了解用于处理色边距的scale_fill_continuous_divergingx参数输入

这个问题是我上一个问题的延续here https://stackoverflow.com/questions/58718527/setting-midpoint-for-continuous-diverging-color-scale-on-a-heatmap.

我有一个带有可用数据集的热图。数据集粘贴如下：

library(ggplot2)
library(colorspace)

bigtest <- structure(list(x = c(-8, -7, -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8,
                                -8, -7, -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, -8,
                                -7, -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, -8, -7,
                                -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, -8, -7, -6,
                                -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, -8, -7, -6, -5,
                                -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8),
                          y = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
                                 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
                                 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3,
                                 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4,
                                 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5,
                                 5, 5, 5, 5, 5, 5, 5, 5),
                          z = c(1281.35043, 576.76381, 403.46607,
                                363.28815, 363.13356, 335.04997, 246.93314, 191.56371, 165.35087,
                                165.35087, 136.33712, 83.91203, 107.5773, 56.91087, 56.91089,
                                54.16559, 54.18172, 1841.60838, 1098.66304, 424.80686, 363.52776,
                                363.13355, 335.04998, 246.93314, 191.69473, 165.35087, 165.35087,
                                136.33712, 83.91204, 107.57729, 56.91087, 56.91088, 54.16421,
                                54.16794, 2012.52217, 1154.7927, 446.79023, 363.31379, 363.13356,
                                335.04997, 246.93314, 191.9613, 165.35087, 165.35087, 136.33712,
                                83.91202, 107.57731, 56.91088, 56.91088, 54.1642, 54.16559, 2077.10354,
                                1217.43403, 450.18301, 363.44225, 363.13357, 363.13363, 253.99753,
                                218.43223, 165.35087, 165.35014, 136.33712, 83.91203, 107.57822,
                                82.87399, 56.91087, 54.1642, 54.1642, 2092.56391, 1229.49925,
                                451.15179, 392.30728, 363.13356, 363.13282, 264.18944, 218.4308,
                                165.35087, 165.35044, 136.33712, 83.91202, 83.92709, 82.87353,
                                82.87406, 56.54491, 54.16421, 2206.93318, 1231.66411, 457.37767,
                                392.41558, 363.13357, 363.13283, 335.06272, 191.95211, 165.35087,
                                165.35014, 136.33712, 136.35211, 112.12755, 82.73634, 82.87353,
                                82.87418, 54.16421)),
                     row.names = c(NA, -102L),
                     class = c("tbl_df", "tbl", "data.frame"))

我正在使用以下代码部分生成热图：

ggplot(bigtest, aes(x = x, y = y)) +
  geom_tile(aes(fill = z)) +
  scale_fill_continuous_divergingx(palette = 'RdBu', rev = TRUE, mid = 347.48, l3 = 54, p3 = 2206, p4 = 325)

我对图中的期望是白色以特定值为中心，而其他渐变则根据高于或低于该值而发散。但是，通过使用不同的参数，我似乎并不完全理解参数的含义l3, p3, and p4所指的。当我在审查文档 https://rdrr.io/cran/colorspace/man/scale_colour_continuous_divergingx.html对于此功能，建议自定义比例的参数来自divergingx_hclcolorspace 包中的函数。

审查时divergingx_hcl 文档 https://rdrr.io/cran/colorspace/man/divergingx_hcl.html它指出它们对应于不同的输入参数的坐标。我完全迷失了，完全不知道这是什么。任何帮助我理解这些参数的指导（不仅仅是l3, p3, and p4但其他参数）将不胜感激。

^{Created on 2019-11-07 by the reprex package https://reprex.tidyverse.org (v0.3.0)}

首先，所有的颜色都被指定为HCL（色调、彩度、亮度），分别对应颜色的类型（红、绿、蓝等）、颜色的鲜艳度（低彩度是灰色，高彩度是非常鲜艳的） ），以及颜色的亮度（高亮度为白色，低亮度为黑色）。

参数l3表示色阶一端颜色的亮度分量。 （l1是另一端的亮度，并且l2是中间的亮度。）亮度从 0 到 100。因此，如果您希望末端的颜色更暗，请将亮度设置为较低的值。参数p3 and p4是控制颜色从中点过渡到端点的速度的指数。一般来说，接近 0 的值意味着更快的转换，大于 1 的值意味着更慢的转换。你不太可能想要p3 or p4值大于 10。

