initProject: Create project directories, initialize a git repo, create README.md

Description

This creates folders for the separate parts of a project. It tries to be clever about which directories are created and where they are placed. Please see details for 3 scenarios for which we have planned. If a directory already exists, it will not be damaged or re-created.

Usage

initProject(dir = NULL, ddir = "data", wdir = "workingdata",
  odir = "output", tdir = "tmp", ldir = "lit", writedir = "writeup",
  rdir = "R", ..., gitArgs = "--shared=group")

Arguments

dir

Default NULL, otherwise a legal directory name to serve as the top level directory for this project

ddir

Data directory, place where "read only" unadjusted data files are kept. Default is "data". If user sets it as NA or NULL, the directory will not be created.

wdir

Working data directory, where recorded, revised, and cleaned data files are kept. Default is "workingdata".

odir

Output directory. Default is "output".

tdir

Temporary directory, where trash files can be kept for safe keeping. Default is "tmp".

ldir

Literature directory, where material about the project can be stored. Default is "lit".

writedir

The folder where the project writeup will be stored. Default is "writeup".

rdir

The name to be used for the R files. Defaults to "R".

...

A list of other directories that the user would like to create. For example, adir = "admin",

cdir =
"client_provided"

, bdir = "codebooks",

sdir =
"Stata"

, mdir = "Mplus". These may be grouped in a named vector or list, if user convenience dictates.

gitArgs

This function tries to run "git init" and in our center we add "--shared=group" on a network file server. If that is undesirable in a user's context, put the argument gitArgs as "".

Value

Name of project top level directory. Leaves the R working directory unchanged.

Details

If the dir argument is NULL, as default, then the current working directory will be the place where new directories and the git repository will be created. Assuming the current working directory's base name is not "R", then folders named "R", "data", and so forth will be created in the current working directory.

If one has a current R working directory with a basename "R" (suppose it is "/tmp/whatever/R"), and the user runs initProject(), something different happens. The function assumes we don't want to create subdirectories inside R. We don't want to end up with "/tmp/whatever/R/R". We don't want "/tmp/whatever/R/data" either. Instead, it assumes we want the new directories created on same level as R, so it creates "/tmp/whatever/data", "/tmp/whatever/workingdata", and so forth. From within the R directory, these new directories are seen as "../data", "../workingdata", and so forth. That is, we should end up with directories and a git repo in "/tmp/whatever".

If the dir argument is provided by the user, then that is used as the folder in which directories "R", "data", "workingdate", and so forth are created. All materials are created in dir, no matter what the current working directory is named (even if it is "R").

The examples demonstrate all three of these scenarios.

Examples

Run this code

tempdir <- tempdir()
setwd(tempdir)
dir.create("test0")
setwd("test0")
initProject()
list.files(all.files = TRUE)
setwd(tempdir)
dir.create("test1")
setwd("test1")
## demonstrate ability to create other directories
initProject(admin = "admin", clientfiles = "client")
list.files(all.files = TRUE)
setwd(tempdir)
dir.create("test2/R", recursive = TRUE)
setwd("test2/R")
initProject(adir = "accounting")
list.files(all.files = TRUE)
list.files("../", all.files = TRUE)
setwd(tempdir)
initProject("test3")
list.files("test3", all.files = TRUE)
setwd(tempdir)
## demonstrate ability to create other directories
initProject(file.path(tempdir, "test4"),
            list(mdir = "Mplus", sasdir = "SAS"))
list.files(file.path(tempdir, "test4"), all.files = TRUE)
setwd(tempdir)
## demonstrate ability to nullify standard directories
initProject("test5", odir = NA, tdir = NA, writedir = NA)
list.files("test5", all.files = TRUE)
unlink(c("test1", "test2", "test3", "test4", "test5"), recursive = TRUE)