Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/evamaerey/ggwipe

remove stat and geom layers from plot; return the last plot but wiped
https://github.com/evamaerey/ggwipe

Last synced: 11 days ago
JSON representation

remove stat and geom layers from plot; return the last plot but wiped

Awesome Lists containing this project

README 

        
---

output: 

  github_document:

    toc: TRUE

---



```{r, include = FALSE}

knitr::opts_chunk$set(

  collapse = TRUE,

  comment = "#>"

)

library(tidyverse, warn.conflicts = F)

ggplot2::theme_set(theme_gray(base_size = 18))

```



# {ggwipe}: print the last plot and remove stat/geom/annotate layers in one step 



In general ggplot2 lets you work in very fluid ways, including undoing your work.  For example facet_null() and aes(fill = NULL) are moves that allow you to 'undo' your faceting and fill decisions.  



However, when it comes to layers, ggplot2 itself does not allow you to remove stat/geom/annotation layers dynamically.



ggwipe::last_plot_wipe() lets you remove all layers or specified layers from the 'last plot'.  



There are other approaches to this problem which I summarize in the other work section.  The difference between ggwipe::last_plot_wipe() and these approaches is that last_plot_wipe doesn't require you to save out your plot as an object.  Instead it favors a more exploratory work flow, so uses ggplot2::last_plot() retrieve the last plot, modifies it by removing layers as specified, before it's returned back to the user for presentation or further manipulation. 



```{r, out.width="49%", fig.show="hold", error = T}

library(tidyverse, warn.conflicts = F)

library(ggwipe)

mtcars %>% 

  ggplot() + 

  aes(am, fill = factor(vs)) +

  geom_bar()



last_plot_wipe() + 

  geom_bar(position = "fill")

```



```{r, out.width="49%", fig.show="hold", error = T}

mtcars %>% 

  ggplot() + 

  aes(am, fill = factor(vs)) +

  geom_bar()



last_plot_wipe() + 

  aes(fill = NULL, 

      y = vs) + 

  geom_count()



last_plot_wipe() 



```



## You can specify the specific layer, with the `index = n` argument



```{r, out.width="49%", fig.show="hold"}

ggplot(data = cars) + 

  aes(x = speed, y = dist) + 

  geom_rug() + 

  geom_point() + 

  geom_smooth()



last_plot_wipe(index = 1)  # removes rug

```



## You can also use it for backtracking - removing the most recent layer with `last_plot_wipe_last()`.



This function automatically detects the last layer.



```{r, out.width="49%", fig.show="hold"}

ggplot(data = cars) + 

  aes(x = speed, y = dist) + 

  geom_rug() + 

  geom_point() + 

  geom_smooth()



last_plot_wipe_last()

last_plot_wipe_last()

last_plot_wipe_last()

```



# Curious about implementation? Details about building these functions



# Step 0. Status quo, manually removing layers



```{r}

p <- mtcars %>% 

  ggplot() + 

  aes(am, fill = factor(vs) ) +

  geom_bar()



p



p[[2]] <- NULL # removes all layers specification



p



last_plot() + 

  geom_bar(position = "fill")

```



# Step 0.b removing a specific layer.



```{r}

p <- mtcars %>% 

  ggplot() + 

  aes(am, fill = factor(vs) ) +

  geom_bar() + 

  stat_count(geom = "point", size = 14) + 

  stat_count(geom = "label", aes(label = after_stat(count)))



p



p[[2]][[2]] <- NULL # removes second layer specification



p



```



# put it in a function: `last_plot_wipe`



```{r last_plot_wipe}

#' Title

#'

#' @param index 

#'

#' @return

#' @export

#'

#' @examples

last_plot_wipe <- function(index = NULL){

  

  

  p <- ggplot2::last_plot()

  

  # removes all layers specification

  if(is.null(index)){ p[[2]] <- NULL }else{p[[2]][[index]] <- NULL}

  

  return(p)

  

}

```



## Try it out.



```{r}

mtcars %>% 

  ggplot() + 

  aes(am, fill = factor(vs)) +

  geom_bar()



last_plot_wipe() + 

  geom_bar(position = "fill")



mtcars %>% 

  ggplot() + 

  aes(am, fill = factor(vs) ) +

  geom_bar() + 

  stat_count(geom = "point", size = 14) + 

  stat_count(geom = "label", aes(label = after_stat(count)))



last_plot_wipe(index = 2)



```



# A convenience function, last_plot_wipe_last



To only wipe the final layer



```{r last_plot_wipe_last}

#' Title

#'

#' @return

#' @export

#'

#' @examples

last_plot_wipe_last <- function(){

  

  

  p <- ggplot2::last_plot()

  

  index <- length(p[[2]])

  

  # removes all layers specification

  p[[2]][[index]] <- NULL

  

  return(p)

  

}

```



## Try it



```{r}

mtcars %>% 

  ggplot() + 

  aes(am, fill = factor(vs) ) +

  geom_bar() + 

  stat_count(geom = "point", size = 14) + 

  stat_count(geom = "label", aes(label = after_stat(count)))



last_plot_wipe_last()

```



# Other work



I think the difference is that other work would have you save the ggplot2 object and then manipulate it, where as ggwipe lends itself to printing the ggplot2 object and then retrieving and printing without a new layer in one step.  



- gghost

- pedro alpho

- Jonathan Sidi



```{r, eval = F, echo = F}

knitr::knit_exit()

```



# Part 2. Packaging and documentation  🚧 ✅ 



## minimal requirements for github package.  Have you:



### Created files for package archetecture with `devtools::create("./ggbarlabs")` ✅ 



### Moved functions R folder? ✅  



```{r}

knitr::knit_code$get() |> names()

```



```{r}

readme2pkg::chunk_to_r(c("last_plot_wipe", 

                         "last_plot_wipe_last"))

```



### Added roxygen skeleton? ✅ 



 for auto documentation and making sure proposed functions are *exported*



### Managed dependencies ? ✅ 



package dependancies managed, i.e. `depend::function()` in proposed functions and declared in the DESCRIPTION



```{r, eval = F}

usethis::use_package("ggplot2")

```



### Chosen a license? ✅ 



```{r, eval = F}

usethis::use_mit_license()

```



### Run `devtools::check()` and addressed errors? 🚧 🚧 🚧 🚧 🚧 



## Listen 🚧 



### Consulted with technical experts 🚧 

### Consulted with potential users 🚧 



Getting started with that!



## Polish.  Have you.



### Settled on examples and put them in the roxygen skeleton? 🚧 



### Written formal tests of functions? 🚧 



### Sent tests in this readme to package via readme2pkg 🚧 



That would look like this...



```

chunk_to_tests_testthat("test_geom_barlab_count")

```



### Have you worked added a description and author information in the DESCRIPTION file? 🚧 



### Addressed *all* notes, warnings and errors. 🚧 



## Promote



### Package website built? 🚧 



### Package website deployed? 🚧 



## Harden



### Submit to CRAN?  🚧 



# Reports, Environment



## Description file extract



```{r}



```



## Environment



Here I just want to print the packages and the versions



```{r}

all <- sessionInfo() |> print() |> capture.output()

all[11:17]



```



## `devtools::check()` report



```{r, error = T, eval = F}

# rm(list = c("geom_barlab_count", "geom_barlab_count_percent"))

devtools::check(pkg = ".")

```