我正在尝试使用ggplot2重现以下 [base R] 图
我已经完成了大部分工作，但目前让我感到困惑的是线段的放置，这些线段连接了带有相应标签的地块右侧的边缘地毯图。标签已被绘制（在下面的第二个图中）anotation_custom()，我使用了@baptiste 关闭剪辑的技巧以允许在绘图边距中绘制。
尽管进行了多次尝试，但我无法将它们放置segmentGrobs()在设备中的所需位置，以便它们加入正确的地毯刻度和标签。
一个可重现的例子是
y <- data.frame(matrix(runif(30*10), ncol = 10))
names(y) <- paste0("spp", 1:10)
treat <- gl(3, 10)
time <- factor(rep(1:10, 3))
require(vegan); require(grid); require(reshape2); require(ggplot2)
mod <- prc(y, treat, time)
如果您没有安装vegan，我dput会在问题末尾附加一个强化对象，fortify()如果您希望运行该示例并fortify()方便地使用ggplot(). 我还包括一个有点冗长的函数 ，myPlt()它说明了我到目前为止所做的工作，如果您加载了包并且可以创建mod.
我已经尝试了很多选项，但我现在似乎在正确放置线段时在黑暗中挣扎。
我不是在寻找为示例数据集绘制标签/段的特定问题的解决方案，而是我可以用来以编程方式放置段和标签的通用解决方案，因为这将autoplot()构成class(mod). 我的标签工作正常，只是线段没有。所以对于问题：
- 当我想放置一个包含从数据坐标到,运行的线的段 grob 时，如何使用
xmin,xmax,ymin,参数？ymaxx0y0x1y1 - 也许以不同的方式问，你如何使用在已知数据坐标之间的绘图区域之外
annotation_custom()绘制线段，到，？x0y0x1y1
我很乐意收到仅在绘图区域中有任何旧绘图但展示了如何在绘图边缘的已知坐标之间添加线段的答案。
我不喜欢使用，annotation_custom()所以如果有更好的解决方案，我也会考虑。我确实想避免在绘图区域中有标签；我想我可以通过参数使用annotate()和扩展刻度中的 x 轴限制来实现这一点expand。
fortify()方法_
fortify.prc <- function(model, data, scaling = 3, axis = 1,
...) {
s <- summary(model, scaling = scaling, axis = axis)
b <- t(coef(s))
rs <- rownames(b)
cs <- colnames(b)
res <- melt(b)
names(res) <- c("Time", "Treatment", "Response")
n <- length(s$sp)
sampLab <- paste(res$Treatment, res$Time, sep = "-")
res <- rbind(res, cbind(Time = rep(NA, n),
Treatment = rep(NA, n),
Response = s$sp))
res$Score <- factor(c(rep("Sample", prod(dim(b))),
rep("Species", n)))
res$Label <- c(sampLab, names(s$sp))
res
}
这dput()
这是来自的输出fortify.prc(mod)：
structure(list(Time = c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 1, 2,
3, 4, 5, 6, 7, 8, 9, 10, NA, NA, NA, NA, NA, NA, NA, NA, NA,
NA), Treatment = c(2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3,
3, 3, 3, 3, 3, 3, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA), Response = c(0.775222658013234,
-0.0374860102875694, 0.100620532505619, 0.17475403767196, -0.736181209242918,
1.18581913245908, -0.235457236665258, -0.494834646295896, -0.22096700738071,
-0.00852429328460645, 0.102286976108412, -0.116035743892094,
0.01054849999509, 0.429857364190398, -0.29619258318138, 0.394303081010858,
-0.456401545475929, 0.391960511587087, -0.218177702859661, -0.174814586471715,
0.424769871360028, -0.0771395073436865, 0.698662414019584, 0.695676522106077,
-0.31659375422071, -0.584947748238806, -0.523065304477453, -0.19259357510277,
-0.0786143714402391, -0.313283220381509), Score = structure(c(1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c("Sample",
"Species"), class = "factor"), Label = c("2-1", "2-2", "2-3",
"2-4", "2-5", "2-6", "2-7", "2-8", "2-9", "2-10", "3-1", "3-2",
"3-3", "3-4", "3-5", "3-6", "3-7", "3-8", "3-9", "3-10", "spp1",
"spp2", "spp3", "spp4", "spp5", "spp6", "spp7", "spp8", "spp9",
"spp10")), .Names = c("Time", "Treatment", "Response", "Score",