要获取调色板的默认参数，您可以使用divergingx_palettes()命令：

library(colorspace)
divergingx_palettes('RdBu')
#> HCL palette
#> Name: RdBu
#> Type: Diverging (flexible)
#> Parameter ranges:
#>  h1 h2  h3 c1 c2 c3 l1 l2 l3  p1
#>  20 NA 230 60  0 50 20 98 15 1.4

^{Created on 2019-11-07 by the reprex package https://reprex.tidyverse.org (v0.3.0)}

这表明由指定的终点处的颜色l3天已经很黑了。改变l3从 15 到 0 会使颜色变暗一点，但不会太多。更远，p2, p3, and p4没有指定，这意味着它们都取自p1，因此是 1.4。因此，颜色插值比线性插值稍慢。

有了这些知识，下面的例子就应该有意义了。要了解更多信息，我建议阅读色彩空间网站上的各种文章：http://colorspace.r-forge.r-project.org/ http://colorspace.r-forge.r-project.org/

首先是数据：

library(ggplot2)
library(colorspace)

bigtest <- structure(list(x = c(-8, -7, -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8,
                                -8, -7, -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, -8,
                                -7, -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, -8, -7,
                                -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, -8, -7, -6,
                                -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, -8, -7, -6, -5,
                                -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8),
                          y = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
                                0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
                                1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3,
                                3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4,
                                4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5,
                                5, 5, 5, 5, 5, 5, 5, 5),
                          z = c(1281.35043, 576.76381, 403.46607,
                                363.28815, 363.13356, 335.04997, 246.93314, 191.56371, 165.35087,
                                165.35087, 136.33712, 83.91203, 107.5773, 56.91087, 56.91089,
                                54.16559, 54.18172, 1841.60838, 1098.66304, 424.80686, 363.52776,
                                363.13355, 335.04998, 246.93314, 191.69473, 165.35087, 165.35087,
                                136.33712, 83.91204, 107.57729, 56.91087, 56.91088, 54.16421,
                                54.16794, 2012.52217, 1154.7927, 446.79023, 363.31379, 363.13356,
                                335.04997, 246.93314, 191.9613, 165.35087, 165.35087, 136.33712,
                                83.91202, 107.57731, 56.91088, 56.91088, 54.1642, 54.16559, 2077.10354,
                                1217.43403, 450.18301, 363.44225, 363.13357, 363.13363, 253.99753,
                                218.43223, 165.35087, 165.35014, 136.33712, 83.91203, 107.57822,
                                82.87399, 56.91087, 54.1642, 54.1642, 2092.56391, 1229.49925,
                                451.15179, 392.30728, 363.13356, 363.13282, 264.18944, 218.4308,
                                165.35087, 165.35044, 136.33712, 83.91202, 83.92709, 82.87353,
                                82.87406, 56.54491, 54.16421, 2206.93318, 1231.66411, 457.37767,
                                392.41558, 363.13357, 363.13283, 335.06272, 191.95211, 165.35087,
                                165.35014, 136.33712, 136.35211, 112.12755, 82.73634, 82.87353,
                                82.87418, 54.16421)),
                     row.names = c(NA, -102L),
                     class = c("tbl_df", "tbl", "data.frame"))

现在的情节：

ggplot(bigtest, aes(x = x, y = y)) +
  geom_tile(aes(fill = z)) +
  scale_fill_continuous_divergingx(
    palette = 'RdBu', rev = TRUE,
    mid = 347.48
  )

ggplot(bigtest, aes(x = x, y = y)) +
  geom_tile(aes(fill = z)) +
  scale_fill_continuous_divergingx(
    palette = 'RdBu', rev = TRUE,
    mid = 347.48,
    p3 = .2,
    p4 = .2
  )

ggplot(bigtest, aes(x = x, y = y)) +
  geom_tile(aes(fill = z)) +
  scale_fill_continuous_divergingx(
    palette = 'RdBu', rev = TRUE,
    mid = 347.48,
    l3 = 0,
    p3 = .2,
    p4 = .2
  )

^{Created on 2019-11-07 by the reprex package https://reprex.tidyverse.org (v0.3.0)}