Closed kylebaron closed 8 months ago
This PR also shuffles the arguments for
pm_grid()
; I don't believe this will change current behavior (?)
I suppose the main change in behavior can come when callers have been passing unnamed args through. wrap_plots()
signature is patchwork::wrap_plots(..., ncol = 3, {more named args})
, so any unnamed args would get absorbed by its ...
.
fa <- function(x, ..., y = 2) {
fc(x, y = y, ...)
}
fb <- function(x, y = 2, ...) {
fc(x, y = y, ...)
}
fc <- function(..., y = 200) {
message("y: ", deparse1(y))
message("...: ", deparse1(list(...)))
}
fa(1, 3)
#> y: 2
#> ...: list(1, 3)
fb(1, 3)
#> y: 3
#> ...: list(1)
But, in this case, it seems very unlikely that callers are passing unnamed args through. patchwork::wrap_plots()
treats ...
as additional ggplot2 objects, and I don't suspect any callers rely on that rather than including the plot in x
. I of course leave that judgement call up to you, but I think the move is fine.
Summary
This PR falls out of #77, where we created panels of plots and potentially annotated them with
tag_levels
. This lets the user to passtag_levels
topm_grid()
to tag levels after arranging. This will makepm_grid()
more useful for other, existing functions that return lists of plots which are typically arranged withpm_grid()
.This PR also shuffles the arguments for
pm_grid()
; I don't believe this will change current behavior (?) but thought it would be safest to push...
to the back.