monads

The goal of monads is to provide a ‘toy’ implementation of some common monads.

See this blog post for some additional background.

Installation

You can install the development version of monads like so:

# install.packages("remotes")
remotes::install_github("jonocarroll/monads")

Example

With the addition of a new ‘flatMap’ pipe %>>=% a monad instance can be passed to regular functions, unpacked, applied, and re-packaged. One example is to create a ‘logger’ which records the function applications at each step.

library(monads)
library(dplyr, warn.conflicts = FALSE)

result <- loggerM(mtcars) %>>=%
  filter(mpg > 10) %>>=%
  select(mpg, cyl, disp) %>>=%
  arrange(desc(mpg)) %>>=%
  head()

value(result)
#>                 mpg cyl  disp
#> Toyota Corolla 33.9   4  71.1
#> Fiat 128       32.4   4  78.7
#> Honda Civic    30.4   4  75.7
#> Lotus Europa   30.4   4  95.1
#> Fiat X1-9      27.3   4  79.0
#> Porsche 914-2  26.0   4 120.3

logger_log(result)
#> ✔ Log of 4 operations:
#> 
#>  mtcars %>%
#>    filter(mpg > 10) %>%
#>    select(mpg, cyl, disp) %>%
#>    arrange(desc(mpg)) %>%
#>    head()

This package also defines Maybe, Result, Timer, and List monads which all work with this new pipe. Additional monads can be added by third-parties to further extend the capabilities.

See the vignette for more details.

Prior Art

• {monads} - a sketched-out implementation that relies on dispatch for flatMap operations.
• {rmonad} - archived on CRAN, but offers a sophisticated ‘funnel’ mechanism and various ways to capture steps of a pipeline.
• {maybe} - a more detailed implementation of Maybe.
• {chronicler} - a way to post-process the result at each step and capture information, such as the runtime (see Timer) or dimensions. Requires an explicit bind() at each step. Associated blog post.