feat(database): shape the library like files, and shrink the catalog

Plans 013 and 014, the album page that prompted them, and the smaller fixes they turned up. Changelog, largest first. ## The local library is shaped like files, not like MusicBrainz `audio_files` carries its own tags and points at `albums` and `artists`; `file_genres` is the one real many-to-many. `recordings`, `release_group_recordings`, `artist_credit`, `artist_credit_artist`, `recording_genres`, `release_groups` and `release_to_rg` are gone from the local side, and with them a six-way join in every read, a `MIN(release_group_id)` subquery in eleven queries and a first-credited-artist subquery in nine. Measured on a real 25,966-file library, every many-to-many that model expressed was 1:1 in the data. - Ownership is a file. `GetFilePathsByRecordingMBIDs`, `LibraryMBIDIndex.CheckMBIDs`, `collectLibraryEntities` and `pruneStaleLocalCrossReferences` all join `audio_files`, so the 812 orphaned recordings, 216 release groups and 260 artists that library carried are now structurally impossible. - One projection: every track query selects from the `track_metadata` view, one row type, one mapper. Nine hand-rolled copies had drifted far enough to report different years on different screens. - `library_id = 0` means every library, so each list query exists once instead of scoped and unscoped with a branch at every call site. - No migration chain. `sql/schemas/` is the one description of the shape; `sql/migrations/`, `applyMigrations` and `schema_migrations` are squashed away, along with the drift between them that had sqlc generating against a stale schema. - `database.InsertTestTrack` is the one test seeder; twenty test files had been assembling the old FK chain each in its own order. ## The catalog stores its ids as bytes `explore_index`'s three 36-char MBID columns and its entity-type text are 16 raw bytes and a small integer. The table and its six indexes go 780 MB to 405 MB on a real 2,052,200-row catalog, which is why a fresh install is ~0.6 GB rather than ~1.0 GB. - `backend/explore/mbid.go` is the only place the encoding is known; everything above it speaks dashed strings. - `CHECK(length(mbid) = 16)` makes a stringly write fail at the insert rather than silently returning no rows, since SQLite does not coerce between TEXT and BLOB. - The importer asks the artifact what encoding it carries and converts on the way in, so the artifact already published keeps working and no format bump is needed. - `indexRowColumns`/`scanIndexRow` replace four copies of a 22-column list, and `TestStoredEncodingRoundTrips` sweeps every read path. ## An album page that says how much of the album is yours - One question, asked once: is there a file. `filePaths` is filled by a single batched lookup when the tracklist settles, and the badge, the Play count, the dimmed rows and every menu item read it — replacing four claims of decreasing confidence that could show a green tick on an album whose every action did nothing. - Play, Play 7 of 12, or no play button at all. - `total_tracks` on `explore_index` (~2 bytes over 400,677 release groups) and on `audio_files` from tags that have always carried it: a complete MBID-matched album now makes no catalog call at all, where it used to spend the most expensive request the app makes. - A merged cluster shows the running order the most releases agree on, and the version list marks the release you own rather than standing a synthetic entry in for it. - `AlbumReleasesFailed`: a slow fetch is no longer reported as a failed one by a 12-second timer. - Rows not in the library are dimmed in place (with `aria-disabled`) instead of the owned ones wearing a green tick and a legend. ## Caches and cover art get ceilings - Only the three tiers of a cover are stored; the full-resolution copy nothing rendered was 1,134 MB of a 1.4 GB covers directory. - One artist portrait is downloaded and the rest are remembered as URLs — 4.1 GB of a 5.3 GB cache was candidates no code path reads. - `browsedArtBudget` and `httpCacheBudget` bound what an age cannot: the same install held art for 5,770 artists in a 1,301-artist library. - `OrphanedArtistImagesJob` joined a bare MBID onto a sharded directory, so it deleted the rows that were the only record of the files it left behind. `explore.ArtistImageDir` is that layout's one definition now. ## The autotag queue asks whether there is work `tagging_items` was a row per album folder, not a queue, and no query read the `tag_status` column that held the answer. The four queue queries ask the files, which matters most where it is least visible: `startPrefetch` was scoring every album in a tagged library against MusicBrainz. ## Phantom playlist tracks resolve in place An M3U8 imported before its files leaves phantom rows; they now match by path and fall back to position, keep their place in the playlist when resolved, and pair best-first so two phantoms cannot claim the same file. ## Playing a track plays the list it is in Double-click, and Play on a single row's menu, queue the list as displayed with `startIndex` on that row — the album page and the track list used to queue one track and discard the album around it. A multi-row selection still plays exactly itself. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AfVYUVExXsx1nSWrXN8mAh
2026-08-16 13:58:15 -04:00
parent 1128881e8d
commit e7748f1fd5
208 changed files with 10944 additions and 12104 deletions
@@ -63,5 +63,24 @@ Never hand-write one. Seeding points `YJ_CORE_INDEX_URL` at a dead
 address on purpose, so no seed depends on what the explore artifact
 server happened to be serving.
-Rebuild a seed after any schema change, or the restored database is
+Rebuild a seed after any schema change. Nothing migrates a restored
-migrated on open in a way the seed's author never saw.
+database: `applySchema` is `CREATE TABLE IF NOT EXISTS`, so an old seed
 keeps its old columns, the app starts, and the first query dies on
 `no such column`.
 **Restoring the seed does not disable the artifact fetch — only
 *building* it does.** `dev-headless` leaves `YJ_CORE_INDEX_URL` alone,
 so on a developer machine the restored app immediately downloads and
 imports the real ~1.1M-row catalog, through the one writer connection,
 while whatever you started it for is running. A full `make e2e` against
 that reported **14 failures** that were all contention; the same suite
 against the same seed with
 ```bash
 YJ_CORE_INDEX_URL='http://127.0.0.1:1/none.tar.zst' make dev-headless SEED=default
 ```
 is the configuration CI runs (`ci.yml` sets exactly that address) and is
 what to use before believing a failure. The tell is in `.dev/app.log` —
 an import logging progress — and in how the failures look: timeouts
 spread across unrelated specs rather than one surface being wrong.
@@ -1,66 +1,76 @@
 # Changing the database schema
-The reasoning — why there are two files, what the old 48-step migration
+The reasoning — why the local library is shaped like files rather than
-chain got wrong, and when squashing is legitimate — is in `CLAUDE.md`
+like MusicBrainz, and what the metadata tables cost before they went —
-under *Backend packages → database*. Read it once. This is the
+is in `CLAUDE.md` under *Backend packages → database*. Read it once.
-checklist.
+This is the checklist.
-**A brand-new table needs one file, not two.** The rule below is about
+**There is one description of the schema and no migration chain.**
-a *column added to a table that already exists*. `applySchema` runs
+`sql/schemas/*.sql` declares the current shape; `applySchema` runs every
-every file in `sql/schemas/` on every open, so a
+file on every open, and `CREATE ... IF NOT EXISTS` makes that idempotent.
-`CREATE TABLE IF NOT EXISTS` reaches an existing install verbatim and a
+`sql/migrations/`, `applyMigrations` and `schema_migrations` were
-migration for it would be a second description of the same table — the
+squashed away with plan 013. So:
-thing the third rule forbids. Its indexes go in the schema file too,
+
-because the column and the index arrive together.
+**Adding a table or a column is one edit to one file.**
 ```bash
 make generate            # sqlc + templ
 go test ./backend/database/ ./backend/datamap/
 make test
 ```
 A new table has a second gate: **`backend/datamap`**. Add an entry
 stating its Kind and Lifetime, or `TestCatalogCoversSchema` fails — and
 if it is `Authored` and cascades, `TestAuthoredCascadesAreDeliberate`
-wants an explicit exemption with a note, because authored data is what
+wants an explicit exemption with a note, because authored data is what a
-a user cannot get back.
+user cannot get back. If a *column* holds a different Kind from its
 table (an authored flag on an owned projection, a fetched value beside a
 tag-derived one), say so in the entry's note; `audio_files` and `lyrics`
 are the worked examples.
-Adding a **column** to an existing table needs **two** files, not one:
+**Existing databases are not migrated.** Nothing upgrades a database
 from an older shape — delete your dev `YJ_HOME` and rescan, and rebuild
 any seed you rely on (`make sandbox-seed NAME=default`). Revisit this
 once real user databases exist in the wild.
-1. **`backend/database/sql/schemas/*.sql`** — `CREATE TABLE ... IF NOT
+**A stale one fails at the first query, not at open**, which is worth
-   EXISTS`, the literal target shape, what sqlc reads and what a fresh
+knowing before you read the error. `applySchema` is
-   install gets verbatim. Add the new column **last** in the
+`CREATE TABLE IF NOT EXISTS`, so an old database keeps its old columns
-   `CREATE TABLE`.
+and gains nothing; the app then starts fine and dies on
-2. **`backend/database/sql/migrations/NNNN_description.sql`** — the
+`no such column: title`. Every tier that does not *run the app* — unit
-   `ALTER TABLE ... ADD COLUMN` (and any index on it) that gets an
+tests, `make ui-test`, `tsc` — is green while this is true, because
-   existing database to the same shape. Schema files are a no-op against
+they build their database from the current schema. `make e2e` and
-   a table that already exists, so without this an upgrade never gets
+`make dev` are the two that will tell you, and only after the seed has
-   the column.
+been rebuilt.
-Then:
+## The four ways this goes wrong
-```bash
+- **A query file must be ASCII.** sqlc's parameter rewriter works on
-make generate                                        # sqlc + templ
+  byte offsets, so a single non-ASCII character in a *query* comment
-go test ./backend/database/         # migration + column-order tests
+  (an em dash, a curly quote) shifts every placeholder and generates
-make test
+  garbage like `SELECid` — a parse error a long way from its cause.
-```
+  Schema files are not rewritten and may contain anything.
 - **A slice and a named parameter do not compose.** `sqlc.slice`
  expands to N placeholders, but `sqlc.arg` is numbered independently,
  so the two in one query bind the wrong values —
  `GetFilePathsByAlbums([1,2], 0)` read album id 2 as the library id.
  Where a query needs both, return the column and filter in Go.
 - **A write wearing a query's shape still needs the writer.**
  `QueryContext`/`QueryRow` route to the query-only read pool, so an
  `INSERT ... RETURNING` through one fails at runtime with "attempt to
  write a readonly database (8)". Use `ExecContext`, or
  `QueryRowWriter`. `TestNoWritesOnTheReadPool` walks the tree for it.
 - **A view is dropped and recreated.** `CREATE VIEW IF NOT EXISTS`
  no-ops against a database holding the old definition, so
  `track_metadata.sql` opens with `DROP VIEW IF EXISTS`.
-Rebuild any seed you rely on (`make sandbox-seed NAME=default`) and
+## Where things go
 delete your own dev `YJ_HOME` if you want to see the fresh-install path
 rather than the migrated one.
 ## The three ways this goes wrong
 - **Column order must match between the two paths.** `ADD COLUMN`
  always appends, so a migrated column declared anywhere but last in
  `CREATE TABLE` leaves fresh and upgraded installs disagreeing on
  order — and sqlc binds `SELECT *` positionally, so one of them
  silently reads the wrong field.
  `TestMigrations_ColumnOrderMatchesFreshInstall` is the regression test.
 - **Do not put an index on a migrated column in `sql/schemas/`.**
  Schema files run *before* migrations, against a database that may not
  have the column yet, and the predicate fails. Declare the index in the
  migration, after the `ALTER TABLE`.
 - **Do not add a third description of the schema anywhere.** A
  migration's `ADD COLUMN` failing with "duplicate column name" against
  an already-current database is expected and tolerated, not an error to
  route around.
 New queries go in `backend/database/sql/queries/`; generated Go lands in
-`backend/database/sql/sqlcgen/`, which is never edited by hand. Tests
+`backend/database/sql/sqlcgen/`, which is never edited by hand. Anything
-use `database.NewTestDB(t)`, built by the same `applySchema` production
+returning a track selects from the `track_metadata` view rather than
-uses, so the two cannot diverge.
+re-joining — that is why there is one row type and one mapper.
 Tests use `database.NewTestDB(t)`, built by the same `applySchema`
 production uses, and seed rows with `database.InsertTestTrack(t, db,
 database.TestTrack{...})` rather than assembling inserts by hand.
@@ -2354,3 +2354,43 @@ Five more things worth keeping:
  child does not prove the icon inside it is `pointer-events: none`.
  Both were checked with a real mouse (`mousemove`/`mousedown`/
  `mouseup`) and a real Tab/Enter before either was believed.
 ## A queue of work has to ask whether there is work
 Reported as "the autotag page has every album in it, even the MB-tagged
 ones". Both halves of that were true and they were two different bugs,
 which is why the first answer found (the pending list) accounted for
 nine rows out of 2172.
 **Every scanned folder gets a `tagging_items` row.**
 `UpsertTaggingItemOnTrackAdd` runs per track with `status = 'pending'`
 and no condition, so the table is a row per album folder, not a queue.
 `saveAudioFile` *does* record the answer — `audio_files.tag_status` is
 `user_confirmed` on import for any file carrying a recording MBID, and
 `idx_audio_files_tag_status_untagged` was declared for the filter — but
 no query in the app read the column. Pending therefore meant "has a
 row". The four queue queries ask the files now
 (`EXISTS … tag_status = 'untagged'`), which matters most where it is
 least visible: `startPrefetch` was scoring every album in a tagged
 library against MusicBrainz.
 **The 2094 were the Completed section, and nobody completed them.** A
 backfill stamped `status = 'confirmed'` on every folder whose files
 already had MBIDs, and the sidebar keeps confirmed rows deliberately —
 so the review page's history became a list of the library. They are
 distinguishable without new state: every status flip the app performs
 goes through `SetTaggingItemStatus` or `SetTaggingItemBestMatch` and
 both stamp `last_checked_at`, so *confirmed with no `last_checked_at`,
 no score and no best match* is the backfill's row and not the user's.
 All 2094 were that shape; the two the app had actually applied were
 not. Migration 0007 sets `cleared_at` on them — the column exists for
 exactly this, and it keeps the row so a rescan cannot reset review
 state.
 Two smaller things fell out. The reviewed states are **exempt** from
 the untagged-files predicate, or an applied folder would vanish from
 Completed the instant it succeeded — the section is history, not work.
 And `tag_status` was only ever written by the *insert* path, so a file
 another tagger stamped after import kept `untagged` for ever and its
 folder kept asking; `updateAudioFile` promotes it now, guarded on
 `untagged` so a deliberate `user_skipped_permanent` survives a rescan.
@@ -0,0 +1,639 @@
 # 013 — The database audit
 **Status:** **complete** (2026-08-16). R1–R10 landed, the album page
 that prompted the audit with them, and the one part of R5 that ships
 *in the artifact* — a per-release-group track denominator — landed as
 plan 014.
 The audit below is unchanged from when it was written — the measurements
 describe the *old* shape and are the reason for the new one.
 **Branch:** none
 **Created:** 2026-08-15
 **Supersedes:** the four-part album-page fix sketched in conversation
 (it survives, reduced, as R1 and R3 below)
 **Related:** 010 (owned albums offline), 011 (owned artists'
 discography), 012 (API call audit), 002 (data lifecycle)
 ---
 ## Method
 Every number here is measured against the **real 25,966-track library**
 at `~/.local/share/yellowjacket/yj.db` (copied read-only), not against
 a fixture and not inferred from the code. Where a claim rests on a
 capability rather than a count — "sqlc can do X" — it was executed, not
 assumed.
 The brief: *efficiency and simplicity — the minimum required to achieve
 our featureset*, with fewer lines and a smaller database as evidence
 rather than as the goal. Two named sources of confusion to resolve:
 **local versus remote** versions of a thing, **files versus tracks**,
 and **indexed versus live** lookups. One added constraint: **avoid
 hitting APIs by storing intelligently, without a ridiculous base
 install.**
 ---
 ## The measurements
 ### The database is 1.00 GB, and 78% of it is one table
 | object | size | rows |
 |---|---|---|
 | `explore_index` | 383 MB | 2,052,200 |
 | its five indexes + `UNIQUE(mbid)` | 395 MB | — |
 | its two FTS tables | 85 MB | 2,052,200 + 96,451 |
 | `recordings` | 38 MB (27 MB of it lyrics) | 26,778 |
 | `lyrics_index` | 18 MB | 24,294 |
 | `artist_metadata` | 12 MB | 7,673 |
 | `http_cache` | 9 MB | 2,930 |
 | `audio_files` | 5 MB | 25,966 |
 | everything else | < 10 MB | — |
 The local library — the part that is *the user's* — is about 50 MB.
 The catalog and its indexes are 780 MB.
 ### Inside `explore_index`, half the bytes are three text columns
 | column | bytes | note |
 |---|---|---|
 | `mbid` | 70 MB | 36-char text; 16 bytes as a blob |
 | `artist_mbid` | 70 MB | same, and it is a foreign key in disguise |
 | `caa_release_mbid` | 62 MB | same |
 | `entity_type` | 18 MB | three distinct values, stored as words |
 | `title` / `artist_name` / `release_name` | 74 MB | real data |
 Five columns are declared, shipped in the artifact, selected in every
 query, and **empty**: `aliases` (0 rows), `sort_name` (0),
 `disambiguation` (0), `country` (69 rows of 2.05 M), `artist_type`
 (72). `aliases` is additionally a column in *both* FTS tables, so the
 tokenizer indexes nothing, twice.
 ### Two 50 MB indexes have a `WHERE` clause that excludes 0.3% of rows
 `idx_explore_title_lower` (53 MB) and `idx_explore_artist_lower`
 (48 MB) are `WHERE popularity > 0`. 2,046,645 of 2,052,200 rows satisfy
 that. They are full indexes wearing a partial index's clothes, and they
 exist to serve one exact-match tier (`ExactMatches`,
 `searchindex.go:1298`) that the champion FTS — 96,451 rows, 2 MB —
 already covers the popular half of.
 ### The local library models many-to-many relationships that are all 1:1
 | claim | measured |
 |---|---|
 | recordings with more than one file | **0** |
 | recordings in more than one release group | **0** |
 | artist credits with more than one artist | **3** of 2,823 |
 | files sharing a recording | **0** |
 `recordings` (26,778) is one row per file. `release_group_recordings`
 (26,778) is one row per file. `artist_credit` (2,823) and
 `artist_credit_artist` (2,826) differ by three.
 ### …and it leaks rows that outlive the files
 | orphan | count |
 |---|---|
 | `recordings` with no `audio_files` row | **812** (218 carry MBIDs) |
 | `release_groups` with no file underneath | **216** |
 | `artists` credited on no file | **260** |
 | `explore_index` rows flagged **`in_library` with no file behind them** | **129** recordings, 2 release groups, 1 artist |
 That last row is the bug reported today, in the user's own data.
 ### The query surface
 | surface | count |
 |---|---|
 | sqlc queries | 235 (7,850 generated Go lines) |
 | raw SQL call sites outside sqlc | 188 |
 | bound IPC methods | 272 |
 | `X` / `XByLibrary` query twins | 14 (8 of them exposed as separate bindings) |
 | copies of the "one row per file with its metadata" projection | **9**, plus the view that already defines it |
 `mapTrackRow` takes **22 positional arguments** and is called from 9
 places, because each duplicated query generates its own row struct.
 ### The data directory is 8.5 GB — the database is the small part
 | path | size | of which |
 |---|---|---|
 | `artist-images/` | 5.4 GB | **4,125 MB is candidate images no code path reads**; 1,222 MB is primaries + tiers for **5,770 artists** in a library with **1,301** |
 | `covers/` | 1.4 GB | **1,134 MB is originals**; all three rendered tiers together are 110 MB |
 | `ffmpeg/` | 283 MB | bundled binary |
 | `yj.db` | 1.0 GB | above |
 | `yj.db.bak` + `.bak.20260309` | 452 MB | nothing deletes these |
 | art caches (`cover-art-cache`, `artist-image-cache`) | 81 MB | catalog art, fine |
 The 4.1 GB of unreachable artist candidates is the bug `CLAUDE.md`
 records as fixed; this install still carries it, so **the janitor jobs
 have never run here**. Worth confirming they run at all before
 declaring that one closed.
 ---
 ## The diagnosis
 Everything below is downstream of one thing.
 **There are three different notions of "a track" in this app, and the
 code keeps asking the wrong one.**
 1. **A file** — a row in `audio_files`. The only thing that is
   unambiguously *yours*: it has a path, it plays.
 2. **A local entity** — a row in `recordings` / `release_groups` /
   `artists`. Created by a scan *from* a file, but with an independent
   lifetime: nothing deletes it when the file goes, and retagging a
   file **creates a new one and abandons the old**
   (`library.go:1722` repoints `audio_files.recording_id` at a fresh
   recording; `pruneOrphanedMetadata` only runs on the scan's
   *deleted-file* branch, `library.go:982`). This is where the 812
   orphans come from — and autotagging is the machine that makes them.
 3. **A catalog entity** — a row in `explore_index`, downloaded, global,
   identical for every user.
 "Is this mine" is asked of **(2)** almost everywhere, and answered by
 **(1)** whenever the user actually does something:
 - `LibraryMBIDIndex.CheckMBIDs` (`librarymbid.go:64`) is literally
  `SELECT mbid FROM recordings WHERE mbid IN (…)`. It sets `inLibrary`
  on every catalog tracklist.
 - `pruneStaleLocalCrossReferences` (`searchindex.go:2480`) clears
  `explore_index.in_library` when the **`recordings` row** disappears —
  not when the file does. Hence 129 phantom "you own this" rows.
 - `albumLibraryStatus()` in `explore-album-details.ts` ORs four claims
  of decreasing confidence, none of which is "a file exists".
 - But `GetFilePathsByRecordingMBIDs`, which every *action* goes
  through, joins `audio_files`. It is the only one that tells the
  truth.
 So a retagged file leaves behind a recording carrying the **old** MBID;
 the catalog matches that MBID; the row renders owned, undimmed, with a
 Play button; and every action on it fails with "could not be found in
 your library" — on a fully-tagged library. The user's instinct that the
 check is fragile is correct, and the fragility is not the live lookup.
 **The live lookup is the only part that is right.**
 The same confusion explains "files vs tracks" and "local vs remote":
 tables (2) exist to be a local mirror of the catalog's shape, so a
 "track" is sometimes a file, sometimes a mirror row, sometimes a
 catalog row, and the three are joined by MBID — a key that **two of the
 three can lack or lie about**.
 ---
 ## Findings and recommendations
 ### R1 — Ownership is "a file exists". Say it once, in SQL.
 *Cheap, immediate, and it fixes the reported bug.*
 - `CheckMBIDs`' `recordings` and `release_groups` branches gain a join
  to `audio_files`. (`artists` too, via credit.)
 - `pruneStaleLocalCrossReferences` tests for a file, not for a local
  row.
 - `pruneOrphanedMetadata` runs after the retag path as well as the
  delete path — or, better, is deleted along with the tables that need
  it (R2).
 - One-shot cleanup of the 812/216/260 existing orphans at open.
 **Effect:** 129 lying rows in this library become honest; the class
 cannot recur while (2) exists.
 ### R2 — Collapse the MusicBrainz-shaped local schema into a file-shaped one
 *The big one. It is what makes R1 structural rather than a patch.*
 The local model imitates MusicBrainz's normalization — `artist_credit`
 is an MB concept — for a dataset in which **every relationship it
 models is 1:1** (measured above). The cost of that imitation:
 - 5 tables (`recordings`, `release_group_recordings`, `artist_credit`,
  `artist_credit_artist`, `release_to_rg` — the last has **0 rows** and
  no schema-file writer) and ~12 indexes.
 - A 6-way join in every read, including a `MIN(release_group_id)`
  subquery repeated in **11 places** to undo a many-to-many that never
  happens, and a "first credited artist" subquery in **9** to undo
  another (the row-multiplication bug class documented at length in
  `CLAUDE.md`, which serves 3 rows).
 - An orphan-cleanup subsystem (`GetOrphaned*IDs` ×3, `Count*References`
  ×2, `pruneOrphanedMetadata`) that exists only because these rows can
  outlive their file — and which does not actually work (812 orphans).
 - The entire phantom-ownership class above.
 Proposed shape:
 ```
 audio_files   id, path, library_id, …, title, track_no, disc_no, year,
              composer, comment, artist_credit TEXT, artist_id→artists,
              album_id→albums, recording_mbid, modified_at, …
 albums        id, name, artist_id, mbid, year, original_year,
              cover_art_id, total_tracks…      (genuinely many files→1)
 artists       id, name, mbid                    (genuinely many→1)
 genres + file_genres                            (genuinely many↔many:
                                                 107k rows / 26k files)
 ```
 `artist_credit` survives as **text on the file** (display: "A feat.
 B") plus `artist_id` (the primary artist, for grouping) — which is
 everything the UI does with it today, minus the join that multiplies
 rows.
 **Effect:** a row exists iff a file exists, so R1 becomes a foreign key
 rather than a rule anyone can forget. Removes 5 tables, ~12 indexes,
 ~30 sqlc queries, the orphan subsystem, both repeated subqueries, and
 the `AUTOMATIC COVERING INDEX` SQLite builds on every library load.
 Estimated −1,500 to −2,500 lines across `backend/library`,
 `backend/database/sql/*` and `sqlcgen`.
 **Cost:** one real migration of user data (not an `ADD COLUMN`), and it
 touches autotag, tagwriter, playlist matching and the explore xref.
 This is the item to sequence carefully; everything else is independent
 of it.
 ### R3 — One projection, one row type, one mapper
 `track_metadata` (the view) already *is* the canonical "one row per
 file" definition, and **only the raw-SQL search paths use it**
 (`search.go`, `lyrics_search.go`). Every sqlc query re-implements it —
 9 copies, which have already drifted: the view prefers
 `rg.original_year` for `year`, `GetAllTracksWithFullMetadata` uses
 `r.year`. The same library shows a different year depending on which
 screen you are on.
 **Verified, not assumed:** sqlc generates cleanly against the view —
 `SELECT * FROM track_metadata WHERE …` yields one `TrackMetadatum`
 struct with correct types (run during this audit).
 And the 14 `X`/`XByLibrary` twins collapse into one query each:
 ```sql
 WHERE (CAST(sqlc.arg(library_id) AS INTEGER) = 0
       OR library_id = CAST(sqlc.arg(library_id) AS INTEGER))
 ```
 **Measured cost of the collapse: none.** Scoped-with-OR 23 ms, scoped
 direct 21 ms, unscoped 145 ms over the full 26k rows.
 **Effect:** −14 queries, −8 bindings, −8 frontend branches, 9 row
 structs → 1, 9 call sites of a 22-argument mapper → 1. Roughly −2,000
 generated lines and −300 hand-written ones, and the year inconsistency
 cannot exist.
 ### R4 — Put `explore_index` on a diet (~200 MB, no feature loss)
 | change | saved |
 |---|---|
 | `mbid`, `artist_mbid`, `caa_release_mbid` as 16-byte blobs | ~110 MB in the table |
 | …and the same keys in `UNIQUE(mbid)` (99 MB) and `idx_explore_index_artist_mbid` (131 MB) | ~70–100 MB |
 | `entity_type` → INTEGER | 18 MB + index |
 | drop `aliases`, `sort_name`, `disambiguation` (0 rows); reconsider `country`/`artist_type` (69/72 rows) | small bytes, real clarity — and one fewer empty FTS column |
 | make the two `LOWER()` indexes' partial predicate *mean* something (`popularity >= championPopThreshold OR in_library`), or retire the tier onto the champion FTS | up to 101 MB |
 Better still for `artist_mbid`: it is a foreign key spelled as text.
 An integer reference to the artist row is 8 bytes instead of 36 and
 makes the 131 MB index a fraction of its size.
 **Also worth separating:** `in_library`, `local_*_id`, `is_similar` and
 `discog_fetched` are *personalization* stored inside the *shipped
 catalog* table, which is why the artifact import has to merge by
 explicit column list and why `artist_enrichment` had to become its own
 table for exactly this reason. Measured: `in_library` and
 `local_*_id IS NOT NULL` agree on **every one of 2,052,200 rows** —
 they are the same fact stored twice. A `library_xref(mbid, kind,
 local_id)` side table would make the catalog table purely the artifact
 and delete the merge-by-column-list rule.
 ### R5 — Ask the network less, without a bigger install
 Present state (from `musicbrainz.go:17-27`): search 24 h, **entity 7
 days**, releases 90 days. MusicBrainz entity data changes on the order
 of *never* for the fields we read, and 251 of 2,930 cache rows are
 already expired on this install — so a fully-populated artist page
 re-fetches itself weekly, forever.
 - **Raise `cacheTTLEntity` to a year** (or drop expiry and revalidate
  in the background). Cost: bytes already stored. Benefit: the
  steady-state network cost of browsing your own library goes to
  roughly zero.
 - **Ship a per-release-group `total_tracks` in the artifact.** 010
  correctly rejects shipping *tracklists* (the per-artist track budget
  would truncate them, and "Play 7 of 9" for a twelve-track album is a
  confident lie). But the **denominator** is one small integer per
  release group — 400,677 rows, ~2 bytes — and it is exactly what
  `albumLibraryStatus`/`ownership()` needs to say complete /
  incomplete / unknown for a catalog album with no local tags. Tiny,
  honest, and it does not depend on coverage.
 - **Keep 010's per-user backfill** for the tracklists themselves; this
  does not replace it, it shrinks what it has to cover.
 - `http_cache` has no size bound and no vacuum beyond expiry. Give it a
  ceiling.
 ### R6 — The 5.3 GB on disk that no feature needs
 - **4,125 MB of artist candidate images** that nothing reads (the
  documented bug — but the janitors have not run on this install;
  verify they run at all).
 - Artist images exist for **5,770 artists** in a **1,301-artist**
  library. Fetching art for artists you do not own is the same
  "prefetch everything" instinct as the discography backfill 011
  corrected.
 - **1,134 MB of cover originals** versus 110 MB for all three rendered
  tiers. Nothing renders the original; and it is re-derivable from the
  audio file itself, which is on disk by definition. Keep `_lg` as the
  largest and drop originals — that is 1.1 GB with no visible change.
 - `yj.db.bak` (394 MB) and `yj.db.bak.20260309` (58 MB) accumulate with
  nothing to clean them.
 This is the largest single win available and it does not touch the
 schema.
 ### R7 — Redundant indexes and dead columns
 Five indexes are prefixes of an existing UNIQUE/PK and can be dropped
 outright (they cost write time on every insert):
 `idx_recording_genres_recording_id` ⊂ `UNIQUE(recording_id, genre_id)` ·
 `idx_similar_artist_map_source` ⊂ `PK(source, similar)` ·
 `idx_artist_credit_artist_artist_id` ⊂ `UNIQUE(artist_id, credit_id)` ·
 `idx_artist_metadata_mbid` ⊂ `PK(mbid, source)` ·
 `idx_artist_images_mbid` ⊂ `UNIQUE(artist_mbid, source, source_url)`.
