I want to use multiple colour values from my dataset.
But I can't.

Labels with negative values passed to with_outer_glow, seem to lose their "-" sign.

See following example...

example contour plots

library(ggplot2)
library(geomtextpath)
library(ggfx)
df <- expand.grid(x = seq(nrow(volcano)), y = seq(ncol(volcano)))
df$z <- as.vector(volcano)-150 #make some values negative

this works - shows negative numbers

ggplot(df, aes(x, y, z = z)) +
geom_contour_filled(bins = 6, alpha = 0.6) +
geom_textcontour(bins = 6, size = 3, linetype = NA)

this doesn't work - negative numbers not shown

ggplot(df, aes(x, y, z = z)) +
geom_contour_filled(bins = 6, alpha = 0.6) +
with_outer_glow(
geom_textcontour(aes(z = z), linetype = NA, size = 3),
colour = 'white',
expand = 5,
sigma = 1
)

• bloom
• various dithering
• motion blur
• Fft blur

Prepare for release:

• git pull
• Check current CRAN check results
• Polish NEWS
• devtools::build_readme()
• urlchecker::url_check()
• devtools::check(remote = TRUE, manual = TRUE)
• devtools::check_win_devel()
• rhub::check_for_cran()
• revdepcheck::cloud_check()
• Update cran-comments.md
• git push

Submit to CRAN:

• usethis::use_version('patch')
• devtools::submit_cran()
• Approve email

Wait for CRAN...

• Accepted 🎉
• git push
• usethis::use_github_release()
• usethis::use_dev_version()
• git push

We are using magick for now but it is slow. The ggfx API should not expose magick specific settings so it can be easier to switch to a more performant backend (shaders) when possible

Hi 👋,

I want to apologize if this is the wrong place for this, since this likely is more of a feature-wish than an issue 😬.

Basically, I love the idea of with_raster(), but currently I am only able to control the resolution (dpi on the final figure) of the rasterized ggplot layer indirectly, by up-scaling the exported figure. Yet, when trying to keep plot configurations consistent over many plots (eg for a scientific publication), this forces one to also correct a zillion of other settings (font size, line widths, symbol sizes etc).

Hence, my question: is it possible to somehow adjust the resolution of with_raster() (now or potentially in the future)?
(Also - apologies if this is a case where I am just to silly to RTFM correctly 😨)

c.f. https://github.com/cansik/processing-bloom-filter#implementation

• get_viewport_area() / set_viewport_area()
• get_viewport_cols_rows()
• get_raster_area() / set_raster_area()

The idea to use blurs for confidence intervals is great. But I fear result shown here : https://ggfx.data-imaginist.com/articles/geoms.html for confidence intervals is not useful (sorry, I don't want to be harsh).

To explain:
Blurring should result in a figure in which regions with the same confidence interval level have the same color density/opacity/brightness.

Some thing like this, just smoother:
p = ggplot(mpg, aes(displ, hwy))
for (l in seq(5,95,5)/100)
p = p + geom_smooth(level = l, alpha = .1, fill = "blue", method = "loess", size = 0)
p

cheers - Guido

Feature Request

Would it be possible to specify a location in axis coordinates and get the pixel index?

This is an example of trying to set the pixel corresponding to x=0, y=0 to red. Note that calculate_pixel_index_of_coordinate() is a placeholder for the function I'm looking for.

zero_zero_dot = function(the_raster, color) {
  viewport_location() %>% cat()
  # find the pixel that corresponds to (x=0, y=0)
  zero_pixel_index = calculate_pixel_index_of_coordinate(the_raster, x=0, y=0)
  # set that pixel's color
  the_raster[zero_pixel_index] <- farver::encode_native(color)
  
  the_raster
}


tibble(x=-1:5, y=-1:5) %>% 
  ggplot(aes(x=x, y=y)) + 
  with_custom(
    geom_point(),
    filter = zero_zero_dot,
    color = "red"
  )

Hi,

Love the package. But I've been having some issue recently when using the with_mask function. The output is quite pixelated, to the point of being unusable for a few use-cases (generative art).

Below is a reproducible example that just plots some random points, where if using a mask, the result is pixelated regardless of how the image is exported. I am using the latest version of ggfx from the github repo and ggplot2 ‘3.4.0’.

