icecream is designed to make print debugging easier. It allows you to print out an expression, its value and (optionally) which function and file the call originated in.

This is an R port of gruns/icecream. All credit for the idea belongs to Ansgar Grunseid.

Installation

Install from CRAN with:

install.packages("icecream")

Or you can install the development version from GitHub with:

devtools::install_github("lewinfox/icecream")

Inspect variables

The ic() function prints its argument and its value. It also returns the value of the evaluated argument, meaning that it is effectively transparent in code - just wrap an expression in ic() to get debugging output.

library(icecream)

is_negative <- function(x) x < 0

ic(is_negative(1))
#> ℹ ic| `is_negative(1)`: logi FALSE

ic(is_negative(-1))
#> ℹ ic| `is_negative(-1)`: logi TRUE

You’re more likely to want to do this within a function:

some_function <- function(x) {
  intermediate_value <- x * 10
  answer <- ic(intermediate_value / 2)
  return(answer)
}

some_function(1)
#> ℹ ic| `intermediate_value / 2`: num 5
#> [1] 5

some_function(10)
#> ℹ ic| `intermediate_value / 2`: num 50
#> [1] 50

More complex inputs like lists and data frames are summarised to avoid cluttering the terminal.

df <- ic(iris)
#> ℹ ic| `iris`: data.frame [150 x 5]: $'Sepal.Length': dbl [150], ...

my_list <- ic(list(a = 1, b = 3, c = 1:100))
#> ℹ ic| `list(a = 1, b = 3, c = 1:100)`: list [3]: $'a': dbl [1], $'b': dbl [1], $'c': int [100]

Inspect execution

Calling ic() with no arguments causes it to print out the file, line and parent function it was called from.

In this example we have a file demo.R that contains two functions. We’ve inserted ic() calls at strategic points so we can track what’s being executed.

# demo.R
f1 <- function(x) {
  ic()
  if (x > 0) {
    f2()
  }
}

f2 <- function() {
  ic()
}

f3 <- function(x) {
  ic(x)
}
source("demo.R")

f1(-1)
#> ℹ ic| `global::f1()` in demo.R:3:2

f1(1)
#> ℹ ic| `global::f1()` in demo.R:3:2
#> ℹ ic| `global::f2()` in demo.R:10:2

In the case of functions that haven’t been source()d or loaded from a package there is no source code to refer to. In these cases the function’s environment will be displayed.

orphan_func <- function() {
  ic()
  TRUE
}

orphan_func()
#> ℹ ic| `global::orphan_func()` in <env: global>
#> [1] TRUE

e <- new.env()
attr(e, "name") <- "icecream_van"
environment(orphan_func) <- e

orphan_func()
#> ℹ ic| `orphan_func()` in <env: icecream_van>
#> [1] TRUE

Enable / disable

The ic_enable() and ic_disable() functions enable or disable the ic() function. If disabled, ic() will return the result of evaluating its input but will not print anything.

ic_enable() # This is TRUE by default

ic(mean(1:100))
#> ℹ ic| `mean(1:100)`: num 50.5

ic_disable()

ic(mean(1:100))
#> [1] 50.5

Convenience functions with_ic_enable() and with_ic_disable() are also provided.

ic_enable()

with_ic_disable(ic(mean(1:100)))
#> [1] 50.5

ic_disable()

with_ic_enable(ic(mean(1:100)))
#> ℹ ic| `mean(1:100)`: num 50.5

Options

The following options can be used to control behaviour:

icecream.enabled

Boolean. If FALSE, calls to ic(foo) simply evaluate and return foo. No output is printed. This option can be set directly or with the ic_enable() and ic_disable() functions.

icecream.prefix

This is printed at the beginning of every line. Defaults to "ic|".

ic(mean(1:5))
#> ℹ ic| `mean(1:5)`: num 3

options(icecream.prefix = "DEBUG:")
ic(mean(1:5))
#> ℹ DEBUG: `mean(1:5)`: num 3

options(icecream.prefix = "\U1F366")
ic(mean(1:5))
#> ℹ 🍦 `mean(1:5)`: num 3

icecream.always.include.context

Boolean. If TRUE, when calling ic(foo) the source file and line will be printed along with the expression and value. If no srcref() is available the function’s environment will be displayed instead. This can be useful for more complicated debugging but produces a lot of output so is disabled by default.

f3(1)
#> ℹ ic| `x`: num 1

options(icecream.always.include.context = TRUE)

f3(1)
#> ℹ ic| `global::f3()` in demo.R:14:2 | `x`: num 1

When ic() is called with no arguments, the context is always printed because showing the location of the call is the only reason to call ic() on its own.

icecream.peeking.function and icecream.max.lines

These two options control how the result of evaluation of an expression is printed. icecream.peeking.function indicates the function that summarizes the object. Default value is ic_autopeek, which works like utils::str for most of the time, but gives more informative output for lists, data.frames and their subclasses in a more compact way. icecream.max.lines determines maximum number of lines that the peek of an object occupies; defaults to 1.

For more complex data you may want to use e.g. head function and 5 lines.

data(iris)

ic(iris) # we would like to see header of the data
#> ℹ ic| `iris`: data.frame [150 x 5]: $'Sepal.Length': dbl [150], ...

options(
  icecream.peeking.function = head,
  icecream.max.lines = 5
)

ic(iris)
#> ℹ ic| `iris`:  Sepal.Length Sepal.Width Petal.Length Petal.Width Species 1          5.1         3.5          1.4         0.2  setosa 2          4.9         3.0          1.4         0.2  setosa 3          4.7         3.2          1.3         0.2  setosa 4          4.6         3.1          1.5         0.2  setosa

NOTE: Due to r-lib/cli#626, newlines are currently being removed from ic()’s multi-line output.

Note that if icecream.max.lines is greater than 1 and summary of an object is longer than 1, the alert occupies one line more due to the header.

icecream.output.function, icecream.arg.to.string.function

Not implemented yet. See the configuration section of the original project docs for details of what they will do.

TODO:

  • Implement ic.format() (see here).
  • Implement ic.output.function. At the moment it uses cli::cli_alert_info()
  • Implement ic.arg.to.string.function