In Bai-Li-NOAA/ghBackup: An R Package that Backs Up your GitHub Repositories

knitr::opts_chunk$set(echo = TRUE)

Installation

This R package helps back up GitHub repositories from an organization using R. To install the package from GitHub:

remotes::install_github("nmfs-ost/ghBackup")

You can read the help file with

?ghBackup::back_up_gh_orgs

The back_up_gh_orgs() function requires a GitHub token, a destination directory, and the organization name of the repositories. It then uses the token to clone repositories from the GitHub organization and save the files in the destination directory. The run will download clones of all public and private repositories the token can access.

Configuration

Create a token

A new personal GitHub token can be created from the GitHub settings -> Developer settings -> Personal access tokens tab. More instructions can be found on the GitHub "creating a personal access token" website.

Once you have a token, you can use it with the gh_token argument in the back_up_gh_orgs() function. There are two options to set up gh_token.

Option 1: simply provide token in an R script and do not share the token with others. For example, gh_token = "abcdefghijklmnopqrstuvwxyz01234567890123"
Option 2 (Recommended): Use R package gitcreds to manage GitHub credentials without printing out the token in an R script. gitcreds::gitcreds_set() adds Git credentials in the credential stores and gh::gh_token() can be used to return the local user's token. Please see gitcreds website and gh website for more details. The token will not be shared when sharing the R script.

Back up repositories

To back up all repositories from, for example, nmfs-fish-tools and nmfs-general-modeling-tools to C:/Users/ghbackup/ using gh::gh_token(), you can use the following R code in an R script:

back_up_gh_orgs(
 gh_token = gh::gh_token(),
 backup_path = "C:/Users/ghbackup/",
 orgs_name = c("nmfs-fish-tools", "nmfs-general-modeling-tools")
)

Clones of all repositories will be saved in the C:/Users/ghbackup/ directory and a zip archives of clones of all repositories will be created under parent directory C:/Users/.

Upload the zip file to a Google Drive folder

Thanks @k-doering-NOAA for providing the solution. This approach uses an auto auth setting so no interaction is needed. See gargle's "Non-interactive auth" vignette for more details:

options(gargle_oauth_email = "*@noaa.gov")

zip_name <- "ghBackup.zip"
zip_path <- file.path(dirname(backup_path), zip_name)
googledrive::drive_put(
  media = zip_path,
  path = googledrive::as_id("https://drive.google.com/drive/folders/1vju2qQ0dPh49Ay7VGaWv6CvGqMZsZKPc"),
  name = zip_name
)

The full R script example can be found here: inst/extdata/examples/example_back_up_gh.R.

Schedule the run

You can schedule the R script/workflow on Windows using the Windows task scheduler. To add a task in an R script, you need to install taskscheduleR package and run the example code below:

install.packages("taskscheduleR")
library(taskscheduleR)

# R script with tasks
task_script <- system.file(
  "extdata", "example_back_up_gh.R",
  package = "ghBackup"
)

# Run every 5 mins, starting from within 60 seconds
taskscheduleR::taskscheduler_create(
  taskname = "ghBackup_5min",
  rscript = task_script,
  schedule = "MINUTE",
  startdate = format(Sys.Date(), "%m/%d/%Y"),
  starttime = format(Sys.time() + 60, "%H:%M"),
  modifier = 5
)