backup

#!/usr/bin/env bash
# backup_under_3GB.sh
# Copies files < 3 GiB from SOURCE to DEST/DD-MM-YY-HH-MM, preserving tree structure.

set -Eeuo pipefail

usage() {
  cat <<EOF
Usage: $(basename "$0") [-n] <SOURCE_DIR> <DEST_DIR>

  -n, --dry-run   Show what would be copied without making changes

Creates a timestamped folder (DD-MM-YY-HH-MM) inside DEST_DIR and copies all
regular files smaller than 3 GiB from SOURCE_DIR into it, preserving directories.

Examples:
  $(basename "$0") /data/projects /backups
  $(basename "$0") -n ~/Documents /Volumes/Backups
EOF
}

# --- Parse args ---
DRY_RUN=0
if [[ $# -lt 2 ]]; then
  if [[ "${1:-}" == "-n" || "${1:-}" == "--dry-run" ]]; then
    echo "Error: missing SOURCE_DIR and DEST_DIR."
  fi
  usage; exit 1
fi

if [[ "${1:-}" == "-n" || "${1:-}" == "--dry-run" ]]; then
  DRY_RUN=1
  shift
fi

SRC=${1:?}; DEST=${2:?}

# --- Preconditions ---
if ! command -v rsync >/dev/null 2>&1; then
  echo "Error: 'rsync' is required but not found in PATH." >&2
  exit 2
fi

if [[ ! -d "$SRC" ]]; then
  echo "Error: SOURCE_DIR '$SRC' does not exist or is not a directory." >&2
  exit 3
fi

mkdir -p "$DEST"

# --- Config ---
# 3 GiB in bytes (3 * 1024^3); rsync --max-size accepts raw bytes.
MAX_SIZE_BYTES=3221225472
STAMP=$(date +%d-%m-%y-%H-%M)
TARGET="$DEST/$STAMP"

mkdir -p "$TARGET"

# --- Do the backup ---
# We want: preserve structure, copy only files < 3 GiB, include dirs & symlinks.
# Strategy: build list of eligible files with find (portable byte-size test) and feed to rsync.
# Using relative paths to preserve tree cleanly.
echo "Starting backup:"
echo "  Source:      $SRC"
echo "  Destination: $TARGET"
echo "  Max size:    < 3 GiB"
[[ $DRY_RUN -eq 1 ]] && echo "  Mode:        DRY RUN"

# Find prints NUL-separated relative file paths for files smaller than 3 GiB.
# The 'c' suffix makes the size check in bytes and is portable (BSD/GNU find).
# Rsync copies only the provided files and recreates the directory tree under $TARGET.
(
  cd "$SRC"
  find . -type f -size -"${MAX_SIZE_BYTES}"c -print0
) | rsync -a --from0 --files-from=- "$SRC"/ "$TARGET"/ \
    --max-size="$MAX_SIZE_BYTES" \
    $([[ $DRY_RUN -eq 1 ]] && printf %s "--dry-run --itemize-changes") \
    --human-readable

# Optional: write a simple manifest for traceability
{
  echo "Backup timestamp: $STAMP"
  echo "Source: $SRC"
  echo "Destination: $TARGET"
  echo "Max file size: < 3 GiB"
  echo "Command: $(basename "$0") $([[ $DRY_RUN -eq 1 ]] && echo -n "--dry-run " )\"$SRC\" \"$DEST\""
  echo "File count:"
  find "$TARGET" -type f | wc -l
} > "$TARGET/backup_manifest.txt"

echo "Backup complete: $TARGET"