Feature request: Set the current `facet_*_paginate` page without calling the facet function again

[billdenney]

In humanpred/ggtibble#2, I'm trying to make a generic conversion from the facet_*_paginate uses to a list of figures (similar to what is in #219). A challenge with this is that I only know how to update the page by calling facet_*_paginate again with a new page number.

I think that it's as simple as updating p$facet$params$page <- 2 (where p is the original figure and 2 is the new page number desired), but I hesitate to reach that deeply into the facet settings in case something changes in the future.

Would it be reasonable to create something like a set_facet_page() function that could possibly be as simple as:

set_facet_page <- function(p, page) {
  p$facet$params$page <- page
  p
}

[pormr]

Currently I am using a repeat loop to render PDF which contains paginated plots spanning arbitary number of pages.

pdf("path_to_pdf.pdf", ...)
i <- 1
repeat {
  p <- ggplot(data = plot_data) +
    ... +
    facet_wrap_paginate(~variable, nrow = nR, ncol = nC, page = i) +
    ...
  
  print(p)

  i <- i + 1
  if (i > ggforce::n_pages(p)) {
    dev.off()
    break
  }
}

Similar approach can be employed to generate a list of ggplot objects. However, ggforce::n_pages can be time-consuming since it calls ggplot_build (compared with ggplot_add/+, which simply put all layers together). So it could be further optimized if you know how many pages are there in advance.

[billdenney]

@pormr , yes, that works and is what I do when I'm the only one writing the code. I want to include this in a package that I develop (ggtibble), so I won't know what the call to facet_wrap_paginate or facet_grid_paginare looks like.

[pormr]

I see. It seems to be fine to modify the value of page, as all arguments passed to the faceting function will be stored in p$facet$params (including default ones). This behavior is unlikely to change unless there is a significant update to {ggplot2}'s underlying faceting system. However, I am uncertain about how to properly create a "clone" of an existing ggplot object and wrap all copies into a list, you may want to consult the package author or the ggplot2 extension community.