PaperclipAi backup rotation script

prune-backups.sh

#!/usr/bin/env bash
set -euo pipefail

BACKUP_DIR="${BACKUP_DIR:-$HOME/.paperclip/instances/default/data/backups}"

# Retention config (override via env)
RETENTION_HOURLY="${RETENTION_HOURLY:-24}"
RETENTION_DAILY_DAYS="${RETENTION_DAILY_DAYS:-14}"
RETENTION_WEEKLY_DAYS="${RETENTION_WEEKLY_DAYS:-28}"
RETENTION_MONTHLY="${RETENTION_MONTHLY:-1}" # 1=enabled, 0=disabled

DRY_RUN=false

usage() {
  cat <<EOF
Backup rotation script

Usage:
  $(basename "$0") [--dry-run] [--help]

Options:
  --dry-run     Show what would be deleted without deleting
  --help        Show this help

Environment variables:
  BACKUP_DIR             Backup directory
  RETENTION_HOURLY       Number of most recent backups to always keep (default: 24)
  RETENTION_DAILY_DAYS   Keep 1 per day for N days (default: 14)
  RETENTION_WEEKLY_DAYS  Keep 1 per week for N days (default: 28)
  RETENTION_MONTHLY      Keep 1 per month forever (1=on, 0=off)

Examples:
  RETENTION_HOURLY=48 $(basename "$0")
  RETENTION_DAILY_DAYS=7 RETENTION_WEEKLY_DAYS=21 $(basename "$0") --dry-run
EOF
}

# Parse args
for arg in "$@"; do
  case "$arg" in
    --dry-run) DRY_RUN=true ;;
    --help|-h) usage; exit 0 ;;
    *) echo "Unknown argument: $arg" >&2; usage; exit 1 ;;
  esac
done

[[ -d "$BACKUP_DIR" ]] || { echo "Backup directory not found: $BACKUP_DIR" >&2; exit 1; }

get_mtime() {
  if [[ "$(uname)" == "Darwin" ]]; then
    stat -f %m "$1"
  else
    stat -c %Y "$1"
  fi
}

fmt_date() {
  if date -d "@$1" +"$2" >/dev/null 2>&1; then
    date -d "@$1" +"$2"
  else
    date -r "$1" +"$2"
  fi
}

shopt -s nullglob
raw_files=( "$BACKUP_DIR"/*.sql )
shopt -u nullglob

(( ${#raw_files[@]} )) || { echo "No backups found."; exit 0; }

# Sort newest first
mapfile -t files < <(
  for f in "${raw_files[@]}"; do
    printf '%s\t%s\n' "$(get_mtime "$f")" "$f"
  done | sort -rn | cut -f2-
)

now=$(date +%s)

declare -A keep daily weekly monthly

kept=0
deleted=0

# Keep N most recent
for i in "${!files[@]}"; do
  (( i < RETENTION_HOURLY )) && keep["${files[$i]}"]=1
done

for f in "${files[@]}"; do
  mtime=$(get_mtime "$f")
  age_days=$(( (now - mtime) / 86400 ))

  day=$(fmt_date "$mtime" "%Y-%m-%d")
  week=$(fmt_date "$mtime" "%Y-%W")
  month=$(fmt_date "$mtime" "%Y-%m")

  if (( age_days <= RETENTION_DAILY_DAYS )) && [[ -z "${daily[$day]:-}" ]]; then
    daily["$day"]="$f"
  fi

  if (( age_days <= RETENTION_WEEKLY_DAYS )) && [[ -z "${weekly[$week]:-}" ]]; then
    weekly["$week"]="$f"
  fi

  if (( RETENTION_MONTHLY == 1 )) && [[ -z "${monthly[$month]:-}" ]]; then
    monthly["$month"]="$f"
  fi
done

for k in "${!daily[@]}"; do keep["${daily[$k]}"]=1; done
for k in "${!weekly[@]}"; do keep["${weekly[$k]}"]=1; done
for k in "${!monthly[@]}"; do keep["${monthly[$k]}"]=1; done

for f in "${files[@]}"; do
  if [[ -n "${keep[$f]:-}" ]]; then
    ((++kept))
  else
    if $DRY_RUN; then
      echo "[dry-run] would delete: $(basename "$f")"
    else
      rm -- "$f"
      echo "deleted: $(basename "$f")"
    fi
    ((++deleted))
  fi
done

suffix=""
$DRY_RUN && suffix=" (dry run)"

echo "$(date): kept=$kept deleted=$deleted total=${#files[@]}$suffix"