/GroveR

Hierarchical dependency management of objects in R, a la "make"

Primary LanguageR

Build Status

GroveR brings Declarative Programming to data-centered applications. For example, suppose your code looks like this:

create_foo <- function(bar, BaZzO) {
  ## create foo stuff here...
}

create_bar <- function() {
  ## create bar stuff here...
}

create_baz <- function() {
  ## create baz stuff here...
}

create_quux <- function(bazz) {
  ## create quux stuff here...
}

x <- create_bar()
BAZ <- create_baz()
result <- create_foo(x, BAZ)

Using the tools of GroveR, that can be transformed to this:

App <- GroveR$new()
`%createdBy%` <- App$auto

foo %createdBy% function(bar, baz) {
  ## create foo stuff here...
}

bar %createdBy% function() {
  ## create bar stuff here...
}

baz %createdBy% function() {
  ## create baz stuff here...
}

quux %createdBy% function(baz) {
  ## create quux stuff here...
}

## This builds any dependencies necessary to create 'foo'
result <- App$getArtifact('foo')

As benefits, the application now has:

  • automatic caching of intermediate artifacts

  • intelligent rebuilding of artifacts when dependencies change

  • standardized naming of data artifacts throughout the application

  • configurable logging of artifact creation using futile.logger

  • graphical tools to show data dependencies (colors indicate what's up-to-date or not):

    data dependencies

Copyright (c) 2016 Windlogics, Inc. See the DESCRIPTION file for licensing information.