cell colour by value and text transformations in gt

I want to make use of {gt}'s nice new-ish data_color revamp but I would like to also make use of text_transformation and to add links and custom text to the table. So for example given this data:

library(dplyr)
library(gt)

dat <- data.frame(
  flag = c("safe", "out", "tie"),
  url = c("https://www.google.com", "https://www.yahoo.com", "https://www.bing.com"),
  value1 = c(1, NA, 3),
  value2 = c(1.1, NA, 8)
)

dat

    dat
  flag                    url value1 value2
1 safe https://www.google.com      1    1.1
2  out  https://www.yahoo.com     NA     NA
3  tie   https://www.bing.com      3    8.0

and then create a table:

gt(dat) %>%
  cols_hide(columns = c("flag", "url")) %>%
  data_color(
    direction = "row",
    palette = "viridis"
  ) %>%
  text_transform(fn = function(x) glue("<a href=\"{dat$url}\" target=\"_blank\">{x}</a> {dat$flag}"))

So I can make a table that has active urls:

But what if I have multiple columns of urls that matches a corresponding vlaue column. So for example this data.frame:

dat <- data.frame(
  flag = c("safe", "out", "tie"),
  url1 = c("https://www.google.com", "https://www.yahoo.com", "https://www.bing.com"),
  url2 = c("https://www.one.com", "https://www.two.com", "https://www.three.com"),
  value1 = c(1, NA, 3),
  value2 = c(1.1, NA, 8)
)

dat

How can I add the values from url_1 to value_1 and the same for _2? I know I can pivot_* to get the right links but then AKAIK the columns become strings and then I can't use them in data_color.

Any ideas how I might merge these two abilities?

Maybe there is a more elegant approach but a simple one would be to use two text_transform, one for each pair of url_ and value_ columns:

library(dplyr)
library(gt)
library(glue)

gt(dat) %>%
  cols_hide(columns = c("flag", "url1", "url2")) %>%
  data_color(
    direction = "row",
    palette = "viridis"
  ) %>%
  text_transform(
    fn = function(x) glue("<a href=\"{dat$url1}\" target=\"_blank\">{x}</a> {dat$flag}"),
    ocation = cells_body("value1")
  ) |>
  text_transform(
    fn = function(x) glue("<a href=\"{dat$url2}\" target=\"_blank\">{x}</a> {dat$flag}"),
    location = cells_body("value2")
  )

EDIT And as an alternative approach to a for you could use purrr::reduce like so:

cols <- grep("value*", names(dat), value = TRUE)

gt(dat) %>%
  cols_hide(columns = c("flag", "url1", "url2")) %>%
  data_color(
    direction = "row",
    palette = "viridis"
  ) %>%
  purrr::reduce(cols, \(x, col) {
    url_col <- gsub("value", "url", col)
    text_transform(x,
      fn = function(x) glue("<a href=\"{dat[[url_col]]}\" target=\"_blank\">{x}</a> {dat$flag}"),
      location = cells_body(col)
    )
  }, .init = .)

@stefan: This was too long to add as a comment but taking your approach and generalizing it, this could work here:

dat <- tibble(
  flag = c("safe", "out", "tie"),
  url1 = c("https://www.google.com", "https://www.yahoo.com", "https://www.bing.com"),
  url2 = c("https://www.one.com", "https://www.two.com", "https://www.three.com"),
  value1 = c(1, NA, 3),
  value2 = c(1.1, NA, 8)
)

gt_table <- gt(dat) %>%
  cols_hide(columns = c("flag", "url1", "url2")) %>%
  data_color(
    direction = "row",
    palette = "viridis"
  )


cols <- grep("value*", names(dat), value = TRUE)

for (i in seq_along(cols)) {
  url_col <- gsub("value", "url", cols[i])
  gt_table <- gt_table %>%
    text_transform(
      fn = function(x) glue("<a href=\"{dat[,url_col, drop = TRUE]}\" target=\"_blank\">{x}</a>"),
      location = cells_body(cols[i])
    )
}

gt_table