 Dead data:
 - **`recordings.genre`** — populated on 25,619 rows at every scan and
  **read by nothing**. Every genre read goes through
  `recording_genres` + `genres`. Write-only column.
 - **`release_groups.total_tracks` / `total_discs`** — 0 rows populated;
  the feature that needed them put the number on
  `release_group_recordings` instead.
 - **`release_to_rg`** — 0 rows, no writer in any schema file.
 - `libraries.sql` carries a doc comment about `download_requests`,
  pasted from another file. Small, but it is the kind of drift the
  two-file schema rule exists to catch.
 ### R8 — One genuine N+1
 `mixCandidates` (`explore/mix.go:181`) issues
 `GetGenreNamesByFilePath` **per candidate path**, inside a loop over
 similar artists, inside a loop over seed artists. Twenty seeds × twenty
 similar × thirty paths is 12,000 single-row queries for one mix. It is
 one query with an `IN` clause, or one query for the whole weighted set.
 (`mixSeedProfile` above it is the same shape, bounded by seed size.)
 Nothing else in the tree matches this pattern — a scan of every query
 issued inside a loop turned up 72 candidates and this is the only real
 one.
 ### R9 — The IPC surface has internals in it
 Bound and reachable from the frontend today: `AcquirePipelineLock`,
 `ReleasePipelineLock`, `SetJobRegistry`, `SetScanHooks`,
 `SetRescanHooks`, `SetRemovalHooks`, `MusicBrainz`, `CAALimiter`,
 `PopulateLocalCrossReferences`. v3's generator binds every exported
 method; these want to be unexported or moved off the service type.
 Free lines, and one less way to wedge the app from a console.
 ### R10 — The test DB is not the shape production runs
 `NewTestDB` shares one in-memory connection and leaves `readDB` nil, so
 `reader()` returns the writer. That is why the read-pool write bug
 (documented in `CLAUDE.md`) reached a user, and why
 `TestNoWritesOnTheReadPool` had to be a tree-walk instead of a test.
 Giving the test DB two handles over one shared in-memory file would let
 that be an ordinary test.
 ---
 ## What I recommend leaving alone
 - **The download subsystem** (requests / downloads / items). Three
  tables, clean lifetimes, well argued in the schema comments. The
  `download_wants` table in this install is the pre-rename name; the
  rename migration will clear it on next launch.
 - **The champion FTS.** 96k rows, 2 MB, a real latency tier.
 - **The dual write/read handle**, WAL, and the persist-writer queues.
  These are recent, measured, and correct.
 - **File paths as the frontend's identity for a track.** Integer ids
  would be cheaper over IPC, but `CLAUDE.md`'s argument (an index goes
  stale on re-sort/refilter, a path does not) is right, and the cost is
  bounded.
 - **Storing lyrics locally** (27 MB + 18 MB index for 24k tracks). That
  is the API-avoidance trade working exactly as intended.
 ---
 ## What landed (2026-08-15 / 16)
 ### The third pass: the album page, which is where the report came from
 The audit started from a user report — a fully-tagged library saying
 "not in your library", on hover rather than on click — and R1 fixed the
 half of that which lives in SQL. The other half was the page: ownership
 was four claims OR'd into a tick, and the context menu asked the backend
 per row, as the menu opened.
 `explore-album-details` now resolves the displayed tracklist's file
 paths **once**, from `updated()`, into one `filePaths` map that the
 badge, the Play count, the dimmed rows and every menu item read. The
 synthesised local tracks carry their own `FilePath`, so a library album
 costs no lookup at all; a catalog tracklist costs one batched
 `GetFilePathsByRecordingMBIDs`. `catalogScope()` no longer returns
 `'library'` here — that was the second complaint in the same report, and
 the artist page keeps it because a library-only *artist* really is
 missing sections.
 Two bugs fell out of doing it this way, and neither is the one that was
 reported:
 - The render loop. Guarding the lookup on `filePaths` (answered) rather
  than on `askedFor` (asked) re-requests every *unowned* MBID forever,
  because an unowned MBID never lands in the map.
 - "No release data available" over a tracklist held in memory.
  `loadLocalTracks` rebuilt the version list only when catalog releases
  existed, but the "Your Library" entry is synthesised *from* the local
  tracks — so the no-releases case was the one case it skipped. Nothing
  caught it because the old ownership check answered from the local
  album id and never needed the tracklist to exist.
 ### The second pass: R5–R10
 | | before | after |
 |---|---|---|
 | the two exact-match indexes | 101 MB | **3 MB** (predicate narrowed to the champion set; plan unchanged, measured) |
 | cover art on disk | original + 3 tiers | **3 tiers** — 1,134 MB of a 1.4 GB directory was the original, and nothing rendered it |
 | browsed artist art | 90-day expiry, no ceiling | expiry **plus a 256 MB budget**, oldest evicted first; owned artists never in it |
 | MusicBrainz entity TTL | 7 days | **1 year**, with a 128 MB ceiling on the response cache |
 | redundant indexes | 5 | **0** (3 dropped here, 2 went with their tables) |
 | internal methods on the IPC surface | 24 | **0** (`//wails:ignore`; 272 → 248 bound methods) |
 | test DB | one handle, `readDB` nil | **two handles**, the shape production runs |
 The catalog line is R4, finished the day after: MBIDs stored as 16 raw
 bytes and entity types as codes, measured by converting the real
 2,052,200-row catalog through the shipped schema. It needed no artifact
 rebuild — the importer asks the artifact which encoding it carries and
 converts the older text form on the way in. Plan 014 has the detail.
 Two of those repaid immediately. Giving the test database its own
 read pool **caught three tests writing through it** on the first run —
 the exact bug class that reached a user as "attempt to write a readonly
 database" and that `TestNoWritesOnTheReadPool` had to walk the source
 tree to find. And the artist-image sweep's own test turned out to seed
 an `artists` row with no file and call it owned: the phantom this whole
 audit is about, sitting in the fixture of the test that guards it.
 **One finding in this audit was wrong.** `aliases`, `sort_name`,
 `disambiguation`, `country` and `artist_type` are not dead columns. They
 are empty on that install because the artist-enrichment pass had barely
 run (which is finding 011's subject), but `indexOneArtist` writes all
 five, and `aliases` is an FTS column that makes an artist findable by
 alias. They stay.
 ### The first pass: R2, carrying R1 and R3
 R2 shipped with R1 and R3 inside it, because the collapse made them
 free rather than separate work. No migration: fresh installs only, by
 the user's decision, so `sql/migrations/` went with it.
 | | before | after |
 |---|---|---|
 | local tables | 9 | 5 (`audio_files`, `albums`, `artists`, `genres`, `file_genres`) |
 | sqlc queries | 235 | 185 |
 | generated Go | 7,850 | 6,023 |
 | bound IPC methods | 272 | 264 |
 | copies of the track projection | 9 + the view | the view |
 | `X`/`XByLibrary` query twins | 14 | 0 |
 | migration files + runner | 7 + ~120 lines | 0 |
 | **net** | | **−5,070 lines** across 122 files |
 Gone: `recordings`, `release_group_recordings`, `artist_credit`,
 `artist_credit_artist`, `pruneOrphanedMetadata`'s four sweeps,
 `RemoveLibrary`'s eight, `mapTrackRow`'s 22 positional arguments, and
 340 lines of `tagwriter/dbsync.go` that existed to relink and then
 un-orphan those tables.
 Ownership is now a file in every one of the places that used to ask a
 metadata table: `CheckMBIDs`, `collectLibraryEntities`,
 `pruneStaleLocalCrossReferences` and `GetFilePathsByRecordingMBIDs`.
 Three things found on the way, each written down where it can be hit
 again (`CLAUDE.md`, `references/schema-change.md`):
 - **sqlc's parameter rewriter is byte-offset based**, so one em dash in
  a *query* comment corrupts generation into `SELECid`.
 - **`sqlc.slice` and `sqlc.arg` do not compose** — slice expansion
  renumbers, so `GetFilePathsByAlbums([1,2], 0)` read album id 2 as the
  library id. Caught by a test, not by a type.
 - **`release_to_rg` looked dead and was not**: 0 rows on any ordinary
  install, because only a local `indexbuild` fills it, and the daily
  incremental refresh reads it. Restored.
 Verified: `make lint` (3 configurations), `go test ./...` plus the
 `indexbuild` and `dev` tag passes, `tsc --noEmit`, `make ui-test`
 (768), and a new end-to-end test that scans the real fixture library
 and asserts no row outlives its file
 (`TestScan_FixtureLibraryLeavesNothingBehind`).
 ---
 ## Sequence
 **Revised 2026-08-15, after the compatibility constraint was lifted:**
 breaking changes are acceptable and the schema may be squashed. That
 inverts the order — R2 was last only because of the migration, and it
 *subsumes* R1 (ownership becomes a foreign key) and reshapes R3 (the
 projection is defined over the new tables). Doing R1 and R3 against the
 old shape first would be work thrown away.
 1. **R2** — the schema collapse, with the rebuild below. It carries R1
   and R3 with it.
 2. **R6** — reclaim the 5.3 GB on disk; confirm the janitors run.
 3. **R7 / R9 / R8 / R10** — the small correctness and hygiene items.
 4. **R4** — the `explore_index` diet. Artifact rebuild + format bump.
 5. **R5** — cache TTLs (trivial) and the shipped denominator (rides
   along with R4's artifact change).
 ### "Break everything" has a floor, and it is not the schema
 Reshaping tables freely is fine. **Dropping the database is not**, and
 the numbers say so — a wipe-and-rescan would destroy:
 | | count | why a rescan does not restore it |
 |---|---|---|
 | files marked `user_confirmed` | **25,014** | the user's autotag review decisions |
 | reviewed tagging folders (`confirmed`/`skipped`) | **2,109** | ditto, plus every `skipped` becomes pending again |
 | rows in `recordings.lyrics` | **24,294** | an unknown share came from **LRCLIB**, not from tags — re-fetching them is precisely the API traffic we are trying to avoid |
 | playlists / playlist tracks | 22 / 1,917 | `Authored`; nothing else has them |
 So the change ships as a **one-shot in-place rebuild**: create the new
 tables, `INSERT … SELECT` across, drop the old ones, in a single
 transaction at open. Seconds on 26k rows, ~40 lines of SQL, no
 migration *chain* and no rollback path — which is the freedom that was
 actually being asked for. `sql/migrations/` gets squashed into
 `sql/schemas/` at the same time (`NOTES.md` already blesses this
 pre-1.0).
 ### Two tables are classified as one Kind and hold another
 `backend/datamap` already encodes what is safe to lose (`Owned` and
 `Derived` rebuild from the files; `Cache` is expensive; `Authored` is
 irreplaceable). The audit found two places where the *column* disagrees
 with the *table's* entry, which is exactly why a wipe looked cheaper
 than it is:
 - **`audio_files.tag_status`** — the table is `Owned` (a projection of
  the files), but `user_confirmed` / `user_skipped_permanent` are
  **`Authored`**: a decision the user made that exists nowhere else.
 - **`recordings.lyrics`** — the table is `Owned`, but lyrics fetched by
  the LRCLIB backfill are **`Cache`**, and nothing records which of the
  24,294 rows came from a tag and which from the network.
 The new schema fixes both by construction: lyrics move to their own
 MBID-keyed table with a `source` column (so they survive any rebuild of
 the owned tables, and the provenance question becomes answerable), and
 `tag_status`' authored values are carried across explicitly rather than
 recomputed.
 **Expected outcome if all of it lands:** database ~1.0 GB → ~0.75 GB,
 data directory 8.5 GB → ~2.5 GB, sqlc queries 235 → ~180, generated Go
 7,850 → ~5,000, bound methods 272 → ~255, and — the part that matters —
 one definition of "this is mine" that a file either satisfies or does
 not.
 ## The open questions, answered
 1. **R2's migration** — the user's call, and it was "just assume this
   new version will only be installed by a new user". So there is no
   in-place rebuild and no chain: `sql/schemas/` is the whole
   description. An existing `YJ_HOME` does not open (its `audio_files`
   has `recording_id` and none of the tag columns, and
   `CREATE TABLE IF NOT EXISTS` cannot add them) — delete and rescan,
   and rebuild any seed with `make sandbox-seed`.
 2. **R4's artifact format** — no break was needed. The importer asks
   the artifact what it carries rather than trusting a version, so the
   published text-form artifact still imports. Plan 014 has it.
 3. **Yes, the janitors run.** `Runner.Start` calls `RunDue` immediately
   and `lastRun` is in-memory, so every launch runs everything due.
   The 4.1 GB survived because `OrphanedArtistImagesJob` joined a bare
   MBID onto a *sharded* directory — deleting the rows and leaving the
   files, which is worse than not running — and because
   `StrayArtistImageFilesJob` did not exist. Both are fixed; it was a
   bug report, not a cleanup.
 ## Measured on the finished refactor
 | | expected | actual |
 |---|---|---|
 | sqlc queries | ~180 | **185** |
 | generated Go | ~5,000 | **6,024** |
 | bound methods | ~255 | **248** |
 | `explore_index` + indexes | — | **780 MB → 405 MB** |
 ## The one recommendation not taken
 R4's "better still" for `artist_mbid`: an integer reference to the
 artist row (8 bytes) rather than the 16 raw bytes it now stores. It is
 a further ~30 MB on `idx_explore_index_artist_mbid`, and the reason to
 stop short is that the *artifact* carries MBIDs and not local ids, so
 the import would have to resolve every row against a table it is in the
 middle of filling. Worth its own argument, not a footnote to this one.
@@ -0,0 +1,99 @@
 # 014 — The catalog's compact encoding, and the denominator it owed
 **Status:** **complete** (2026-08-16). The encoding landed first; the
 per-release-group `total_tracks` denominator landed with the album page
 that spends it.
 **Branch:** none
 **Created:** 2026-08-16
 **Depends on:** nothing
 **Related:** 013 (the database audit, which measured all of this), 010
 (owned albums offline), 001 (ship core index)
 ---
 ## The encoding
 Measured on the real 2,052,200-row catalog, converting it through the
 shipped schema (not a projection):
 | object | before | after |
 |---|---|---|
 | `explore_index` | 383 MB | **242 MB** |
 | `idx_explore_index_artist_mbid` | 131 MB | **65 MB** |
 | `UNIQUE(mbid)` | 99 MB | **54 MB** |
 | `idx_explore_index_entity_pop` | 47 MB | **28 MB** |
 | `idx_explore_caa_release` | 17 MB | **11 MB** |
 | the two `LOWER()` indexes | 101 MB | **3 MB** (013) |
 | **total** | **780 MB** | **405 MB** |
 Every row converted with the `CHECK` constraints live, which is also a
 result: no MBID in a real 2 M-row catalog is malformed.
 **No format bump, and no rebuilt artifact needed.** The importer asks
 the artifact what encoding it carries (`typeof(mbid)`) and converts on
 the way in if it is the old text form, so the artifact already
 published keeps working and the exporter switches whenever CI next
 runs. That is strictly better than the version negotiation this plan
 originally proposed.
 The silent-failure risk the plan was written around was handled by
 making the failure loud instead of by avoiding the change: a `CHECK` on
 the column turns a stringly write into an error at the insert, the
 22-column projection became one constant and one scanner instead of
 four copies, and `TestStoredEncodingRoundTrips` sweeps every read path
 in the package. It found one real bug on its first run — the artifact
 probe was asking the read pool, where the attached artifact does not
 exist.
 ## The denominator
 `total_tracks` on `explore_index`, ~2 bytes across 400,677 release
 groups. It makes "do I have all of this" answerable offline for an
 album whose **files declared no total**, which is a great deal of any
 untagged library and the one thing `GetAlbumCompleteness` cannot answer
 from tags. 010 rightly rejected shipping whole tracklists — the
 per-artist track budget truncates them, and a truncated tracklist is a
 confident lie about which tracks exist. A denominator has no such
 problem, and the album page spends it as one: the numerator stays
 local (distinct track numbers on disk), only the denominator is
 borrowed, and only where the tags have none.
 Four things about it are load-bearing.
 **It is counted before the popularity filter.** `cmd/indexbuild` counts
 the canonical dump's rows per kept release, which is that release's
 track count because the dump carries one row per recording per
 canonical release. Counting the *kept* recordings instead would say
 "9" about a twelve-track album whose other three nobody has played —
 worse than saying nothing, and the same class of lie as the truncated
 tracklist. `TestDumpImportEndToEnd` has an unplayed track on a fixture
 album for exactly this: three tracks in the total, two indexed as
 recordings.
 **Zero means "the catalog does not say"**, which is the same third
 state the local answer already has. An album neither side can total
 wears no ring rather than a wrong one.
 **Adding a column to the importer's SELECT is how you break every
 artifact already published.** `artifactHasTotals()` asks the attached
 artifact whether the column exists, the same way and on the same handle
 as `artifactStoresText()`, and selects a literal `0` when it does not.
 Verified by forcing the probe true: the older shape then fails with
 `no such column: total_tracks`, which is what a shipped build would
 have done to a file nobody can re-cut retroactively.
 **A test seeder that binds the upsert's parameters by hand is not
 "breaking where the app breaks".** Three of them did, on the argument
 that a schema change should fail the tests in the same place — and what
 it actually produced was `missing argument with index 25`, three files
 at a time, for a column none of them cares about. They go through
 `upsertBatch` now, which is the one writer, and keep the property they
 wanted: a field written to the wrong column still fails there.
 ## Done when
 - [x] `GetAlbumCompleteness`'s gap is answerable for a catalog album the
      library has no tags for, with no network call.
 - [x] The artifact grows by less than a megabyte (~800 kB at 400,677
      release groups).
 - [x] An artifact published before the column still imports.
@@ -223,80 +223,101 @@ rather than renaming them.
 - `library` — Concurrent library scanning, metadata extraction, cover art deduplication, incremental rescan. Also **removal**, below.
 - `database` — SQLite via pure-Go driver. Schema in `database/sql/schemas/`, queries in `database/sql/queries/`. **sqlc** generates Go code into `database/sql/sqlcgen/` — never edit that directory by hand.
-  **Schema changes need two things, not one.** `sql/schemas/*.sql` is
+  **The schema is one description, and there is no migration chain.**
-  `CREATE ... IF NOT EXISTS` and is what sqlc reads — it's the single
+  `sql/schemas/*.sql` is `CREATE ... IF NOT EXISTS`, declares the
-  source of truth for "what the schema looks like right now", and it's
+  current shape of every table, and is what sqlc reads and what an
-  what a fresh install gets verbatim. But it's a no-op against a
+  install gets verbatim. That is the whole mechanism: `sql/migrations/`,
-  database that already has the table, so an existing install needs a
+  `applyMigrations` and `schema_migrations` were squashed away with the
-  matching file in `sql/schemas/../migrations/` (e.g.
+  file-shaped rewrite (plan 013). A schema change is one edit to one
-  `NNNN_description.sql`, `ALTER TABLE ... ADD COLUMN ...` /
+  file plus `make generate`. Reintroducing a chain means reintroducing
-  `CREATE INDEX ...`) to actually reach that shape. Both run on every
+  the drift it caused before — `sql/schemas/` and the migrations
-  open, migrations after schema files, tracked in `schema_migrations`
+  disagreed, and sqlc generated against the stale one.
  so each applies once; a migration's `ALTER TABLE ADD COLUMN` failing
  with "duplicate column name" on an already-current database is
  expected and tolerated, not an error.
-  A few things that bite if forgotten:
+  **The local library is shaped like files, not like MusicBrainz.**
-  - **Column order must match between the two paths.** `ALTER TABLE
+  `audio_files` carries its own tags — title, artist credit, track and
-    ADD COLUMN` always appends at the end, so a migrated column must
+  disc numbers, year, composer, the recording MBID — and points at two
-    also be declared *last* in the `CREATE TABLE` in `sql/schemas/`
+  shared rows: `albums` (many files to one) and `artists` (many albums
-    — otherwise a fresh install and an upgraded install disagree on
+  to one). `file_genres` is the one genuine many-to-many. That is the
-    column order, and a `SELECT *` query (sqlc binds those
+  entire local model.
-    positionally) silently reads the wrong field on one of them. See
+
-    `backend/database/migrations_test.go`'s
+  It used to be MusicBrainz's: `recordings`, `release_group_recordings`,
-    `TestMigrations_ColumnOrderMatchesFreshInstall`, which is the
+  `artist_credit` and `artist_credit_artist` sat between a file and its
-    regression test for exactly this.
+  own tags. Measured on a real 25,966-file library, **every**
-  - **Don't put an index on a migrated column in `sql/schemas/`.**
+  many-to-many that model expressed was 1:1 in the data — no recording
-    Schema files run before migrations, against a database that may
+  had two files, none belonged to two release groups, and 3 credits of
-    not have that column yet — the index's predicate would fail
+  2,823 listed more than one artist. What it cost was a six-way join in
-    (this is precisely the bug an earlier session shipped and a user
+  every read, a `MIN(release_group_id)` subquery in eleven queries and a
-    hit at `make sandbox`). Declare it in the migration file instead,
+  first-credited-artist subquery in nine to undo fan-outs that never
-    after the `ALTER TABLE` that adds the column.
+  happened, and a class of bugs where a metadata row **outlived the file
-  - This project **had** a 48-step migration chain before and tore it
+  that created it**: retagging a file created a new recording and
-    out (see `.planning/NOTES.md`, "No migration chain") because
+  abandoned the old one, so that library carried 812 orphaned
-    `sql/schemas/` had drifted from what the migrations actually
+  recordings, 216 release groups and 260 artists — and 129 catalog rows
-    produced and sqlc silently generated against the stale version.
+  that confidently claimed to be owned by files that no longer existed.
-    The design here avoids that by keeping `sql/schemas/` as the
+
-    literal target shape (not a hand-maintained description of it)
+  A few things that follow, and bite if forgotten:
-    and letting migrations replay tolerantly against it — but the
+  - **Ownership is a file.** "Do I own this" is asked of `audio_files`
-    same drift is possible again if a schema change ships without
+    and nothing else — `GetFilePathsByRecordingMBIDs`,
-    updating both files. Don't reintroduce a *second* description of
+    `LibraryMBIDIndex.CheckMBIDs`, `collectLibraryEntities` and
-    the schema anywhere else.
+    `pruneStaleLocalCrossReferences` all join it. A metadata row is not
    ownership; that was the bug, and it is now structurally impossible
    for a row to exist without its file.
  - **The projection is defined once, in the `track_metadata` view.**
    Every query that returns a track selects from it, which is why
    there is one row type (`sqlcgen.TrackMetadatum`) and one mapper
    (`trackFromRow`). It existed before and only the raw-SQL search
    paths used it, so nine hand-rolled copies had already drifted: the
    view preferred the album's original year and one copy used the
    track's, and the same library reported different years on different
    screens. The FTS searches cannot be sqlc queries (MATCH is not in
    its grammar) and spell the column list out in `search.go` — that is
    the one exception and it is one constant.
  - **`library_id = 0` means every library.** Each list query used to
    exist twice, scoped and unscoped, with a branch at every call site
    and a separate binding for each. One query answers both, and the
    scoped form costs nothing measurable (23 ms against 21 ms over 26k
    rows).
  - **A slice and a named parameter do not compose in sqlc.**
    `sqlc.slice` expands to N placeholders but a named argument is
    numbered independently, so `GetFilePathsByAlbums([1,2], 0)` read
    album id 2 as the library id. Where a query needs both, return
    `library_id` and filter in Go (`inLibrary`).
  - **A cache without a ceiling is a leak with a schedule.** Every
    store that grows with use declares a budget beside its retention
    (`browsedArtBudget`, `httpCacheBudget`), because an age bound does
    not bound anything a user can outrun in an afternoon.
  - **A query file must be ASCII.** sqlc's parameter rewriter works on
    byte offsets, so one non-ASCII character in a *query* comment
    corrupts the generated Go into garbage like `SELECid`. Schema files
    are not rewritten and may contain anything.
  - **A view is dropped and recreated, not migrated.**
    `CREATE VIEW IF NOT EXISTS` no-ops against a database holding the
    old definition, so `track_metadata.sql` opens with
    `DROP VIEW IF EXISTS`; a view holds no data, so rebuilding it on
    every open costs nothing.
  - **A write wearing a query's shape still needs the writer.**
    `DB.QueryContext`/`QueryContextWith`/`QueryRow` route to a
    *query-only* read pool (a second `sql.DB` over the same file), so
    an `INSERT ... RETURNING` issued through one fails at runtime with
-    "attempt to write a readonly database (8)" — which is exactly what
+    "attempt to write a readonly database (8)". Use `ExecContext`, or
-    `CreateSmartPlaylist` did, meaning no smart playlist could be
+    `QueryRowWriter` when the statement really does return a row.
-    created at all. Use `ExecContext`, or `QueryRowWriter` when the
+    Nothing caught this because `NewTestDB` shares one in-memory
-    statement really does return a row. Nothing caught this because
+    connection and leaves `readDB` nil, so `reader()` returns the
-    `NewTestDB` shares one in-memory connection and leaves `readDB`
+    *writer* under test. `TestNoWritesOnTheReadPool` walks the tree for
-    nil, so `reader()` returns the *writer* under test and the unit
+    it, in the same spirit as `TestNoDirectRuntimeEmits`.
    tests exercised a handle the app does not have.
    `TestNoWritesOnTheReadPool` walks the tree for it, in the same
    spirit as `TestNoDirectRuntimeEmits` and for the same reason — a
    lint pass only sees one build configuration.
  - **A new table needs one file, not two.** The two-file rule is
    about a column added to a table that already exists.
    `applySchema` runs every file in `sql/schemas/` on *every* open,
    so a `CREATE TABLE IF NOT EXISTS` reaches an existing install
    verbatim and a migration for it would be a second description of
    the same table — which is exactly what the third rule forbids.
    `excluded_paths` is the worked example, index included, since the
    column and its index arrive together.
  - **A new table has to say what kind of data it holds.**
    `backend/datamap` is a catalogue of every table's Kind and
    Lifetime, and `TestCatalogCoversSchema` fails on a table missing
    from it. `TestAuthoredCascadesAreDeliberate` then makes an
    *authored* table that cascades an explicit, argued exemption —
-    authored data is what a user cannot get back.
+    authored data is what a user cannot get back. Two entries say
-  - **Squashing is fine pre-1.0.** While this hasn't shipped to real
+    **MIXED KIND** and mean it: `audio_files` is an owned projection
-    users, periodically folding `sql/migrations/` into `sql/schemas/`
+    except for `play_count`, `last_played` and `tag_status`, which are
-    and deleting the migration files (then wiping your own dev/sandbox
+    authored; `lyrics` carries a `source` column because a lyric read
-    DB) is a legitimate way to keep the migrations directory from
+    from a tag is free to rebuild and one fetched from LRCLIB is not.
-    accumulating dev-only churn — same effect as the old "just nuke
+  - **Test data has one seeder.** `database.InsertTestTrack` inserts a
-    it" workflow, opt-in instead of mandatory. Stop doing that once
+    file with its artist, album and genres. Twenty test files used to
-    real user databases exist in the wild.
+    carry their own, each assembling the old FK chain in a slightly
    different order.
 - `metadata` — Tag extraction (ID3v2, Vorbis Comments, FLAC).
 - `jobs` — The registry every long-running operation reports through:
  progress, pause/cancel, a global indicator and (for scans) a pause
@@ -391,6 +412,67 @@ work happens **once, centrally**, and users download the result:
  (`dumpincremental.go`), and resolves artists outside the artifact's
  coverage lazily on first view.
 **The catalog stores ids as bytes, and that is a size decision.**
 `explore_index` is 2,052,200 rows, and its MBIDs and entity types were
 half of it: three 36-character text columns and one storing the words
 "artist", "release_group", "recording" two million times. They are 16
 raw bytes and a small integer now. Measured on a real catalog, the
 table and its six indexes went **780 MB to 405 MB** — the largest
 single saving available in this app, and the reason a fresh install is
 ~0.6 GB rather than ~1.0 GB.
 `backend/explore/mbid.go` is the only place that encoding is known.
 Everything above it speaks dashed strings and entity names —
 `SearchIndexResult`, the bindings, the frontend — and `dbMBID` /
 `dbEntityType` convert at the SQL boundary. That confinement is the
 point: the alternative is blobs reaching code that has no use for them.
 Four things about it are load-bearing, and they exist because of *how*
 this fails when it fails: **SQLite does not coerce between TEXT and
 BLOB**, so a query comparing the column against a 36-character string
 returns no rows rather than an error, and a scan into a plain string
 yields sixteen bytes of mojibake. Neither is visible except as a result
 that is quietly empty.
 - **The column checks itself.** `CHECK(length(mbid) = 16)` means a
  stringly *write* fails at the insert that made it. It also caught
  every fixture that had been using `"rh"` as an MBID; `testMBID()`
  hashes a label into a real one so they stay readable.
 - **The projection is one constant and one scanner.**
  `indexRowColumns` / `scanIndexRow` replaced four copies of a 22-column
  list and four matching `Scan` calls — four chances to decode wrongly.
  `indexRowColumnsFor("i")` is the same list qualified, for the FTS join
  where both sides have a `title`.
 - **A query that names an entity type inline writes the code with the
  name beside it** — `entity_type = 1 /* artist */`. Splicing a Go
  constant in would keep them in step automatically but makes every such
  query a concatenation; `TestEntityCodesAreStable` pins the mapping
  instead, because it is a storage format and changing one is not a
  refactor.
 - **`TestStoredEncodingRoundTrips` sweeps every read path** — lookup,
  top-N, exact match, FTS search, popularity batch, the CAA map — and
  asserts each returns something with a dashed id. A missed conversion
  site shows up there and essentially nowhere else.
 **The artifact is read in either encoding.** A published artifact
 carries whichever form the exporter that built it used, and there is one
 already out there in the old text form. `artifactStoresText` asks the
 artifact (`typeof(mbid)`) rather than trusting a version number, and
 `artifactSelectColumns` converts on the way in — one `unhex` per row on
 a once-a-month import, against requiring a rebuilt artifact before a new
 build can read anything. That probe **must** run on the writer:
 `core` is attached to that one connection, so `QueryContext` asks a
 pool where the artifact does not exist, and the error would silently
 select the conversion path for an artifact that needs none.
 **Its shape is the pattern for every column added after the fact.**
 `artifactHasTotals` is the same question about `total_tracks`, on the
 same handle: an artifact built before a column existed is still a
 perfectly good catalog, so it is *asked* and the missing column is
 selected as a literal `0`. Adding the column to the importer's SELECT
 list without that is how a published artifact — which nobody can re-cut
 retroactively — starts failing with `no such column`.
 **Background work yields, and says so in the context.** The post-scan
 backfills share MusicBrainz's rate limiters with every page the user
 can open, and both were FIFO — so a thousand-artist enrichment put an
@@ -453,7 +535,20 @@ is what every other failure here already does. SQLite's writer pool is
 `MaxOpenConns(1)`, so the workers queue at the Go level rather than
 racing for the file.
