CODEWITHSUNDEEP
rr6
Brandon Loudermilk
Brandon Loudermilk

Reputation: 970

Using set to dynamically construct R6 class and instance

I'd like to create a function create_thresholds that dynamically creates an R6 class and an instance object. The idea is to dynamically create a data structure that has read-only active fields that client-objects can access similar to the example illustrated by Wickham (https://adv-r.hadley.nz/r6.html - 16.3.2 Active fields). I've looked at dynamically add function to r6 class instance but I am looking for a solution that modifies the class definition, not add a function to an instance object.

The function create_thresholds reads in a data.frame of var/val pairs, then creates an R6Class object on which it tries to set various private fields and create read-only active fields. The function terminates by creating a new instance of the Thresholds object.

As illustrated in the code below, calling the active methods appears to evaluate the method expressions in the incorrect environment.

library("R6")


create_thresholds <- function(df) {
  require(glue)
  Thresholds <- R6Class("Thresholds")

    for (i in 1:nrow(df)){
      e <- environment()
      print(e)
      Thresholds$set("private", glue(".{df$var[i]}"), df$val[i])
      Thresholds$set("active", glue("{df$var[i]}"), function() {
        eval(parse(text = glue("private$.{df$var[i]}")))
      })
    }

  th <- Thresholds$new()

  return(th)
}

df <- tibble::tribble(~var, ~val,
                      "min_units", 100,
                      "min_discount", 999)


th <- create_thresholds(df)
th$min_discount ## expect 999
th$min_units ## OOPS! expect 100

Upvotes: 2

Views: 475

Answers (1)

Brandon Loudermilk
Brandon Loudermilk

Reputation: 970

The following solution entails dynamically creating a string form of the entire function/method definition, and then calling eval(parse()) on it.

library("R6")


df <- tibble::tribble(~var, ~val,
                      "min_units", 100,
                      "min_discount", 999)

create_thresholds <- function(df) {
  Thresholds <- R6Class("Thresholds")

  for (i in 1:nrow(df)){
    mthd_name <- df$var[i]
    mthd_def <- glue::glue("function() private$.{mthd_name}")
    Thresholds$set("private", glue(".{mthd_name}"), df$val[i])
    Thresholds$set("active", mthd_name, eval(parse(text = mthd_def)))
  }

  hh <- Thresholds$new()
  return(hh)
}

hh <- create_thresholds(df)
hh$min_discount # expect 999!
hh$min_units #expect 100

Upvotes: 1

Related Questions