如何在R绘图中嵌入自定义矢量图并设置位置、缩放与旋转

Yep, you absolutely can embed vector graphics (AI, EPS, PDF) into R plots with full control over their position, scaling, and rotation. Below are the most reliable methods using R's grid system and dedicated packages, which work for both base R plots and ggplot2:

Prerequisites #

First, install and load the core packages we'll use:

install.packages(c("grImport2", "grid", "ggplot2", "gridBase"))
library(grImport2)
library(grid)
library(ggplot2)
library(gridBase)

Importing Vector Files #

grImport2 is designed to handle vector graphics directly, preserving their crispness. Here's how to load different formats:

• EPS/PDF: Directly use readPicture() (for multi-page PDFs, specify the page number):

  # Load an EPS file
  vec_img <- readPicture("your_vector_file.eps")
  
  # Load page 2 of a PDF file
  vec_img <- readPicture("your_vector_file.pdf", page = 2)
  

• AI Files: Adobe Illustrator files are essentially PDF variants, but sometimes have compatibility quirks. Save your AI file as an EPS or standard PDF (via Illustrator's Save As > choose EPS/PDF with "Preserve Vector Data" enabled) for best results, then load it like the examples above.

Embedding in Base R Plots #

Base R uses a separate graphics system, so we'll use gridBase to bridge it with the grid system (where our vector graphics live):

# First, create your base plot
plot(1:10, 1:10, main = "Base Plot with Embedded Vector", xlab = "X", ylab = "Y")

# Switch to the grid viewport tied to your base plot
vps <- baseViewports()
pushViewport(vps$inner, vps$figure, vps$plot)

# Draw the vector graphic: customize position, size, and rotation
# Use npc coordinates (0 = bottom/left, 1 = top/right) for plot-relative positioning
grid.picture(
  vec_img,
  x = 0.7, y = 0.3,  # Position: 70% right, 30% up from bottom-left
  width = 0.2, height = 0.2,  # Size: 20% of plot width/height
  angle = 45,  # Rotate 45 degrees clockwise
  hjust = 0.5, vjust = 0.5  # Center the graphic on the x/y point
)

# Return to the base graphics system
popViewport(3)

Embedding in ggplot2 #

Since ggplot2 is built on the grid system, embedding vector graphics is more straightforward with annotation_custom():

# Create your base ggplot
p <- ggplot(mtcars, aes(wt, mpg)) +
  geom_point(color = "steelblue") +
  labs(title = "ggplot2 with Embedded Vector Graphic")

# Convert the imported vector to a grid grob (graphical object)
vec_grob <- pictureGrob(
  vec_img,
  width = unit(2, "cm"), height = unit(2, "cm"),  # Fixed size in centimeters
  angle = 30  # Rotate 30 degrees clockwise
)

# Add the vector graphic to the plot
# Use data coordinates (matches your plot's x/y axes) here
p + annotation_custom(
  vec_grob,
  xmin = 3, xmax = 5,  # X range for the graphic
  ymin = 20, ymax = 25  # Y range for the graphic
)

# Alternatively, use npc coordinates for plot-relative positioning:
# p + annotation_custom(
#   vec_grob,
#   xmin = 0.7, xmax = 1, ymin = 0, ymax = 0.3,
#   coords = "npc"
# )

Advanced Customization Tips #

• Positioning: Use unit() to mix units (e.g., unit(1, "inch") for fixed size, unit(0.5, "npc") for plot-relative positioning).
• Scaling: Adjust width/height in grid.picture() or pictureGrob(), or use the scale parameter to resize proportionally.
• Rotation: The angle parameter uses clockwise degrees (e.g., angle = -90 for counter-clockwise 90 degrees).
• Alignment: Use hjust (horizontal) and vjust (vertical) to fine-tune where the graphic is anchored relative to its x/y position.

Notes to Avoid Headaches #

• Always export your final plot as a vector format (like PDF) to preserve the embedded vector graphic's crispness. Use ggsave("final_plot.pdf", p, device = "pdf") for ggplot2, or pdf("final_plot.pdf"); plot(...); dev.off() for base plots.
• Avoid using bitmap-focused packages (like magick) for this task—they'll rasterize your vector graphic, losing its scalability.
• If your AI/PDF has transparency, ensure you export it with transparency enabled; grImport2 supports alpha channels.

内容的提问来源于stack exchange，提问作者passiflora