388 W -0.152007 W
0.150186 W 0.0115928 W 0.149515 W -0.00274168 W
0.00379128 W -0.134802 W 0.00312049 W -0.149136 W
0.149627 W 0.0110334 W … 0.148956 W -0.00330104 W
0.000670795 W -0.137922 W 0.0 W -0.152257 W
0.152928 W 0.0143344 W 0.152257 W 0.0 W
julia> p_measures = heatmap(
program_names,
program_names,
distance_matrix_measures,
)
and the plot has the string as labels. In Makie I can’t figure out a way to do it. The best I could do is to tweak xtickformat = v -> program_names, but then the labels are not aligned. Is it possible to do better? Thanks