mccarthy-m-g / palettes Goto Github PK

Add a ggpalette() function that takes a palette or color as input and returns a ggplot themed like a swatch as output. Something like: https://images.app.goo.gl/j9Vu9NRJuC5vvMHr7

Doing this would also require creating as.dataframe() methods for the palette and color classes.

As far as handling multiple colours, facets would probably work well

This can be accomplished using the crayon package; however, this may be unnecessary given that RStudio 1.6 might support this natively rstudio/rstudio#5424.

Edit: RStudio support seems to be for the editor pane, not output, so maybe still worth implementing myself

Where you supply a palette and it colours theme elements based on the number of colours in the palette (e.g., if 1 colour, all elements take on that colour, if 2 then colour 1 is the primary and colour 2 is the secondary, etc.).

theme_prism from ggprism does something similar (picking a theme based on a palette): https://cran.r-project.org/web/packages/ggprism/vignettes/themes.html, https://github.com/csdaw/ggprism/blob/master/R/theme_prism.R

These are issues or features that I am considering for v0.2.0. Some of these may be pushed back to later versions or not be included at all.

If you have suggestions, please create a new issue.

If you would like to contribute, please comment on the relevant issue.

Items

• #9
• #26
• #27
• #3
• #16
• #28
• #29

https://vctrs.r-lib.org/articles/pillar.html

Prepare for release:

• git pull
• Check current CRAN check results
• Check if any deprecation processes should be advanced, as described in Gradual deprecation
• Polish NEWS
• devtools::build_readme()
• urlchecker::url_check()
• devtools::check(remote = TRUE, manual = TRUE)
• devtools::check_win_devel()
• devtools::check_mac_release()
• rhub::check_for_cran()
• revdepcheck::revdep_check(num_workers = 4)
• Update cran-comments.md
• git push
• Draft blog post

Submit to CRAN:

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

Wait for CRAN...

• Accepted 🎉
• git push
• usethis::use_github_release()
• usethis::use_dev_version()
• git push
• Finish blog post
• Tweet
• Add link to blog post in pkgdown news menu

I notice that the code base has functions for lightening/darkening colour palettes and vectors but these are not exported. Any reason why not? They would be nice to have!

See: https://pkgdown.r-lib.org/articles/pkgdown.html#reference

Since coloured cli output doesn’t just work in R Markdown and Quarto, it might be nice to show how to get coloured colour previews in these outputs using {fansi}, like I did here: https://tidytales.ca/posts/2022-12-20_palettes/

e.g., pal_color()

e.g., lighten.palettes_palette()

Example here https://blog.earo.me/2019/11/03/practical-guide-to-s3/

New function for creating bivariate colour palettes that can be used with the biscale package. Main points of consideration are:

• Should the function take 4 colours as arguments (one for each corner of the palette), or 3 colours as arguments (with the 4th being interpolated based on the x and y colours)?
• Should the number of dimensions be infinite, or should they be restricted? (this is more a practical question based on how well I can program something that generalizes well to higher dimensions, or if I can only create something nice for smaller numbers of dimensions)

Additive vs subtractive mixing https://www.mathscareers.org.uk/colour-calculations/ (maybe I should use these methods separately for + and – operators)

Modifying RGB colour space to an additive space https://stackoverflow.com/a/41598885/16844576

Function to invert a colour, and round it to black or white https://stackoverflow.com/a/35970186

Function to mix colours

Colour quantization https://www.r-bloggers.com/2016/01/color-quantization-in-r/

Since colours in colour vectors can be named, it might be nice to also output those names when they’re there (maybe as a muted grey colour in parentheses next to the hex code?).

I would probably exclude this when printing in tibbles though, because the output would be too wide there.

and add alias for "color" instead of that being the default

https://www.tidyverse.org/blog/2022/12/purrr-1-0-0/

map_dfc() and map_dfr() (and their map2 and pmap variants) have been superseded in favour of using the appropriate map function along with list_rbind() or list_cbind().