"Label"), row.names = c("1", "2", "3", "4", "5", "6", "7", "8",
"9", "10", "11", "12", "13", "14", "15", "16", "17", "18", "19",
"20", "spp1", "spp2", "spp3", "spp4", "spp5", "spp6", "spp7",
"spp8", "spp9", "spp10"), class = "data.frame")
我试过的：
myPlt <- function(x, air = 1.1) {
## fortify PRC model
fx <- fortify(x)
## samples and species scores
sampScr <- fx[fx$Score == "Sample", ]
sppScr <- fx[fx$Score != "Sample", ]
ord <- order(sppScr$Response)
sppScr <- sppScr[ord, ]
## base plot
plt <- ggplot(data = sampScr,
aes(x = Time, y = Response,
colour = Treatment, group = Treatment),
subset = Score == "Sample")
plt <- plt + geom_line() + # add lines
geom_rug(sides = "r", data = sppScr) ## add rug
## species labels
sppLab <- sppScr[, "Label"]
## label grobs
tg <- lapply(sppLab, textGrob, just = "left")
## label grob widths
wd <- sapply(tg, function(x) convertWidth(grobWidth(x), "cm",
valueOnly = TRUE))
mwd <- max(wd) ## largest label
## add some space to the margin, move legend etc
plt <- plt +
theme(plot.margin = unit(c(0, mwd + 1, 0, 0), "cm"),
legend.position = "top",
legend.direction = "horizontal",
legend.key.width = unit(0.1, "npc"))
## annotate locations
## - Xloc = new x coord for label
## - Xloc2 = location at edge of plot region where rug ticks met plot box
Xloc <- max(fx$Time, na.rm = TRUE) +
(2 * (0.04 * diff(range(fx$Time, na.rm = TRUE))))
Xloc2 <- max(fx$Time, na.rm = TRUE) +
(0.04 * diff(range(fx$Time, na.rm = TRUE)))
## Yloc - where to position the labels in y coordinates
yran <- max(sampScr$Response, na.rm = TRUE) -
min(sampScr$Response, na.rm = TRUE)
## This is taken from vegan:::linestack
## attempting to space the labels out in the y-axis direction
ht <- 2 * (air * (sapply(sppLab,
function(x) convertHeight(stringHeight(x),
"npc", valueOnly = TRUE)) *
yran))
n <- length(sppLab)
pos <- numeric(n)
mid <- (n + 1) %/% 2
pos[mid] <- sppScr$Response[mid]
if (n > 1) {
for (i in (mid + 1):n) {
pos[i] <- max(sppScr$Response[i], pos[i - 1] + ht[i])
}
}
if (n > 2) {
for (i in (mid - 1):1) {
pos[i] <- min(sppScr$Response[i], pos[i + 1] - ht[i])
}
}
## pos now contains the y-axis locations for the labels, spread out
## Loop over label and add textGrob and segmentsGrob for each one
for (i in seq_along(wd)) {
plt <- plt + annotation_custom(tg[[i]],
xmin = Xloc,
xmax = Xloc,
ymin = pos[i],
ymax = pos[i])
seg <- segmentsGrob(Xloc2, pos[i], Xloc, pos[i])
## here is problem - what to use for ymin, ymax, xmin, xmax??
plt <- plt + annotation_custom(seg,
## xmin = Xloc2,
## xmax = Xloc,
## ymin = pos[i],
## ymax = pos[i])
xmin = Xloc2,
xmax = Xloc,
ymin = min(pos[i], sppScr$Response[i]),
ymax = max(pos[i], sppScr$Response[i]))
}
## Build the plot
p2 <- ggplot_gtable(ggplot_build(plt))
## turn off clipping
p2$layout$clip[p2$layout$name=="panel"] <- "off"
## draw plot
grid.draw(p2)
}
图基于我尝试过的myPlt()
这是我myPlt()从上面所做的。请注意通过标签绘制的小水平刻度 - 这些应该是上图第一个图中的倾斜线段。
