make map_signif_level more flexible #52
Conversation
|
Hi, thank you very much for the pull request. I am currently travelling, but will take a look next week. |
draw_key_point leads to weird errors when trying to use manual xmin/xmax/annotation with data grouped by multiple aesthetics
R/significance_annotation.R
Outdated
| }else{ | ||
| if(is.numeric(p_value)){ | ||
| if(p_value < .Machine$double.eps){ | ||
| sprintf("p < %.2e", .Machine$double.eps) | ||
| }else{ | ||
| as.character(sprrintf("p = %.2g"), p_value) | ||
| as.character(sprintf("p = %.2g"), p_value) |
There was a problem hiding this comment.
I think printing just the number without the "p = " in front is the better default, can you change that back?
There was a problem hiding this comment.
done. I just thought it would be more consistent with "p < 2e-16".
| @@ -215,7 +215,7 @@ GeomSignif <- ggplot2::ggproto("GeomSignif", ggplot2::Geom, | |||
| required_aes = c("x", "xend", "y", "yend", "annotation"), | |||
| default_aes = ggplot2::aes(shape = 19, colour = "black", textsize = 3.88, angle = 0, hjust = 0.5, | |||
| vjust = 0, alpha = NA, family = "", fontface = 1, lineheight = 1.2, linetype=1, size=0.5), | |||
| draw_key = ggplot2::draw_key_point, | |||
| draw_key = function(...){grid::nullGrob()}, | |||
There was a problem hiding this comment.
completely prevents drawing of a legend key for the ggsignif layer. In my case, I was trying to use ggsignif with geom_sina from ggforce and kept getting weird errors until I made this change.
| @@ -145,6 +145,7 @@ StatSignif <- ggplot2::ggproto("StatSignif", ggplot2::Stat, | |||
| #' @param map_signif_level boolean value, if the p-value are directly written as annotation or asterisks are used instead. | |||
| #' Alternatively one can provide a named numeric vector to create custom mappings from p-values to annotation: | |||
| #' For example: c("***"=0.001, "**"=0.01, "*"=0.05) | |||
| #' Alternatively, one can provide a function that takes a numeric argument (the p-value) and returns a string | |||
There was a problem hiding this comment.
Could you maybe also add an example that demonstrates how to use the custom function, for example to show how to print the p-values as "p = 0.03"?
|
Thank you again, I have merged the PR and will submit the updated package to CRAN soon. |
map_signif_level can now take a user-supplied function to format the p-value. Also, this pull-request changes the default threshold for printing exact p-values to .Machine$double.eps (should still be 2e-16 on most computers)