DVD Backup Pipeline

DVD Backup Pipeline automates turning personal DVD backups into a Plex-ready streaming library. Rip and encode run as separate workers connected by a disk-backed queue — the drive ejects as soon as raws land, encoding continues in the background, and the human does the only thing machines can’t: load the tray.

Source on GitHub

Highlights

  • Decoupled two-worker design: rip worker (identify → classify movie/TV → Plex metadata → MakeMKV → queue → eject) and encode worker (claim oldest → HandBrake/ffmpeg → atomic move into the movie share or Show/Season NN layout).
  • A queue you can operate by hand: pending/, claimed/, failed/ JSON items — failed raws park for retry by moving a file back to pending; no database, no daemon.
  • Metadata without API keys: movie vs. TV-season classification, per-title episode ripping, and missing years resolved from Wikipedia/Wikidata with caching (year-cache.json), plus optional movie-map.csv / tv-map.csv overrides.
  • One streaming baseline: bin/transcode.py is the single source of truth — MKV, x264, 720p ceiling with no upscaling, 24 fps cap, CRF 20 or two-pass target sizes (1.4 GB movie / 450 MB episode); the PowerShell drive-facing watcher (dvdbot.ps1) mirrors it with HandBrakeCLI, with the divergences documented setting-by-setting.
  • Composed with downscale-videos: directory watching and _downscale.py inherit that tool’s decide-then-act standardization, so the whole library converges on one household prescription.
  • Library tooling included: scrub-titles (year backfill with dry-run), combine, conform (transcode anything off-baseline), sync.

Related posts