密谋-不同表面的不同颜色 [英] plotly - different colours for different surfaces

问题描述

使用plotly我希望每个表面都具有不同的颜色.

Using plotly I would like to have each surface to have different colour.

library(plotly)
t1 <- seq(-3, 3, 0.1); t2 <- seq(-3, 3, 0.1)

p1 <- matrix(nrow = length(t1), ncol = length(t2))
p2 <- matrix(nrow = length(t1), ncol = length(t2))

p8a1 <- 1.2
p8a2 <- 1
p8d <- -1
p8b1 <- 0.7
p8b2 <- 0.6

for (i in 1:length(t2)) {
   for (j in 1:length(t1)) {
      p1[i, j] <- 1 / (1 + exp(-1.7 * (p8a1 * t1[j] + p8a2 * t2[i] + p8d)))
      p2[i, j] <- (1 / (1 + exp(-1.7 * p8a1 * (t1[j]- p8b1)))) * 
                  (1 / (1 + exp(-1.7 * p8a2 * (t2[j]- p8b2))))
   }
}

df1 <- list(t1, t2, p1)
df2 <- list(t1, t2, p2)

names(df1) <- c("t1", "t2", "p1")
names(df2) <- c("t1", "t2", "p2")
m <- list(l = 10, r = 10, b = 5, t = 0, pad = 3)

p <- plot_ly(color = c("red", "blue")) %>%
     add_surface(x = df1$t1,
                 y = df1$t2,
                 z = df1$p1,
                 opacity = 0.8) %>%
     add_surface(x = df2$t1,
                 y = df2$t2,
                 z = df2$p2,
                 opacity = 1) %>%
     layout(autosize = F, width = 550, height = 550, margin = m,
            scene = list(xaxis = list(title = "Theta 1"),
                         yaxis = list(title = "Theta 2"),
                         zaxis = list(title = "P")),
            dragmode = "turntable")
p

不幸的是，我无法更改这两个表面的颜色.我尝试将color = I("red")和color = I("blue")参数添加到add_surface中，但这只是将两个表面的色标从红色更改为蓝色.

Unfortunately, I'm not able to change colours of these two surfaces. I tried to add color = I("red") and color = I("blue") arguments into add_surface but this just changed colour scale from red to blue for both surfaces.

我还尝试将color = "red"添加到plot_ly()并将inherit = F添加到第二个add_surface.这仅将第一个曲面更改为红色，但仅将黄色默认颜色更改为红色.我很想让一个表面红色，第二个表面蓝色.

I also tried to add color = "red" into plot_ly() and add inherit = F into second add_surface. This changed the first surface only, but only the yellow default color into red. I would love to have one surface red and second one blue.

推荐答案

听起来微不足道，但在Plotly中有点棘手.表面图的颜色是从z值或与z尺寸相同的数组中得出的.此颜色数组仅接受数字值，不接受颜色字符串或RGB值.

Sounds trivial but it's a bit tricky in Plotly. The color of a surface plot is either derived from the z values or from an array with the same dimensions as z. This color array only accepts numerical values, no color strings or RGB values.

所以我们为颜色定义一个数组

So let's define an array for our colors

color <- rep(0, length(df1$p1))
dim(color) <- dim(df1$p1)

接下来，我们需要欺骗Plotly忽略色标.

Next we need to trick Plotly into ignoring the colorscale.

surfacecolor=color,
             cauto=F,
             cmax=1,
             cmin=0

等等，我们有一个颜色统一的情节.

et voilà, we have a uniformely colored plot.

library(plotly)
t1 <- seq(-3, 3, 0.1); t2 <- seq(-3, 3, 0.1)

p1 <- matrix(nrow = length(t1), ncol = length(t2))
p2 <- matrix(nrow = length(t1), ncol = length(t2))

p8a1 <- 1.2
p8a2 <- 1
p8d <- -1
p8b1 <- 0.7
p8b2 <- 0.6

for (i in 1:length(t2)) {
  for (j in 1:length(t1)) {
    p1[i, j] <- 1 / (1 + exp(-1.7 * (p8a1 * t1[j] + p8a2 * t2[i] + p8d)))
    p2[i, j] <- (1 / (1 + exp(-1.7 * p8a1 * (t1[j]- p8b1)))) * 
      (1 / (1 + exp(-1.7 * p8a2 * (t2[j]- p8b2))))
  }
}

df1 <- list(t1, t2, p1)
df2 <- list(t1, t2, p2)

names(df1) <- c("t1", "t2", "p1")
names(df2) <- c("t1", "t2", "p2")
m <- list(l = 10, r = 10, b = 5, t = 0, pad = 3)

color <- rep(0, length(df1$p1))
dim(color) <- dim(df1$p1)
p <- plot_ly(colors = c('red', 'blue')) %>%
  add_surface(x = df1$t1,
              y = df1$t2,
              z = df1$p1,
              opacity = 0.8,
              #surfacecolor=c('red')
              surfacecolor=color,
              cauto=F,
              cmax=1,
              cmin=0
  )
color2 <- rep(1, length(df2$p2))
dim(color2) <- dim(df2$p2 )

p <-  add_surface(p,
              x = df2$t1,
              y = df2$t2,
              z = df2$p2,
              opacity = 1,
              surfacecolor=color2,
              cauto=F,
              cmax=1,
              cmin=0)
p

这篇关于密谋-不同表面的不同颜色的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！