In pbiecek/archivist2: Tools for Storing, Restoring and Searching for R Objects

Let's prepare some set to be archived.

library(ggplot2)
library(ggthemes)
library(archivist2)
createEmptyRepo( "SETS" )

setLocalRepo( "SETS" )
data(iris)

plotArtifact <- ggplot( iris, aes(x = Sepal.Length, y = Species)) +
  geom_point()+
  theme_wsj()

plotData <- iris
plotFunctions <- list( ggplot, geom_point, theme_wsj)

Such set now can be archived using saveSetToRepo function

saveSetToRepo( artifact = plotArtifact,
                           data = plotData,
                           functions = plotFunctions)

Description

saveSetToRepo function saves desired set of artifacts to the local Repository in a given directory. To learn more about artifacts visit http://pbiecek.github.io/archivist/. Set is a collection containing

• an artifact,
• data needed to create the artifact,
• list of functions needed to create the artifact.

Details

saveSetToRepo archives artifact, data and functions using saveToRepo function but additionally it adds Tags to every part of a set in convention as: set:md5hashOfArtifact to remember that all objects came originally from one set. This additional tag helps to restore a set from a Repository.

Value

As a result of this function a character strings is returned, which determines the md5hash of the archived artifact.

Some technical solutions

saveSetToRepo uses saveToRepo function so normally names of elements of a set can not be archived as a name tag. The names of all archived elements from a set are archived under tag named set:name: .

showLocalRepo

One can check archived elements that come only from sets.

showLocalRepo(method = "sets")

loadSetFromLocalRepo

Having a tag named set:name and a tag named set that specify to which artifacts this element is linked one can restore already archived set with loadSetFromLocalRepo function (not yet in the archivist package) by giving artifact's md5hash that is an output of a saveSetToRepo function.

loadSetFromLocalRepo <- function( md5hash, repoDir = NULL ){

  stopifnot( is.character( md5hash ), length( md5hash ) == 1 )

  # get hashes
  artifactHash <- md5hash
  SetElementsHashes <- searchInLocalRepo( paste0("set:", md5hash ), fixed = FALSE,
                                          repoDir = repoDir) 

  # get names
  artifactName <- getTagsLocal( artifactHash, tag="set:name", repoDir = repoDir)
  artifactName <- sub(x = artifactName, pattern = "set:name:", replacement = "")

  SetElementsNames <- sapply( SetElementsHashes, function( element ){
    nameElem <- getTagsLocal( md5hash = element, tag="set:name", repoDir = repoDir)
    sub(x = nameElem, pattern = "set:name:", replacement = "")
  })


  # assign artifact
  assign( x = artifactName, value = loadFromLocalRepo(artifactHash, repoDir = repoDir, value = TRUE), 
          envir = parent.frame(1))
  # assign elements of a set
  for( i in seq_along(SetElementsNames)){
    assign( x = SetElementsNames[i], value = loadFromLocalRepo(SetElementsHashes[i], repoDir = repoDir, value = TRUE), 
            envir = parent.frame(1))
  }

}

Example

For archived set, let's remove the set from global environment and then let's load a set from Repository

ls()

rm( list = c("plotData", "plotFunctions", "plotArtifact") )
ls()

loadSetFromLocalRepo("f0745effbbedbf160243a154c6339574")
ls()

pbiecek/archivist2 documentation built on May 24, 2019, 10:36 p.m.