-Four things about the marks are load-bearing. The marks are **a table, not
+Five things about the marks are load-bearing. **A mark records that the
 upstream was asked, not that it answered with something.** ListenBrainz
 returns 200 and `[]` for an artist it has no popularity data for — which
 is most of a long-tail library, and the same is true of an artist whose
 every row falls under `indexMinPopularity` — and keying
 `discog_fetched` on "did rows come back" made those artists permanent
 candidates: "Filling in artist details" re-ran for up to
 `discogBackfillMaxPerRun` of them on **every launch**, forever, doing
 the same two fetches to the same empty answer. So `indexOneArtist` sets
 the mark when both fetches *succeeded* (`fetchTopReleaseGroups` and
 `fetchTopRecordings` return an error for that reason), and only a real
 failure — transport, non-2xx, unreadable body — leaves the artist for
 the next run. `browseFullDiscography` already had this right: an artist
 with genuinely no release groups is still marked browsed. The marks are **a table, not
 more `explore_index` columns**, because `artifactimport.go` merges the
 downloaded catalog by column list — a flag added there is a second
 place to remember, and forgetting it silently wipes every mark on the
@@ -505,6 +600,33 @@ that invents its own flat layout agrees with the bug.
 keeping only `explore.ArtistImageKeepNames()` and refusing an empty
 keep set for the reason the covers sweep refuses an empty live set.
 **An age is not a ceiling, and a cache needs one.** Art for an artist
 the user owns is kept indefinitely; everything else aged out after 90
 days and nothing counted it, so the same install held portraits for
 **5,770 artists in a 1,301-artist library** — every artist page opened
 in Explore fetches one, and a browsing afternoon is entirely inside the
 retention window. `browsedArtBudget` (256 MB) is the second pass:
 oldest browsed artist first, until what is left fits, with owned
 artists outside the budget entirely. `httpCacheBudget` is the same rule
 one cache over, and it is what makes the year-long entity TTL below
 safe — once answers stop expiring, expiry stops being a bound.
 **"Owned" is a file here too.** The sweep's live set used to be "there
 is an `artists` row", which the file-shaped schema made meaningless;
 it joins `audio_files` now, like every other ownership question. Its
 test had seeded an artists row with no file and called it owned — the
 exact phantom, in the fixture of the test that guards it.
 **Only the tiers of a cover are stored.** `saveCoverArt` writes
 `_sm`/`_md`/`_lg` and records the largest as `cover_art.file_path`;
 `coverart.ResolveURLs` reports that one as `Original` too, because it
 is the largest kept. The full-resolution image used to be written
 beside them and was **1,134 MB of a 1.4 GB covers directory** against
 110 MB for all three tiers — with nothing rendering it, since the grid
 caps at 350 px and the largest tier is 400. The bytes are still in the
 audio file, which is where they came from, so the repair pass that
 regenerated tiers *from the stored original* went with it.
 **Frontend** (`frontend/`): Lit 3.2 web components + Web Awesome UI library + HTMX. State management via singleton reactive stores in `src/store/`. Wails bindings auto-generated as TypeScript in `frontend/bindings/`, nested by Go import path — don't edit by hand. The `@go` alias absorbs the constant prefix, so a call site imports `@go/library/library.js`.
 **One seam states what the generated types get wrong, rather than 78 patches.** v3's generator is honest where v2's lied: a Go `nil` slice marshals to JSON `null` and always has (v2 typed it `T[]`), and a Go named string type is a closed set (v2 typed it `string`). There is no flag to turn either off, correctly. So `utils/binding.ts` states the app's actual contract at the only place it is true — `list` yields `[]` for a nil slice, `dict`/`dictByName` yield `{}` for a nil map and drop null-valued keys (which loses nothing: `noUncheckedIndexedAccess` already makes every read `V | undefined`), and `compact` is the same for a map arriving as a *field*. Where a nullable slice is a model field there is no boundary to put it at, and those are `?? []` at the point of use.
@@ -958,31 +1080,47 @@ when the whole release is owned, **Play 7 of 12** when some of it is,
 and **no play button at all** when none is, because a Play button that
 plays nothing (or seven tracks of forty) is worse than none.
-`albumLibraryStatus()` is deliberately *not* what decides that. It is
+**The page asks one question, once, and it is "is there a file".**
-four claims of decreasing confidence OR'd into one tick — a local album
+Ownership used to be several claims of decreasing confidence OR'd
-id, the backend's cross-reference, a cached MBID match, and finally
+together — a local album id, the backend's cross-reference, a cached
-*any single track* marked `inLibrary` — which is a fine answer to "is
+MBID match, and finally *any single track* flagged `inLibrary` — none
-any of this mine" and a useless basis for a button. `ownership()`
+of which is "there is a file to play", which is why the tick could be
-counts the displayed tracklist instead, whose `inLibrary` flags the
+green on an album whose every action did nothing, and why the
-backend sets per recording MBID.
+tracklist's context menu asked the backend on **hover** whether the row
 it was drawing was owned. `filePaths` is the one answer: a map from a
 displayed track to its path, filled once from `updated()` by a single
 batched `GetFilePathsByRecordingMBIDs`, and read by the badge, the Play
 button's count, the dimmed rows and every menu item. `askedFor` is a
 separate set from `filePaths` because the guard has to be *asked*, not
 *answered* — an unowned MBID never lands in the map, so guarding on the
 map re-requests it on every render, forever.
 The other half is that the *displayed* tracklist is also one thing.
 `buildVersionEntries` synthesises the "Your Library" entry from
 `localTracks`, so an album the catalog cannot answer for is exactly the
 case that needs the version list rebuilt — `loadLocalTracks` guarded
 that rebuild on `releases.length > 0` and so rendered "No release data
 available" over a tracklist it was holding in memory.
 **And the key it plays by is not the key it looks owned by.** The local
 album id is used wherever there is one, because a library-only album
 has *no* recording MBIDs — its tracks are synthesised from
 `GetAlbumTracks` with `mbid: RecordingMBID || ''` — so an MBID-keyed
 lookup on an untagged library resolves to nothing and Play queues
-nothing while looking entirely correct.
+nothing while looking entirely correct. Those synthesised tracks carry
-`GetFilePathsByRecordingMBIDs` is the catalog-only fallback and the
+their `FilePath` from the same rows, so they populate `filePaths`
-third member of the `GetFilePathsBy…` family: one query, paths only,
+directly and cost no lookup at all; the batched
-grouped so the caller keeps the tracklist's order. It exists rather
+`GetFilePathsByRecordingMBIDs` is what answers for a *catalog*
-than a lookup by track id because **`MBTrack.LocalID` is declared and
+tracklist. It is the third member of the `GetFilePathsBy…` family: one
-nothing in the backend ever writes it**.
+query, paths only, grouped so the caller keeps the tracklist's order.
 It exists rather than a lookup by track id because **`MBTrack.LocalID`
 is declared and nothing in the backend ever writes it**.
 **How much of an album is here is a question the files can answer.**
 `ownership()` above counts the *displayed* tracklist, which for a
-library copy is a tautology — every local track is `inLibrary: true`,
+library copy is a tautology — every local track has a file, so owned
-so owned always equals total and "do I have all of this" had no local
+always equals total and "do I have all of this" had no local answer.
-answer. The album page therefore asked MusicBrainz, and
+The album page therefore asked MusicBrainz, and
 `BrowseReleases` is the most expensive call the app makes: releases
 plus every version's full tracklist, on a 1 req/s limiter shared with
 `PrefetchReleases`, which fires up to eight when an artist page
@@ -990,8 +1128,7 @@ renders.
 The denominator was already on disk. `metadata` has read the "5/12"
 totals off every file since forever (`m.Track()`, `m.Disc()`) and
-discarded them; they persist to
+discarded them; they persist to `audio_files.total_tracks` now, and
 `release_group_recordings.total_tracks` now, and
 `GetAlbumCompleteness` sums them. **A complete, MBID-matched album
 makes no catalog call at all** — identity from the MBID, tracklist from
 the tags, which between them are what the browse was being spent on.
@@ -1010,9 +1147,30 @@ Owned counts *distinct track numbers* for the same reason in reverse:
 this app detects duplicates, and counting two files of track 3 twice
 would report a short album as complete.
-What tags cannot give is *which* tracks are missing, only how many — so
+**Where the tags have no total, the catalog does.** `explore_index`
-an incomplete album still browses, and that is now the exception rather
+carries a per-release-group `total_tracks` — ~2 bytes across 400,677
-than every album load. Two smaller consequences: existing databases
+rows, about 800 kB of artifact — and `completenessAnswer()` merges the
 two: the numerator stays local (distinct track numbers on disk) and
 only the denominator is borrowed, because a catalog total describes the
 canonical release while the files' own total, where they declare one,
 describes the release the user actually has. Zero still means "the
 catalog does not say", so an album neither side can total keeps the
 plain tick.
 Two rules hold up the column itself. **It is counted before the
 popularity filter**: `cmd/indexbuild` counts the canonical dump's rows
 per kept release, and counting only the *kept recordings* would say "9"
 about a twelve-track album whose other three nobody has played — the
 same confident lie that kept whole tracklists out of the artifact.
 And **adding a column to the importer's SELECT is how you break every
 artifact already published**, so `artifactHasTotals()` asks the
 attached artifact whether the column is there (on the writer, where
 `core` is attached) and selects a literal `0` when it is not — the same
 shape as the encoding probe beside it.
 What neither side can give is *which* tracks are missing, only how many
 — so an incomplete album still browses, and that is now the exception
 rather than every album load. Two smaller consequences: existing databases
 read "unknown" until a rescan repopulates the column (which degrades to
 exactly the old behaviour, so nothing breaks), and our own `tagwriter`
 writes track and disc *numbers* but not totals, so autotagging a folder
@@ -1095,6 +1253,27 @@ missing half; `catalogFailed` is the only route to `unavailable` now,
 and the timer is a 60 s backstop for a genuine hang rather than the
 verdict.
 **Activating a row plays the list the row is in, from that row.** A
 double-click — and Play on a single row's context menu — queues the
 list as *displayed* with `startIndex` on that row, not a queue of one
 track that stops when the song ends. The two playlist views and
 `cover-grid`'s album dropdown always did this; the album page and
 `track-list` did not, so playing anything from the two largest
 tracklists in the app discarded the album around it. Three rules come
 with it. **A menu asks how much is selected**: one row is a position
 and means "from here", several rows are an explicit choice of *those*
 tracks and become the queue on their own (which is also the only case
 where `shuffleStart` still applies, since no one row was named as the
 place to start). **The index is into the paths, not into the rows** —
 `explore-album-details` queues `ownedFilePaths()` and its dimmed rows
 are not in it, so an index taken from the tracklist starts an album
 somewhere else entirely, or past its end. And **the index is looked up
 when it is used, not remembered**: selection keys are file paths
 because those survive a re-sort, a re-filter and a refetch, and an
 index survives none of the three — `displayIndexOf` is that lookup,
 against `cachedSortedTracks`, which is the only order the user can see
 and therefore the only one they can mean.
 **Ask for what the caller uses, once.** "Play this artist" resolved
 file paths with one `GetAlbumTracks` per album, sequentially, and every
 one of the four sites doing that asked for whole track rows to read
@@ -328,43 +328,42 @@ func (a *Applier) Apply(
 func (a *Applier) syncDBMBIDs(
 	ctx context.Context, tr TrackApply, cand Candidate,
 ) error {
 	// Look up recording row via audio_file.
 	af, err := a.q.GetAudioFile(ctx, tr.Local.AudioFileID)
 	if err != nil {
 		return fmt.Errorf("get audio_file: %w", err)
 	}
 	if tr.CandidateTrack.MBID != "" {
-		if err := a.q.SetRecordingMBID(ctx, sqlcgen.SetRecordingMBIDParams{
+		if err := a.q.SetFileRecordingMBID(ctx, sqlcgen.SetFileRecordingMBIDParams{
-			Mbid: sql.NullString{String: tr.CandidateTrack.MBID, Valid: true},
+			RecordingMbid: sql.NullString{String: tr.CandidateTrack.MBID, Valid: true},
-			ID:   af.RecordingID,
+			ID:            tr.Local.AudioFileID,
 		}); err != nil {
 			return fmt.Errorf("set recording mbid: %w", err)
 		}
 	}
-	if cand.ReleaseGroupMBID != "" {
+	if cand.ReleaseGroupMBID == "" {
-		rgID, err := a.q.GetRecordingReleaseGroupID(ctx, af.RecordingID)
+		return nil
-		if err == nil && rgID > 0 {
+	}
 			if err := a.q.SetReleaseGroupMBID(ctx, sqlcgen.SetReleaseGroupMBIDParams{
 				Mbid: sql.NullString{String: cand.ReleaseGroupMBID, Valid: true},
 				ID:   rgID,
 			}); err != nil {
 				return fmt.Errorf("set release group mbid: %w", err)
 			}
-			// Stamp the release-group's original-release year too —
+	// The album is reached through the file rather than through two
-			// this is what the tracklist / smart-playlist year rule
+	// join tables; SetFileAlbumMBID takes the file id and does the
-			// surfaces by default once the user accepts a candidate.
+	// lookup in one statement.
-			if year := parseYear(cand.OriginalDate); year > 0 {
+	if err := a.q.SetFileAlbumMBID(ctx, sqlcgen.SetFileAlbumMBIDParams{
-				if err := a.q.SetReleaseGroupOriginalYear(
+		Mbid: sql.NullString{String: cand.ReleaseGroupMBID, Valid: true},
-					ctx, sqlcgen.SetReleaseGroupOriginalYearParams{
+		ID:   tr.Local.AudioFileID,
-						OriginalYear: sql.NullInt64{Int64: int64(year), Valid: true},
+	}); err != nil {
-						ID:           rgID,
+		return fmt.Errorf("set album mbid: %w", err)
-					},
+	}
-				); err != nil {
+
-					return fmt.Errorf("set release group original year: %w", err)
+	// Stamp the album's original-release year too - this is what the
-				}
+	// tracklist and the smart-playlist year rule surface by default
 	// once the user accepts a candidate.
 	if year := parseYear(cand.OriginalDate); year > 0 {
 		af, err := a.q.GetAudioFile(ctx, tr.Local.AudioFileID)
 		if err == nil && af.AlbumID.Valid {
 			if err := a.q.SetAlbumOriginalYear(
 				ctx, sqlcgen.SetAlbumOriginalYearParams{
 					OriginalYear: sql.NullInt64{Int64: int64(year), Valid: true},
 					ID:           af.AlbumID.Int64,
 				},
 			); err != nil {
 				return fmt.Errorf("set album original year: %w", err)
 			}
 		}
 	}
@@ -2,7 +2,6 @@ package autotag_test
 import (
 	"context"
 	"database/sql"
 	"log/slog"
 	"sync"
 	"testing"
@@ -87,51 +86,22 @@ func seedAudioFiles(
 	q := db.Queries
 	ctx := db.Ctx
 	ac, err := q.UpsertArtistCredit(ctx, "Test Artist")
 	if err != nil {
 		t.Fatalf("upsert artist credit: %v", err)
 	}
 	rg, err := q.UpsertReleaseGroup(ctx, sqlcgen.UpsertReleaseGroupParams{
 		Name:                "Test Album",
 		AlbumArtistCreditID: sql.NullInt64{Int64: ac.ID, Valid: true},
 	})
 	if err != nil {
 		t.Fatalf("upsert rg: %v", err)
 	}
 	out := make([]sqlcgen.AudioFile, 0, len(paths))
 	for i, p := range paths {
-		rec, err := q.CreateRecordingFull(ctx, sqlcgen.CreateRecordingFullParams{
+		id := database.InsertTestTrack(t, db, database.TestTrack{
-			Name:           p,
+			FilePath:    p,
-			ArtistCreditID: ac.ID,
+			Title:       p,
-			TrackNumber:    sql.NullInt64{Int64: int64(i + 1), Valid: true},
+			Artist:      "Test Artist",
 			Album:       "Test Album",
 			TrackNumber: int64(i + 1),
 			LengthMs:    100000,
 			GroupKey:    groupKey,
 		})
 		if err != nil {
 			t.Fatalf("create recording: %v", err)
 		}
-		if _, err := q.CreateReleaseGroupRecording(ctx, sqlcgen.CreateReleaseGroupRecordingParams{
+		af, err := q.GetAudioFile(ctx, id)
 			ReleaseGroupID: rg.ID,
 			RecordingID:    rec.ID,
 			TrackNumber:    sql.NullInt64{Int64: int64(i + 1), Valid: true},
 		}); err != nil {
 			t.Fatalf("link rg recording: %v", err)
 		}
 		af, err := q.CreateAudioFileWithGroupKey(ctx, sqlcgen.CreateAudioFileWithGroupKeyParams{
 			FilePath:           p,
 			LengthMilliseconds: 100000,
 			FileTypeID:         0,
 			RecordingID:        rec.ID,
 			Basename:           p,
 			LibraryID:          0,
 			GroupKey:           groupKey,
 			TagStatus:          "untagged",
 		})
 		if err != nil {
-			t.Fatalf("create audio file: %v", err)
+			t.Fatalf("read seeded audio file: %v", err)
 		}
 		out = append(out, af)
@@ -56,7 +56,7 @@ func (r *LocalResolver) LocalTracksForGroup(
 }
 // ResolveLocal returns candidate releases sourced from the local
-// DB's release_groups rows (filtered to those carrying an MBID)
+// DB's albums (filtered to those carrying an MBID)
 // whose normalized name matches the tagging item's album name.
 // No network calls.  Candidates carry all tracks flat; caller runs
 // AlignTracks on each to produce per-track alignments.
@@ -67,7 +67,7 @@ func (r *LocalResolver) ResolveLocal(
 		return nil, nil
 	}
-	rows, err := r.q.ListLocalReleaseGroupCandidates(ctx, albumName)
+	rows, err := r.q.ListLocalAlbumCandidates(ctx, albumName)
 	if err != nil {
 		return nil, fmt.Errorf("list local candidates: %w", err)
 	}
@@ -84,12 +84,12 @@ func (r *LocalResolver) ResolveLocal(
 			continue
 		}
-		if _, ok := byID[row.ReleaseGroupID]; !ok {
+		if _, ok := byID[row.AlbumID]; !ok {
-			byID[row.ReleaseGroupID] = localCandidate(row)
+			byID[row.AlbumID] = localCandidate(row)
 		}
-		tracksByID[row.ReleaseGroupID] = append(
+		tracksByID[row.AlbumID] = append(
-			tracksByID[row.ReleaseGroupID],
+			tracksByID[row.AlbumID],
 			CandidateTrack{
 				Position:     int(row.TrackNumber),
 				DiscNumber:   int(row.DiscNumber),
@@ -113,15 +113,15 @@ func (r *LocalResolver) ResolveLocal(
 // localCandidate converts one sqlc row (minus track-level fields)
 // into a Candidate shell.  Track fields and alignments are filled
 // in by the caller.
-func localCandidate(row sqlcgen.ListLocalReleaseGroupCandidatesRow) *Candidate {
+func localCandidate(row sqlcgen.ListLocalAlbumCandidatesRow) *Candidate {
 	date := ""
 	if row.Year > 0 {
 		date = fmt.Sprintf("%04d", row.Year)
 	}
 	mbid := ""
-	if row.ReleaseGroupMbid.Valid {
+	if row.AlbumMbid.Valid {
-		mbid = row.ReleaseGroupMbid.String
+		mbid = row.AlbumMbid.String
 	}
 	return &Candidate{
@@ -2,13 +2,11 @@ package autotag_test
 import (
 	"context"
 	"database/sql"
 	"log/slog"
 	"testing"
 	"yellowjacket/backend/autotag"
 	"yellowjacket/backend/database"
 	"yellowjacket/backend/database/sql/sqlcgen"
 )
 // seedAlbum drops a minimal release_group + recordings + audio_files
@@ -33,70 +31,19 @@ type seededTrack struct {
 func seed(t *testing.T, db *database.DB, album seededAlbum) {
 	t.Helper()
 	ctx := db.Ctx
 	q := db.Queries
 	ac, err := q.UpsertArtistCredit(ctx, "Test Artist")
 	if err != nil {
 		t.Fatalf("upsert ac: %v", err)
 	}
 	rg, err := q.UpsertReleaseGroup(ctx, sqlcgen.UpsertReleaseGroupParams{
 		Name:                album.albumName,
 		AlbumArtistCreditID: sql.NullInt64{Int64: ac.ID, Valid: true},
 	})
 	if err != nil {
 		t.Fatalf("upsert rg: %v", err)
 	}
 	if album.releaseMBID != "" {
 		if _, err := db.ExecContext(
 			`UPDATE release_groups SET mbid = ? WHERE id = ?`,
 			album.releaseMBID, rg.ID,
 		); err != nil {
 			t.Fatalf("set rg mbid: %v", err)
 		}
 	}
 	for _, tr := range album.tracks {
-		rec, err := q.CreateRecordingFull(ctx, sqlcgen.CreateRecordingFullParams{
+		database.InsertTestTrack(t, db, database.TestTrack{
-			Name:           tr.title,
+			FilePath:      tr.filePath,
-			ArtistCreditID: ac.ID,
+			Title:         tr.title,
-			TrackNumber:    sql.NullInt64{Int64: int64(tr.trackNumber), Valid: true},
+			Artist:        "Test Artist",
 			Album:         album.albumName,
 			AlbumMBID:     album.releaseMBID,
 			RecordingMBID: tr.recordingMBID,
 			TrackNumber:   int64(tr.trackNumber),
 			LengthMs:      tr.lengthMillis,
 			LibraryID:     album.libraryID,
 			GroupKey:      album.groupKey,
 		})
 		if err != nil {
 			t.Fatalf("create recording: %v", err)
 		}
 		if tr.recordingMBID != "" {
 			if _, err := db.ExecContext(
 				`UPDATE recordings SET mbid = ? WHERE id = ?`,
 				tr.recordingMBID, rec.ID,
 			); err != nil {
 				t.Fatalf("set recording mbid: %v", err)
 			}
 		}
 		if _, err := q.CreateReleaseGroupRecording(ctx, sqlcgen.CreateReleaseGroupRecordingParams{
 			ReleaseGroupID: rg.ID,
 			RecordingID:    rec.ID,
 			TrackNumber:    sql.NullInt64{Int64: int64(tr.trackNumber), Valid: true},
 		}); err != nil {
 			t.Fatalf("link rg recording: %v", err)
 		}
 		if _, err := q.CreateAudioFileWithGroupKey(ctx, sqlcgen.CreateAudioFileWithGroupKeyParams{
 			FilePath:           tr.filePath,
 			LengthMilliseconds: tr.lengthMillis,
 			FileTypeID:         0,
 			RecordingID:        rec.ID,
 			Basename:           tr.filePath,
 			LibraryID:          album.libraryID,
 			GroupKey:           album.groupKey,
 			TagStatus:          "untagged",
 		}); err != nil {
 			t.Fatalf("create audio file: %v", err)
 		}
 	}
 	if _, err := db.ExecContext(`
@@ -19,6 +19,8 @@ const applyJobPrefix = "autotag:"
 // registry gets progress, cancel and the global indicator for free; the
 // three subsystems that lacked them were the three that were not
 // registered.
 //
 //wails:ignore // internal wiring, not part of the app's IPC surface.
 func (s *Service) SetJobRegistry(reg *jobs.Registry) {
 	s.mu.Lock()
 	s.jobsReg = reg
@@ -3,12 +3,12 @@ package autotagservice
 import (
 	"database/sql"
 	"errors"
 	"fmt"
 	"log/slog"
 	"testing"
 	"yellowjacket/backend/autotag"
 	"yellowjacket/backend/database"
 	"yellowjacket/backend/database/sql/sqlcgen"
 )
 // newTestService builds a Service with just enough wired up for
@@ -36,62 +36,18 @@ func newTestService(t *testing.T, db *database.DB) *Service {
 func seedMixedBagFolder(t *testing.T, db *database.DB, groupKey string, libraryID int64) {
 	t.Helper()
 	ctx := db.Ctx
 	q := db.Queries
 	addTrack := func(filePath, title, artist, album, albumArtist string, trackNum int) {
-		ac, err := q.UpsertArtistCredit(ctx, artist)
+		database.InsertTestTrack(t, db, database.TestTrack{
-		if err != nil {
+			FilePath:    filePath,
-			t.Fatalf("upsert artist credit: %v", err)
+			Title:       title,
-		}
+			Artist:      artist,
-
+			Album:       album,
-		rec, err := q.CreateRecordingFull(ctx, sqlcgen.CreateRecordingFullParams{
+			AlbumArtist: albumArtist,
-			Name:           title,
+			TrackNumber: int64(trackNum),
-			ArtistCreditID: ac.ID,
+			LengthMs:    200000,
-			TrackNumber:    sql.NullInt64{Int64: int64(trackNum), Valid: true},
+			LibraryID:   libraryID,
 			GroupKey:    groupKey,
 		})
 		if err != nil {
 			t.Fatalf("create recording: %v", err)
 		}
 		if album != "" {
 			albumArtistAC, err := q.UpsertArtistCredit(ctx, albumArtist)
 			if err != nil {
 				t.Fatalf("upsert album artist credit: %v", err)
 			}
 			rg, err := q.UpsertReleaseGroup(ctx, sqlcgen.UpsertReleaseGroupParams{
 				Name:                album,
 				AlbumArtistCreditID: sql.NullInt64{Int64: albumArtistAC.ID, Valid: true},
 			})
 			if err != nil {
 				t.Fatalf("upsert release group: %v", err)
 			}
 			if _, err := q.CreateReleaseGroupRecording(
 				ctx,
 				sqlcgen.CreateReleaseGroupRecordingParams{
 					ReleaseGroupID: rg.ID,
 					RecordingID:    rec.ID,
 					TrackNumber:    sql.NullInt64{Int64: int64(trackNum), Valid: true},
 				},
 			); err != nil {
 				t.Fatalf("link release group recording: %v", err)
 			}
 		}
 		if _, err := q.CreateAudioFileWithGroupKey(ctx, sqlcgen.CreateAudioFileWithGroupKeyParams{
 			FilePath:           filePath,
 			LengthMilliseconds: 200000,
 			FileTypeID:         0,
 			RecordingID:        rec.ID,
 			Basename:           filePath,
 			LibraryID:          libraryID,
 			GroupKey:           groupKey,
 			TagStatus:          "untagged",
 		}); err != nil {
 			t.Fatalf("create audio file: %v", err)
 		}
 	}
 	addTrack("/junk/01.mp3", "Song A1", "Artist One", "Album One", "Artist One", 1)
@@ -113,58 +69,22 @@ func seedMixedBagFolder(t *testing.T, db *database.DB, groupKey string, libraryI
 func seedCoherentAlbum(t *testing.T, db *database.DB, groupKey string, libraryID int64) {
 	t.Helper()
-	ctx := db.Ctx
+	for i, title := range []string{"Come Together", "Something", "Maxwell's Silver Hammer"} {
-	q := db.Queries
+		database.InsertTestTrack(t, db, database.TestTrack{
-
+			FilePath:    fmt.Sprintf("/beatles/%02d.mp3", i+1),
-	ac, err := q.UpsertArtistCredit(ctx, "The Beatles")
+			Title:       title,
-	if err != nil {
+			Artist:      "The Beatles",
-		t.Fatalf("upsert artist credit: %v", err)
+			Album:       "Abbey Road",
-	}
+			TrackNumber: int64(i + 1),
-
+			LengthMs:    200000,
-	rg, err := q.UpsertReleaseGroup(ctx, sqlcgen.UpsertReleaseGroupParams{
+			LibraryID:   libraryID,
-		Name:                "Abbey Road",
+			GroupKey:    groupKey,
 		AlbumArtistCreditID: sql.NullInt64{Int64: ac.ID, Valid: true},
 	})
 	if err != nil {
 		t.Fatalf("upsert release group: %v", err)
 	}
 	titles := []string{"Come Together", "Something", "Maxwell's Silver Hammer", "Oh! Darling"}
 	for i, title := range titles {
 		rec, err := q.CreateRecordingFull(ctx, sqlcgen.CreateRecordingFullParams{
 			Name:           title,
 			ArtistCreditID: ac.ID,
 			TrackNumber:    sql.NullInt64{Int64: int64(i + 1), Valid: true},
 		})
 		if err != nil {
 			t.Fatalf("create recording: %v", err)
 		}
 		if _, err := q.CreateReleaseGroupRecording(ctx, sqlcgen.CreateReleaseGroupRecordingParams{
 			ReleaseGroupID: rg.ID,
 			RecordingID:    rec.ID,
 			TrackNumber:    sql.NullInt64{Int64: int64(i + 1), Valid: true},
 		}); err != nil {
 			t.Fatalf("link release group recording: %v", err)
 		}
 		if _, err := q.CreateAudioFileWithGroupKey(ctx, sqlcgen.CreateAudioFileWithGroupKeyParams{
 			FilePath:           groupKey + "/" + title + ".mp3",
 			LengthMilliseconds: 200000,
 			FileTypeID:         0,
 			RecordingID:        rec.ID,
 			Basename:           title + ".mp3",
 			LibraryID:          libraryID,
 			GroupKey:           groupKey,
 			TagStatus:          "untagged",
 		}); err != nil {
 			t.Fatalf("create audio file: %v", err)
 		}
 	}
 	if _, err := db.ExecContext(`
 		INSERT INTO tagging_items (group_key, library_id, track_count, album_name, album_artist, disc_number, status)
-		VALUES (?, ?, 4, 'Abbey Road', 'The Beatles', 0, 'pending')
+		VALUES (?, ?, 3, 'Abbey Road', 'The Beatles', 0, 'pending')
 	`, groupKey, libraryID); err != nil {
 		t.Fatalf("insert tagging item: %v", err)
 	}
@@ -293,20 +213,11 @@ func TestSplitMixedFolder_NothingToClusterErrors(t *testing.T) {
 	db := database.NewTestDB(t)
-	if _, err := db.Queries.CreateAudioFileWithGroupKey(
+	database.InsertTestTrack(t, db, database.TestTrack{
-		db.Ctx,
+		FilePath: "/coherent/01.mp3",
-		sqlcgen.CreateAudioFileWithGroupKeyParams{
+		Title:    "Track",
-			FilePath:    "/coherent/01.mp3",
+		GroupKey: "g-coherent",
-			FileTypeID:  0,
+	})
 			RecordingID: mustCreateRecording(t, db, "Track"),
 			Basename:    "01.mp3",
 			LibraryID:   0,
 			GroupKey:    "g-coherent",
 			TagStatus:   "untagged",
 		},
 	); err != nil {
 		t.Fatalf("create audio file: %v", err)
 	}
 	if _, err := db.ExecContext(`
 		INSERT INTO tagging_items (group_key, library_id, track_count, album_name, album_artist, disc_number, status)
@@ -328,20 +239,11 @@ func TestListPendingFolders_PrunesOrphanedEntries(t *testing.T) {
 	db := database.NewTestDB(t)
 	// A real, live folder — must survive.
