codeexplode

flipbooks – go-at-your-own-pace, code and output animations – have up to this point been crafted for education and demonstration purposes.

The flipbook format doesn’t make much sense for getting actual work done. It is inefficient. We repeat steps (step 1, steps 1-2, steps 1-3) to create the animation.

However, code that isn’t developed with a flipbook end presentation in mind might also might be easier to consume in the flipbook format. It might be transformed into a educational or code-review tool on occasion.

The goal of codeexplode is to take a .Rmd input (maybe a .R file later), and convert this to a flipbook product.

This is done by creating a new .Rmd file from a source .Rmd file where

  1. the original yaml is replaced with a flipbook-ready yaml (output to xaringan)
  2. flipbookr::chunk_reveal() is paired with each existing code chunks to ‘explode’ them; the original source code chunks are reset to include = FALSE.

Right now, {codeexplode} is not that clever. It work best for auto-type reveal, where code is written in a sequential way and you would be able to see outputs as you go. (i.e. right assignment).

Installation

And the development version from GitHub with devtools::install_github("EvaMaeRey/codeexplode")

Code that plays nice with codeexplode

The following code works well because it is piped, and most pause points (where parentheses are balanced at the end of a line) will produce an output

summary(cars) ->
  summary_cars

library(ggplot2) 
cars %>% 
  ggplot() + 
  aes(x = speed, y = dist) + 
  geom_point(size = 4) + 
  aes(color = speed) + 
  scale_color_viridis_c()

codeexplode in action

Now let’s also look at what codeexplode does with this .Rmd.

codeexplode::rmd_code_explode("README.Rmd",
  rmd_output = "docs/readme_exploded.Rmd")

I rendered the exploded and rendered the .Rmd outside this file, because if felt to mind-bendy to do here. The resultant flipbook (.html) is shown here.