There’s also a new map_vec() function that might be helpful

For developers to create a new palettes package with all the boilerplate set up.

Can probably wrap usethis::create_from_github() and hard code it to a template repository. Only catch is this would require an Internet connection.

Or I can try to build everything manually using the appropriate use_ functions. use_template() looks like it will fit the bill.

For example, pal_palette() instead of palette() to avoid masking base R functions

Add function to check the accessibility of a palette. The cvd argument in colorspace::swatchplot() does this. See http://colorspace.r-forge.r-project.org/articles/palette_visualization.html

Maybe also consider extending the S3 classes to contain info on whether a palette is accessible (as an attribute or something)

TODO

With testthat: https://vctrs.r-lib.org/articles/s3-vector.html#testing

The info text would provide a hint on how to fix the error. Something like (as separate “i” bullets):

• Is/Are your hexadecimal string/strings of the form “#rrggbb” or “#rrggbbaa”?
• Is/Are your colour name/names included in grDevices::colors()?

It would be nice to make this context aware, so only the relevant hint shows up if invalid colours are specified in a single way, and both show up if there are both invalid hex strings and colour names.

The skittle/smartie swatches for colour palettes in this paper look nice and it would be cool to have the option to print palettes in the console like on page 8, for users who might want a compact display.

It would be controlled with a new option (palettes.print_style?), and the implementation would basically just involve not printing the hex codes—only symbols—and restricting the number of colours to 10 per line before a line break.

The print styles could have the following names: “default” (the current style), “compact”. Or the default could just be NULL rather than an explicit value. Either way, using strings here would be a good idea rather than a boolean in case I want to add new print styles in the future.

https://cli.r-lib.org/reference/num_ansi_colors.html

In RStudio and pkgdown colours just work, but for other formats (e.g., Quarto) the number of colours needs to be increased from the default of 8. It’s easy for a user to do this, but it might be nice to set these options/environment variables up internally.

The documentation says it can be error prone, so will need some testing to avoid problems.

e.g., rgb_to_colour(), hsl_to_colour()

Like in https://davidchall.github.io/ipaddress/articles/ipaddress-classes.html

Edit: Need to decide if this is necessary, or if it’s unnecessary bloat

{palettes} could use a better hex sticker! Open to contributions for this 😄

I made the current one using some lightly modified free to use vector graphics. It’s okay, but could be a lot better.

Some examples here:

https://meghan.rbind.io/blog/2022-10-11-creating-custom-color-palettes-with-ggplot2/

https://www.garrickadenbuie.com/blog/custom-discrete-color-scales-for-ggplot2/

https://drsimonj.svbtle.com/creating-corporate-colour-palettes-for-ggplot2

These can also be used as default options with:

options(
  ggplot2.discrete.colour = function(...) scale_colour_viridis_d(..., begin = 0.15, end = 0.85),
  ggplot2.discrete.fill = function(...) scale_fill_viridis_d(..., begin = 0.15, end = 0.85),
  ggplot2.continuous.colour = function(...) scale_colour_viridis_c(..., begin = 0.15, end = 0.85),
  ggplot2.continuous.fill = function(...) scale_fill_viridis_c(..., begin = 0.15, end = 0.85)
)

Right now the output in the README only shows the colour symbol and hex code, but it doesn’t show the <palettes_colour[1]> class above that like it’s supposed to (e.g., it should look how the pal_palette() output looks).

Hi Michael,

I'm working through the code in the Creating a Colour Package vignette and you make use of a pal_brewer function. When I run the code, I get an error: could not find function "pal_brewer". I can't find the function anywhere in the palettes code base nor in any other R package. What am I missing?

Thanks!

I’m just using the symbols from cli: https://cli.r-lib.org/reference/symbol.html. So any of these would be available.

Code-wise it could be as simple as using getOption() with the bullet as a default. I would also need to add something in the creating a package vignette to show developers how to change the symbol their package uses though.