Compare all MKV files in a directory to find the real movie among near-identical Blu-ray rip playlists

compare-mkv.sh

#!/usr/bin/env bash
#
# compare-mkv.sh — Compare all MKV files in a directory to find the real movie
# among near-identical Blu-ray rip playlists.
#
# Usage: ./compare-mkv.sh   (operates on the current directory)
#
# Requires: ffprobe, ffmpeg (from FFmpeg). On macOS: brew install ffmpeg

set -euo pipefail

if ! command -v ffprobe >/dev/null 2>&1; then
  echo "Error: ffprobe not found. Install FFmpeg (brew install ffmpeg)." >&2
  exit 1
fi

shopt -s nullglob nocaseglob
mkvs=(./*.mkv)
shopt -u nullglob nocaseglob

if [ ${#mkvs[@]} -eq 0 ]; then
  echo "No MKV files found in the current directory." >&2
  exit 1
fi

printf '%-30s %12s %12s %9s %s\n' "FILE" "DURATION" "SIZE(GiB)" "CHAPTERS" "VIDEO MD5"
printf '%s\n' "--------------------------------------------------------------------------------------------"

declare -A md5_seen

for f in "${mkvs[@]}"; do
  name=$(basename "$f")

  # Duration in seconds, formatted as H:MM:SS
  dur_raw=$(ffprobe -v error -show_entries format=duration \
            -of default=nw=1:nk=1 "$f" 2>/dev/null || echo 0)
  dur_int=${dur_raw%.*}
  dur_fmt=$(printf '%d:%02d:%02d' $((dur_int/3600)) $(((dur_int%3600)/60)) $((dur_int%60)))

  # Size in GiB
  size_bytes=$(ffprobe -v error -show_entries format=size \
               -of default=nw=1:nk=1 "$f" 2>/dev/null || echo 0)
  size_gib=$(awk "BEGIN{printf \"%.2f\", $size_bytes/1073741824}")

  # Chapter count
  chapters=$(ffprobe -v error -show_chapters "$f" 2>/dev/null | grep -c '\[CHAPTER\]' || true)

  # MD5 of the video stream only (ignores audio/subtitle ordering differences)
  vmd5=$(ffmpeg -v error -i "$f" -map 0:v -f md5 - 2>/dev/null | sed 's/^MD5=//')

  dupe=""
  if [ -n "${md5_seen[$vmd5]:-}" ]; then
    dupe="  <-- DUPLICATE of ${md5_seen[$vmd5]}"
  else
    md5_seen[$vmd5]="$name"
  fi

  printf '%-30s %12s %12s %9s %s%s\n' \
    "$name" "$dur_fmt" "$size_gib" "$chapters" "${vmd5:0:12}..." "$dupe"
done

echo
echo "Tip: the real feature is usually the longest, largest file with the"
echo "expected chapter count. Files sharing a video MD5 are identical features."