-	if _, err := db.Queries.CreateAudioFileWithGroupKey(
+	database.InsertTestTrack(t, db, database.TestTrack{
-		db.Ctx,
+		FilePath: "/live/01.mp3",
-		sqlcgen.CreateAudioFileWithGroupKeyParams{
+		Title:    "Track",
-			FilePath:    "/live/01.mp3",
+		GroupKey: "g-live",
-			FileTypeID:  0,
+	})
 			RecordingID: mustCreateRecording(t, db, "Track"),
 			Basename:    "01.mp3",
 			LibraryID:   0,
 			GroupKey:    "g-live",
 			TagStatus:   "untagged",
 		},
 	); err != nil {
 		t.Fatalf("create audio file: %v", err)
 	}
 	if _, err := db.ExecContext(`
 		INSERT INTO tagging_items (group_key, library_id, track_count, album_name, album_artist, disc_number, status)
@@ -385,22 +287,3 @@ func TestListPendingFolders_PrunesOrphanedEntries(t *testing.T) {
 		t.Errorf("expected g-orphan row to be deleted from tagging_items, got err=%v", err)
 	}
 }
 func mustCreateRecording(t *testing.T, db *database.DB, title string) int64 {
 	t.Helper()
 	ac, err := db.Queries.UpsertArtistCredit(db.Ctx, "Artist")
 	if err != nil {
 		t.Fatalf("upsert artist credit: %v", err)
 	}
 	rec, err := db.Queries.CreateRecordingFull(db.Ctx, sqlcgen.CreateRecordingFullParams{
 		Name:           title,
 		ArtistCreditID: ac.ID,
 	})
 	if err != nil {
 		t.Fatalf("create recording: %v", err)
 	}
 	return rec.ID
 }
@@ -12,7 +12,15 @@ import (
 // PathPrefix is the URL path prefix for cover art served by the asset handler.
 const PathPrefix = "/covers/"
-// URLs holds the resolved URL paths for all cover art size variants.
+// URLs holds the resolved URL paths for a cover's size variants.
 //
 // Original is the largest variant kept, which is the Large one: the
 // full-resolution image is no longer stored.  It was 1,134 MB of a
 // 1.4 GB covers directory on a real 2,057-album library against 110 MB
 // for all three rendered tiers, and nothing rendered it - the grid caps
 // at 350 px and the largest tier is 400.  The field keeps its name
 // because it is what a caller means by "the cover", and the bytes it
 // came from are still in the audio file if a bigger one is ever wanted.
 type URLs struct {
 	Original string
 	Small    string
@@ -36,25 +44,41 @@ func CoversDir() (string, error) {
 	return filepath.Join(dataDir, dirName), nil
 }
-// SizedFilename derives a sized-variant filename from an original cover art
+// Suffixes are the size variants a cover is stored as, largest last.
-// filename and a size suffix.
+var Suffixes = []string{"_sm", "_md", "_lg"}
 // For example, SizedFilename("a1b2c3d4.jpg", "_sm") returns "a1b2c3d4_sm.jpg".
 func SizedFilename(originalFilename, suffix string) string {
 	ext := filepath.Ext(originalFilename)
 	name := strings.TrimSuffix(originalFilename, ext)
-	return name + suffix + ".jpg"
+// SizedFilename derives a sized-variant filename from a cover art
 // filename and a size suffix.  The input may itself be a variant, so
 // its suffix is stripped first: SizedFilename("a1b2_lg.jpg", "_sm")
 // and SizedFilename("a1b2.jpg", "_sm") both return "a1b2_sm.jpg".
 func SizedFilename(filename, suffix string) string {
 	return BaseName(filename) + suffix + ".jpg"
 }
 // BaseName strips the extension and any size suffix from a cover art
 // filename, leaving the content hash that identifies the cover.
 func BaseName(filename string) string {
 	name := strings.TrimSuffix(filename, filepath.Ext(filename))
 	for _, suffix := range Suffixes {
 		if strings.HasSuffix(name, suffix) {
 			return strings.TrimSuffix(name, suffix)
 		}
 	}
 	return name
 }
 // ResolveURLs converts a cover art filesystem path into URL paths
 // for the original and all size variants (small, medium, large).
 func ResolveURLs(filesystemPath string) URLs {
 	base := filepath.Base(filesystemPath)
 	large := PathPrefix + SizedFilename(base, "_lg")
 	return URLs{
-		Original: PathPrefix + base,
+		Original: large,
 		Small:    PathPrefix + SizedFilename(base, "_sm"),
 		Medium:   PathPrefix + SizedFilename(base, "_md"),
-		Large:    PathPrefix + SizedFilename(base, "_lg"),
+		Large:    large,
 	}
 }
@@ -108,8 +108,10 @@ func TestResolveURLs(t *testing.T) {
 	t.Parallel()
 	tests := []struct {
-		name     string
+		name string
-		path     string
+		path string
 		// Original is the largest kept variant: the full-resolution
 		// image is not stored (see URLs).
 		wantOrig string
 		wantSm   string
 		wantMd   string
@@ -118,7 +120,7 @@ func TestResolveURLs(t *testing.T) {
 		{
 			name:     "absolute path",
 			path:     "/home/user/.local/share/yellowjacket/covers/a1b2c3d4.jpg",
-			wantOrig: "/covers/a1b2c3d4.jpg",
+			wantOrig: "/covers/a1b2c3d4_lg.jpg",
 			wantSm:   "/covers/a1b2c3d4_sm.jpg",
 			wantMd:   "/covers/a1b2c3d4_md.jpg",
 			wantLg:   "/covers/a1b2c3d4_lg.jpg",
@@ -126,7 +128,7 @@ func TestResolveURLs(t *testing.T) {
 		{
 			name:     "bare filename",
 			path:     "abcdef01.png",
-			wantOrig: "/covers/abcdef01.png",
+			wantOrig: "/covers/abcdef01_lg.jpg",
 			wantSm:   "/covers/abcdef01_sm.jpg",
 			wantMd:   "/covers/abcdef01_md.jpg",
 			wantLg:   "/covers/abcdef01_lg.jpg",
@@ -5,13 +5,10 @@ import (
 	"context"
 	"database/sql"
 	"embed"
 	"errors"
 	"fmt"
 	"io/fs"
 	"log/slog"
 	"path"
 	"sort"
 	"strconv"
 	"strings"
 	_ "modernc.org/sqlite" // Register sqlite driver.
@@ -26,9 +23,6 @@ import (
 //go:embed sql/schemas/*.sql
 var schemas embed.FS
 //go:embed sql/migrations/*.sql
 var migrations embed.FS
 // DB wraps the SQLite database connection and queries.
 //
 // Two handles back a single database file.  db is the single-writer
@@ -301,20 +295,19 @@ func (d *DB) ResumeExploreIndexFTS() error {
 	return nil
 }
-// applySchema creates the full schema on a fresh database and brings
+// applySchema creates the full schema.
 // an existing one up to date via sql/migrations.
 //
-// The schema files under sql/schemas are CREATE ... IF NOT EXISTS,
+// The schema files under sql/schemas are CREATE ... IF NOT EXISTS and
-// so on a genuinely new database they create every table already at
+// declare the current, latest shape of every table — so running them
-// its current, latest shape — that's the fast path new installs
+// against a fresh database produces exactly that shape, and running
-// take.  A database that already has an older shape (e.g. a
+// them against a database already at that shape does nothing.  That is
-// tagging_items missing a column a later build added) needs the gap
+// the whole mechanism; there is no migration chain and no
-// closed, which IF NOT EXISTS can't do: it silently no-ops on a
+// schema_migrations table.
-// table that already exists, columns and all.  sql/migrations holds
+//
-// small, additive, numbered files (ALTER TABLE, CREATE INDEX, etc.)
+// There was one, and it was squashed (see .planning/plans/013): a chain
-// for exactly that gap, tracked in schema_migrations so each applies
+// only earns its keep once real user databases exist in the wild, and
-// at most once — see applyMigrations for how a fresh database's
+// until then it is a second description of the schema that can drift
-// already-current tables tolerate replaying them anyway.
+// from the first — which this project has already been bitten by once.
 func applySchema(ctx context.Context, db *sql.DB) error {
 	dirEntries, err := schemas.ReadDir("sql/schemas")
 	if err != nil {
@@ -344,171 +337,9 @@ func applySchema(ctx context.Context, db *sql.DB) error {
 		return fmt.Errorf("could not create explore FTS triggers: %w", err)
 	}
 	if err := applyMigrations(ctx, db); err != nil {
 		return fmt.Errorf("could not apply migrations: %w", err)
 	}
 	// The download subsystem's Want/Request rename reuses table names
 	// (download_requests names a different table before and after), so
 	// it cannot be a plain sql/migrations file the way an ADD COLUMN
 	// migration can; see download_rename_migration.go for why.
 	if err := migrateDownloadRename(ctx, db); err != nil {
 		return fmt.Errorf("could not migrate download rename: %w", err)
 	}
 	return nil
 }
 // schemaMigrationsTable tracks which sql/migrations files have run,
 // by their leading numeric prefix.
 const schemaMigrationsTable = `
 CREATE TABLE IF NOT EXISTS schema_migrations (
  version    INTEGER PRIMARY KEY,
  applied_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
 )`
 // applyMigrations runs every sql/migrations file not yet recorded in
 // schema_migrations, in filename order (numeric prefix), one
 // statement at a time.
 //
 // Every migration runs on EVERY database, fresh or old — there is no
 // "skip on fresh install" branch. A fresh database's tables already
 // carry a migration's effect (sql/schemas declares the target shape
 // directly), so its statements are expected to sometimes be no-ops
 // there: "duplicate column name" from an ALTER TABLE ADD COLUMN is
 // tolerated and treated as "already applied", the same way
 // createExploreIndexFTSTriggers tolerates "already exists". Any
 // other error is fatal. This is deliberately simpler than detecting
 // "is this database fresh" — every migration converges both a fresh
 // and an upgraded database to the identical final schema (including
 // column order — ALTER TABLE ADD COLUMN always appends at the end,
 // so sql/schemas must declare a migrated column last too; see the
 // comment on tagging_items.sql and the regression test in
 // migrations_test.go).
 func applyMigrations(ctx context.Context, db *sql.DB) error {
 	if _, err := db.ExecContext(ctx, schemaMigrationsTable); err != nil {
 		return fmt.Errorf("create schema_migrations: %w", err)
 	}
 	dirEntries, err := migrations.ReadDir("sql/migrations")
 	if err != nil {
 		return fmt.Errorf("could not read migrations directory: %w", err)
 	}
 	sort.Slice(dirEntries, func(i, j int) bool {
 		return dirEntries[i].Name() < dirEntries[j].Name()
 	})
 	for _, dirEntry := range dirEntries {
 		if dirEntry.IsDir() {
 			continue
 		}
 		version, err := migrationVersion(dirEntry.Name())
 		if err != nil {
 			return err
 		}
 		applied, err := migrationApplied(ctx, db, version)
 		if err != nil {
 			return err
 		}
 		if applied {
 			continue
 		}
 		filePath := path.Join("sql/migrations", dirEntry.Name())
 		sqlContent, err := fs.ReadFile(migrations, filePath)
 		if err != nil {
 			return fmt.Errorf("could not read file %s: %w", filePath, err)
 		}
 		if err := execMigrationStatements(ctx, db, string(sqlContent)); err != nil {
 			return fmt.Errorf("error executing migration %s: %w", dirEntry.Name(), err)
 		}
 		if _, err := db.ExecContext(
 			ctx, `INSERT INTO schema_migrations (version) VALUES (?)`, version,
 		); err != nil {
 			return fmt.Errorf("record migration %d applied: %w", version, err)
 		}
 	}
 	return nil
 }
 // execMigrationStatements runs a migration file one statement at a
 // time — NOT as one multi-statement Exec — so that one statement
 // being a tolerable no-op (ALTER TABLE ADD COLUMN on a fresh
 // database) doesn't abort the statements after it in the same file
 // (e.g. a trailing CREATE INDEX that a fresh database still needs,
 // since sql/schemas deliberately doesn't declare an index on a
 // migrated column — see the comment on tagging_items.sql).
 //
 // Splitting on ";" is safe for the simple ALTER/CREATE TABLE/CREATE
 // INDEX statements migrations are expected to contain; it is NOT
 // safe for statements embedding a literal semicolon (e.g. a CREATE
 // TRIGGER body) — write those with executeContext calls in Go
 // instead of a sql/migrations file, the same way the explore FTS
 // triggers already are.
 func execMigrationStatements(ctx context.Context, db *sql.DB, script string) error {
 	for stmt := range strings.SplitSeq(script, ";") {
 		stmt = strings.TrimSpace(stmt)
 		if stmt == "" {
 			continue
 		}
 		if _, err := db.ExecContext(ctx, stmt); err != nil {
 			if strings.Contains(err.Error(), "duplicate column name") {
 				continue
 			}
 			return fmt.Errorf("statement %q: %w", stmt, err)
 		}
 	}
 	return nil
 }
 // migrationVersion extracts the leading integer prefix from a
 // migration filename, e.g. "0001_tagging_items_synthetic.sql" -> 1.
 func migrationVersion(filename string) (int, error) {
 	prefix, _, ok := strings.Cut(filename, "_")
 	if !ok {
 		return 0, fmt.Errorf("%w: %s", errMigrationFilename, filename)
 	}
 	version, err := strconv.Atoi(prefix)
 	if err != nil {
 		return 0, fmt.Errorf("%w: %s", errMigrationFilename, filename)
 	}
 	return version, nil
 }
 var errMigrationFilename = errors.New(
 	"migration filename must start with a numeric prefix followed by '_' (e.g. 0001_description.sql)",
 )
 func migrationApplied(ctx context.Context, db *sql.DB, version int) (bool, error) {
 	var v int
 	err := db.QueryRowContext(
 		ctx, `SELECT version FROM schema_migrations WHERE version = ?`, version,
 	).Scan(&v)
 	switch {
 	case errors.Is(err, sql.ErrNoRows):
 		return false, nil
 	case err != nil:
 		return false, fmt.Errorf("check migration %d: %w", version, err)
 	default:
 		return true, nil
 	}
 }
 // applyPRAGMAs configures SQLite connection settings. Called by both
 // NewDB and NewTestDB to ensure identical behavior.
 func applyPRAGMAs(ctx context.Context, db *sql.DB) error {
@@ -312,31 +312,13 @@ func TestPhantomPlaylistTracksAreCleaned(t *testing.T) {
 	// Create prerequisite data: artist_credit, recording,
 	// audio_file.
-	_, err := db.ExecContext(
+	InsertTestTrack(t, db, TestTrack{
-		"INSERT INTO artist_credit (id, text) VALUES (1, 'Test Artist')",
+		FilePath:  "/test/music/song.mp3",
-	)
+		Title:     "Test Song",
-	if err != nil {
+		Artist:    "Test Artist",
-		t.Fatalf("insert artist_credit: %v", err)
+		LengthMs:  180000,
-	}
+		LibraryID: libID,
-
+	})
 	_, err = db.ExecContext(
 		"INSERT INTO recordings (id, name, artist_credit_id) " +
 			"VALUES (1, 'Test Song', 1)",
 	)
 	if err != nil {
 		t.Fatalf("insert recording: %v", err)
 	}
 	_, err = db.ExecContext(
 		"INSERT INTO audio_files "+
 			"(id, file_path, length_milliseconds, file_type_id, "+
 			"recording_id, library_id) "+
 			"VALUES (1, '/test/music/song.mp3', 180000, 0, 1, ?)",
 		libID,
 	)
 	if err != nil {
 		t.Fatalf("insert audio_file: %v", err)
 	}
 	// Create playlist.
 	playlist, err := db.Queries.CreatePlaylist(
@@ -442,39 +424,18 @@ func TestAudioFilesLibraryForeignKey(t *testing.T) {
 	db, libID := NewTestDBWithLibrary(t, "Test", "/test/fk-lib")
 	// Insert prerequisite recording.
 	InsertTestTrack(t, db, TestTrack{
 		FilePath:  "/test/track.mp3",
 		Title:     "Track",
 		Artist:    "Test",
 		LibraryID: libID,
 	})
 	// Insert audio file with invalid library_id - should fail FK.
 	_, err := db.ExecContext(
 		"INSERT INTO artist_credit (id, text) VALUES (1, 'Test')",
 	)
 	if err != nil {
 		t.Fatalf("insert artist_credit: %v", err)
 	}
 	_, err = db.ExecContext(
 		"INSERT INTO recordings (id, name, artist_credit_id) " +
 			"VALUES (1, 'Track', 1)",
 	)
 	if err != nil {
 		t.Fatalf("insert recording: %v", err)
 	}
 	// Insert audio file with valid library_id — should succeed.
 	_, err = db.ExecContext(
 		"INSERT INTO audio_files "+
 			"(id, file_path, length_milliseconds, file_type_id, "+
 			"recording_id, library_id) "+
 			"VALUES (1, '/test/song.mp3', 180000, 0, 1, ?)",
 		libID,
 	)
 	if err != nil {
 		t.Fatalf("insert audio_file with valid library: %v", err)
 	}
 	// Insert audio file with invalid library_id — should fail FK.
 	_, err = db.ExecContext(
 		"INSERT INTO audio_files " +
-			"(id, file_path, length_milliseconds, file_type_id, " +
+			"(id, file_path, length_milliseconds, file_type_id, library_id) " +
-			"recording_id, library_id) " +
+			"VALUES (2, '/test/song2.mp3', 200000, 0, 999)",
 			"VALUES (2, '/test/song2.mp3', 200000, 0, 1, 999)",
 	)
 	if err == nil {
 		t.Error(
@@ -483,16 +444,16 @@ func TestAudioFilesLibraryForeignKey(t *testing.T) {
 	}
 	// Count files by library.
-	count, err := db.Queries.CountAudioFilesByLibrary(
+	count, err := db.Queries.CountAudioFiles(
 		db.Ctx, libID,
 	)
 	if err != nil {
-		t.Fatalf("CountAudioFilesByLibrary: %v", err)
+		t.Fatalf("CountAudioFiles: %v", err)
 	}
 	if count != 1 {
 		t.Errorf(
-			"CountAudioFilesByLibrary = %d, want 1", count,
+			"CountAudioFiles = %d, want 1", count,
 		)
 	}
 }
@@ -503,31 +464,13 @@ func TestTrackMetadataViewHasLibraryID(t *testing.T) {
 	db, libID := NewTestDBWithLibrary(t, "Test", "/test/view-lib")
 	// Insert prerequisites.
-	_, err := db.ExecContext(
+	InsertTestTrack(t, db, TestTrack{
-		"INSERT INTO artist_credit (id, text) VALUES (1, 'View Artist')",
+		FilePath:  "/test/view.mp3",
-	)
+		Title:     "View Track",
-	if err != nil {
+		Artist:    "View Artist",
-		t.Fatalf("insert artist_credit: %v", err)
+		LengthMs:  200000,
-	}
+		LibraryID: libID,
-
+	})
 	_, err = db.ExecContext(
 		"INSERT INTO recordings (id, name, artist_credit_id) " +
 			"VALUES (1, 'View Track', 1)",
 	)
 	if err != nil {
 		t.Fatalf("insert recording: %v", err)
 	}
 	_, err = db.ExecContext(
 		"INSERT INTO audio_files "+
 			"(id, file_path, length_milliseconds, file_type_id, "+
 			"recording_id, library_id) "+
 			"VALUES (1, '/test/view.mp3', 200000, 0, 1, ?)",
 		libID,
 	)
 	if err != nil {
 		t.Fatalf("insert audio_file: %v", err)
 	}
 	// Query track_metadata VIEW and verify library_id is present
 	// with the correct value.
@@ -842,29 +785,13 @@ func TestPlayHistoryTable(t *testing.T) {
 	// Round-trip: insert a play_history row and verify play_count update.
 	// First, set up test data. The test DB already has library id=0.
-	_, err = db.ExecContext(
+	InsertTestTrack(t, db, TestTrack{
-		"INSERT OR IGNORE INTO artist_credit (id, text) VALUES (1, 'Test Artist')",
+		FilePath:    "/test/play_history.mp3",
-	)
+		Title:       "Test Track",
-	if err != nil {
+		Artist:      "Test Artist",
-		t.Fatalf("insert artist_credit: %v", err)
+		TrackNumber: 1,
-	}
+		DiscNumber:  1,
-
+	})
 	_, err = db.ExecContext(
 		`INSERT OR IGNORE INTO recordings (id, name, artist_credit_id, track_number, disc_number)
 		 VALUES (1, 'Test Track', 1, 1, 1)`,
 	)
 	if err != nil {
 		t.Fatalf("insert recording: %v", err)
 	}
 	_, err = db.ExecContext(
 		`INSERT INTO audio_files
 		 (id, file_path, length_milliseconds, file_type_id, recording_id, library_id)
 		 VALUES (1, '/test/track.mp3', 180000, 0, 1, 0)`,
 	)
 	if err != nil {
 		t.Fatalf("insert audio_file: %v", err)
 	}
 	// Verify default play_count is 0.
 	var playCount int64
@@ -1,141 +0,0 @@
 package database
 import (
 	"context"
 	"database/sql"
 	"errors"
 	"fmt"
 )
 // migrateDownloadRename performs the download subsystem's table rename
 // for existing databases that still carry the old table names: the
 // durable "I asked for this" record moved from download_wants to
 // download_requests, and the one-shot search-and-grab attempt moved
 // from download_requests to download_downloads (see CLAUDE.md and
 // .planning/NOTES.md for the full Want->Request / Request->Download
 // rename).
 //
 // This cannot be a plain sql/migrations file the way an ADD COLUMN
 // migration is. That pattern's tolerance for "duplicate column name"
 // works because a fresh database's sql/schemas pass already produces
 // the identical target shape under the identical table name, so
 // replaying the ALTER TABLE against it is a safe no-op. Here the name
 // "download_requests" is reused for a different table before and after
 // the rename, so a fresh database's schema pass creates a real, empty,
 // correctly-shaped download_downloads AND a real, empty,
 // correctly-shaped (new) download_requests before this ever runs.
 // Blindly replaying "ALTER TABLE download_requests RENAME TO
 // download_downloads" against that fresh database would rename the new,
 // empty Request table into Download's place, destroying the fresh
 // install rather than no-opping. Gating on whether the OLD
 // download_wants table still exists — a name nothing creates or
 // references once this has run — is what tells an old database and a
 // fresh (or already migrated) one apart without executing anything
 // destructive on the fresh path.
 func migrateDownloadRename(ctx context.Context, db *sql.DB) error {
 	var name string
 	err := db.QueryRowContext(
 		ctx,
 		`SELECT name FROM sqlite_master WHERE type = 'table' AND name = 'download_wants'`,
 	).Scan(&name)
 	switch {
 	case errors.Is(err, sql.ErrNoRows):
 		// Nothing to migrate: either a fresh install (sql/schemas
 		// already produced the target shape) or a database this has
 		// already run against.
 	case err != nil:
 		return fmt.Errorf("check for download_wants table: %w", err)
 	default:
 		if err := runDownloadRename(ctx, db); err != nil {
 			return err
 		}
 	}
 	return ensureDownloadIndexes(ctx, db)
 }
 // runDownloadRename performs the actual rename dance against a
 // database confirmed to still have the old download_wants table.
 func runDownloadRename(ctx context.Context, db *sql.DB) error {
 	stmts := []string{
 		// The schema pass already created an empty, correctly-shaped
 		// download_downloads placeholder under this name (it never
 		// existed under the old naming), which would otherwise collide
 		// with the rename below.
 		`DROP TABLE IF EXISTS download_downloads`,
 		// 1. Free the "download_requests" name: the old one-shot
 		// attempt table becomes download_downloads.
 		`ALTER TABLE download_requests RENAME TO download_downloads`,
 		`ALTER TABLE download_downloads RENAME COLUMN want_id TO request_id`,
 		// 2. Claim the now-free "download_requests" name for the
 		// durable-intent table.
 		`ALTER TABLE download_wants RENAME TO download_requests`,
 		// 3. The transfer table's FK now points at download_downloads.
 		`ALTER TABLE download_items RENAME COLUMN request_id TO download_id`,
 		// Named indexes survive a table/column rename attached to their
 		// old name, so drop them here; ensureDownloadIndexes recreates
 		// them under the names sql/schemas' comments describe.
 		`DROP INDEX IF EXISTS idx_download_requests_created`,
 		`DROP INDEX IF EXISTS idx_download_requests_state`,
 		`DROP INDEX IF EXISTS idx_download_wants_due`,
 		`DROP INDEX IF EXISTS idx_download_wants_entity`,
 		`DROP INDEX IF EXISTS idx_download_wants_parent`,
 		`DROP INDEX IF EXISTS idx_download_items_request`,
 	}
 	tx, err := db.BeginTx(ctx, nil)
 	if err != nil {
 		return fmt.Errorf("begin download rename migration: %w", err)
 	}
 	defer func() { _ = tx.Rollback() }()
 	for _, stmt := range stmts {
 		if _, err := tx.ExecContext(ctx, stmt); err != nil {
 			return fmt.Errorf("download rename migration %q: %w", stmt, err)
 		}
 	}
 	if err := tx.Commit(); err != nil {
 		return fmt.Errorf("commit download rename migration: %w", err)
 	}
 	return nil
 }
 // ensureDownloadIndexes creates the indexes sql/schemas deliberately
 // omits inline for the renamed table/columns (see
 // migrateDownloadRename), under their final names. Safe to call
 // unconditionally: IF NOT EXISTS makes it a no-op once created, and by
 // the time this runs every column/table involved is guaranteed to be
 // in its final shape on both a fresh and a migrated database.
 func ensureDownloadIndexes(ctx context.Context, db *sql.DB) error {
 	stmts := []string{
 		`CREATE INDEX IF NOT EXISTS idx_download_downloads_created
 		    ON download_downloads(created_at DESC)`,
 		`CREATE INDEX IF NOT EXISTS idx_download_downloads_state
 		    ON download_downloads(state)`,
 		`CREATE INDEX IF NOT EXISTS idx_download_requests_due
 		    ON download_requests(next_try_at) WHERE state = 'wanted'`,
 		`CREATE INDEX IF NOT EXISTS idx_download_requests_entity
 		    ON download_requests(entity, state)`,
 		`CREATE INDEX IF NOT EXISTS idx_download_requests_parent
 		    ON download_requests(parent_id)`,
 		`CREATE INDEX IF NOT EXISTS idx_download_items_download
 		    ON download_items(download_id)`,
 	}
 	for _, stmt := range stmts {
 		if _, err := db.ExecContext(ctx, stmt); err != nil {
 			return fmt.Errorf("ensure download index: %w", err)
 		}
 	}
 	return nil
 }
@@ -1,365 +0,0 @@
 package database
 import (
 	"database/sql"
 	"errors"
 	"testing"
 )
 // oldDownloadRequestsDDL, oldDownloadWantsDDL and oldDownloadItemsDDL
 // are frozen snapshots of the download subsystem's tables exactly as
 // they read before the Want/Request rename (see
 // download_rename_migration.go) — i.e. what a real user's existing
 // database looks like today, before upgrading to a build that includes
 // this migration.
 const oldDownloadRequestsDDL = `
 CREATE TABLE IF NOT EXISTS download_requests (
    id                 TEXT PRIMARY KEY,
    library_id         INTEGER NOT NULL,
    source             TEXT    NOT NULL DEFAULT 'manual',
    want_id            INTEGER REFERENCES download_wants(id) ON DELETE SET NULL,
    release_mbid       TEXT,
    release_group_mbid TEXT,
    recording_mbid     TEXT,
    artist             TEXT    NOT NULL DEFAULT '',
    album              TEXT    NOT NULL DEFAULT '',
    query              TEXT    NOT NULL DEFAULT '',
    expected           TEXT    NOT NULL DEFAULT '[]',
    state              TEXT    NOT NULL DEFAULT 'searching'
        CHECK(state IN ('searching', 'found', 'queued', 'grabbing',
                        'verifying', 'tagging', 'importing',
                        'complete', 'cancelled', 'failed')),
    error              TEXT    NOT NULL DEFAULT '',
    created_at         DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
    updated_at         DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
    FOREIGN KEY(library_id) REFERENCES libraries(id) ON DELETE CASCADE
 );
 CREATE INDEX IF NOT EXISTS idx_download_requests_created
    ON download_requests(created_at DESC);
 CREATE INDEX IF NOT EXISTS idx_download_requests_state
    ON download_requests(state);
 `
 const oldDownloadWantsDDL = `
 CREATE TABLE IF NOT EXISTS download_wants (
    id         INTEGER PRIMARY KEY AUTOINCREMENT,
    mbid       TEXT    NOT NULL,
    entity     TEXT    NOT NULL
        CHECK(entity IN ('artist', 'release-group', 'release', 'recording')),
    library_id INTEGER NOT NULL,
    artist     TEXT    NOT NULL DEFAULT '',
    title      TEXT    NOT NULL DEFAULT '',
    scope      TEXT    NOT NULL DEFAULT 'future'
        CHECK(scope IN ('future', 'all')),
    secondary  INTEGER NOT NULL DEFAULT 0,
    state      TEXT    NOT NULL DEFAULT 'wanted'
        CHECK(state IN ('wanted', 'satisfied', 'paused')),
    parent_id  INTEGER,
    attempts     INTEGER  NOT NULL DEFAULT 0,
    last_error   TEXT     NOT NULL DEFAULT '',
    last_tried_at DATETIME,
    next_try_at  DATETIME,
    external_ids TEXT    NOT NULL DEFAULT '{}',
    created_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
    updated_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
    UNIQUE(mbid, library_id),
    FOREIGN KEY(library_id) REFERENCES libraries(id) ON DELETE CASCADE,
    FOREIGN KEY(parent_id) REFERENCES download_wants(id) ON DELETE CASCADE
 );
 CREATE INDEX IF NOT EXISTS idx_download_wants_due
    ON download_wants(next_try_at)
    WHERE state = 'wanted';
 CREATE INDEX IF NOT EXISTS idx_download_wants_entity
    ON download_wants(entity, state);
 CREATE INDEX IF NOT EXISTS idx_download_wants_parent
    ON download_wants(parent_id);
 `
 const oldDownloadItemsDDL = `
 CREATE TABLE IF NOT EXISTS download_items (
    id           TEXT PRIMARY KEY,
    request_id   TEXT    NOT NULL,
    provider_id  INTEGER NOT NULL,
    transport_id INTEGER,
    external_id  TEXT    NOT NULL DEFAULT '',
    candidate    TEXT    NOT NULL DEFAULT '{}',
    state        TEXT    NOT NULL DEFAULT 'queued'
        CHECK(state IN ('searching', 'found', 'queued', 'grabbing',
                        'verifying', 'tagging', 'importing',
                        'complete', 'cancelled', 'failed')),
    staging_dir  TEXT    NOT NULL DEFAULT '',
    bytes_done   INTEGER NOT NULL DEFAULT 0,
    bytes_total  INTEGER NOT NULL DEFAULT 0,
    imported_paths TEXT  NOT NULL DEFAULT '[]',
    error        TEXT    NOT NULL DEFAULT '',
    created_at   DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
    updated_at   DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
    FOREIGN KEY(request_id) REFERENCES download_requests(id) ON DELETE CASCADE
 );
 CREATE INDEX IF NOT EXISTS idx_download_items_live
    ON download_items(state)
    WHERE state NOT IN ('complete', 'cancelled', 'failed');
 CREATE INDEX IF NOT EXISTS idx_download_items_request
    ON download_items(request_id);
 CREATE INDEX IF NOT EXISTS idx_download_items_state
    ON download_items(state);
 `
 // seedOldDownloadSchema builds the pre-rename download tables and
 // inserts one row of real data into each, standing in for a real
 // user's database at the moment it upgrades.
