In daroczig/logger: A Lightweight, Modern and Flexible Logging Utility

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

logger

A lightweight, modern and flexible logging utility for R -- heavily inspired by the futile.logger R package and logging Python module.

Installation

#| eval: false
install.packages("logger")

The most recent, development version of logger can also be installed from GitHub:

#| eval: false
# install.packages("pak")
pak::pak("daroczig/logger")

Quick example

Setting the log level threshold to something low and logging various messages in ad-hoc and programmatic ways:

#| include: false
library(logger)
log_appender(appender_stdout)

library(logger)
log_threshold(DEBUG)
log_info("Script starting up...")

pkgs <- available.packages()
log_info("There are {nrow(pkgs)} R packages hosted on CRAN!")

for (letter in letters) {
  lpkgs <- sum(grepl(letter, pkgs[, "Package"], ignore.case = TRUE))
  log_level(
    if (lpkgs < 5000) TRACE else DEBUG,
    "{lpkgs} R packages including the {shQuote(letter)} letter"
  )
}

log_warn("There might be many, like {1:2} or more warnings!!!")

You can even use a custom log layout to render the log records with colors, as you can see in layout_glue_colors():

colored log output

But you could set up any custom colors and layout, eg using custom colors only for the log levels, make it grayscale, include the calling function or R package namespace with specific colors etc. For more details, see vignette("write_custom_extensions").

Related work

There are many other logging packages available on CRAN:

futile.logger: probably the most popular log4j variant (and I'm a big fan)
logging: just like Python's logging package
lgr: built on top of R6.
loggit: capture message, warning and stop function messages in a JSON file
log4r: log4j-based, object-oriented logger
rsyslog: logging to syslog on 'POSIX'-compatible operating systems
lumberjack: provides a special operator to log changes in data

Why use logger? I decided to write the n+1th extensible log4j logger that fits my liking --- and hopefully yours as well --- with the aim to:

Keep it close to log4j.
Respect the modern function/variable naming conventions and general R coding style.
By default, rely on glue() when it comes to formatting / rendering log messages, but keep it flexible if others prefer sprintf() (e.g. for performance reasons) or other functions.
Support vectorization (eg passing a vector to be logged on multiple lines).
Make it easy to extend with new features (e.g. custom layouts, message formats and output).
Prepare for writing to various services, streams etc
Provide support for namespaces, preferably automatically finding and creating a custom namespace for all R packages writing log messages, each with optionally configurable log level threshold, message and output formats.
Allow stacking loggers to implement logger hierarchy -- even within a namespace, so that the very same log call can write all the TRACE log messages to the console, while only pushing ERRORs to DataDog and eg INFO messages to CloudWatch.
Optionally colorize log message based on the log level.
Make logging fun!