In EvaMaeRey/ggmemory: What the Package Does (One Line, Title Case)

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

ggmemory

The goal of ggmemory is to present code and output as it was executed in an interactive session; support is currently focused on data manipulation pipelines and data visualization.

---

How.

• Code is harvested from the history. This can be done via RStudio IDE or other means like savehistory("temp_history.R").

• Then the user can format the code to respect indentation conventions for pipelines. Recommended is using RStudio and rstudioapi::executeCommand("reformatCode") or menu options to do the same. On wish list is to do this reformatting step programmatically formatR::tidy_source() seems promising, but does not work on code containing errors. Error free code for interactive sessions does not seem realistic.

• Then, the ggmemory function r_code_extract_code_list can be used on the reformated .R file containing the history. This function further parses code into a list containing individual pipelines or single lines of code based on indentation that compose the entire history.

library(magrittr)
'library(ggplot2)
ggplot(data = cars) +
  aes(x = speed) +
  aes(y = dist) +
  geom_point(color = "magenta")
ggplot(data = cars) +
  aes(x = speed) +
  aes(y = dist) +
  geom_point(color = "magenta",
             size = 8)
ggplot(data = cars) +
  aes(x = speed) +
  aes(y = dist) +
  geom_point(color = "magenta",
             size = 8,
             alpha = .5)' %>% 
  ggmemory::r_code_extract_code_list(code = .)

• Finally, this list with the code sequence can be entered as an argument to the function chunk_reveal from the {flipbookr} package, used with {xaringan}. Each slide in the Xaringan slide show will display a code execution moment of the interactive session by the user.
  `

'library(ggplot2)
ggplot(data = cars) +
  aes(x = speed) +
  aes(y = dist) +
  geom_point(color = "magenta")
ggplot(data = cars) +
  aes(x = speed) +
  aes(y = dist) +
  geom_point(color = "magenta",
             size = 8)' %>% 
  ggmemory::r_code_extract_code_list(code = .) ->
my_executions

flipbookr::chunk_reveal(code_seq = my_executions)

• A template is provided with the ideal set up for the {xaringan} slides; i.e. error = TRUE

Steps 3-5 might be bundled up in the future.

Precidents and Motivation.

flipbookr was designed to present curated code pipelines or code experiences (rotating arguments). Consumers can see how you might build up a chart or data manipulation pipeline. However, the experiences are rather curated, and not true to how data manipulation is actually conduction -- probably not as linear as the 'classic flipbookr' presentations.

Moreover, action within lines -- for example changing color of a point geom or size: geom_point() -> geom_point(size = 12) -- is tricky to demonstrate using classic {flipbookr} presentations. Motivation for working on {memory} are other packages, whose within-line changes are where important points of decision: {ggstamp} and {ggsample}.

Related work.

A cool package also doing work in this area -- data analytics evolution -- is {camcorder}. Once the 'camcorder' is turned on, this package saves all capture all data visualizations produced in an interactive session. Then, these are reproduced together in a gif.

EvaMaeRey/ggmemory documentation built on Dec. 17, 2021, 7:25 p.m.