 func seedOldDownloadSchema(t *testing.T, db *sql.DB) {
 	t.Helper()
 	for _, ddl := range []string{
 		oldDownloadWantsDDL, oldDownloadRequestsDDL, oldDownloadItemsDDL,
 	} {
 		if _, err := db.ExecContext(t.Context(), ddl); err != nil {
 			t.Fatalf("create old download schema: %v", err)
 		}
 	}
 	if _, err := db.ExecContext(
 		t.Context(),
 		`INSERT INTO libraries (id, name, path) VALUES (1, 'Test', '/music')`,
 	); err != nil {
 		t.Fatalf("seed library: %v", err)
 	}
 	if _, err := db.ExecContext(
 		t.Context(),
 		`INSERT INTO download_wants
 		    (id, mbid, entity, library_id, artist, title, state)
 		 VALUES (1, 'artist-mbid', 'artist', 1, 'Radiohead', 'Radiohead', 'wanted')`,
 	); err != nil {
 		t.Fatalf("seed download_wants: %v", err)
 	}
 	if _, err := db.ExecContext(
 		t.Context(),
 		`INSERT INTO download_requests
 		    (id, library_id, source, want_id, release_group_mbid, artist, album, state)
 		 VALUES ('dl-1', 1, 'wanted', 1, 'rg-mbid', 'Radiohead', 'OK Computer', 'complete')`,
 	); err != nil {
 		t.Fatalf("seed download_requests: %v", err)
 	}
 	if _, err := db.ExecContext(
 		t.Context(),
 		`INSERT INTO download_items
 		    (id, request_id, provider_id, state)
 		 VALUES ('item-1', 'dl-1', 1, 'complete')`,
 	); err != nil {
 		t.Fatalf("seed download_items: %v", err)
 	}
 }
 // TestDownloadRename_FreshInstallUntouched confirms applySchema on a
 // brand-new database produces the target shape directly and that
 // migrateDownloadRename's gate (checking for the old download_wants
 // table) is a no-op there — the destructive path this test guards
 // against is exactly the one described in download_rename_migration.go:
 // blindly replaying the rename against a fresh database's already-
 // correct, empty download_requests/download_downloads tables.
 func TestDownloadRename_FreshInstallUntouched(t *testing.T) {
 	t.Parallel()
 	db := openMemDB(t)
 	if err := applySchema(t.Context(), db); err != nil {
 		t.Fatalf("apply schema (fresh): %v", err)
 	}
 	for _, table := range []string{"download_downloads", "download_requests", "download_items"} {
 		var name string
 		err := db.QueryRowContext(
 			t.Context(),
 			`SELECT name FROM sqlite_master WHERE type = 'table' AND name = ?`,
 			table,
 		).Scan(&name)
 		if err != nil {
 			t.Errorf("expected table %q to exist on a fresh install: %v", table, err)
 		}
 	}
 	var stray string
 	err := db.QueryRowContext(
 		t.Context(),
 		`SELECT name FROM sqlite_master WHERE type = 'table' AND name = 'download_wants'`,
 	).Scan(&stray)
 	if !errors.Is(err, sql.ErrNoRows) {
 		t.Errorf("old download_wants table should not exist on a fresh install, err=%v", err)
 	}
 	// Both auto-download guardrail indexes sql/schemas deliberately
 	// omits (see ensureDownloadIndexes) must still exist.
 	for _, idx := range []string{
 		"idx_download_requests_due",
 		"idx_download_requests_entity",
 		"idx_download_requests_parent",
 		"idx_download_items_download",
 	} {
 		var name string
 		err := db.QueryRowContext(
 			t.Context(),
 			`SELECT name FROM sqlite_master WHERE type = 'index' AND name = ?`,
 			idx,
 		).Scan(&name)
 		if err != nil {
 			t.Errorf("expected index %q to exist on a fresh install: %v", idx, err)
 		}
 	}
 }
 // TestDownloadRename_UpgradesExistingDatabase is the regression test
 // for the rename itself: an old-shaped database (download_wants +
 // old-style download_requests, both with real rows) must end up with
 // the same table names, column names, and data a fresh install would
 // have — nothing dropped, nothing silently emptied.
 func TestDownloadRename_UpgradesExistingDatabase(t *testing.T) {
 	t.Parallel()
 	fresh := openMemDB(t)
 	if err := applySchema(t.Context(), fresh); err != nil {
 		t.Fatalf("apply schema (fresh): %v", err)
 	}
 	upgraded := openMemDB(t)
 	librariesDDL, err := schemas.ReadFile("sql/schemas/libraries.sql")
 	if err != nil {
 		t.Fatalf("read libraries schema: %v", err)
 	}
 	if _, err := upgraded.ExecContext(t.Context(), string(librariesDDL)); err != nil {
 		t.Fatalf("create libraries table: %v", err)
 	}
 	seedOldDownloadSchema(t, upgraded)
 	if err := applySchema(t.Context(), upgraded); err != nil {
 		t.Fatalf("apply schema (upgrade path): %v", err)
 	}
 	// Column order must match a fresh install's, for the same reason
 	// TestMigrations_ColumnOrderMatchesFreshInstall checks tagging_items:
 	// sqlc's `SELECT *` binds positionally.
 	for _, table := range []string{"download_downloads", "download_requests", "download_items"} {
 		freshCols := tableColumns(t, fresh, table)
 		upgradedCols := tableColumns(t, upgraded, table)
 		if len(freshCols) != len(upgradedCols) {
 			t.Fatalf(
 				"%s: column count mismatch: fresh has %d (%v), upgraded has %d (%v)",
 				table, len(freshCols), freshCols, len(upgradedCols), upgradedCols,
 			)
 		}
 		for i := range freshCols {
 			if freshCols[i] != upgradedCols[i] {
 				t.Errorf(
 					"%s: column order mismatch at %d: fresh %q, upgraded %q\nfresh:    %v\nupgraded: %v",
 					table,
 					i,
 					freshCols[i],
 					upgradedCols[i],
 					freshCols,
 					upgradedCols,
 				)
 			}
 		}
 	}
 	// The seeded rows survived the rename under their new names.
 	var (
 		requestMBID   string
 		requestEntity string
 	)
 	err = upgraded.QueryRowContext(
 		t.Context(), `SELECT mbid, entity FROM download_requests WHERE id = 1`,
 	).Scan(&requestMBID, &requestEntity)
 	if err != nil {
 		t.Fatalf("seeded request row missing after rename: %v", err)
 	}
 	if requestMBID != "artist-mbid" || requestEntity != "artist" {
 		t.Errorf("request row corrupted: mbid=%q entity=%q", requestMBID, requestEntity)
 	}
 	var (
 		downloadRequestID sql.NullInt64
 		downloadAlbum     string
 	)
 	err = upgraded.QueryRowContext(
 		t.Context(),
 		`SELECT request_id, album FROM download_downloads WHERE id = 'dl-1'`,
 	).Scan(&downloadRequestID, &downloadAlbum)
 	if err != nil {
 		t.Fatalf("seeded download row missing after rename: %v", err)
 	}
 	if !downloadRequestID.Valid || downloadRequestID.Int64 != 1 {
 		t.Errorf("download.request_id = %v, want 1 (renamed from want_id)", downloadRequestID)
 	}
 	if downloadAlbum != "OK Computer" {
 		t.Errorf("download.album = %q, want OK Computer", downloadAlbum)
 	}
 	var itemDownloadID string
 	err = upgraded.QueryRowContext(
 		t.Context(),
 		`SELECT download_id FROM download_items WHERE id = 'item-1'`,
 	).Scan(&itemDownloadID)
 	if err != nil {
 		t.Fatalf("seeded item row missing after rename: %v", err)
 	}
 	if itemDownloadID != "dl-1" {
 		t.Errorf("item.download_id = %q, want dl-1 (renamed from request_id)", itemDownloadID)
 	}
 	// The old table is gone, not just emptied.
 	var stray string
 	err = upgraded.QueryRowContext(
 		t.Context(),
 		`SELECT name FROM sqlite_master WHERE type = 'table' AND name = 'download_wants'`,
 	).Scan(&stray)
 	if !errors.Is(err, sql.ErrNoRows) {
 		t.Errorf("old download_wants table should be gone after migration, err=%v", err)
 	}
 	// Running the whole thing again (as a second app startup would) is
 	// a no-op: the gate sees no download_wants table and does nothing
 	// further, so this must not error or duplicate anything.
 	if err := applySchema(t.Context(), upgraded); err != nil {
 		t.Fatalf("apply schema a second time: %v", err)
 	}
 	var count int
 	if err := upgraded.QueryRowContext(
 		t.Context(), `SELECT COUNT(*) FROM download_requests`,
 	).Scan(&count); err != nil {
 		t.Fatalf("count download_requests: %v", err)
 	}
 	if count != 1 {
 		t.Errorf("download_requests has %d rows after a second migration pass, want 1", count)
 	}
 }
@@ -1,18 +1,26 @@
 package database
 import (
 	"crypto/sha256"
 	"strings"
 	"testing"
 )
 // seedExploreRow inserts one explore_index row.
 //
 // The catalog stores an MBID as 16 raw bytes and an entity type as a
 // code (see backend/explore/mbid.go), and the column says so, so the
 // label these tests use as an id is hashed into something the table
 // will accept.  What they actually assert on is the FTS text.
 func seedExploreRow(t *testing.T, db *DB, mbid, title, artist string) {
 	t.Helper()
 	sum := sha256.Sum256([]byte(mbid))
 	if _, err := db.ExecContext(`
 		INSERT INTO explore_index (entity_type, mbid, title, artist_name, artist_mbid)
-		VALUES ('recording', ?, ?, ?, '')
+		VALUES (3 /* recording */, ?, ?, ?, x'')
-	`, mbid, title, artist); err != nil {
+	`, sum[:16], title, artist); err != nil {
 		t.Fatalf("seed %s: %v", mbid, err)
 	}
 }
@@ -202,7 +210,8 @@ func TestExploreFTSUpdateSkipsUnchangedText(t *testing.T) {
 	// A popularity refresh: an FTS column is not named at all.
 	if _, err := db.ExecContext(
-		"UPDATE explore_index SET popularity = 42 WHERE mbid = 'mbid-1'",
+		"UPDATE explore_index SET popularity = 42 WHERE title = ?",
 		"Unchanged Title",
 	); err != nil {
 		t.Fatalf("popularity update: %v", err)
 	}
@@ -212,8 +221,8 @@ func TestExploreFTSUpdateSkipsUnchangedText(t *testing.T) {
 	if _, err := db.ExecContext(`
 		UPDATE explore_index
 		SET title = 'Unchanged Title', artist_name = 'Steady Artist', popularity = 43
-		WHERE mbid = 'mbid-1'
+		WHERE title = ?
-	`); err != nil {
+	`, "Unchanged Title"); err != nil {
 		t.Fatalf("no-op text update: %v", err)
 	}
@@ -237,7 +246,8 @@ func TestExploreFTSUpdateReindexesChangedText(t *testing.T) {
 	seedExploreRow(t, db, "mbid-2", "Original Title", "Some Artist")
 	if _, err := db.ExecContext(
-		"UPDATE explore_index SET title = 'Corrected Title' WHERE mbid = 'mbid-2'",
+		"UPDATE explore_index SET title = 'Corrected Title' WHERE title = ?",
 		"Original Title",
 	); err != nil {
 		t.Fatalf("rename: %v", err)
 	}
@@ -252,7 +262,8 @@ func TestExploreFTSUpdateReindexesChangedText(t *testing.T) {
 	// The same for the other two indexed columns.
 	if _, err := db.ExecContext(
-		"UPDATE explore_index SET artist_name = 'Renamed Artist', aliases = 'AKA Thing' WHERE mbid = 'mbid-2'",
+		"UPDATE explore_index SET artist_name = 'Renamed Artist', aliases = 'AKA Thing' WHERE title = ?",
 		"Corrected Title",
 	); err != nil {
 		t.Fatalf("artist rename: %v", err)
 	}
@@ -1,15 +1,26 @@
 package database
 import (
 	"database/sql"
 	"errors"
 	"fmt"
 	"strings"
 	"unicode"
 )
 // toNullString treats an empty string as NULL.
 func toNullString(v string) sql.NullString {
 	if v == "" {
 		return sql.NullString{}
 	}
 	return sql.NullString{String: v, Valid: true}
 }
 // LyricsHit is a single result from a lyric-fragment search: the
-// matched recording plus enough metadata to render and play it.
+// matched file plus enough metadata to render and play it.
 type LyricsHit struct {
-	RecordingID        int64
+	AudioFileID        int64
 	FilePath           string
 	LengthMilliseconds int64
 	Title              string
@@ -37,27 +48,22 @@ func (d *DB) SearchLyrics(query string, limit int) ([]LyricsHit, error) {
 		return nil, nil
 	}
-	// Map the matched recording (lyrics_index.rowid == recordings.id)
+	// lyrics_index.rowid is the audio file's id, so the hit is already
-	// to a representative playable file via the lowest audio_files id,
+	// a playable file - it used to be a recording id, which then had to
-	// then to the track_metadata VIEW for display fields.
+	// be mapped back to "some file of that recording" by a grouped
 	// subquery.
 	//
 	// SAFETY: FTS5 MATCH syntax unsupported by sqlc. Query is parameterized; no string interpolation.
-	rows, err := d.db.QueryContext(d.Ctx, `
+	rows, err := d.reader().QueryContext(d.Ctx, `
 		SELECT
-			r.id,
+			tm.id,
 			tm.file_path,
 			tm.length_milliseconds,
 			tm.title,
 			tm.artist_name,
 			tm.album
 		FROM lyrics_index li
-		JOIN recordings r ON r.id = li.rowid
+		JOIN track_metadata tm ON tm.id = li.rowid
 		JOIN (
 			SELECT recording_id, MIN(id) AS af_id
 			FROM audio_files
 			GROUP BY recording_id
 		) af ON af.recording_id = r.id
 		JOIN track_metadata tm ON tm.id = af.af_id
 		WHERE lyrics_index MATCH ?
 		ORDER BY rank
 		LIMIT ?
@@ -73,7 +79,7 @@ func (d *DB) SearchLyrics(query string, limit int) ([]LyricsHit, error) {
 	for rows.Next() {
 		var h LyricsHit
 		if err := rows.Scan(
-			&h.RecordingID,
+			&h.AudioFileID,
 			&h.FilePath,
 			&h.LengthMilliseconds,
 			&h.Title,
@@ -93,43 +99,64 @@ func (d *DB) SearchLyrics(query string, limit int) ([]LyricsHit, error) {
 	return results, nil
 }
-// GetRecordingLyrics returns the stored lyrics for a recording, or
+// GetLyrics returns the stored lyrics for a file, or "" if none.
-// an empty string if none are stored.
+func (d *DB) GetLyrics(audioFileID int64) (string, error) {
 func (d *DB) GetRecordingLyrics(recordingID int64) (string, error) {
 	var lyrics string
-	err := d.db.QueryRowContext(d.Ctx,
+	err := d.reader().QueryRowContext(d.Ctx,
-		"SELECT COALESCE(lyrics, '') FROM recordings WHERE id = ?",
+		"SELECT text FROM lyrics WHERE audio_file_id = ?", audioFileID,
 		recordingID,
 	).Scan(&lyrics)
 	if errors.Is(err, sql.ErrNoRows) {
 		return "", nil
 	}
 	if err != nil {
-		return "", fmt.Errorf("could not read recording lyrics: %w", err)
+		return "", fmt.Errorf("could not read lyrics: %w", err)
 	}
 	return lyrics, nil
 }
-// SetRecordingLyrics writes lyrics onto a recording and keeps the FTS
+// SetLyrics writes lyrics for a file and keeps the FTS index in sync.
-// lyrics_index in sync (delete + reinsert the single row).  Used by
+//
-// the LRCLIB backfill to persist fetched lyrics.  Passing an empty
+// `source` says where they came from, which is the question the old
-// string clears both the column and the index entry.
+// column could not answer: lyrics read from a USLT frame are rebuilt
-func (d *DB) SetRecordingLyrics(recordingID int64, lyrics string) error {
+// free by any rescan, and lyrics fetched from LRCLIB are network
-	if _, err := d.db.ExecContext(d.Ctx,
+// traffic nobody wants to repeat.  Passing an empty string clears both
-		"UPDATE recordings SET lyrics = ? WHERE id = ?",
+// the row and the index entry.
-		lyrics, recordingID,
+func (d *DB) SetLyrics(audioFileID int64, lyrics, source, recordingMBID string) error {
-	); err != nil {
+	if strings.TrimSpace(lyrics) == "" {
-		return fmt.Errorf("could not update recording lyrics: %w", err)
+		if _, err := d.db.ExecContext(d.Ctx,
 			"DELETE FROM lyrics WHERE audio_file_id = ?", audioFileID,
 		); err != nil {
 			return fmt.Errorf("could not delete lyrics: %w", err)
 		}
 		return d.upsertLyricsIndex(audioFileID, "")
 	}
-	return d.upsertLyricsIndex(recordingID, lyrics)
+	if _, err := d.db.ExecContext(d.Ctx, `
 		INSERT INTO lyrics (audio_file_id, text, source, recording_mbid)
 		VALUES (?, ?, ?, ?)
 		ON CONFLICT(audio_file_id) DO UPDATE SET
 			text = excluded.text,
 			source = excluded.source,
 			recording_mbid = COALESCE(excluded.recording_mbid, lyrics.recording_mbid),
 			fetched_at = CURRENT_TIMESTAMP
 	`, audioFileID, lyrics, source, toNullString(recordingMBID)); err != nil {
 		return fmt.Errorf("could not write lyrics: %w", err)
 	}
 	return d.upsertLyricsIndex(audioFileID, lyrics)
 }
-// upsertLyricsIndex refreshes a single recording's entry in the
+// upsertLyricsIndex refreshes a single file's entry in the contentless
-// contentless lyrics_index.  contentless_delete=1 makes the DELETE
+// lyrics_index.  contentless_delete=1 makes the DELETE valid; an empty
-// valid; an empty lyrics string leaves the row deleted.
+// lyrics string leaves the row deleted.
-func (d *DB) upsertLyricsIndex(recordingID int64, lyrics string) error {
+func (d *DB) upsertLyricsIndex(audioFileID int64, lyrics string) error {
 	if _, err := d.db.ExecContext(d.Ctx,
-		"DELETE FROM lyrics_index WHERE rowid = ?", recordingID,
+		"DELETE FROM lyrics_index WHERE rowid = ?", audioFileID,
 	); err != nil {
 		return fmt.Errorf("could not delete lyrics_index row: %w", err)
 	}
@@ -141,7 +168,7 @@ func (d *DB) upsertLyricsIndex(recordingID int64, lyrics string) error {
 	// SAFETY: FTS5 virtual table INSERT unsupported by sqlc. All values parameterized.
 	if _, err := d.db.ExecContext(d.Ctx,
 		"INSERT INTO lyrics_index(rowid, lyrics) VALUES (?, ?)",
-		recordingID, lyrics,
+		audioFileID, lyrics,
 	); err != nil {
 		return fmt.Errorf("could not insert lyrics_index row: %w", err)
 	}
@@ -149,22 +176,16 @@ func (d *DB) upsertLyricsIndex(recordingID int64, lyrics string) error {
 	return nil
 }
-// RebuildLyricsIndex repopulates lyrics_index from scratch using the
+// RebuildLyricsIndex repopulates lyrics_index from the lyrics table.
 // current recordings table.  Cheap for a personal library and safe to
 // run after every scan.
 func (d *DB) RebuildLyricsIndex() error {
-	if _, err := d.db.ExecContext(d.Ctx,
+	if _, err := d.db.ExecContext(d.Ctx, "DELETE FROM lyrics_index"); err != nil {
 		"DELETE FROM lyrics_index",
 	); err != nil {
 		return fmt.Errorf("could not clear lyrics_index: %w", err)
 	}
-	// SAFETY: FTS5 virtual table INSERT unsupported by sqlc. Values sourced from recordings; no user input.
+	// SAFETY: FTS5 virtual table INSERT. Values sourced from lyrics; no user input.
 	if _, err := d.db.ExecContext(d.Ctx, `
 		INSERT INTO lyrics_index(rowid, lyrics)
-		SELECT id, lyrics
+		SELECT audio_file_id, text FROM lyrics WHERE text != ''
 		FROM recordings
 		WHERE lyrics IS NOT NULL AND lyrics != ''
 	`); err != nil {
 		return fmt.Errorf("could not rebuild lyrics_index: %w", err)
 	}
@@ -172,39 +193,35 @@ func (d *DB) RebuildLyricsIndex() error {
 	return nil
 }
-// RecordingsMissingLyrics returns recordings that have no stored
+// LyricsCandidate identifies a file that needs its lyrics fetched and
-// lyrics but do carry the artist/title/duration needed to look them
+// carries the fields an external provider matches on.
-// up from an external provider.  Used by the LRCLIB backfill.  The
+type LyricsCandidate struct {
-// limit bounds each batch so the backfill can be run incrementally.
+	AudioFileID        int64
-func (d *DB) RecordingsMissingLyrics(limit int) ([]LyricsCandidate, error) {
+	Title              string
 	Artist             string
 	Album              string
 	RecordingMBID      string
 	LengthMilliseconds int64
 }
 // FilesMissingLyrics returns files with no stored lyrics that carry
 // the artist/title/duration needed to look them up.  Used by the
 // LRCLIB backfill; the limit bounds each batch.
 func (d *DB) FilesMissingLyrics(limit int) ([]LyricsCandidate, error) {
 	if limit <= 0 {
 		limit = 200
 	}
-	rows, err := d.db.QueryContext(d.Ctx, `
+	rows, err := d.reader().QueryContext(d.Ctx, `
-		SELECT
+		SELECT tm.id, tm.title, tm.artist_name, tm.album,
-			r.id,
+		       tm.recording_mbid, tm.length_milliseconds
-			COALESCE(r.name, ''),
+		FROM track_metadata tm
-			COALESCE(ac.text, ''),
+		WHERE NOT EXISTS (SELECT 1 FROM lyrics l WHERE l.audio_file_id = tm.id)
-			COALESCE(rg.name, ''),
+		  AND tm.title != '' AND tm.artist_name != ''
 			MIN(af.length_milliseconds)
 		FROM recordings r
 		JOIN audio_files af ON af.recording_id = r.id
 		LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 		LEFT JOIN (
 			SELECT recording_id, MIN(release_group_id) AS release_group_id
 			FROM release_group_recordings
 			GROUP BY recording_id
 		) rgr ON rgr.recording_id = r.id
 		LEFT JOIN release_groups rg ON rg.id = rgr.release_group_id
 		WHERE (r.lyrics IS NULL OR r.lyrics = '')
 		  AND r.name IS NOT NULL AND r.name != ''
 		  AND ac.text IS NOT NULL AND ac.text != ''
 		GROUP BY r.id
 		LIMIT ?
 	`, limit)
 	if err != nil {
-		return nil, fmt.Errorf("could not query recordings missing lyrics: %w", err)
+		return nil, fmt.Errorf("could not query files missing lyrics: %w", err)
 	}
 	defer func() { _ = rows.Close() }()
@@ -214,7 +231,8 @@ func (d *DB) RecordingsMissingLyrics(limit int) ([]LyricsCandidate, error) {
 	for rows.Next() {
 		var c LyricsCandidate
 		if err := rows.Scan(
-			&c.RecordingID, &c.Title, &c.Artist, &c.Album, &c.LengthMilliseconds,
+			&c.AudioFileID, &c.Title, &c.Artist, &c.Album,
 			&c.RecordingMBID, &c.LengthMilliseconds,
 		); err != nil {
 			return nil, fmt.Errorf("could not scan lyrics candidate: %w", err)
 		}
@@ -229,44 +247,23 @@ func (d *DB) RecordingsMissingLyrics(limit int) ([]LyricsCandidate, error) {
 	return out, nil
 }
-// LyricsCandidate identifies a recording that needs its lyrics fetched
+// FileLyricLookup returns the provider-match fields for one file, so
-// and carries the fields an external provider matches on.
+// lyrics can be fetched on demand.  Returns nil if the file has no
-type LyricsCandidate struct {
+// artist/title to match on.
-	RecordingID        int64
+func (d *DB) FileLyricLookup(audioFileID int64) (*LyricsCandidate, error) {
 	Title              string
 	Artist             string
 	Album              string
 	LengthMilliseconds int64
 }
 // RecordingLyricLookup returns the provider-match fields (artist,
 // title, album, duration) for a single recording, so lyrics can be
 // fetched on demand.  Returns nil if the recording has no audio file
 // or no artist/title to match on.
 func (d *DB) RecordingLyricLookup(recordingID int64) (*LyricsCandidate, error) {
 	var c LyricsCandidate
-	err := d.db.QueryRowContext(d.Ctx, `
+	err := d.reader().QueryRowContext(d.Ctx, `
-		SELECT
+		SELECT tm.id, tm.title, tm.artist_name, tm.album,
-			r.id,
+		       tm.recording_mbid, tm.length_milliseconds
-			COALESCE(r.name, ''),
+		FROM track_metadata tm
-			COALESCE(ac.text, ''),
+		WHERE tm.id = ?
-			COALESCE(rg.name, ''),
+	`, audioFileID).Scan(
-			COALESCE(MIN(af.length_milliseconds), 0)
+		&c.AudioFileID, &c.Title, &c.Artist, &c.Album,
-		FROM recordings r
+		&c.RecordingMBID, &c.LengthMilliseconds,
-		JOIN audio_files af ON af.recording_id = r.id
+	)
 		LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 		LEFT JOIN (
 			SELECT recording_id, MIN(release_group_id) AS release_group_id
 			FROM release_group_recordings
 			GROUP BY recording_id
 		) rgr ON rgr.recording_id = r.id
 		LEFT JOIN release_groups rg ON rg.id = rgr.release_group_id
 		WHERE r.id = ?
 		GROUP BY r.id
 	`, recordingID).Scan(&c.RecordingID, &c.Title, &c.Artist, &c.Album, &c.LengthMilliseconds)
 	if err != nil {
-		return nil, fmt.Errorf("could not look up recording for lyrics: %w", err)
+		return nil, fmt.Errorf("could not look up file for lyrics: %w", err)
 	}
 	if c.Title == "" || c.Artist == "" {
@@ -4,59 +4,33 @@ import (
 	"testing"
 )
-// seedLyricsTrack inserts the minimal FK chain (artist_credit →
+// seedLyricsTrack inserts one file with the given lyrics, so lyric
-// recording → audio_file → release_group link) for one track with the
+// searches have something realistic to join against.  It used to
-// given lyrics, so lyric-search tests have realistic joins.
+// insert a four-row FK chain by hand.
 func seedLyricsTrack(
 	t *testing.T,
 	db *DB,
 	id int64,
 	title, artist, album, lyrics string,
 	lenMs int64,
-) {
+) int64 {
 	t.Helper()
-	if _, err := db.ExecContext(
+	fileID := InsertTestTrack(t, db, TestTrack{
-		"INSERT OR IGNORE INTO artist_credit (id, text) VALUES (?, ?)", id, artist,
+		FilePath: "/music/track" + itoa(id) + ".mp3",
-	); err != nil {
+		Title:    title,
-		t.Fatalf("insert artist_credit: %v", err)
+		Artist:   artist,
 		Album:    album,
 		LengthMs: lenMs,
 	})
 	if lyrics != "" {
 		if err := db.SetLyrics(fileID, lyrics, "tag", ""); err != nil {
 			t.Fatalf("seed lyrics: %v", err)
 		}
 	}
-	if _, err := db.ExecContext(
+	return fileID
 		"INSERT OR IGNORE INTO release_groups (id, name) VALUES (?, ?)", id, album,
 	); err != nil {
 		t.Fatalf("insert release_group: %v", err)
 	}
 	if _, err := db.ExecContext(
 		"INSERT INTO recordings (id, name, artist_credit_id, lyrics) VALUES (?, ?, ?, ?)",
 		id, title, id, nullableLyrics(lyrics),
 	); err != nil {
 		t.Fatalf("insert recording: %v", err)
 	}
 	if _, err := db.ExecContext(
 		"INSERT INTO audio_files (id, file_path, length_milliseconds, file_type_id, recording_id) "+
 			"VALUES (?, ?, ?, ?, ?)",
 		id, "/music/track"+itoa(id)+".mp3", lenMs, 0, id,
 	); err != nil {
 		t.Fatalf("insert audio_file: %v", err)
 	}
 	if _, err := db.ExecContext(
 		"INSERT INTO release_group_recordings (release_group_id, recording_id) VALUES (?, ?)",
 		id, id,
 	); err != nil {
 		t.Fatalf("insert release_group_recordings: %v", err)
 	}
 }
 func nullableLyrics(l string) any {
 	if l == "" {
 		return nil
 	}
 	return l
 }
 func itoa(v int64) string {
@@ -111,8 +85,8 @@ func TestSearchLyrics(t *testing.T) {
 		}
 		h := hits[0]
-		if h.RecordingID != 1 {
+		if h.AudioFileID != 1 {
-			t.Errorf("RecordingID = %d, want 1", h.RecordingID)
+			t.Errorf("RecordingID = %d, want 1", h.AudioFileID)
 		}
 		if h.Title != "The Sound of Silence" {
@@ -191,11 +165,11 @@ func TestSetRecordingLyricsUpdatesIndex(t *testing.T) {
 	// Backfill lyrics — should update both the column and the FTS index.
 	const lyrics = "Yesterday all my troubles seemed so far away"
-	if err := db.SetRecordingLyrics(1, lyrics); err != nil {
+	if err := db.SetLyrics(1, lyrics, "lrclib", ""); err != nil {
 		t.Fatalf("SetRecordingLyrics: %v", err)
 	}
-	stored, err := db.GetRecordingLyrics(1)
+	stored, err := db.GetLyrics(1)
 	if err != nil {
 		t.Fatalf("GetRecordingLyrics: %v", err)
 	}
@@ -209,7 +183,7 @@ func TestSetRecordingLyricsUpdatesIndex(t *testing.T) {
 		t.Fatalf("SearchLyrics: %v", err)
 	}
-	if len(hits) != 1 || hits[0].RecordingID != 1 {
+	if len(hits) != 1 || hits[0].AudioFileID != 1 {
 		t.Fatalf("expected recording 1 after backfill, got %+v", hits)
 	}
 }
@@ -222,7 +196,7 @@ func TestRecordingsMissingLyrics(t *testing.T) {
 	seedLyricsTrack(t, db, 1, "Has Lyrics", "Artist A", "Album A", "some words here", 100000)
 	seedLyricsTrack(t, db, 2, "No Lyrics", "Artist B", "Album B", "", 200000)
-	missing, err := db.RecordingsMissingLyrics(50)
+	missing, err := db.FilesMissingLyrics(50)
 	if err != nil {
 		t.Fatalf("RecordingsMissingLyrics: %v", err)
 	}
@@ -232,7 +206,7 @@ func TestRecordingsMissingLyrics(t *testing.T) {
 	}
 	c := missing[0]
-	if c.RecordingID != 2 || c.Title != "No Lyrics" || c.Artist != "Artist B" {
+	if c.AudioFileID != 2 || c.Title != "No Lyrics" || c.Artist != "Artist B" {
 		t.Errorf("unexpected candidate: %+v", c)
 	}
@@ -241,7 +215,7 @@ func TestRecordingsMissingLyrics(t *testing.T) {
 	}
 	// Single-recording lookup mirrors the batch fields.