Thanks in advance for any help

library(tidyverse)
library(ggfx)

test_data <- tibble(
  x = rnorm(2000, 0, 10),
  y = rnorm(2000, 0, 10),
)

circle <- tibble(
  x = cos(seq(0, 2*pi, length.out = 360)),
  y = sin(seq(0, 2*pi, length.out = 360))
) |> 
  mutate(across(everything(), ~ .x * 25))

# the output here is very pixelated
ggplot() +
  lims(x = c(-30, 30), y = c(-30, 30)) +
  coord_equal() +
  theme_void() +
  as_reference(
    geom_polygon(aes(x = x, y = y), circle),
    id = 'circle'
  ) +
  with_mask(
    geom_point(data = test_data, aes(x,y)),
    mask = ch_alpha('circle'),
    invert = FALSE,
    ignore_background = FALSE
  )
  
  # however if I just plot without the mask as a usual ggplot, it is fine:
  ggplot(test_data, aes(x,y)) +
  geom_point() +
  lims(x = c(-30, 30), y = c(-30, 30)) +
  coord_equal() +
  theme_void() 

I have only tried this with with_outer_glow():
colour works, but not color 🙂

ggplot() +
  with_outer_glow(
    geom_point(aes(x = 0, y = 0), size = 10),
    color = "red",
    expand = 10
  )

First release:

• usethis::use_cran_comments()
• Proof read Title: and Description:
• Check that all exported functions have @returns and @examples
• Check that Authors@R: includes a copyright holder (role 'cph')
• Check licensing of included files
• Review https://github.com/DavisVaughan/extrachecks

Prepare for release:

• devtools::build_readme()
• urlchecker::url_check()
• devtools::check(remote = TRUE, manual = TRUE)
• devtools::check_win_devel()
• rhub::check_for_cran()
• Review pkgdown reference index for, e.g., missing topics
• Draft blog post

Submit to CRAN:

• usethis::use_version('major')
• devtools::submit_cran()
• Approve email

Wait for CRAN...

• Accepted 🎉
• usethis::use_github_release()
• usethis::use_news_md()
• usethis::use_dev_version()
• Update install instructions in README
• Finish blog post
• Tweet
• Add link to blog post in pkgdown news menu

Hi, I found what I think is a bug when you combine with_mask and with_shadow. I think the shadow is being put on top of the used path.

# Suspected bug
ggplot(df) +
  geom_path(aes(x,y),
            color="#4bc4ff",linewidth=20,lineend="round" ) +
  with_shadow(
    geom_path(data=df[-3,],aes(x,y),
            color="blue",linewidth=20,lineend="round" ),
    stack=TRUE,x_offset = 0,y_offset = 0,sigma = 20,id="bar") +
  with_mask(
    geom_path(data=df,
              aes(x,y),
              linewidth=20,lineend='round'),
    mask=ch_alpha('bar')
  )

# Desired output
ggplot(df) +
  geom_path(aes(x,y),
            color="#4bc4ff",linewidth=20,lineend="round" ) +
  with_shadow(
    geom_path(data=df[-3,],aes(x,y),
              color="blue",linewidth=20,lineend="round" ),
    stack=FALSE,x_offset = 0,y_offset = 0,sigma = 20,id="bar") +
  with_mask(
    geom_path(data=df,
              aes(x,y),
              linewidth=20,lineend='round'),
    mask=ch_alpha('bar')
  ) +
  geom_path(data=df[-3,],aes(x,y),
            color="blue",linewidth=20,lineend="round" )

Expected output:

Actual output:

Adding theme_void() to a plot with a custom filter using Magick results in an error. When testing with other theme functions such as theme_minimal() or theme_classic() it works fine. Small reprex based on the example in the documentation:

library(ggplot2)
library(ggfx)

# Example from documentation
implode <- function(x, factor = 0.5) {
  vp <- magick::image_read(get_viewport_area(x))
  vp <- magick::image_implode(vp, factor = factor)
  set_viewport_area(x, as.raster(vp, native = TRUE))
}

ggplot(mtcars, aes(x = factor(gear), y = disp)) +
  with_custom(
    geom_boxplot(aes(fill = as.factor(gear))),
    filter = implode
  ) 

# Adding theme_void() returns an error
# Other theme_*() functions don't seem to have this issue
ggplot(mtcars, aes(x = factor(gear), y = disp)) +
  with_custom(
    geom_boxplot(aes(fill = as.factor(gear))),
    filter = implode
  ) +
  theme_void()
#> Error in .subset(raster, index): only 0's may be mixed with negative subscripts

^{Created on 2023-08-25 with reprex v2.0.2}

Thanks for working on this - it's a great idea. I suspect i'm doing something wrong and I know this is in early stages so wouldn't expect everything to work but wanted to flag - just incase it might be useful. I have tried a few different geoms with the example plot in the documentation but it didn't seem to like geom_line (others worked ok).

