phono-junk

Audio CD rip identification, cataloging, verification, and export. Designed around identifying uncommon and foreign-language discs (Thai, Korean, Chinese, Japanese, etc.) where the user may not even know how to type the album title.

Sibling of retro-junk; shares disc-parsing infrastructure via junk-libs.

What it does

• Identify CD rips (CUE/BIN, CHD) by computing canonical disc IDs from the TOC and querying multiple databases in parallel — MusicBrainz, Discogs, and additional providers as pluggable trait impls.
• Verify rip quality against AccurateRip (and CUETools DB, eventually). Per-track confidence scores expose whether the rip is bit-identical to other submitters'.
• Catalog everything in SQLite with user-editable YAML overrides. Disagreements between providers are recorded, not silently resolved.
• Export selected discs as per-track FLAC with embedded Vorbis tags and cover art, organized into a standard music-library tree (<AlbumArtist>/<Album> (<Year>)/NN - Title.flac).
• Scrape album art from Cover Art Archive, Discogs, iTunes Search API, and Amazon (ASIN-direct or PA-API).

CLI and GUI stay in feature sync. GUI ships with a pan-script font bundle (NotoSans + NotoSansCJK + NotoSansThai + NotoSansArabic + NotoSansDevanagari) loaded unconditionally — foreign scripts are the whole point.

Status

Early development. Workspace skeleton compiles; algorithms and providers are scaffolded with traits in place and most implementations pending.

• CLAUDE.md — architecture, dependency graph, conventions
• TODO.md — ordered work queue for the MVP, deferred items, and open questions
• .claude/skills/phono-archive/ — disc-identification and verification algorithm references with upstream citations

Build

cargo build
cargo test
cargo run -p phono-junk-cli -- --help

On first build, Cargo fetches junk-libs over git. For faster iteration when you're also developing junk-libs, clone both repos side-by-side and uncomment the [patch] section at the bottom of the root Cargo.toml.

CLI usage

# Scan a directory tree for rips and identify each against every provider.
phono-junk scan ~/rips

# Identify a single disc from its CUE/CHD.
phono-junk identify ~/rips/pinkerton.cue

# Verify a rip against AccurateRip. Accepts --disc-id or a path.
phono-junk verify --disc-id 17

# Export disc(s) as FLAC.
phono-junk export --disc-ids 17,18 --out ~/Music/library

# Filter the catalog.
phono-junk list --artist weezer --year 1990-1999
phono-junk --format json list --country JP

Global flags (valid on every subcommand):

• --db <path> — library database path. Default: $PHONO_JUNK_DB, else XDG data_dir()/phono-junk/library.db.
• --user-agent <string> — HTTP User-Agent for provider calls. Default identifies phono-junk and links to the repo. MusicBrainz requires a descriptive UA with contact info; override it with your own contact if you plan to run scans at volume.
• --format <human|json> — output shape.
• -v / -vv — log verbosity (INFO / DEBUG). Default is WARN.

Architecture

See CLAUDE.md for the full workspace architecture, dependency graph, crate responsibilities, and development conventions.

Short version: 14 crates organized into analysis foundation (core, toc, accuraterip), pluggable providers (identify + musicbrainz, discogs, itunes, amazon), catalog (catalog, db), cross-cutting (extract, lib), and presentation (cli, gui).

Sibling projects

• retro-junk — same architectural patterns, applied to retro game ROMs and disc images.
• junk-libs — shared disc I/O and generic utilities consumed by both phono-junk and retro-junk.

License

MIT — see LICENSE.