library(ggdiagram)
library(ggplot2)
library(dplyr)
#>
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#>
#> filter, lag
#> The following objects are masked from 'package:base':
#>
#> intersect, setdiff, setequal, union
library(ggtext)
library(ggarrow)
library(arrowheadr)Setup
Packages
Base Plot
To avoid repetitive code, we make a base plot:
my_font <- "Roboto Condensed"
my_font_size <- 20
my_point_size <- 2
# my_colors <- viridis::viridis(2, begin = .25, end = .5)
my_colors <- c("#3B528B", "#21908C")
theme_set(
theme_minimal(base_size = my_font_size, base_family = my_font) +
theme(axis.title.y = element_text(angle = 0, vjust = 0.5))
)
bp <- ggdiagram(
font_family = my_font,
font_size = my_font_size,
point_size = my_point_size,
linewidth = 0.5,
theme_function = theme_minimal,
axis.title.x = element_text(face = "italic"),
axis.title.y = element_text(
face = "italic",
angle = 0,
hjust = 0.5,
vjust = 0.5
)
) +
scale_x_continuous(labels = signs_centered, limits = c(-4, 4)) +
scale_y_continuous(labels = signs::signs, limits = c(-4, 4))Specifying a segment
A segment is a portion of a line between two points.
bp +
{p1 <- ob_point(-3, -1)} +
{p2 <- ob_point(3, 3)} +
{s1 <- ob_segment(p1, p2)}
Features of a segment
Distance between points
s1@distance
#> [1] 7.211103Alternately:
distance(s1)
#> [1] 7.211103Segment resection
Sometimes a segment looks better when it reaches close to its endpoint but is resected a little. As seen in Figure 2, the resect parameter sets the amount of resection in millimeters. If needed, the resect_fins and resect_head parameters control the start and and end point resection separately.
bp +
p1 +
p2 +
ob_segment(p1, p2, resect = 3)
Line passing through the segment
The line that passes through the segment contains information about the segment, such as its slope, intercept, or angle.
To access the line that passes between both points:
s1@line
#>
#> ── <ob_line>
#> # A tibble: 1 × 7
#> slope intercept xintercept a b c color
#> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <chr>
#> 1 0.667 1 -1.5 -4 6 -6 black
s1@line@slope
#> [1] 0.6666667
s1@line@intercept
#> [1] 1
s1@line@angle
#> [1] "34°"Code
bp +
s1@line |> set_props(color = "royalblue4") +
s1@midpoint(position = c(0, 1))@label(
polar_just = ob_polar(s1@line@angle + degree(90), 1.5),
plot_point = TRUE) +
ob_label(c(equation(s1@line),
paste0("Distance = ",
round(s1@distance, 2))),
center = midpoint(s1),
vjust = c(-0.2, 1.1),
angle = s1@line@angle)
Midpoints
By default, the midpoint function’s position argument is .5, which finds the point halfway between the point of a segment:
s1@midpoint()
#>
#> ── <ob_point>
#> # A tibble: 1 × 3
#> x y color
#> <dbl> <dbl> <chr>
#> 1 0 1 blackCode
bp +
s1 +
s1@midpoint()@label("Midpoint", hjust = 0, vjust = 1) +
s1@midpoint(c(0, 0.5, 1))@label(
plot_point = TRUE,
hjust = 1,
vjust = 0)
To find the midpoint 25% of the distance between the endpoints of segment:
s1@midpoint(position = 0.25)
#>
#> ── <ob_point>
#> # A tibble: 1 × 3
#> x y color
#> <dbl> <dbl> <chr>
#> 1 -1.5 0 blackCode
Multiple midpoints can be specified:
bp +
s1 +
s1@midpoint(seq(0, 1, 0.2))
A quick way to get the endpoints of a segment is to specify “midpoints” at positions 0 and 1:
bp +
s1 +
s1@midpoint(c(0, 1))
midpoint property.
Segment chains
If a point object with multiple points is placed in the p1 slot, but the p2 slot is left empty, a series of segments chained together will be created.
bp +
ob_segment(
ob_point(x = c(-4, -2, -1, 2, 4),
y = c(-2, 1, -2, 3, -3)))
Arrowheads and fins
In the ob_segment function, the arrow_head and arrow_fins arguments can be set with any arrowhead function from the ggarrow or the arrowheadr packages. Any 2-column matrix will also be acceptable, as shown here. The default arrowhead can be used via the arrowhead() function. The size of the arrowhead and fins can also be controlled independently with the length_head and length_fins arguments.
bp + ob_segment(p1, p2,
arrow_head = arrowhead(),
arrow_fins = ggarrow::arrow_fins_feather(),
length_head = 5,
length_fins = 7,
linewidth = 2)
Labeling segments
By default, any label is set at the midpoint of the segment (see Figure 10). If it is plain text, it will be just above the segment. Its angle will match that of the segment, but this can be turned off by setting label_sloped to FALSE.
bp + ob_segment(p1, p2, label = "My label")
As shown Figure 11, the ob_label function gives much more control over how and where the label will appear.
bp +
ob_segment(
p1,
p2,
label = ob_label(
label = "My label",
position = 0.25,
color = "blue",
fill = class_color("blue")@lighten(),
family = "serif",
vjust = -0.1,
label.padding = ggplot2::margin(
t = 5,
r = 4,
b = 3,
l = 4
)
)
)
Color Gradients
The color_gradient parameter makes a smooth color gradient from the fins of an arrow to its head. It takes a vector of 2 or more colors. The gradient is a grid gradient fill, so it requires a graphics device that supports gradients (e.g., ragg, svg, or pdf). Because the gradient replaces the color of the arrow’s body, set stroke_color = NA to hide the arrow’s outline.
ggdiagram() +
{c1 <- ob_circle(fill = "dodgerblue4", color = NA)} +
{c2 <- ob_circle(fill = "firebrick4", color = NA) |>
place(from = c1, where = "right", sep = 2)} +
ob_segment(c1@east,
c2@west,
arrow_head = arrowhead(),
arrowhead_length = 4,
linewidth = 2,
gradient = c("dodgerblue4", "firebrick4")) 
Gradients are interpolated in the OKLab color space, which is more perceptually even than RGB. A different space can be chosen for all gradients with options(ggdiagram.gradient_space = "oklch") (or "lab", "lch", or "rgb").
Resection works as usual with gradients. Here, the segment is shortened by 3 mm at each end, and the gradient spans the shortened arrow.
ggdiagram() +
c1 +
c2 +
ob_segment(c1@east,
c2@west,
arrow_head = arrowhead(),
arrowhead_length = 4,
resect = 3,
linewidth = 2,
gradient = c("dodgerblue4", "orchid4", "firebrick4"))