-	one, err := db.RecordingLyricLookup(2)
+	one, err := db.FileLyricLookup(2)
 	if err != nil {
 		t.Fatalf("RecordingLyricLookup: %v", err)
 	}
@@ -1,196 +0,0 @@
 package database
 import (
 	"database/sql"
 	"testing"
 )
 // oldTaggingItemsDDL is a frozen snapshot of tagging_items exactly as
 // it read before sql/migrations/0001_tagging_items_synthetic.sql —
 // i.e. what a real user's existing database looks like today, before
 // upgrading to a build that includes that migration.
 const oldTaggingItemsDDL = `
 CREATE TABLE IF NOT EXISTS tagging_items (
  group_key               TEXT PRIMARY KEY,
  library_id              INTEGER NOT NULL,
  track_count             INTEGER NOT NULL DEFAULT 0,
  album_name              TEXT NOT NULL DEFAULT '',
  album_artist            TEXT NOT NULL DEFAULT '',
  disc_number              INTEGER NOT NULL DEFAULT 0,
  best_match_release_mbid TEXT,
  score                   REAL,
  last_checked_at         DATETIME,
  status                  TEXT NOT NULL DEFAULT 'pending'
    CHECK(status IN ('pending', 'matched', 'confirmed', 'skipped')),
  cleared_at              DATETIME,
  created_at              DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
  FOREIGN KEY(library_id) REFERENCES libraries(id)
 );
 CREATE INDEX IF NOT EXISTS idx_tagging_items_library_status
    ON tagging_items(library_id, status);
 CREATE INDEX IF NOT EXISTS idx_tagging_items_status_pending
    ON tagging_items(library_id) WHERE status = 'pending';
 `
 // tableColumns returns the column names of a table in on-disk
 // (positional) order, via PRAGMA table_info — the order sqlc's
 // generated `SELECT *` scans bind to positionally.
 func tableColumns(t *testing.T, db *sql.DB, table string) []string {
 	t.Helper()
 	rows, err := db.QueryContext(t.Context(), "PRAGMA table_info("+table+")")
 	if err != nil {
 		t.Fatalf("PRAGMA table_info(%s): %v", table, err)
 	}
 	defer func() { _ = rows.Close() }()
 	var cols []string
 	for rows.Next() {
 		var (
 			cid        int
 			name       string
 			ctype      string
 			notnull    int
 			dfltValue  sql.NullString
 			primaryKey int
 		)
 		if err := rows.Scan(&cid, &name, &ctype, &notnull, &dfltValue, &primaryKey); err != nil {
 			t.Fatalf("scan table_info row: %v", err)
 		}
 		cols = append(cols, name)
 	}
 	if err := rows.Err(); err != nil {
 		t.Fatalf("iterate table_info: %v", err)
 	}
 	return cols
 }
 func openMemDB(t *testing.T) *sql.DB {
 	t.Helper()
 	db, err := sql.Open("sqlite", ":memory:?_busy_timeout=5000&_journal_mode=WAL")
 	if err != nil {
 		t.Fatalf("open in-memory db: %v", err)
 	}
 	db.SetMaxOpenConns(1)
 	t.Cleanup(func() { _ = db.Close() })
 	if err := applyPRAGMAs(t.Context(), db); err != nil {
 		t.Fatalf("apply pragmas: %v", err)
 	}
 	return db
 }
 // TestMigrations_ColumnOrderMatchesFreshInstall is the regression
 // test for the exact failure mode that got the old 48-step migration
 // chain torn out (see .planning/NOTES.md, "No migration chain"):
 // sql/schemas drifting from what migrations actually produce, so
 // sqlc-generated code silently reads the wrong thing.
 //
 // A fresh install takes tagging_items straight from sql/schemas
 // (CREATE TABLE, columns in file order). An existing database takes
 // it from sql/schemas (the base shape, unchanged since the table
 // already existed) plus sql/migrations/0001 (`ALTER TABLE ADD
 // COLUMN`, which SQLite always appends at the END of the column
 // list, regardless of where the column sits in the CREATE TABLE
 // statement). If sql/schemas ever declares a migrated column
 // somewhere other than last, the two paths produce tables with the
 // SAME columns in a DIFFERENT order — invisible until a `SELECT *`
 // (e.g. GetTaggingItem) silently binds a value to the wrong field.
 func TestMigrations_ColumnOrderMatchesFreshInstall(t *testing.T) {
 	t.Parallel()
 	fresh := openMemDB(t)
 	if err := applySchema(t.Context(), fresh); err != nil {
 		t.Fatalf("apply schema (fresh): %v", err)
 	}
 	upgraded := openMemDB(t)
 	librariesDDL, err := schemas.ReadFile("sql/schemas/libraries.sql")
 	if err != nil {
 		t.Fatalf("read libraries schema: %v", err)
 	}
 	if _, err := upgraded.ExecContext(t.Context(), string(librariesDDL)); err != nil {
 		t.Fatalf("create libraries table: %v", err)
 	}
 	if _, err := upgraded.ExecContext(t.Context(), oldTaggingItemsDDL); err != nil {
 		t.Fatalf("create pre-migration tagging_items: %v", err)
 	}
 	// sql/schemas no-ops on the pre-existing tagging_items (IF NOT
 	// EXISTS), then sql/migrations/0001's ALTER TABLE statements
 	// actually add the missing columns for real this time.
 	if err := applySchema(t.Context(), upgraded); err != nil {
 		t.Fatalf("apply schema (upgrade path): %v", err)
 	}
 	freshCols := tableColumns(t, fresh, "tagging_items")
 	upgradedCols := tableColumns(t, upgraded, "tagging_items")
 	if len(freshCols) != len(upgradedCols) {
 		t.Fatalf(
 			"column count mismatch: fresh install has %d (%v), upgraded has %d (%v)",
 			len(freshCols), freshCols, len(upgradedCols), upgradedCols,
 		)
 	}
 	for i := range freshCols {
 		if freshCols[i] != upgradedCols[i] {
 			t.Errorf(
 				"column order mismatch at position %d: fresh install has %q, upgraded has %q\nfresh:    %v\nupgraded: %v",
 				i,
 				freshCols[i],
 				upgradedCols[i],
 				freshCols,
 				upgradedCols,
 			)
 		}
 	}
 }
 // TestMigrations_FreshDatabaseStillRecordsAndGetsIndex confirms a
 // brand-new database runs migration 0001 (tolerating "duplicate
 // column name" from its ALTER TABLE statements, since sql/schemas
 // already declared those columns), records it applied, AND still
 // gets the trailing CREATE INDEX statement sql/schemas deliberately
 // omits for migrated columns.
 func TestMigrations_FreshDatabaseStillRecordsAndGetsIndex(t *testing.T) {
 	t.Parallel()
 	fresh := openMemDB(t)
 	if err := applySchema(t.Context(), fresh); err != nil {
 		t.Fatalf("apply schema: %v", err)
 	}
 	var version int
 	err := fresh.QueryRowContext(
 		t.Context(), "SELECT version FROM schema_migrations WHERE version = 1",
 	).Scan(&version)
 	if err != nil {
 		t.Fatalf("expected migration 1 to be recorded as applied on a fresh db: %v", err)
 	}
 	var indexName string
 	err = fresh.QueryRowContext(
 		t.Context(),
 		"SELECT name FROM sqlite_master WHERE type = 'index' AND name = 'idx_tagging_items_parent_group_key'",
 	).Scan(&indexName)
 	if err != nil {
 		t.Fatalf("expected idx_tagging_items_parent_group_key to exist on a fresh db: %v", err)
 	}
 }
@@ -0,0 +1,86 @@
 package database
 import (
 	"testing"
 )
 // TestOneRowPerTrackForAMultiArtistCredit pins what is left of the
 // multi-artist problem, which is now much smaller than it was.
 //
 // It used to be possible for one file to produce several rows: an
 // artist credit was a row in its own table linking *many* artists, so
 // any query that joined artist_credit_artist to read the artist MBID
 // returned the same track once per credited artist.  The playlist, the
 // queue, the library list and the phantom resolver all did, and all
 // showed collaborations twice.  Nine queries carried a
 // first-credited-artist subquery to work around it.
 //
 // The join is gone: a file carries its credit as text and points at one
 // primary artist, so the fan-out has nothing to fan out from.  What is
 // still worth pinning is that the credit text survives intact - a
 // collaboration must still *read* as one - and that the file resolves
 // to exactly one row wherever it is asked for.
 func TestOneRowPerTrackForAMultiArtistCredit(t *testing.T) {
 	t.Parallel()
 	db := NewTestDB(t)
 	id := InsertTestTrack(t, db, TestTrack{
 		FilePath:   "/lib/collab.mp3",
 		Title:      "Collab Song",
 		Artist:     "A feat. B",
 		ArtistMBID: "mbid-a",
 		Album:      "An Album",
 		LengthMs:   200000,
 	})
 	t.Run("one row in the view", func(t *testing.T) {
 		var n int
 		if err := db.QueryRowWriter(
 			`SELECT COUNT(*) FROM track_metadata WHERE id = ?`, id,
 		).Scan(&n); err != nil {
 			t.Fatalf("count: %v", err)
 		}
 		if n != 1 {
 			t.Errorf("track_metadata rows = %d, want 1", n)
 		}
 	})
 	t.Run("the credit is preserved and the artist resolved", func(t *testing.T) {
 		rows, err := db.Queries.GetTracks(db.Ctx, 0)
 		if err != nil {
 			t.Fatalf("get tracks: %v", err)
 		}
 		if len(rows) != 1 {
 			t.Fatalf("tracks = %d, want 1", len(rows))
 		}
 		if rows[0].ArtistName != "A feat. B" {
 			t.Errorf("artist credit = %q, want %q", rows[0].ArtistName, "A feat. B")
 		}
 		if rows[0].ArtistMbid != "mbid-a" {
 			t.Errorf("artist mbid = %q, want %q", rows[0].ArtistMbid, "mbid-a")
 		}
 	})
 	t.Run("one row per album track", func(t *testing.T) {
 		var albumID int64
 		if err := db.QueryRowWriter(
 			`SELECT album_id FROM audio_files WHERE id = ?`, id,
 		).Scan(&albumID); err != nil {
 			t.Fatalf("album id: %v", err)
 		}
 		rows, err := db.Queries.GetTracks(db.Ctx, 0)
 		if err != nil {
 			t.Fatalf("album tracks: %v", err)
 		}
 		if len(rows) != 1 {
 			t.Errorf("album tracks = %d, want 1", len(rows))
 		}
 	})
 }
@@ -5,6 +5,8 @@ import (
 	"database/sql"
 	"fmt"
 	"strings"
 	"yellowjacket/backend/database/sql/sqlcgen"
 )
 // SearchRow holds a single result from an FTS5 or basename search.
@@ -183,203 +185,80 @@ func (d *DB) RebuildSearchIndex() error {
 	return nil
 }
-// SearchTrackRow holds a full track result from an FTS5 search,
+// trackMetadataColumns is the column list of the track_metadata view,
-// matching all 16 columns returned by GetAllTracksWithFullMetadata.
+// in the order sqlc generates TrackMetadatum's fields.  The FTS
-type SearchTrackRow struct {
+// searches below cannot be sqlc queries (MATCH is not in its grammar),
-	FilePath           string
+// so this is the one place the view's shape is written out by hand.
-	LengthMilliseconds int64
+const trackMetadataColumns = `
-	Title              string
+	tm.id, tm.file_path, tm.length_milliseconds, tm.title, tm.artist_name,
-	ArtistName         string
+	tm.track_number, tm.disc_number, tm.album, tm.genre, tm.year,
-	TrackNumber        sql.NullInt64
+	tm.release_year, tm.composer, tm.file_type, tm.sample_rate,
-	DiscNumber         sql.NullInt64
+	tm.bit_depth, tm.channels, tm.bitrate, tm.file_size, tm.library_id,
-	Album              string
+	tm.play_count, tm.last_played, tm.cover_art_path, tm.artist_mbid,
-	Genre              string
+	tm.release_group_mbid, tm.recording_mbid, tm.album_id, tm.artist_id`
-	Year               int64
+
-	Composer           string
+// scanTrackMetadata reads track_metadata rows into the generated row
-	FileType           string
+// type, so an FTS hit and an ordinary query produce the same Track.
-	SampleRate         int64
+func scanTrackMetadata(rows *sql.Rows) ([]sqlcgen.TrackMetadatum, error) {
-	BitDepth           int64
+	var out []sqlcgen.TrackMetadatum
-	Channels           int64
+
-	Bitrate            int64
+	for rows.Next() {
-	FileSize           int64
+		var r sqlcgen.TrackMetadatum
 		if err := rows.Scan(
 			&r.ID, &r.FilePath, &r.LengthMilliseconds, &r.Title, &r.ArtistName,
 			&r.TrackNumber, &r.DiscNumber, &r.Album, &r.Genre, &r.Year,
 			&r.ReleaseYear, &r.Composer, &r.FileType, &r.SampleRate,
 			&r.BitDepth, &r.Channels, &r.Bitrate, &r.FileSize, &r.LibraryID,
 			&r.PlayCount, &r.LastPlayed, &r.CoverArtPath, &r.ArtistMbid,
 			&r.ReleaseGroupMbid, &r.RecordingMbid, &r.AlbumID, &r.ArtistID,
 		); err != nil {
 			return nil, fmt.Errorf("scan track metadata: %w", err)
 		}
 		out = append(out, r)
 	}
 	if err := rows.Err(); err != nil {
 		return nil, fmt.Errorf("iterate track metadata: %w", err)
 	}
 	return out, nil
 }
-// SearchFTSTracks performs a full-text search and returns full track
+// SearchFTSTracks performs a full-text search and returns whole tracks.
-// metadata for each match.  Unlike SearchFTS (which returns only 5
+//
-// columns), this includes all 16 fields needed for library.Track.
+// A library id of 0 means every library.  There were two of these, one
 // per case, each with its own copy of a sixteen-column projection that
 // silently dropped the MBIDs and the play count - which is why the
 // caller used to pass zeros for them.
 func (d *DB) SearchFTSTracks(
-	query string, limit int,
+	query string, libraryID int64, limit int,
-) ([]SearchTrackRow, error) {
+) ([]sqlcgen.TrackMetadatum, error) {
 	query = strings.TrimSpace(query)
 	if query == "" {
 		return nil, nil
 	}
 	ftsQuery := buildFTSQuery(query)
 	// SAFETY: FTS5 MATCH syntax unsupported by sqlc. Query is parameterized; no string interpolation.
-	rows, err := d.db.QueryContext(d.Ctx, `
+	rows, err := d.reader().QueryContext(d.Ctx, `
-		SELECT
+		SELECT`+trackMetadataColumns+`
 			tm.file_path,
 			tm.length_milliseconds,
 			tm.title,
 			tm.artist_name,
 			tm.track_number,
 			tm.disc_number,
 			tm.album,
 			tm.genre,
 			tm.year,
 			tm.composer,
 			tm.file_type,
 			tm.sample_rate,
 			tm.bit_depth,
 			tm.channels,
 			tm.bitrate,
 			tm.file_size
 		FROM search_index si
 		JOIN track_metadata tm ON tm.id = si.rowid
 		WHERE search_index MATCH ?
 		  AND (? = 0 OR tm.library_id = ?)
 		ORDER BY rank
 		LIMIT ?
-	`, ftsQuery, limit)
+	`, buildFTSQuery(query), libraryID, libraryID, limit)
 	if err != nil {
-		return nil, fmt.Errorf(
+		return nil, fmt.Errorf("FTS track search failed: %w", err)
 			"FTS track search failed: %w", err,
 		)
 	}
 	defer func() { _ = rows.Close() }()
-	var results []SearchTrackRow
+	return scanTrackMetadata(rows)
 	for rows.Next() {
 		var r SearchTrackRow
 		if err := rows.Scan(
 			&r.FilePath,
 			&r.LengthMilliseconds,
 			&r.Title,
 			&r.ArtistName,
 			&r.TrackNumber,
 			&r.DiscNumber,
 			&r.Album,
 			&r.Genre,
 			&r.Year,
 			&r.Composer,
 			&r.FileType,
 			&r.SampleRate,
 			&r.BitDepth,
 			&r.Channels,
 			&r.Bitrate,
 			&r.FileSize,
 		); err != nil {
 			return nil, fmt.Errorf(
 				"could not scan search track row: %w",
 				err,
 			)
 		}
 		results = append(results, r)
 	}
 	if err := rows.Err(); err != nil {
 		return nil, fmt.Errorf(
 			"search track row iteration error: %w",
 			err,
 		)
 	}
 	return results, nil
 }
 // SearchFTSTracksByLibrary performs a full-text search scoped to a
 // specific library and returns full track metadata for each match.
 func (d *DB) SearchFTSTracksByLibrary(
 	query string, limit int, libraryID int64,
 ) ([]SearchTrackRow, error) {
 	query = strings.TrimSpace(query)
 	if query == "" {
 		return nil, nil
 	}
 	ftsQuery := buildFTSQuery(query)
 	// SAFETY: FTS5 MATCH syntax unsupported by sqlc. Query is parameterized; no string interpolation.
 	rows, err := d.db.QueryContext(d.Ctx, `
 		SELECT
 			tm.file_path,
 			tm.length_milliseconds,
 			tm.title,
 			tm.artist_name,
 			tm.track_number,
 			tm.disc_number,
 			tm.album,
 			tm.genre,
 			tm.year,
 			tm.composer,
 			tm.file_type,
 			tm.sample_rate,
 			tm.bit_depth,
 			tm.channels,
 			tm.bitrate,
 			tm.file_size
 		FROM search_index si
 		JOIN track_metadata tm ON tm.id = si.rowid
 		WHERE search_index MATCH ? AND tm.library_id = ?
 		ORDER BY rank
 		LIMIT ?
 	`, ftsQuery, libraryID, limit)
 	if err != nil {
 		return nil, fmt.Errorf(
 			"FTS library track search failed: %w", err,
 		)
 	}
 	defer func() { _ = rows.Close() }()
 	var results []SearchTrackRow
 	for rows.Next() {
 		var r SearchTrackRow
 		if err := rows.Scan(
 			&r.FilePath,
 			&r.LengthMilliseconds,
 			&r.Title,
 			&r.ArtistName,
 			&r.TrackNumber,
 			&r.DiscNumber,
 			&r.Album,
 			&r.Genre,
 			&r.Year,
 			&r.Composer,
 			&r.FileType,
 			&r.SampleRate,
 			&r.BitDepth,
 			&r.Channels,
 			&r.Bitrate,
 			&r.FileSize,
 		); err != nil {
 			return nil, fmt.Errorf(
 				"could not scan library search track row: %w",
 				err,
 			)
 		}
 		results = append(results, r)
 	}
 	if err := rows.Err(); err != nil {
 		return nil, fmt.Errorf(
 			"library search track row iteration error: %w",
 			err,
 		)
 	}
 	return results, nil
 }
 // scanSearchRows reads all rows from a query result into a slice.
 func scanSearchRows(
 	rows interface {
 		Next() bool
@@ -3,6 +3,8 @@ package database
 import (
 	"fmt"
 	"testing"
 	"yellowjacket/backend/database/sql/sqlcgen"
 )
 // seedSearchData inserts ~7 tracks with the full FK chain required for
@@ -84,128 +86,44 @@ func seedSearchData(t *testing.T, db *DB) {
 		},
 	}
 	// Build unique sets.
 	artistMap := map[string]int64{}
 	albumMap := map[string]int64{}
 	var artistID, albumID int64
 	for _, tr := range tracks {
-		if _, ok := artistMap[tr.artist]; !ok {
+		var genres []string
 			artistID++
 			artistMap[tr.artist] = artistID
 		}
 		if _, ok := albumMap[tr.album]; !ok {
 			albumID++
 			albumMap[tr.album] = albumID
 		}
 	}
 	// Insert artist_credit rows.
 	for text, id := range artistMap {
 		_, err := db.ExecContext(
 			"INSERT INTO artist_credit (id, text) VALUES (?, ?)",
 			id, text,
 		)
 		if err != nil {
 			t.Fatalf("insert artist_credit %q: %v", text, err)
 		}
 	}
 	// Insert release_groups.
 	for name, id := range albumMap {
 		_, err := db.ExecContext(
 			"INSERT INTO release_groups (id, name) VALUES (?, ?)",
 			id, name,
 		)
 		if err != nil {
 			t.Fatalf("insert release_group %q: %v", name, err)
 		}
 	}
 	// Insert genres + recording_genres.
 	genreMap := map[string]int64{}
 	var genreID int64
 	for _, tr := range tracks {
 		if tr.genre == "" {
 			continue
 		}
 		if _, ok := genreMap[tr.genre]; !ok {
 			genreID++
 			genreMap[tr.genre] = genreID
 			_, err := db.ExecContext(
 				"INSERT INTO genres (id, name) VALUES (?, ?)",
 				genreID, tr.genre,
 			)
 			if err != nil {
 				t.Fatalf("insert genre %q: %v", tr.genre, err)
 			}
 		}
 	}
 	for _, tr := range tracks {
 		acID := artistMap[tr.artist]
 		rgID := albumMap[tr.album]
 		// Insert recording.
 		_, err := db.ExecContext(
 			"INSERT INTO recordings (id, name, artist_credit_id, "+
 				"track_number, disc_number, year, genre, composer) "+
 				"VALUES (?, ?, ?, ?, ?, ?, ?, ?)",
 			tr.id, tr.title, acID, tr.trackNum, tr.discNum,
 			tr.year, tr.genre, tr.composer,
 		)
 		if err != nil {
 			t.Fatalf("insert recording %d %q: %v", tr.id, tr.title, err)
 		}
 		// Insert audio_files.
 		_, err = db.ExecContext(
 			"INSERT INTO audio_files (id, file_path, "+
 				"length_milliseconds, file_type_id, recording_id, "+
 				"sample_rate, bit_depth, channels, bitrate, file_size) "+
 				"VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?)",
 			tr.id, tr.filePath, tr.lenMs, tr.ftID, tr.id,
 			tr.sr, tr.bd, tr.ch, tr.br, tr.fsize,
 		)
 		if err != nil {
 			t.Fatalf("insert audio_file %d: %v", tr.id, err)
 		}
 		// Link recording to release_group.
 		_, err = db.ExecContext(
 			"INSERT INTO release_group_recordings "+
 				"(release_group_id, recording_id, track_number, disc_number) "+
 				"VALUES (?, ?, ?, ?)",
 			rgID, tr.id, tr.trackNum, tr.discNum,
 		)
 		if err != nil {
 			t.Fatalf("insert release_group_recordings %d→%d: %v", rgID, tr.id, err)
 		}
 		// Insert search_index entry (rowid must match audio_files.id).
 		if err := db.InsertSearchIndex(
 			tr.id, tr.filePath, tr.title, tr.artist, tr.album,
 		); err != nil {
 			t.Fatalf("insert search_index for %d: %v", tr.id, err)
 		}
 		// Insert recording_genres link.
 		if tr.genre != "" {
-			gID := genreMap[tr.genre]
+			genres = []string{tr.genre}
 		}
-			_, err = db.ExecContext(
+		var trackNum, discNum int64
-				"INSERT INTO recording_genres (recording_id, genre_id) VALUES (?, ?)",
+		if tr.trackNum != nil {
-				tr.id, gID,
+			trackNum = *tr.trackNum
-			)
+		}
-			if err != nil {
+
-				t.Fatalf("insert recording_genres %d→%d: %v", tr.id, gID, err)
+		if tr.discNum != nil {
-			}
+			discNum = *tr.discNum
 		}
 		id := InsertTestTrack(t, db, TestTrack{
 			FilePath:    tr.filePath,
 			Title:       tr.title,
 			Artist:      tr.artist,
 			Album:       tr.album,
 			Genres:      genres,
 			TrackNumber: trackNum,
 			DiscNumber:  discNum,
 			Year:        tr.year,
 			LengthMs:    tr.lenMs,
 		})
 		// The fixtures assert on audio properties and the composer,
 		// which InsertTestTrack does not carry - they are not part of
 		// what a seeder should have to know about a track.
 		if _, err := db.ExecContext(
 			`UPDATE audio_files
 			 SET file_type_id = ?, sample_rate = ?, bit_depth = ?,
 			     channels = ?, bitrate = ?, file_size = ?, composer = ?
 			 WHERE id = ?`,
 			tr.ftID, tr.sr, tr.bd, tr.ch, tr.br, tr.fsize, tr.composer, id,
 		); err != nil {
 			t.Fatalf("set audio properties for %q: %v", tr.filePath, err)
 		}
 	}
 }
@@ -553,7 +471,7 @@ func TestSearchFTSTracks(t *testing.T) {
 	db := NewTestDB(t)
 	seedSearchData(t, db)
-	results, err := db.SearchFTSTracks("queen", 10)
+	results, err := db.SearchFTSTracks("queen", 0, 10)
 	if err != nil {
 		t.Fatalf("SearchFTSTracks: %v", err)
 	}
@@ -563,7 +481,7 @@ func TestSearchFTSTracks(t *testing.T) {
 	}
 	// Find the Bohemian Rhapsody result and verify all 16 fields.
-	var br *SearchTrackRow
+	var br *sqlcgen.TrackMetadatum
 	for i, r := range results {
 		if r.Title == "Bohemian Rhapsody" {
@@ -635,26 +553,12 @@ func TestInsertAndDeleteSearchIndex(t *testing.T) {
 	db := NewTestDB(t)
 	// Set up minimal FK chain for a single track.
-	_, err := db.ExecContext(
+	InsertTestTrack(t, db, TestTrack{
-		"INSERT INTO artist_credit (id, text) VALUES (1, 'Test Artist')",
+		FilePath: "/test/track.mp3",
-	)
+		Title:    "Test Track",
-	if err != nil {
+		Artist:   "Test Artist",
-		t.Fatalf("insert artist_credit: %v", err)
+		LengthMs: 180000,
-	}
+	})
 	_, err = db.ExecContext(
 		"INSERT INTO recordings (id, name, artist_credit_id) VALUES (1, 'Test Track', 1)",
 	)
 	if err != nil {
 		t.Fatalf("insert recording: %v", err)
 	}
 	_, err = db.ExecContext(
 		"INSERT INTO audio_files (id, file_path, length_milliseconds, file_type_id, recording_id) VALUES (1, '/test/track.mp3', 180000, 0, 1)",
 	)
 	if err != nil {
 		t.Fatalf("insert audio_file: %v", err)
 	}
 	// Insert into search index.
 	if err := db.InsertSearchIndex(
@@ -698,41 +602,15 @@ func TestRebuildSearchIndex(t *testing.T) {
 	db := NewTestDB(t)
-	// Seed the full entity graph WITHOUT inserting into search_index.
+	// Seed the file WITHOUT putting it in search_index.
-	_, err := db.ExecContext(
+	InsertTestTrack(t, db, TestTrack{
-		"INSERT INTO artist_credit (id, text) VALUES (1, 'Rebuild Artist')",
+		FilePath:        "/rebuild/track.mp3",
-	)
+		Title:           "Rebuild Track",
-	if err != nil {
+		Artist:          "Rebuild Artist",
-		t.Fatalf("insert artist_credit: %v", err)
+		Album:           "Rebuild Album",
-	}
+		LengthMs:        200000,
-
+		SkipSearchIndex: true,
-	_, err = db.ExecContext(
+	})
 		"INSERT INTO recordings (id, name, artist_credit_id) VALUES (1, 'Rebuild Track', 1)",
 	)
 	if err != nil {
 		t.Fatalf("insert recording: %v", err)
 	}
 	_, err = db.ExecContext(
 		"INSERT INTO audio_files (id, file_path, length_milliseconds, file_type_id, recording_id) VALUES (1, '/rebuild/track.mp3', 200000, 0, 1)",
 	)
 	if err != nil {
 		t.Fatalf("insert audio_file: %v", err)
 	}
 	_, err = db.ExecContext(
 		"INSERT INTO release_groups (id, name) VALUES (1, 'Rebuild Album')",
 	)
 	if err != nil {
 		t.Fatalf("insert release_group: %v", err)
 	}
 	_, err = db.ExecContext(
 		"INSERT INTO release_group_recordings (release_group_id, recording_id) VALUES (1, 1)",
 	)
 	if err != nil {
 		t.Fatalf("insert release_group_recordings: %v", err)
 	}
 	// Search should return nothing before rebuild.
 	results, err := db.SearchFTS("Rebuild", 10)
@@ -887,36 +765,22 @@ func TestSearchIndexUpdateCycle(t *testing.T) {
 	db := NewTestDB(t)
 	// Set up minimal FK chain for a single track at rowid 100.
