boomer and grouped operations

[krlmlr]

options(conflicts.policy = list(warn = FALSE))
library(dplyr)
tibble(a = 1:3) |>
  mutate(.by = a, b = a + 1) |>
  boomer::boom()
#> 💣 mutate(tibble(a = 1:3), .by = a, b = a + 1) 
#> · 💣 tibble(a = 1:3) 
#> · · 💣 💥 1:3 
#> · · [1] 1 2 3
#> · · 
#> · 💥 tibble(a = 1:3) 
#> · # A tibble: 3 × 1
#> ·       a
#> ·   <int>
#> · 1     1
#> · 2     2
#> · 3     3
#> · 
#> · 💣 💥 a + 1 
#> · [1] 2
#> · 
#> · 💣 💥 a + 1 
#> · [1] 3
#> · 
#> · 💣 💥 a + 1 
#> · [1] 4
#> · 
#> 💥 mutate(tibble(a = 1:3), .by = a, b = a + 1) 
#> # A tibble: 3 × 2
#>       a     b
#>   <int> <dbl>
#> 1     1     2
#> 2     2     3
#> 3     3     4
#> # A tibble: 3 × 2
#>       a     b
#>   <int> <dbl>
#> 1     1     2
#> 2     2     3
#> 3     3     4

^{Created on 2024-01-18 with reprex v2.0.2}

I was only somewhat surprised to see the multiple invocations of a + 1 shown here, but for larger data this becomes difficult to review. If we collected the results (as suggested in #72), we could render as an HTML document or show in a Shiny app, and collapse/expand/filter as needed.

Alternatively, if there was a way to stop booming everything inside mutate(), those wouldn't be shown. options(boomer.ignore = "mutate") doesn't do the right thing here, do we need options(boomer.ignore.inside) ?

[moodymudskipper]

+ is the printing function here, that's why ignoring mutate doesn't work.

Maybe with options(boomer.ignore.inside = TRUE) when the function is entered the environment containing the shims could be dropped and set back on.exit.

Meanwhile maybe boom_on() and boom_off() can help

options(conflicts.policy = list(warn = FALSE))
library(dplyr)
library(boomer)
fun <- function() {
  boom_on()
  x <- tibble(a = 1:3)
  boom_off()
  mutate(x, .by = a, b = a + 1)
}
fun()
#> 💣 tibble(a = 1:3) 
#> · 💣 💥 1:3 
#> · [1] 1 2 3
#> · 
#> 💥 tibble(a = 1:3) 
#> # A tibble: 3 × 1
#>       a
#>   <int>
#> 1     1
#> 2     2
#> 3     3
#> 
#> 💣 💥 boom_off() 
#> NULL
#> # A tibble: 3 × 2
#>       a     b
#>   <int> <dbl>
#> 1     1     2
#> 2     2     3
#> 3     3     4

^{Created on 2024-01-18 with reprex v2.0.2}

[krlmlr]

Hm... I do want to see the results of the mutate() call.

[moodymudskipper]

OK I think I know how to do it, however, in the above call using the base pipe it's worth noting that the tibble call is also "inside", with the magrittr pipe it would be outside