library(ggplot2)
library(ggfx)

ggplot(mtcars, aes(mpg, disp)) + 
    with_shadow(geom_line(colour = 'red', size = 3), sigma = 3) +
    geom_line()
#> Error in validDetails.polyline(x): invalid 'arrow' argument

It may have something to do with this in {grid} (though it may not):

######################################
# POLYLINES primitive
######################################
# Very similar to LINES primitive, but allows
# multiple polylines via 'id' and 'id.lengths' args
# as per POLYGON primitive
validDetails.polyline <- function(x) {
  if (!is.unit(x$x) ||
      !is.unit(x$y))
      stop("'x' and 'y' must be units")
  if (!is.null(x$id) && !is.null(x$id.lengths))
      stop("it is invalid to specify both 'id' and 'id.lengths'")
  if (length(x$x) != length(x$y))
      stop("'x' and 'y' must be same length")
  if (!is.null(x$id) && (length(x$id) != length(x$x)))
      stop("'x' and 'y' and 'id' must all be same length")
  if (!is.null(x$id))
      x$id <- as.integer(x$id)
  if (!is.null(x$id.lengths) && (sum(x$id.lengths) != length(x$x)))
      stop("'x' and 'y' and 'id.lengths' must specify same overall length")
  if (!is.null(x$id.lengths))
      x$id.lengths <- as.integer(x$id.lengths)
  if (!(is.null(x$arrow) || inherits(x$arrow, "arrow")))
      stop("invalid 'arrow' argument")
  x
}

When we use ggfx::with_outer_glow (and presumably other ggfx functions) the resulting output is converted to bitmap. Is this intended behaviour?

require(tidyverse)
require(sf)

j0 = jsonlite::fromJSON('https://petition.parliament.uk/petitions/554276.json')
d0 = j0$data$attributes$signatures_by_constituency |> as_tibble()
g = sf::read_sf('https://petitionmap.unboxedconsulting.com/json/uk/uk/topo_wpc.json')
d = g |> left_join(d0, by = c(id = 'ons_code')) |> st_set_crs(4326) |> st_transform("EPSG:27700")

# vector output
p1 = ggplot() + 
  geom_sf(aes(fill = signature_count), data = d, col = 'white', linewidth=0.03) +
  coord_sf(ylim = c(5e4, 9.6e5)) +
  scale_fill_viridis_c(direction = -1) +
  theme_void() +
  # theme(plot.background = element_rect(fill = 'grey90', colour = NA)) +
  labs(fill = 'Signatures')

ggsave(p1, filename = 'map1.pdf', width = 30, height = 45); browseURL('map1.pdf')

# When we use with_outer_glow the resulting map bitmap
p2 = ggplot() + 
  ggfx::with_outer_glow(
    geom_sf(aes(fill = signature_count), data = d, col = 'white', linewidth=0.03),
    colour = 'lightblue1',
    sigma = 10,
    expand = 2
  ) +
  coord_sf(ylim = c(5e4, 9.6e5)) +
  scale_fill_viridis_c(direction = -1) +
  theme_void() +
  labs(fill = 'Signatures')

ggsave(p2, filename = 'map2.pdf', width = 30, height = 45); browseURL('map2.pdf')

Output of map2 looks like this up close:

See related SO post:

• How to blur part of a plot in ggplot?

Something like below:

library(ggplot2)
library(ggfx)

ggplot(mtcars, aes(x = disp, y = mpg)) +
  with_blur(
    geom_line(),
    sigma = unit(1, 'mm'),
    xmin = 100, xmax = 200, ymin = 15, ymax = 30 # specify rectangular area
  )

Currently no reference raster is created

Very possible I've missed some key understanding, but at first pass it appears that some functions, specifically as_reference() and with_shadow() do not work with sf objects/geom_sf(). Looks to be some incompatibility with sf's geometry list columns? I keep getting Error in UseMethod("with_shadow") : no applicable method for 'with_shadow' applied to an object of class "list"

Reprex:

library(tidyverse)
library(sf)
library(ggfx)

nc <- st_read(system.file("shape/nc.shp", package="sf"))

ggplot(nc) +
  as_reference(
    geom_sf(data = nc,
            aes(fill = BIR79)),
    id = "nc"
)

ggplot(nc) +
  with_shadow(
    geom_sf(aes(fill = BIR79)),
    sigma = 3
)