-	_, err := db.ExecContext(
+	id := InsertTestTrack(t, db, TestTrack{
-		"INSERT INTO artist_credit (id, text) VALUES (100, 'Old Artist')",
+		FilePath:        "/test/update_cycle.mp3",
-	)
+		Title:           "Old Title",
-	if err != nil {
+		Artist:          "Old Artist",
-		t.Fatalf("insert artist_credit: %v", err)
+		LengthMs:        200000,
-	}
+		SkipSearchIndex: true,
-
+	})
 	_, err = db.ExecContext(
 		"INSERT INTO recordings (id, name, artist_credit_id) VALUES (100, 'Old Title', 100)",
 	)
 	if err != nil {
 		t.Fatalf("insert recording: %v", err)
 	}
 	_, err = db.ExecContext(
 		"INSERT INTO audio_files (id, file_path, length_milliseconds, file_type_id, recording_id) " +
 			"VALUES (100, '/test/update_cycle.mp3', 200000, 0, 100)",
 	)
 	if err != nil {
 		t.Fatalf("insert audio_file: %v", err)
 	}
 	// 1. Insert with old metadata.
 	if err := db.InsertSearchIndex(
-		100, "/test/update_cycle.mp3", "Old Title", "Old Artist", "Old Album",
+		id, "/test/update_cycle.mp3", "Old Title", "Old Artist", "Old Album",
 	); err != nil {
 		t.Fatalf("InsertSearchIndex (old): %v", err)
 	}
-	// Verify search for "Old Title" returns rowid 100.
+	// Verify search for "Old Title" finds it.
 	results, err := db.SearchFTS("Old Title", 10)
 	if err != nil {
 		t.Fatalf("SearchFTS(Old Title): %v", err)
@@ -926,9 +790,9 @@ func TestSearchIndexUpdateCycle(t *testing.T) {
 		t.Fatal("SearchFTS(Old Title): got 0 results after insert")
 	}
-	// 2. Delete rowid 100.
+	// 2. Delete the row.
-	if err := db.DeleteSearchIndex(100); err != nil {
+	if err := db.DeleteSearchIndex(id); err != nil {
-		t.Fatalf("DeleteSearchIndex(100): %v", err)
+		t.Fatalf("DeleteSearchIndex(%d): %v", id, err)
 	}
 	// Verify "Old Title" no longer found.
@@ -944,25 +808,17 @@ func TestSearchIndexUpdateCycle(t *testing.T) {
 		)
 	}
-	// 3. Update the recording name in the DB to simulate tag edit.
+	// 3. Update the file's title in the DB to simulate a tag edit.
 	_, err = db.ExecContext(
-		"UPDATE recordings SET name = 'New Title' WHERE id = 100",
+		"UPDATE audio_files SET title = 'New Title' WHERE file_path = '/test/update_cycle.mp3'",
 	)
 	if err != nil {
-		t.Fatalf("update recording: %v", err)
+		t.Fatalf("update title: %v", err)
 	}
-	// Also add a new artist_credit for the new artist.
+	// 4. Re-insert the row with new metadata.
 	_, err = db.ExecContext(
 		"INSERT INTO artist_credit (id, text) VALUES (101, 'New Artist')",
 	)
 	if err != nil {
 		t.Fatalf("insert new artist_credit: %v", err)
 	}
 	// 4. Re-insert rowid 100 with new metadata.
 	if err := db.InsertSearchIndex(
-		100, "/test/update_cycle.mp3", "New Title", "New Artist", "New Album",
+		id, "/test/update_cycle.mp3", "New Title", "New Artist", "New Album",
 	); err != nil {
 		t.Fatalf("InsertSearchIndex (new): %v", err)
 	}
@@ -1079,25 +935,13 @@ func TestSearchIndexSchema(t *testing.T) {
 		t.Fatalf("insert artist: %v", err)
 	}
 	// The credit tables this used to assert a UNIQUE constraint on are
 	// gone; a file names its artist directly, and artists are unique by
 	// name, which is asserted below.
 	_, err = db.ExecContext(
-		"INSERT INTO artist_credit (id, text) VALUES (1, 'Test Credit')",
+		"INSERT INTO artists (id, name) VALUES (2, 'Test')",
 	)
 	if err != nil {
 		t.Fatalf("insert artist_credit: %v", err)
 	}
 	_, err = db.ExecContext(
 		"INSERT INTO artist_credit_artist (artist_id, credit_id) VALUES (1, 1)",
 	)
 	if err != nil {
 		t.Fatalf("first insert artist_credit_artist: %v", err)
 	}
 	// Duplicate insert should fail with UNIQUE constraint.
 	_, err = db.ExecContext(
 		"INSERT INTO artist_credit_artist (artist_id, credit_id) VALUES (1, 1)",
 	)
 	if err == nil {
-		t.Error("duplicate artist_credit_artist insert should fail, got nil error")
+		t.Error("duplicate artist name should fail, got nil error")
 	}
 }
@@ -1,10 +0,0 @@
 -- Adds SplitMixedFolder's synthetic-group bookkeeping to an
 -- existing tagging_items table.  A fresh database never runs this
 -- file: sql/schemas/tagging_items.sql already declares these
 -- columns, so applySchema's isFreshDatabase check stamps this
 -- version as applied without executing it.
 ALTER TABLE tagging_items ADD COLUMN synthetic INTEGER NOT NULL DEFAULT 0;
 ALTER TABLE tagging_items ADD COLUMN parent_group_key TEXT NOT NULL DEFAULT '';
 CREATE INDEX IF NOT EXISTS idx_tagging_items_parent_group_key
    ON tagging_items(parent_group_key) WHERE parent_group_key != '';
@@ -1,24 +0,0 @@
 -- Repairs tagging_items rows left behind by a library-scan bug: the
 -- rescan's orphan-cleanup phase deleted audio_files rows for files
 -- removed from disk without decrementing/clearing their tagging
 -- group, so a folder whose contents were fully replaced kept a
 -- phantom entry (stale track_count, no matching audio_files) in the
 -- autotag queue forever. The library scan code no longer has this
 -- gap, but a database written before the fix still carries the
 -- damage — this is a one-time repair, not ongoing bookkeeping.
 --
 -- Drop groups with no audio_files left at all.
 DELETE FROM tagging_items
 WHERE group_key NOT IN (
  SELECT DISTINCT group_key FROM audio_files WHERE group_key != ''
 );
 -- Reconcile track_count for groups that are still alive but drifted
 -- (some, not all, of their tracks were removed without decrementing).
 UPDATE tagging_items
 SET track_count = (
  SELECT COUNT(*) FROM audio_files WHERE audio_files.group_key = tagging_items.group_key
 )
 WHERE track_count != (
  SELECT COUNT(*) FROM audio_files WHERE audio_files.group_key = tagging_items.group_key
 );
@@ -1 +0,0 @@
 ALTER TABLE tagging_items ADD COLUMN album_artist_conflict INTEGER NOT NULL DEFAULT 0;
@@ -1,3 +0,0 @@
 ALTER TABLE queue ADD COLUMN source_type TEXT NOT NULL DEFAULT '';
 ALTER TABLE queue ADD COLUMN source_id INTEGER NOT NULL DEFAULT 0;
 ALTER TABLE queue ADD COLUMN source_label TEXT NOT NULL DEFAULT '';
@@ -1 +0,0 @@
 ALTER TABLE release_groups ADD COLUMN pending_release_mbid TEXT;
@@ -1 +0,0 @@
 ALTER TABLE release_group_recordings ADD COLUMN total_tracks INTEGER;
@@ -0,0 +1,137 @@
 -- Queries over albums (formerly release_groups).
 --
 -- The two-copy pattern is gone here too: one query answers both the
 -- whole-library and the single-library case.  The `fallback_ac`
 -- subquery every album read used to carry -- "if the album has no album
 -- artist credit, borrow one from any of its recordings" -- is gone with
 -- it, because the album carries its own credit text now.
 -- name: UpsertAlbum :one
 INSERT INTO albums (name, artist_credit, artist_id, year, cover_art_id)
 VALUES (?, ?, ?, ?, ?)
 ON CONFLICT(name, artist_credit) DO UPDATE SET
    artist_id    = COALESCE(excluded.artist_id, albums.artist_id),
    year         = COALESCE(excluded.year, albums.year),
    cover_art_id = COALESCE(excluded.cover_art_id, albums.cover_art_id)
 RETURNING *;
 -- name: GetAlbum :one
 SELECT * FROM albums WHERE id = ? LIMIT 1;
 -- name: SetAlbumMBID :exec
 UPDATE albums SET mbid = ? WHERE id = ?;
 -- name: SetAlbumOriginalYear :exec
 UPDATE albums SET original_year = ? WHERE id = ?;
 -- name: SetAlbumCoverArt :exec
 UPDATE albums SET cover_art_id = ? WHERE id = ?;
 -- name: SetAlbumPendingReleaseMBID :exec
 UPDATE albums SET pending_release_mbid = ? WHERE id = ?;
 -- name: ResolveAlbumPendingReleaseMBID :exec
 -- Clears the pending marker once the release-group MBID it stood in for
 -- has been resolved.  Guarded so a real MBID is never overwritten.
 UPDATE albums
 SET mbid = ?, pending_release_mbid = NULL
 WHERE id = ? AND (mbid IS NULL OR mbid = '');
 -- name: GetAlbumsWithPendingReleaseMBID :many
 SELECT id, pending_release_mbid FROM albums
 WHERE pending_release_mbid IS NOT NULL AND pending_release_mbid != ''
  AND (mbid IS NULL OR mbid = '');
 -- name: DeleteAlbum :exec
 DELETE FROM albums WHERE id = ?;
 -- name: DeleteAllAlbums :exec
 DELETE FROM albums;
 -- name: GetEmptyAlbumIDs :many
 -- Albums with no file left behind them.  Under the old schema this was
 -- one of three orphan sweeps that had to run by hand and did not;
 -- audio_files is the only thing that can leave an album empty now, so
 -- this is the whole of it.
 SELECT id FROM albums al
 WHERE NOT EXISTS (
    SELECT 1 FROM audio_files af WHERE af.album_id = al.id
 );
 -- name: GetAlbums :many
 SELECT
    al.id,
    al.name,
    COALESCE(al.original_year, al.year) AS year,
    COALESCE(al.year, 0) AS release_year,
    al.mbid,
    al.artist_credit AS artist_name,
    CAST(COALESCE(ar.mbid, '') AS TEXT) AS artist_mbid,
    COALESCE(ca.file_path, '') AS cover_art_path
 FROM albums al
 LEFT JOIN artists ar   ON ar.id = al.artist_id
 LEFT JOIN cover_art ca ON ca.id = al.cover_art_id
 WHERE EXISTS (
    SELECT 1 FROM audio_files af
    WHERE af.album_id = al.id
      AND af.library_id = COALESCE(NULLIF(CAST(sqlc.arg(library_id) AS INTEGER), 0), af.library_id)
 )
 ORDER BY al.name;
 -- name: GetAlbumsByArtistName :many
 SELECT
    al.id,
    al.name,
    COALESCE(al.original_year, al.year) AS year,
    COALESCE(al.year, 0) AS release_year,
    al.mbid,
    al.artist_credit AS artist_name,
    CAST(COALESCE(ar.mbid, '') AS TEXT) AS artist_mbid,
    COALESCE(ca.file_path, '') AS cover_art_path
 FROM albums al
 LEFT JOIN artists ar   ON ar.id = al.artist_id
 LEFT JOIN cover_art ca ON ca.id = al.cover_art_id
 WHERE (al.artist_credit = sqlc.arg(artist) OR ar.name = sqlc.arg(artist))
  AND EXISTS (
      SELECT 1 FROM audio_files af
      WHERE af.album_id = al.id
        AND af.library_id = COALESCE(NULLIF(CAST(sqlc.arg(library_id) AS INTEGER), 0), af.library_id)
  )
 ORDER BY year, al.name;
 -- name: GetAlbumCompleteness :one
 -- "Do I have all of this album", answered from the tags on disk.
 --
 -- The expectation is a **sum over discs**, not one number: totals are
 -- declared per disc ("5/12" on disc 2 means 12 tracks on disc 2), so a
 -- multi-disc album's expectation is the sum of each disc's declared
 -- total.  A disc whose files declared nothing leaves the whole album
 -- unknowable rather than being covered by the discs that did -- which is
 -- what `known` reports.
 --
 -- Owned counts DISTINCT track numbers: this app detects duplicates, and
 -- counting two files of track 3 twice would report a short album as
 -- complete.
 SELECT
    -- Distinct (disc, track) pairs: this app detects duplicates, and
    -- counting two files of track 3 twice would report a short album as
    -- complete.  A file with no track number falls back to its own id,
    -- because three untagged files are three tracks, not one.
    CAST(COUNT(DISTINCT CAST(COALESCE(a.disc_number, 1) AS TEXT) || ':' ||
                        COALESCE(CAST(a.track_number AS TEXT), 'f' || a.id)
    ) AS INTEGER) AS owned,
    CAST(COALESCE((
        SELECT SUM(per_disc.total)
        FROM (
            SELECT MAX(b.total_tracks) AS total
            FROM audio_files b
            WHERE b.album_id = sqlc.arg(album_id) AND b.total_tracks IS NOT NULL
            GROUP BY COALESCE(b.disc_number, 1)
        ) per_disc
    ), 0) AS INTEGER) AS expected,
    CAST((
        SELECT COUNT(*) = 0 FROM audio_files c
        WHERE c.album_id = sqlc.arg(album_id) AND c.total_tracks IS NULL
    ) AS INTEGER) AS known
 FROM audio_files a
 WHERE a.album_id = sqlc.arg(album_id);
@@ -1,42 +0,0 @@
 -- name: CreateArtistCredit :one
 INSERT INTO artist_credit (text) VALUES (?)
 RETURNING *;
 -- name: GetArtistCredit :one
 SELECT * FROM artist_credit 
 WHERE id = ? LIMIT 1;
 -- name: GetArtistCreditByText :one
 SELECT * FROM artist_credit
 WHERE text = ? LIMIT 1;
 -- name: UpsertArtistCredit :one
 INSERT INTO artist_credit (text) VALUES (?)
 ON CONFLICT(text) DO UPDATE SET text = excluded.text
 RETURNING *;
 -- name: UpdateArtistCredit :exec
 UPDATE artist_credit 
 SET text = ?
 WHERE id = ?;
 -- name: DeleteArtistCredit :exec
 DELETE FROM artist_credit 
 WHERE id = ?;
 -- name: DeleteAllArtistCredits :exec
 DELETE FROM artist_credit;
 -- name: CountArtistCreditReferences :one
 SELECT
  (SELECT COUNT(*) FROM recordings WHERE artist_credit_id = ?1) +
  (SELECT COUNT(*) FROM release_groups WHERE album_artist_credit_id = ?1)
 AS total;
 -- name: GetOrphanedArtistCreditIDs :many
 -- Artist credits no longer used by any recording or release group - run
 -- after orphaned recordings/release groups are deleted, so a credit
 -- that only existed for now-removed tracks is cleaned up too.
 SELECT ac.id FROM artist_credit ac
 WHERE NOT EXISTS (SELECT 1 FROM recordings r WHERE r.artist_credit_id = ac.id)
  AND NOT EXISTS (SELECT 1 FROM release_groups rg WHERE rg.album_artist_credit_id = ac.id);
@@ -1,24 +0,0 @@
 -- name: CreateArtistCreditArtist :one
 INSERT INTO artist_credit_artist (artist_id, credit_id) VALUES (?, ?)
 RETURNING *;
 -- name: GetArtistCreditArtist :one
 SELECT * FROM artist_credit_artist 
 WHERE id = ? LIMIT 1;
 -- name: UpdateArtistCreditArtist :exec
 UPDATE artist_credit_artist 
 SET artist_id = ?, credit_id = ?
 WHERE id =?;
 -- name: DeleteArtistCreditArtist :exec
 DELETE FROM artist_credit_artist 
 WHERE id =?;
 -- name: DeleteAllArtistCreditArtists :exec
 DELETE FROM artist_credit_artist;
 -- name: DeleteArtistCreditArtistByCredit :exec
 DELETE FROM artist_credit_artist
 WHERE credit_id = ?;
@@ -1,67 +1,55 @@
-- name: CreateArtist :one
+-- Queries over artists.
-INSERT INTO artists (name) VALUES (?)
+--
 -- An artist row is reachable two ways: as a file's primary artist
 -- (audio_files.artist_id) and as an album's artist (albums.artist_id).
 -- Both used to route through artist_credit + artist_credit_artist,
 -- which is how "which album artists are in library 2" came to be a
 -- five-join subquery inside a three-join query.
 -- name: UpsertArtist :one
 INSERT INTO artists (name, mbid) VALUES (?, ?)
 ON CONFLICT(name) DO UPDATE SET
    mbid = COALESCE(excluded.mbid, artists.mbid)
 RETURNING *;
 -- name: GetArtist :one
-SELECT * FROM artists 
+SELECT * FROM artists WHERE id = ? LIMIT 1;
 WHERE id = ? LIMIT 1;
 -- name: GetArtistByName :one
-SELECT * FROM artists
+SELECT * FROM artists WHERE name = ? LIMIT 1;
 WHERE name = ? LIMIT 1;
-- name: UpsertArtist :one
+-- name: SetArtistMBID :exec
-INSERT INTO artists (name) VALUES (?)
+UPDATE artists SET mbid = ? WHERE id = ?;
 ON CONFLICT(name) DO UPDATE SET name = excluded.name
 RETURNING *;
 -- name: UpdateArtist :exec
 UPDATE artists 
 SET name = ?
 WHERE id = ?;
 -- name: DeleteArtist :exec
-DELETE FROM artists 
+DELETE FROM artists WHERE id = ?;
 WHERE id = ?;
 -- name: DeleteAllArtists :exec
 DELETE FROM artists;
 -- name: GetUnreferencedArtistIDs :many
 -- Artists no file and no album points at any more.
 SELECT id FROM artists a
 WHERE NOT EXISTS (SELECT 1 FROM audio_files af WHERE af.artist_id = a.id)
  AND NOT EXISTS (SELECT 1 FROM albums al WHERE al.artist_id = a.id);
 -- name: GetAllArtists :many
-SELECT * FROM artists
+SELECT * FROM artists ORDER BY name;
 ORDER BY name;
 -- name: GetAlbumArtists :many
 SELECT DISTINCT a.id, a.name, a.mbid
 FROM artists a
-JOIN artist_credit_artist aca ON aca.artist_id = a.id
+JOIN albums al ON al.artist_id = a.id
-JOIN artist_credit ac ON ac.id = aca.credit_id
+WHERE EXISTS (
-JOIN release_groups rg ON rg.album_artist_credit_id = ac.id
+    SELECT 1 FROM audio_files af
-ORDER BY a.name;
+    WHERE af.album_id = al.id
-
+      AND af.library_id = COALESCE(NULLIF(CAST(sqlc.arg(library_id) AS INTEGER), 0), af.library_id)
 -- name: GetOrphanedArtistIDs :many
 -- Artists no longer credited on any recording or release group - left
 -- behind when a scan's orphan cleanup removes the audio_files that used
 -- to justify them, since deleting an audio_files row doesn't cascade.
 SELECT a.id FROM artists a
 WHERE NOT EXISTS (
    SELECT 1 FROM artist_credit_artist aca WHERE aca.artist_id = a.id
 );
 -- name: GetAlbumArtistsByLibrary :many
 SELECT DISTINCT a.id, a.name, a.mbid
 FROM artists a
 JOIN artist_credit_artist aca ON aca.artist_id = a.id
 JOIN artist_credit ac ON ac.id = aca.credit_id
 JOIN release_groups rg ON rg.album_artist_credit_id = ac.id
 WHERE a.id IN (
    SELECT DISTINCT aca2.artist_id
    FROM artist_credit_artist aca2
    JOIN artist_credit ac2 ON ac2.id = aca2.credit_id
    JOIN release_groups rg2 ON rg2.album_artist_credit_id = ac2.id
    JOIN release_group_recordings rgr2 ON rgr2.release_group_id = rg2.id
    JOIN recordings r2 ON r2.id = rgr2.recording_id
    JOIN audio_files af2 ON af2.recording_id = r2.id
    WHERE af2.library_id = ?
 )
 ORDER BY a.name;
 -- name: GetArtistByFilePath :one
 SELECT COALESCE(a.name, '') AS artist_name, COALESCE(a.mbid, '') AS artist_mbid
 FROM audio_files af
 LEFT JOIN artists a ON a.id = af.artist_id
 WHERE af.file_path = ?
 LIMIT 1;
@@ -1,39 +1,60 @@
 -- Queries over audio_files and the track_metadata view above it.
 --
 -- Every query that returns "a track" selects from `track_metadata`,
 -- which is the one place the projection is defined.  The scoped and
 -- unscoped variants that used to be written twice are one query now:
 -- library_id 0 means "every library", and `(:id = 0 OR library_id = :id)`
 -- costs nothing measurable (23 ms vs 21 ms over 26k rows) because these
 -- queries scan either way.
 -- ---------------------------------------------------------------------
 -- Writes
 -- ---------------------------------------------------------------------
 -- name: CreateAudioFile :one
 INSERT INTO audio_files (file_path, length_milliseconds, file_type_id, recording_id, sample_rate, bit_depth, channels, bitrate, file_size, basename, library_id) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
 RETURNING *;
 -- name: CreateAudioFileWithGroupKey :one
 INSERT INTO audio_files (
-  file_path, length_milliseconds, file_type_id, recording_id,
+  file_path, library_id, file_type_id,
-  sample_rate, bit_depth, channels, bitrate, file_size, basename,
+  length_milliseconds, sample_rate, bit_depth, channels, bitrate, file_size,
-  library_id, group_key, tag_status, modified_at
+  title, artist_credit, artist_id, album_id,
-) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+  track_number, disc_number, total_tracks, year, composer, comment,
  recording_mbid, basename, group_key, modified_at, tag_status
 ) VALUES (
  ?, ?, ?,
  ?, ?, ?, ?, ?, ?,
  ?, ?, ?, ?,
  ?, ?, ?, ?, ?, ?,
  ?, ?, ?, ?, ?
 )
 RETURNING *;
-- name: GetAudioFileGroupKey :one
+-- name: UpdateAudioFileTags :exec
-SELECT group_key FROM audio_files
+-- A rescan of a file whose mtime moved: the tags are re-read and
-WHERE id = ? LIMIT 1;
+-- written over the same row.  Under the old schema this created a
 -- *new* recording and repointed the file at it, abandoning the old one
 -- -- which is where 812 orphaned rows and every phantom "you own this"
 -- came from.  There is nothing to orphan now.
 UPDATE audio_files
 SET title = ?, artist_credit = ?, artist_id = ?, album_id = ?,
    track_number = ?, disc_number = ?, total_tracks = ?, year = ?,
    composer = ?, comment = ?, recording_mbid = ?,
    sample_rate = ?, bit_depth = ?, channels = ?, bitrate = ?,
    file_size = ?, length_milliseconds = ?, modified_at = ?
 WHERE id = ?;
 -- name: SetAudioFileGroupKey :exec
 UPDATE audio_files SET group_key = ? WHERE id = ?;
-- name: GetAudioFile :one
+-- name: PromoteAudioFileTagStatusIfUntagged :exec
-SELECT * FROM audio_files 
+-- A rescan re-reads the tags of a file whose mtime moved, so a file
-WHERE id = ? LIMIT 1;
+-- another tagger stamped with MBIDs since import arrives here still
-
+-- carrying the 'untagged' status it was created with (only the insert
-- name: GetAudioFileByPath :one
+-- path sets it).  Promote it the same way saveAudioFile does.
-SELECT * FROM audio_files
+-- Guarded on 'untagged' so it cannot overwrite a deliberate
-WHERE file_path = ? LIMIT 1;
+-- 'user_skipped_permanent', and so a file losing its MBIDs is left
-
+-- alone -- demotion is the scan's judgement, not this statement's.
 -- name: UpdateAudioFile :exec
 UPDATE audio_files
-SET file_path = ?, length_milliseconds = ?, file_type_id = ?, recording_id = ?, sample_rate = ?, bit_depth = ?, channels = ?, bitrate = ?, file_size = ?, basename = ?
+SET tag_status = 'user_confirmed'
-WHERE id = ?;
+WHERE id = ? AND tag_status = 'untagged';
 -- name: UpdateAudioFileRecording :exec
 UPDATE audio_files
 SET recording_id = ?, sample_rate = ?, bit_depth = ?, channels = ?, bitrate = ?, file_size = ?, length_milliseconds = ?, modified_at = ?
 WHERE id = ?;
 -- name: UpdateAudioFileStat :exec
 -- Records the on-disk mtime/size without re-reading tags.  Used to
@@ -43,307 +64,146 @@ UPDATE audio_files
 SET modified_at = ?, file_size = ?
 WHERE id = ?;
 -- name: SetAudioFileRecordingMBID :exec
 UPDATE audio_files SET recording_mbid = ? WHERE id = ?;
 -- name: DeleteAudioFile :exec
 DELETE FROM audio_files WHERE id = ?;
 -- name: DeleteAllAudioFiles :exec
 DELETE FROM audio_files;
 -- ---------------------------------------------------------------------
 -- Reads: the file row itself
 -- ---------------------------------------------------------------------
 -- name: GetAudioFile :one
 SELECT * FROM audio_files WHERE id = ? LIMIT 1;
 -- name: GetAudioFileByPath :one
 SELECT * FROM audio_files WHERE file_path = ? LIMIT 1;
 -- name: GetAudioFileGroupKey :one
 SELECT group_key FROM audio_files WHERE id = ? LIMIT 1;
 -- name: GetAllAudioFilePaths :many
 SELECT id, file_path FROM audio_files;
 -- name: GetAudioFilesByPaths :many
 SELECT id, library_id, file_path, group_key FROM audio_files
 WHERE file_path IN (sqlc.slice('paths'));
 -- name: GetRandomAudioFilePath :one
 SELECT file_path FROM audio_files ORDER BY RANDOM() LIMIT 1;
 -- name: CountAudioFiles :one
 SELECT COUNT(*) AS count FROM audio_files
 WHERE library_id = COALESCE(NULLIF(CAST(sqlc.arg(library_id) AS INTEGER), 0), library_id);
 -- name: GetLibraryMaxModifiedAt :one
 -- Newest recorded mtime in a library, for the startup soft scan.  0 when
 -- the library is empty or no row has a baseline yet.
 SELECT CAST(COALESCE(MAX(modified_at), 0) AS INTEGER) FROM audio_files
 WHERE library_id = ?;
-- name: DeleteAudioFile :exec
+-- ---------------------------------------------------------------------
-DELETE FROM audio_files 
+-- Reads: tracks
-WHERE id = ?;
+-- ---------------------------------------------------------------------
-- name: CountAudioFiles :one
+-- name: GetTracks :many
-SELECT count(*) FROM audio_files;
+SELECT * FROM track_metadata
 WHERE library_id = COALESCE(NULLIF(CAST(sqlc.arg(library_id) AS INTEGER), 0), library_id);
-- name: GetRandomAudioFilePath :one
+-- name: GetTrackByPath :one
-SELECT file_path FROM audio_files
+SELECT * FROM track_metadata WHERE file_path = ? LIMIT 1;
 ORDER BY RANDOM()
 LIMIT 1;
-- name: GetAllAudioFiles :many
+-- name: GetTracksByAlbum :many
-SELECT * FROM audio_files;
+SELECT * FROM track_metadata
 WHERE album_id = sqlc.arg(album_id)
  AND library_id = COALESCE(NULLIF(CAST(sqlc.arg(library_id) AS INTEGER), 0), library_id)
 ORDER BY disc_number, track_number;
-- name: GetAllAudioFilePaths :many
+-- name: GetTracksByGenre :many
-SELECT id, file_path FROM audio_files;
+SELECT tm.* FROM track_metadata tm
-
+JOIN file_genres fg ON fg.audio_file_id = tm.id
-- name: GetAudioFilesNeedingMetadata :many
+JOIN genres g ON g.id = fg.genre_id
-SELECT * FROM audio_files
+WHERE g.name = sqlc.arg(genre)
-WHERE recording_id = 0;
+  AND tm.library_id = COALESCE(NULLIF(CAST(sqlc.arg(library_id) AS INTEGER), 0), tm.library_id);
 -- name: GetAllAudioFilesWithArtist :many
 SELECT 
    af.id,
    af.file_path,
    af.length_milliseconds,
    af.file_type_id,
    af.recording_id,
    COALESCE(ac.text, '') AS artist_name,
    COALESCE(r.name, '') AS title
 FROM audio_files af
 JOIN recordings r ON af.recording_id = r.id
 JOIN artist_credit ac ON r.artist_credit_id = ac.id;
 -- name: GetTrackMetadataByPath :one
 SELECT 
    af.file_path,
    af.length_milliseconds,
    COALESCE(r.name, '') AS title,
    COALESCE(ac.text, '') AS artist,
    COALESCE(rg.name, '') AS album,
    COALESCE(ca.file_path, '') AS cover_art_path,
    COALESCE(a.mbid, '') AS artist_mbid,
    COALESCE(rg.mbid, '') AS release_group_mbid,
    COALESCE(r.mbid, '') AS recording_mbid
 FROM audio_files af
 LEFT JOIN recordings r ON af.recording_id = r.id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN artists a ON a.id = aca.artist_id
 LEFT JOIN release_group_recordings rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 WHERE af.file_path = ?
 LIMIT 1;
 -- name: GetAllTracksWithFullMetadata :many
 SELECT
    af.file_path,
    af.length_milliseconds,
    COALESCE(r.name, '') AS title,
    COALESCE(ac.text, '') AS artist_name,
    r.track_number,
    r.disc_number,
    COALESCE(rg.name, '') AS album,
    CAST(COALESCE(
        (SELECT GROUP_CONCAT(g.name, '||')
         FROM recording_genres rg_sub
         JOIN genres g ON rg_sub.genre_id = g.id
         WHERE rg_sub.recording_id = r.id),
        ''
    ) AS TEXT) AS genre,
    COALESCE(r.year, 0) AS year,
    COALESCE(r.composer, '') AS composer,
    COALESCE(ft.extension, '') AS file_type,
    af.sample_rate,
    af.bit_depth,
    af.channels,
    af.bitrate,
    af.file_size,
    af.play_count,
    af.last_played,
    COALESCE(ca.file_path, '') AS cover_art_path,
    COALESCE(a.mbid, '') AS artist_mbid,
    COALESCE(rg.mbid, '') AS release_group_mbid,
    COALESCE(r.mbid, '') AS recording_mbid
 FROM audio_files af
 JOIN recordings r ON af.recording_id = r.id
 JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN artists a ON a.id = aca.artist_id
 LEFT JOIN release_group_recordings rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 LEFT JOIN file_types ft ON af.file_type_id = ft.id;
 -- name: SearchAudioFilesByBasename :many
 SELECT
    af.file_path,
    af.length_milliseconds,
    COALESCE(r.name, '') AS title,
    COALESCE(ac.text, '') AS artist,
    COALESCE(rg.name, '') AS album
 FROM audio_files af
 LEFT JOIN recordings r ON af.recording_id = r.id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN (
    SELECT recording_id, MIN(release_group_id) AS release_group_id
    FROM release_group_recordings
    GROUP BY recording_id
 ) rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 WHERE af.basename = ?
