Layouts

Format Objects

library(dyn.log)

configs <- dyn.log::get_configurations(pkgname = "dyn.log")

init_logger(configs$knitr)
#> TRACE [03/26/22 18:17:49 +0000] dyn.log loaded 'knitr' configuration successfully.

Overview

Log Layouts are the mechanism that ties everything in the package together with a composition based design. A log layout is essentially a container for formats, which defines exactly how the log message will get rendered. Layouts also have additional metadata attributes which allow for overrides to the default format separators, spacing, etc., as well as mechanism for associating a particular log layout with a type of R6 class.

Defaults

There are two basic layouts that come with the standard configuration (default.yaml) bundled with the package, default and log_level.

layouts:
- association: default
  seperator: ' '
  new_line: \n
  formats: new_fmt_log_level(),
           new_fmt_timestamp(crayon::silver$italic),
           new_fmt_log_msg()
- association: level_msg
  seperator: ' '
  new_line: \n
  formats: new_fmt_log_level(),
           new_fmt_log_msg()

The R equivalent of creating the default layout is essentially the same as the yaml definition:

new_log_layout(
  format = list(
    new_fmt_log_level(),
    new_fmt_timestamp(crayon::silver$italic),
    new_fmt_log_msg()    
  ),
  association = "default_via_r"
)

An info level log message generated with the package default config:

Logger$info("this is the default layout")

INFO [03/26/22 18:17:50 +0000] this is the default layout

Now with the new layout defined above:

Logger$info("this is the custom layout object", layout = "default_via_r")

INFO [03/26/22 18:17:50 +0000] this is the custom layout object

which render the exact same output, the idea here is that defining layouts via custom R code or via yaml in the configuration are exactly the same.

Customization

To create a customized log layout you only need to specify what formats you want rendered (and their arguments) in the formats argument, as seen above. See the Formats vignette for more detail on formats.

new_log_layout(
  format = list(
    new_fmt_metric(crayon::green$bold, "sysname"),
    new_fmt_metric(crayon::yellow$bold, "release"),
    new_fmt_line_break(),
    new_fmt_log_level(),
    new_fmt_timestamp(crayon::silver$italic),
    new_fmt_exec_scope(crayon::magenta$bold, "calling_fn"),
    new_fmt_literal(crayon::blue$italic, "literal text"),
    new_fmt_log_msg(),
    new_fmt_line_break(),
    new_fmt_exec_scope(crayon::cyan$bold, "call_stack")
  ),
  seperator = '-',
  association = "log-with-callstack"
)

log_fn <- function() {
  outer <- function() {
    inner <- function() {
      
      var1 <- "abc"; var2 <- 123; var3 <- round(runif(1), digits = 6)
      
      Logger$debug("my log message - var1: '{var1}', var2: '{var2}', var3: '{var3}'", 
                   layout = 'log-with-callstack')
    }
    inner()
  }
  outer()
}

log_fn()

Linux-5.11.0-1028-azure
DEBUG-[03/26/22 18:17:50 +0000]-get_call_stack-literal text-my log message - var1: 'abc', var2: '123', var3: '0.84236'
NA-NA-NA-NA-NA-NA-NA-NA-NA-NA-NA-NA-NA-NA-get_call_stack

Brandon Moretz

2022-03-26

Format Objects

Overview

Defaults

Customization