Merge pull request #38 from VariantSync/benjamin/replication

Create a script for archiving source repositories

Author: pmbittner

docs/replication/README.md

@@ -0,0 +1,44 @@
+# Replication
+
+The script `create-forks-on-github.sh` can be used to create forks of all the
+repositories in the `../../../DiffDetectiveMining/` folder. It marks them as a
+fork and replicates the local state in this fork. The indended use case is to
+provide immutable copies of all source data to make all results of this research
+reproducible.
+
+## Usage
+First run:
+```bash
+./create-forks-on-github.sh
+```
+
+It will ask some verification questions and prints all commands that it will
+perform. If the source repositories are not located in
+`../../../DiffDetectiveMining/` (relative to this file) or
+`../DiffDetectiveMining` (relative to the root directory of this repository) the
+comment character of the variable `PATH_TO_REPOSITORIES` in the second paragraph
+can be removed and the actual path can be set.
+
+To execute the printed commands uncomment the last line of the second paragraph
+which disables `DRY_RUN`.
+
+## Login
+You can either login beforehand with `gh auth login` or just run the script
+which will run this command for you. This script will *not* log you out once
+it's finished because you may want to run additional commands (or you where
+previously logged in).
+
+## Additional notes
+This script should be idempotent as `gh` detects an already forked project and
+the performed force push should not change anything (unless the remote
+repository was changed).
+
+## Requirements
+This script requires a standard UNIX environment with at least the following
+additional tools:
+- `bash`, the GNU POSIX shell
+- `gh`, the github command line interface
+- GNU `find`, the POSIX `find` with some additional options (mainly `-print0`,
+  but this was not tested with other `find` implementations)
+
+This script was tested under GNU+Linux.

docs/replication/create-forks-on-github.sh

@@ -0,0 +1,73 @@
+#!/usr/bin/env bash
+
+# Default settings
+PATH_TO_REPOSITORIES="$(dirname "${BASH_SOURCE[0]}")/../../../DiffDetectiveMining"
+DRY_RUN=y
+
+# Override settings
+#PATH_TO_REPOSITORIES="absolute/path/to/directory/containing/the/respositories"
+# uncomment to following line, to actually run th 'gh' commands
+#DRY_RUN=n
+
+continue-with() {
+  echo
+  read -p "Do you want to continue with $1? [y/N] " answer
+  [ "$answer" == "y" ]
+}
+
+run() {
+  echo "\$ $*"
+  if [ "$DRY_RUN" = "n" ]
+  then
+    "$@"
+  fi
+}
+
+repos() {
+  find "$PATH_TO_REPOSITORIES" -mindepth 1 -maxdepth 1 -type d "$@"
+}
+
+PATH_TO_REPOSITORIES="$(realpath "$PATH_TO_REPOSITORIES")"
+cd "$PATH_TO_REPOSITORIES"
+echo "The following repos in '$PATH_TO_REPOSITORIES' will be forked:"
+repos
+continue-with "these $(repos -print0 | tr -d -c '\0' | tr '\0' '\n' | wc -l) repos" || exit 1
+
+if gh auth status |& grep -q 'You are not logged into any GitHub hosts.' &>/dev/null
+then
+  run gh auth login || exit 1
+  was_logged_in=0
+else
+  echo
+  gh auth status
+
+  continue-with "this account" ||
+  {
+    run gh auth logout &&
+    run gh auth login || exit 1
+  }
+  was_logged_in=1
+fi
+
+repos -print0 |
+while IFS= read -d '' -r repository
+do
+  echo
+  run cd "$repository"
+  run gh repo fork --remote
+  run git push -f origin
+done
+
+if [ "$was_logged_in" = "1" ]
+then
+  cat <<EOF
+
+Warning: 'gh' is still logged in, to log out use
+
+    gh auth logout
+
+EOF
+else
+  echo
+  run gh auth logout
+fi