#!/usr/bin/env bash
# Prune old rows from logs/storage-growth/history.csv to cap file size.
# Keeps header and either last KEEP_DAYS (as proxy: days*200 rows) or last MAX_ROWS.
# Usage: ./scripts/monitoring/prune-storage-history.sh [--days N] [--max-rows N] [--dry-run]

set -euo pipefail

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
PROJECT_ROOT="$(cd "$SCRIPT_DIR/../.." && pwd)"
LOG_DIR="${LOG_DIR:-$PROJECT_ROOT/logs/storage-growth}"
HISTORY_CSV="${HISTORY_CSV:-$LOG_DIR/history.csv}"
KEEP_DAYS="${KEEP_DAYS:-90}"
MAX_ROWS="${MAX_ROWS:-0}"
ROWS_PER_DAY=200
DRY_RUN=0

while [ $# -gt 0 ]; do
  case "$1" in
    --days)
      KEEP_DAYS="${2:-90}"
      shift 2
      ;;
    --max-rows)
      MAX_ROWS="${2:-0}"
      shift 2
      ;;
    --dry-run)
      DRY_RUN=1
      shift
      ;;
    *)
      shift
      ;;
  esac
done

if [ ! -f "$HISTORY_CSV" ]; then
  echo "File $HISTORY_CSV does not exist; nothing to prune."
  exit 0
fi

header=$(head -1 "$HISTORY_CSV")
total_lines=$(wc -l < "$HISTORY_CSV")
data_lines=$((total_lines - 1))

if [ "$data_lines" -le 0 ]; then
  echo "No data rows in $HISTORY_CSV."
  exit 0
fi

if [ "$MAX_ROWS" -le 0 ]; then
  MAX_ROWS=$((KEEP_DAYS * ROWS_PER_DAY))
fi

if [ "$data_lines" -le "$MAX_ROWS" ]; then
  echo "Data rows ($data_lines) <= keep ($MAX_ROWS); no prune needed."
  exit 0
fi

remove=$((data_lines - MAX_ROWS))
if [ "$DRY_RUN" -eq 1 ]; then
  echo "[dry-run] Would trim to last $MAX_ROWS rows (remove $remove rows)."
  exit 0
fi

{ echo "$header"; tail -n "$MAX_ROWS" "$HISTORY_CSV"; } > "${HISTORY_CSV}.$$"
mv "${HISTORY_CSV}.$$" "$HISTORY_CSV"
echo "Trimmed $HISTORY_CSV to last $MAX_ROWS data rows (removed $remove)."