cache_to_filearray: Generate and automatically cache a file array

In beauchamplab/raveio: File-System Toolbox for RAVE Project

Generate and automatically cache a file array

Description

Avoid repeating yourself

Usage

cache_to_filearray(
  fun,
  filebase,
  globals,
  dimension,
  type = "auto",
  partition_size = 1L,
  verbose = FALSE,
  ...
)

Arguments

fun	function that can be called with no mandatory arguments; the result should be in a matrix or an array
filebase	where to store the array
globals	names of variables such that any changes should trigger a new evaluation of fun. This argument is highly recommended to be set explicitly (with atomic variables) though the function automatically calculates the global variables
dimension	what is the supposed dimension, default is automatically calculated from array. If specified explicitly and the file array dimension is inconsistent, a new calculation will be triggered.
type	file array type, default is "auto"; can be explicitly specified; see filearray_create. Inconsistent type will trigger a new calculation.
partition_size	file array partition size; default is 1; set it to NA to generate it automatically. Notice inconsistent partition size will not trigger calculation if the key variables remain the same
verbose	whether to verbose debug information
...	passed to findGlobals

Examples

c <- 2
b <- list(d = matrix(1:9,3))
filebase <- tempfile()

f <- function() {
  message("New calculation")
  re <- c + b$d
  dimnames(re) <- list(A=1:3, B = 11:13)

  # `extra` attribute will be saved
  attr(re, "extra") <- "extra meta data"
  re
}

# first time running
arr <- cache_to_filearray( f, filebase = filebase )

# cached, no re-run
arr <- cache_to_filearray( f, filebase = filebase )

# file array object
arr

# read into memory
arr[]

# read extra data
arr$get_header("extra")

# get digest results
arr$get_header("raveio::filearray_cache")

## Clean up this example
unlink(filebase, recursive = TRUE)

beauchamplab/raveio documentation built on April 24, 2024, 11:20 p.m.