#!/usr/bin/env bash
# rmev - clean up videos that mpv integrity-check flagged as corrupt.
#
# The corrupt list comes from integrity_cache.tsv (current status). A file is
# only eligible for deletion when its on-disk mtime/size exactly match what the
# cache recorded. (If the file changed after the scan it is skipped as
# "changed" and a re-scan is suggested.)
#
# Default behaviour:
#   * Show the target list and ask for y/N confirmation before deleting.
#   * Use trash-put (trash) if available, otherwise fall back to rm.
#   * After deletion, remove the entries from integrity_cache.tsv and corrupted.log.

set -euo pipefail

CONFIG_DIR="${MPV_CONFIG_DIR:-$HOME/.config/mpv}"
CACHE="$CONFIG_DIR/integrity_cache.tsv"
LOG="$CONFIG_DIR/corrupted.log"

DRY=0
FORCE=0
MODE=auto # MODE: auto|trash|rm

usage() {
  cat <<'EOF'
Usage: rmev [options]

Delete video files that mpv integrity-check flagged as corrupt.

Options:
  -n, --dry-run   Show the targets without deleting anything
  -f, --force     Delete without the confirmation prompt
  -t, --trash     Move to trash (requires trash-put)
  -r, --rm        Permanently delete (rm) instead of trashing
  -h, --help      Show this help

Default: trash if trash-put exists, otherwise rm. Confirms the list first.
Deleted files are also removed from integrity_cache.tsv and corrupted.log.
EOF
}

while [ $# -gt 0 ]; do
  case "$1" in
  -n | --dry-run) DRY=1 ;;
  -f | --force) FORCE=1 ;;
  -t | --trash) MODE=trash ;;
  -r | --rm) MODE=rm ;;
  -h | --help)
    usage
    exit 0
    ;;
  *)
    printf 'Unknown option: %s\n\n' "$1" >&2
    usage >&2
    exit 2
    ;;
  esac
  shift
done

if [ ! -f "$CACHE" ]; then
  printf 'Cache file not found: %s\n' "$CACHE" >&2
  exit 1
fi

# Decide deletion method
if [ "$MODE" = auto ]; then
  if command -v trash-put >/dev/null 2>&1; then MODE=trash; else MODE=rm; fi
fi
if [ "$MODE" = trash ] && ! command -v trash-put >/dev/null 2>&1; then
  printf 'trash-put not found. Use --rm to delete permanently, or install trash-cli.\n' >&2
  exit 1
fi

# Extract entries whose last cached status is corrupt (mtime, size, path).
# The cache is append-only, so a path may appear on several lines; last wins.
mapfile -t entries < <(
  awk -F'\t' '
        NF>=5 {
            p=$5; for (i=6; i<=NF; i++) p = p "\t" $i
            mt[p]=$1; sz[p]=$2; st[p]=$3
        }
        END { for (p in st) if (st[p]=="corrupt") printf "%s\t%s\t%s\n", mt[p], sz[p], p }
    ' "$CACHE"
)

if [ "${#entries[@]}" -eq 0 ]; then
  printf 'No files flagged as corrupt.\n'
  exit 0
fi

# Select real targets: must exist and match the cached mtime/size.
targets=()
skipped=()
for line in "${entries[@]}"; do
  mt=${line%%$'\t'*}
  rest=${line#*$'\t'}
  sz=${rest%%$'\t'*}
  path=${rest#*$'\t'}

  if [ ! -e "$path" ]; then
    skipped+=("missing  | $path")
    continue
  fi
  cur_mt=$(stat -c %Y -- "$path" 2>/dev/null || echo -1)
  cur_sz=$(stat -c %s -- "$path" 2>/dev/null || echo -1)
  if [ "$cur_mt" != "$mt" ] || [ "$cur_sz" != "$sz" ]; then
    skipped+=("changed  | $path  (re-scan needed)")
    continue
  fi
  targets+=("$path")
done

printf 'Corrupt files (%d):\n' "${#targets[@]}"
for p in "${targets[@]}"; do printf '  %s\n' "$p"; done

if [ "${#skipped[@]}" -gt 0 ]; then
  printf '\nSkipped (%d):\n' "${#skipped[@]}"
  for s in "${skipped[@]}"; do printf '  %s\n' "$s"; done
fi

if [ "${#targets[@]}" -eq 0 ]; then
  exit 0
fi

if [ "$DRY" -eq 1 ]; then
  printf '\n(dry-run) Nothing deleted.\n'
  exit 0
fi

method_label=$([ "$MODE" = trash ] && echo 'move to trash' || echo 'permanently delete')
if [ "$FORCE" -ne 1 ]; then
  printf '\n%s %d file(s)? [y/N] ' "$method_label" "${#targets[@]}"
  read -r ans
  case "$ans" in
  y | Y | yes | YES) ;;
  *)
    printf 'Cancelled.\n'
    exit 0
    ;;
  esac
fi

deleted=()
for p in "${targets[@]}"; do
  if [ "$MODE" = trash ]; then
    if trash-put -- "$p"; then deleted+=("$p"); else printf 'Failed: %s\n' "$p" >&2; fi
  else
    if rm -f -- "$p"; then deleted+=("$p"); else printf 'Failed: %s\n' "$p" >&2; fi
  fi
done

past_label=$([ "$MODE" = trash ] && echo 'moved to trash' || echo 'deleted')
printf '%d file(s) %s.\n' "${#deleted[@]}" "$past_label"

# Remove deleted entries from the cache and log
if [ "${#deleted[@]}" -gt 0 ]; then
  tmpset=$(mktemp)
  printf '%s\n' "${deleted[@]}" >"$tmpset"

  prune() { # $1: file, $2: field number where the path starts
    local f="$1" start="$2" tmp
    [ -f "$f" ] || return 0
    tmp=$(mktemp)
    awk -F'\t' -v start="$start" '
            NR==FNR { del[$0]=1; next }
            {
                p=$start; for (i=start+1; i<=NF; i++) p = p "\t" $i
                if (!(p in del)) print
            }
        ' "$tmpset" "$f" >"$tmp"
    mv "$tmp" "$f"
  }

  prune "$CACHE" 5 # integrity_cache.tsv: mtime size status errors PATH
  prune "$LOG" 2   # corrupted.log: timestamp PATH
  rm -f "$tmpset"
fi