geom_text_contour with geom_density_2d

[dhicks]

Context: I'm interested in using geom_text_contour() to direct label some calculated density contours, which I'm using to hack some group highlighting that can accommodate lots of outliers.

I'm having trouble combining geom_text_contour() with the density contours calculated as in geom_density_2d(). MWE:

library(tidyverse)
library(metR)
#> 
#> Attaching package: 'metR'
#> The following object is masked from 'package:purrr':
#> 
#>     cross
library(palmerpenguins)

data('penguins')

ggplot(penguins, aes(bill_length_mm, 
                     flipper_length_mm, 
                     color = species)) +
    geom_point(alpha = 0.5) +
    geom_density_2d(data = ~ filter(.x, 
                                    species %in% c('Adelie', 'Gentoo')),
                    breaks = .001) +
    geom_text_contour(aes(label = species), 
                      data = ~ filter(.x, 
                                      species %in% c('Adelie', 'Gentoo')),
                      stat = 'density_2d',
                      breaks = c(.001))
#> Warning: Removed 2 rows containing non-finite values (`stat_density2d()`).
#> Removed 2 rows containing non-finite values (`stat_density2d()`).
#> Warning: Removed 2 rows containing missing values (`geom_point()`).

^{Created on 2023-03-21 with reprex v2.0.2}

geom_density_2d() has the behavior I'm looking for, drawing a single contour around most of the data for each of the groups of interest. What I need to do is direct label those contours, ideally with one label on each contour. I can't seem to get geom_text_contour() to pick up those contours, though. The default "text_contour" stat needs a separate z aesthetic. So I thought I'd pass it the "density_2d" stat used by geom_density_2d(). But this results in the group labels placed directly on top of each other and not particularly close to the contours.

[eliocamp]

This is weird, but I'd never used geom_text_contour() with a different stat. I'll look into it. In the meantime, though, you might want to check out the ggpath package.

[eliocamp]

Hi! This took forever, sorry!

Looking at the code, the issue is that geom_text_contour() computes the contours in the stat. But the density_2d stat doesn't output contours, but a 2D density grid. I don't think geom_text_contour() can help here.