 LIMIT ?;
 -- name: LookupTrackMetaByPaths :many
-SELECT id, file_path, title, artist_name, album, cover_art_path, artist_mbid, release_group_mbid, recording_mbid
+SELECT id, file_path, title, artist_name, album, cover_art_path,
       artist_mbid, release_group_mbid, recording_mbid
 FROM track_metadata
 WHERE file_path IN (sqlc.slice('paths'));
-- name: GetAudioFilesByLibrary :many
+-- name: SearchTracksByBasename :many
-SELECT * FROM audio_files WHERE library_id = ?;
+SELECT id, file_path, length_milliseconds, title, artist_name, album
 FROM track_metadata
 WHERE file_path IN (
    SELECT file_path FROM audio_files WHERE basename = sqlc.arg(basename)
 )
 LIMIT sqlc.arg(lim);
-- name: CountAudioFilesByLibrary :one
+-- ---------------------------------------------------------------------
-SELECT COUNT(*) AS count FROM audio_files WHERE library_id = ?;
+-- Reads: file paths, grouped by whatever the caller asked about
 -- ---------------------------------------------------------------------
 -- These answer "what can I play" and they all ask audio_files, because
 -- that is the only table whose rows are files.  Grouped rather than
 -- flattened because the caller owns the order.
-- name: DeleteAllAudioFiles :exec
+-- name: GetFilePathsByAlbums :many
-DELETE FROM audio_files;
+-- The library filter is applied in Go rather than here: sqlc numbers a
-
+-- named parameter (?2) but expands a slice into N placeholders, so the
-- name: GetAllTracksWithFullMetadataByLibrary :many
+-- two together bind the wrong values - GetFilePathsByAlbums([1,2], 0)
-SELECT
+-- read album id 2 as the library id.  Returning library_id and
-    af.file_path,
+-- filtering the (small) result is the version that cannot be wrong.
-    af.length_milliseconds,
+SELECT album_id, library_id, file_path FROM audio_files
-    COALESCE(r.name, '') AS title,
+WHERE album_id IN (sqlc.slice('album_ids'))
-    COALESCE(ac.text, '') AS artist_name,
+ORDER BY disc_number, track_number;
    r.track_number,
    r.disc_number,
    COALESCE(rg.name, '') AS album,
    CAST(COALESCE(
        (SELECT GROUP_CONCAT(g.name, '||')
         FROM recording_genres rg_sub
         JOIN genres g ON rg_sub.genre_id = g.id
         WHERE rg_sub.recording_id = r.id),
        ''
    ) AS TEXT) AS genre,
    COALESCE(r.year, 0) AS year,
    COALESCE(r.composer, '') AS composer,
    COALESCE(ft.extension, '') AS file_type,
    af.sample_rate,
    af.bit_depth,
    af.channels,
    af.bitrate,
    af.file_size,
    af.play_count,
    af.last_played,
    COALESCE(ca.file_path, '') AS cover_art_path,
    COALESCE(a.mbid, '') AS artist_mbid,
    COALESCE(rg.mbid, '') AS release_group_mbid,
    COALESCE(r.mbid, '') AS recording_mbid
 FROM audio_files af
 JOIN recordings r ON af.recording_id = r.id
 JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN artists a ON a.id = aca.artist_id
 LEFT JOIN release_group_recordings rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 LEFT JOIN file_types ft ON af.file_type_id = ft.id
 WHERE af.library_id = ?;
 -- name: GetAudioFilesByReleaseGroup :many
 SELECT
    af.file_path,
    af.length_milliseconds,
    COALESCE(r.name, '') AS title,
    COALESCE(ac.text, '') AS artist_name,
    rgr.track_number,
    rgr.disc_number,
    COALESCE(rg.name, '') AS album,
    CAST(COALESCE(
        (SELECT GROUP_CONCAT(g.name, '||')
         FROM recording_genres rg_sub
         JOIN genres g ON rg_sub.genre_id = g.id
         WHERE rg_sub.recording_id = r.id),
        ''
    ) AS TEXT) AS genre,
    COALESCE(r.year, 0) AS year,
    COALESCE(r.composer, '') AS composer,
    COALESCE(ft.extension, '') AS file_type,
    af.sample_rate,
    af.bit_depth,
    af.channels,
    af.bitrate,
    af.file_size,
    COALESCE(a.mbid, '') AS artist_mbid,
    COALESCE(rg.mbid, '') AS release_group_mbid,
    COALESCE(r.mbid, '') AS recording_mbid
 FROM release_group_recordings rgr
 JOIN recordings r ON rgr.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN artists a ON a.id = aca.artist_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 LEFT JOIN file_types ft ON af.file_type_id = ft.id
 WHERE rgr.release_group_id = ?
 ORDER BY rgr.disc_number, rgr.track_number;
 -- name: GetAudioFilesByReleaseGroupByLibrary :many
 SELECT
    af.file_path,
    af.length_milliseconds,
    COALESCE(r.name, '') AS title,
    COALESCE(ac.text, '') AS artist_name,
    rgr.track_number,
    rgr.disc_number,
    COALESCE(rg.name, '') AS album,
    CAST(COALESCE(
        (SELECT GROUP_CONCAT(g.name, '||')
         FROM recording_genres rg_sub
         JOIN genres g ON rg_sub.genre_id = g.id
         WHERE rg_sub.recording_id = r.id),
        ''
    ) AS TEXT) AS genre,
    COALESCE(r.year, 0) AS year,
    COALESCE(r.composer, '') AS composer,
    COALESCE(ft.extension, '') AS file_type,
    af.sample_rate,
    af.bit_depth,
    af.channels,
    af.bitrate,
    af.file_size,
    COALESCE(a.mbid, '') AS artist_mbid,
    COALESCE(rg.mbid, '') AS release_group_mbid,
    COALESCE(r.mbid, '') AS recording_mbid
 FROM release_group_recordings rgr
 JOIN recordings r ON rgr.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN artists a ON a.id = aca.artist_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 LEFT JOIN file_types ft ON af.file_type_id = ft.id
 WHERE rgr.release_group_id = ? AND af.library_id = ?
 ORDER BY rgr.disc_number, rgr.track_number;
 -- "Play this artist" and "play these albums" wanted file paths and asked
 -- for whole track rows to get them, one round trip per album (perf.m2).
 -- These answer the same question in one query and carry only what the
 -- caller uses; the release group id comes back so the caller can keep
 -- its own album ordering.
 -- name: GetFilePathsByReleaseGroups :many
 SELECT rgr.release_group_id, af.file_path
 FROM release_group_recordings rgr
 JOIN recordings r ON rgr.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 WHERE rgr.release_group_id IN (sqlc.slice('release_group_ids'))
 ORDER BY rgr.disc_number, rgr.track_number;
 -- name: GetFilePathsByReleaseGroupsByLibrary :many
 SELECT rgr.release_group_id, af.file_path
 FROM release_group_recordings rgr
 JOIN recordings r ON rgr.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 WHERE rgr.release_group_id IN (sqlc.slice('release_group_ids'))
  AND af.library_id = ?
 ORDER BY rgr.disc_number, rgr.track_number;
 -- Same shape again, keyed on recording MBID, for the catalog side.
 -- An Explore album page knows which of its tracks the user owns only
 -- as a set of recording MBIDs -- that is exactly how the backend
 -- decides `inLibrary` (markReleasesInLibrary -> CheckMBIDs) -- and
 -- MBTrack.LocalID is declared but never written by anything, so there
 -- is no id to ask by. Grouped by MBID because a recording can have
 -- more than one file (the duplicate fixtures are precisely that) and
 -- because the caller owns the order: the tracklist's, not the
 -- database's.
 -- name: GetFilePathsByRecordingMBIDs :many
-SELECT r.mbid AS recording_mbid, af.file_path
+-- The ownership question in its only honest form: which of these
-FROM recordings r
+-- catalog recordings has a *file* behind it.  Asked of audio_files, so
-JOIN audio_files af ON af.recording_id = r.id
+-- a metadata row with no file cannot answer yes.
-WHERE r.mbid IN (sqlc.slice('mbids'))
+SELECT recording_mbid, library_id, file_path FROM audio_files
-ORDER BY af.file_path;
+WHERE recording_mbid IN (sqlc.slice('mbids'))
 ORDER BY file_path;
-- name: GetFilePathsByRecordingMBIDsByLibrary :many
+-- name: GetFilePathsByGenres :many
-SELECT r.mbid AS recording_mbid, af.file_path
+SELECT g.name AS genre, af.library_id, af.file_path
-FROM recordings r
+FROM audio_files af
-JOIN audio_files af ON af.recording_id = r.id
+JOIN file_genres fg ON fg.audio_file_id = af.id
-WHERE r.mbid IN (sqlc.slice('mbids'))
+JOIN genres g ON g.id = fg.genre_id
-  AND af.library_id = ?
+WHERE g.name IN (sqlc.slice('genres'))
-ORDER BY af.file_path;
+ORDER BY af.disc_number, af.track_number;
-- name: GetAudioFilesByPaths :many
+-- name: GetFilePathsByArtistMBID :many
-SELECT id, library_id, file_path, group_key FROM audio_files
+SELECT DISTINCT af.file_path
-WHERE file_path IN (sqlc.slice('paths'));
+FROM audio_files af
 JOIN artists a ON a.id = af.artist_id
 WHERE a.mbid = ?;
 -- ---------------------------------------------------------------------
 -- Ownership, asked in bulk
 -- ---------------------------------------------------------------------
 -- name: OwnedRecordingMBIDs :many
 -- Which of these recording MBIDs are actually in the library.  This is
 -- what marks a catalog tracklist owned; it used to be
 -- `SELECT mbid FROM recordings`, which answered yes for 129 tracks in a
 -- real library that had no file at all.
 SELECT DISTINCT recording_mbid FROM audio_files
 WHERE recording_mbid IN (sqlc.slice('mbids'));
 -- name: OwnedAlbumMBIDs :many
 SELECT DISTINCT al.mbid FROM albums al
 JOIN audio_files af ON af.album_id = al.id
 WHERE al.mbid IN (sqlc.slice('mbids'));
 -- name: OwnedArtistMBIDs :many
 SELECT DISTINCT a.mbid FROM artists a
 JOIN audio_files af ON af.artist_id = a.id
 WHERE a.mbid IN (sqlc.slice('mbids'));
 -- name: GetAudioFilesInLibrary :many
 SELECT * FROM audio_files WHERE library_id = ?;
@@ -1,150 +1,50 @@
 -- Queries over genres and file_genres.
 --
 -- The track-returning ones live in audio_files.sql with the rest of the
 -- track_metadata reads; what is left here is the genre list itself and
 -- the link table's writes.
 -- name: UpsertGenre :one
 INSERT INTO genres (name) VALUES (?)
-ON CONFLICT(name) DO UPDATE SET name = name
+ON CONFLICT(name) DO UPDATE SET name = excluded.name
 RETURNING *;
-- name: CreateRecordingGenre :exec
+-- name: LinkFileGenre :exec
-INSERT OR IGNORE INTO recording_genres (recording_id, genre_id)
+INSERT OR IGNORE INTO file_genres (audio_file_id, genre_id) VALUES (?, ?);
 VALUES (?, ?);
-- name: DeleteRecordingGenres :exec
+-- name: DeleteFileGenres :exec
-DELETE FROM recording_genres
+DELETE FROM file_genres WHERE audio_file_id = ?;
 WHERE recording_id = ?;
-- name: GetGenresByRecordingID :many
+-- name: GetGenreNamesByFile :many
-SELECT g.*
+SELECT g.name FROM genres g
-FROM genres g
+JOIN file_genres fg ON fg.genre_id = g.id
-JOIN recording_genres rg ON g.id = rg.genre_id
+WHERE fg.audio_file_id = ?;
 WHERE rg.recording_id = ?;
-- name: DeleteAllRecordingGenres :exec
+-- name: GetGenreNamesByFilePaths :many
-DELETE FROM recording_genres;
+-- Genres for many files at once.  The mix builder asked this one file
-
+-- at a time, inside two nested loops -- twelve thousand single-row
-- name: DeleteAllGenres :exec
+-- queries to assemble one mix.
-DELETE FROM genres;
+SELECT af.file_path, g.name
-
+FROM audio_files af
-- name: GetTracksByGenre :many
+JOIN file_genres fg ON fg.audio_file_id = af.id
-SELECT
+JOIN genres g ON g.id = fg.genre_id
-    af.file_path,
+WHERE af.file_path IN (sqlc.slice('paths'));
    af.length_milliseconds,
    COALESCE(r.name, '') AS title,
    COALESCE(ac.text, '') AS artist_name,
    r.track_number,
    r.disc_number,
    COALESCE(rlg.name, '') AS album,
    CAST(COALESCE(
        (SELECT GROUP_CONCAT(g2.name, '||')
         FROM recording_genres rg2
         JOIN genres g2 ON rg2.genre_id = g2.id
         WHERE rg2.recording_id = r.id),
        ''
    ) AS TEXT) AS genre,
    COALESCE(r.year, 0) AS year,
    COALESCE(r.composer, '') AS composer,
    COALESCE(ft.extension, '') AS file_type,
    af.sample_rate,
    af.bit_depth,
    af.channels,
    af.bitrate,
    af.file_size
 FROM genres g
 JOIN recording_genres rg ON g.id = rg.genre_id
 JOIN recordings r ON rg.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN (
    SELECT recording_id,
        MIN(release_group_id) AS release_group_id
    FROM release_group_recordings
    GROUP BY recording_id
 ) rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rlg ON rgr.release_group_id = rlg.id
 LEFT JOIN file_types ft ON af.file_type_id = ft.id
 WHERE g.name = ?
 ORDER BY r.name;
 -- name: GetTracksByGenreByLibrary :many
 SELECT
    af.file_path,
    af.length_milliseconds,
    COALESCE(r.name, '') AS title,
    COALESCE(ac.text, '') AS artist_name,
    r.track_number,
    r.disc_number,
    COALESCE(rlg.name, '') AS album,
    CAST(COALESCE(
        (SELECT GROUP_CONCAT(g2.name, '||')
         FROM recording_genres rg2
         JOIN genres g2 ON rg2.genre_id = g2.id
         WHERE rg2.recording_id = r.id),
        ''
    ) AS TEXT) AS genre,
    COALESCE(r.year, 0) AS year,
    COALESCE(r.composer, '') AS composer,
    COALESCE(ft.extension, '') AS file_type,
    af.sample_rate,
    af.bit_depth,
    af.channels,
    af.bitrate,
    af.file_size
 FROM genres g
 JOIN recording_genres rg ON g.id = rg.genre_id
 JOIN recordings r ON rg.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN (
    SELECT recording_id,
        MIN(release_group_id) AS release_group_id
    FROM release_group_recordings
    GROUP BY recording_id
 ) rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rlg ON rgr.release_group_id = rlg.id
 LEFT JOIN file_types ft ON af.file_type_id = ft.id
 WHERE g.name = ? AND af.library_id = ?
 ORDER BY r.name;
 -- name: CountGenreReferences :one
 SELECT COUNT(*) FROM recording_genres WHERE genre_id = ?;
 -- name: DeleteGenre :exec
 DELETE FROM genres WHERE id = ?;
 -- name: DeleteAllGenres :exec
 DELETE FROM genres;
 -- name: GetUnusedGenreIDs :many
 SELECT id FROM genres g
 WHERE NOT EXISTS (SELECT 1 FROM file_genres fg WHERE fg.genre_id = g.id);
 -- name: GetAllGenresWithCounts :many
-SELECT g.name, COUNT(rg.recording_id) AS track_count
+SELECT g.name, COUNT(fg.audio_file_id) AS track_count
 FROM genres g
-JOIN recording_genres rg ON g.id = rg.genre_id
+JOIN file_genres fg ON fg.genre_id = g.id
 JOIN audio_files af ON af.id = fg.audio_file_id
 WHERE af.library_id = COALESCE(NULLIF(CAST(sqlc.arg(library_id) AS INTEGER), 0), af.library_id)
 GROUP BY g.id, g.name
 ORDER BY g.name;
 -- name: GetAllGenresWithCountsByLibrary :many
 SELECT g.name, COUNT(rg.recording_id) AS track_count
 FROM genres g
 JOIN recording_genres rg ON g.id = rg.genre_id
 JOIN recordings r ON rg.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 WHERE af.library_id = ?
 GROUP BY g.id, g.name
 ORDER BY g.name;
 -- Same as GetFilePathsByReleaseGroups, for "play these genres" (perf.m2):
 -- one query instead of one per genre, and file paths instead of whole
 -- track rows, which was 6 MB over the IPC for five genres.
 -- name: GetFilePathsByGenres :many
 SELECT g.name AS genre_name, af.file_path
 FROM genres g
 JOIN recording_genres rg ON g.id = rg.genre_id
 JOIN recordings r ON rg.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 WHERE g.name IN (sqlc.slice('genre_names'))
 ORDER BY r.name;
 -- name: GetFilePathsByGenresByLibrary :many
 SELECT g.name AS genre_name, af.file_path
 FROM genres g
 JOIN recording_genres rg ON g.id = rg.genre_id
 JOIN recordings r ON rg.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 WHERE g.name IN (sqlc.slice('genre_names'))
  AND af.library_id = ?
 ORDER BY r.name;
@@ -2,16 +2,15 @@
 --
 -- Every one of these returns album ids and nothing else.  The display
 -- columns (cover art, artist credit, year) already have exactly one
-- correct expression of them, in GetAllAlbumsWithDetails, and a second
+-- correct expression of them, in GetAlbums, and a second
 -- copy per shelf would be six more places for that to drift.  The home
 -- service joins the ids back to that one album list in Go.
 -- name: HomeRecentlyPlayedAlbums :many
 -- Albums with the most recent play, newest first.
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
 JOIN audio_files af ON af.recording_id = rgr.recording_id
 WHERE af.last_played IS NOT NULL
 GROUP BY rg.id
 ORDER BY MAX(af.last_played) DESC
@@ -22,9 +21,8 @@ LIMIT ?;
 -- stands in for one: it is monotonic and assigned at import, which is
 -- the same ordering an added_at column would give.
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
 JOIN audio_files af ON af.recording_id = rgr.recording_id
 GROUP BY rg.id
 ORDER BY MAX(af.id) DESC
 LIMIT ?;
@@ -32,9 +30,8 @@ LIMIT ?;
 -- name: HomeMostPlayedAlbums :many
 -- Albums by total plays across their tracks.
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
 JOIN audio_files af ON af.recording_id = rgr.recording_id
 GROUP BY rg.id
 HAVING SUM(af.play_count) > 0
 ORDER BY SUM(af.play_count) DESC
@@ -45,9 +42,8 @@ LIMIT ?;
 -- shelf is a different suggestion each time rather than the same
 -- alphabetical head of the list forever.
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
 JOIN audio_files af ON af.recording_id = rgr.recording_id
 GROUP BY rg.id
 HAVING SUM(af.play_count) = 0
 ORDER BY RANDOM()
@@ -56,9 +52,8 @@ LIMIT ?;
 -- name: HomeStaleAlbums :many
 -- Played before, but not for a long while.
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
 JOIN audio_files af ON af.recording_id = rgr.recording_id
 WHERE af.last_played IS NOT NULL
 GROUP BY rg.id
 HAVING MAX(af.last_played) < datetime('now', ?)
@@ -67,9 +62,8 @@ LIMIT ?;
 -- name: HomeRandomAlbums :many
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
 JOIN audio_files af ON af.recording_id = rgr.recording_id
 GROUP BY rg.id
 ORDER BY RANDOM()
 LIMIT ?;
@@ -78,10 +72,10 @@ LIMIT ?;
 -- A random sample of albums carrying a genre, so the same genre shelf
 -- is not the same ten albums every time the page opens.
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
-JOIN recording_genres rgen ON rgen.recording_id = rgr.recording_id
+JOIN file_genres fg ON fg.audio_file_id = af.id
-JOIN genres g ON g.id = rgen.genre_id
+JOIN genres g ON g.id = fg.genre_id
 WHERE g.name = ?
 GROUP BY rg.id
 ORDER BY RANDOM()
@@ -93,10 +87,10 @@ LIMIT ?;
 -- album carries is a shelf about that one album.
 SELECT
    g.name AS genre,
-    COUNT(DISTINCT rgr.release_group_id) AS album_count
+    COUNT(DISTINCT af.album_id) AS album_count
 FROM genres g
-JOIN recording_genres rgen ON rgen.genre_id = g.id
+JOIN file_genres fg ON fg.genre_id = g.id
-JOIN release_group_recordings rgr ON rgr.recording_id = rgen.recording_id
+JOIN audio_files af ON af.id = fg.audio_file_id
 GROUP BY g.id
 HAVING album_count >= 3
 ORDER BY album_count DESC
@@ -106,14 +100,12 @@ LIMIT ?;
 -- Artists by total plays, as the album-artist credit text the album
 -- list already displays.
 SELECT
-    COALESCE(ac.text, '') AS artist_name,
+    rg.artist_credit AS artist_name,
    SUM(af.play_count) AS plays
-FROM release_groups rg
+FROM albums rg
-JOIN artist_credit ac ON ac.id = rg.album_artist_credit_id
+JOIN audio_files af ON af.album_id = rg.id
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+WHERE rg.artist_credit <> ''
-JOIN audio_files af ON af.recording_id = rgr.recording_id
+GROUP BY rg.artist_credit
 WHERE ac.text <> ''
 GROUP BY ac.text
 HAVING plays > 0
 ORDER BY plays DESC
 LIMIT ?;
@@ -1,30 +0,0 @@
 -- Queries backing the dynamic-mix queue fallback (backend/explore/mix.go):
 -- expanding a seed selection into a candidate pool by artist similarity
 -- and genre overlap, restricted to what is actually in the library.
 -- name: GetFilePathsByArtistMBID :many
 SELECT DISTINCT af.file_path
 FROM audio_files af
 JOIN recordings r ON af.recording_id = r.id
 JOIN artist_credit ac ON r.artist_credit_id = ac.id
 JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 JOIN artists a ON a.id = aca.artist_id
 WHERE a.mbid = ?;
 -- name: GetGenreNamesByFilePath :many
 SELECT DISTINCT g.name
 FROM genres g
 JOIN recording_genres rg ON g.id = rg.genre_id
 JOIN recordings r ON rg.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 WHERE af.file_path = ?;
 -- name: GetArtistByFilePath :one
 SELECT COALESCE(a.name, '') AS artist_name, COALESCE(a.mbid, '') AS artist_mbid
 FROM audio_files af
 JOIN recordings r ON af.recording_id = r.id
 JOIN artist_credit ac ON r.artist_credit_id = ac.id
 JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 JOIN artists a ON a.id = aca.artist_id
 WHERE af.file_path = ?
 LIMIT 1;
@@ -47,29 +47,18 @@ SELECT
    pt.playlist_id,
    pt.audio_file_id,
    pt.position,
-    COALESCE(af.file_path, '') AS file_path,
+    COALESCE(tm.file_path, '') AS file_path,
-    COALESCE(af.length_milliseconds, 0) AS length_milliseconds,
+    COALESCE(tm.length_milliseconds, 0) AS length_milliseconds,
-    COALESCE(r.name, pt.phantom_title, '') AS title,
+    COALESCE(tm.title, pt.phantom_title, '') AS title,
-    COALESCE(ac.text, pt.phantom_artist, '') AS artist,
+    COALESCE(tm.artist_name, pt.phantom_artist, '') AS artist,
-    COALESCE(rg.name, pt.phantom_album, '') AS album,
+    COALESCE(tm.album, pt.phantom_album, '') AS album,
-    COALESCE(ca.file_path, pt.phantom_cover_art_path, '') AS cover_art_path,
+    COALESCE(NULLIF(tm.cover_art_path, ''), pt.phantom_cover_art_path, '') AS cover_art_path,
    CASE WHEN pt.audio_file_id IS NULL THEN 1 ELSE 0 END AS is_phantom,
-    COALESCE(a.mbid, '') AS artist_mbid,
+    CAST(COALESCE(tm.artist_mbid, '') AS TEXT) AS artist_mbid,
-    COALESCE(rg.mbid, '') AS release_group_mbid,
+    COALESCE(tm.release_group_mbid, '') AS release_group_mbid,
-    COALESCE(r.mbid, '') AS recording_mbid
+    COALESCE(tm.recording_mbid, '') AS recording_mbid
 FROM playlist_tracks pt
-LEFT JOIN audio_files af ON pt.audio_file_id = af.id
+LEFT JOIN track_metadata tm ON tm.id = pt.audio_file_id
 LEFT JOIN recordings r ON af.recording_id = r.id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN artists a ON a.id = aca.artist_id
 LEFT JOIN (
    SELECT recording_id, MIN(release_group_id) AS release_group_id
    FROM release_group_recordings
    GROUP BY recording_id
 ) rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 WHERE pt.playlist_id = ?
 ORDER BY pt.position;
@@ -79,29 +68,18 @@ SELECT
    pt.playlist_id,
    pt.audio_file_id,
    pt.position,
-    COALESCE(af.file_path, '') AS file_path,
+    COALESCE(tm.file_path, '') AS file_path,
-    COALESCE(af.length_milliseconds, 0) AS length_milliseconds,
+    COALESCE(tm.length_milliseconds, 0) AS length_milliseconds,
-    COALESCE(r.name, pt.phantom_title, '') AS title,
+    COALESCE(tm.title, pt.phantom_title, '') AS title,
-    COALESCE(ac.text, pt.phantom_artist, '') AS artist,
+    COALESCE(tm.artist_name, pt.phantom_artist, '') AS artist,
-    COALESCE(rg.name, pt.phantom_album, '') AS album,
+    COALESCE(tm.album, pt.phantom_album, '') AS album,
-    COALESCE(ca.file_path, pt.phantom_cover_art_path, '') AS cover_art_path,
+    COALESCE(NULLIF(tm.cover_art_path, ''), pt.phantom_cover_art_path, '') AS cover_art_path,
    CASE WHEN pt.audio_file_id IS NULL THEN 1 ELSE 0 END AS is_phantom,
-    COALESCE(a.mbid, '') AS artist_mbid,
+    CAST(COALESCE(tm.artist_mbid, '') AS TEXT) AS artist_mbid,
-    COALESCE(rg.mbid, '') AS release_group_mbid,
+    COALESCE(tm.release_group_mbid, '') AS release_group_mbid,
-    COALESCE(r.mbid, '') AS recording_mbid
+    COALESCE(tm.recording_mbid, '') AS recording_mbid
 FROM playlist_tracks pt
-LEFT JOIN audio_files af ON pt.audio_file_id = af.id
+LEFT JOIN track_metadata tm ON tm.id = pt.audio_file_id
 LEFT JOIN recordings r ON af.recording_id = r.id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN artists a ON a.id = aca.artist_id
 LEFT JOIN (
    SELECT recording_id, MIN(release_group_id) AS release_group_id
    FROM release_group_recordings
    GROUP BY recording_id
 ) rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 ORDER BY pt.playlist_id, pt.position;
 -- name: DeleteAllPlaylistTracks :exec
@@ -132,27 +110,8 @@ WHERE playlist_id = ? AND audio_file_id = (
 );
 -- name: GetTrackPhantomMetadata :one
-SELECT
+-- The display fields a playlist row keeps after its file goes away.
-    COALESCE(r.name, '') AS title,
+SELECT title, artist_name AS artist, album,
-    COALESCE(ac.text, '') AS artist,
+       length_milliseconds AS duration_ms, genre, cover_art_path
-    COALESCE(rg.name, '') AS album,
+FROM track_metadata
-    af.length_milliseconds AS duration_ms,
+WHERE id = ?;
    CAST(COALESCE(
        (SELECT GROUP_CONCAT(g.name, '||')
         FROM recording_genres rg_sub
         JOIN genres g ON rg_sub.genre_id = g.id
         WHERE rg_sub.recording_id = r.id),
        ''
    ) AS TEXT) AS genre,
    COALESCE(ca.file_path, '') AS cover_art_path
 FROM audio_files af
 LEFT JOIN recordings r ON af.recording_id = r.id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN (
    SELECT recording_id, MIN(release_group_id) AS release_group_id
    FROM release_group_recordings
    GROUP BY recording_id
 ) rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 WHERE af.id = ?;
@@ -13,27 +13,12 @@ SET current_position = ?
 WHERE id = 1;
 -- name: GetQueueTracks :many
-SELECT qt.id, qt.audio_file_id, qt.position, af.file_path,
+-- The queue's rows, joined to the one track projection.
-    COALESCE(r.name, '') AS title,
+SELECT qt.id, qt.audio_file_id, qt.position, tm.file_path,
-    COALESCE(ac.text, '') AS artist,
+    tm.title, tm.artist_name AS artist, tm.album, tm.cover_art_path,
-    COALESCE(rg.name, '') AS album,
+    tm.artist_mbid, tm.release_group_mbid, tm.recording_mbid
    COALESCE(ca.file_path, '') AS cover_art_path,
    COALESCE(a.mbid, '') AS artist_mbid,
    COALESCE(rg.mbid, '') AS release_group_mbid,
    COALESCE(r.mbid, '') AS recording_mbid
 FROM queue_tracks qt
-JOIN audio_files af ON qt.audio_file_id = af.id
+JOIN track_metadata tm ON tm.id = qt.audio_file_id
 LEFT JOIN recordings r ON af.recording_id = r.id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN artists a ON a.id = aca.artist_id
 LEFT JOIN (
    SELECT recording_id, MIN(release_group_id) AS release_group_id
    FROM release_group_recordings
    GROUP BY recording_id
 ) rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 ORDER BY qt.position;
 -- name: GetQueueTrackCount :one
@@ -1,47 +0,0 @@
 -- name: CreateRecording :one
 INSERT INTO recordings (name, artist_credit_id) VALUES (?, ?)
 RETURNING *;
 -- name: CreateRecordingFull :one
 INSERT INTO recordings (
  name, artist_credit_id, track_number, disc_number,
  year, genre, composer, lyrics, comment
 ) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)
 RETURNING *;
 -- name: GetRecording :one
 SELECT * FROM recordings 
 WHERE id = ? LIMIT 1;
 -- name: UpdateRecording :exec
 UPDATE recordings 
 SET name = ?, artist_credit_id = ?
 WHERE id = ?;
 -- name: UpdateRecordingFull :exec
 UPDATE recordings
 SET name = ?, artist_credit_id = ?, track_number = ?, disc_number = ?,
    year = ?, genre = ?, composer = ?, lyrics = ?, comment = ?
 WHERE id = ?;
 -- name: DeleteRecording :exec
 DELETE FROM recordings 
 WHERE id = ?;
 -- name: DeleteAllRecordings :exec
 DELETE FROM recordings;
 -- name: GetAllRecordings :many
 SELECT * FROM recordings
 ORDER BY name;
 -- name: CountRecordingsByArtistCredit :one
 SELECT COUNT(*) FROM recordings WHERE artist_credit_id = ?;
 -- name: GetOrphanedRecordingIDs :many
 -- Recordings no longer backed by any audio_files row - left behind
 -- when a scan's orphan cleanup deletes the file that used to own them,
 -- since deleting audio_files doesn't cascade to recordings.
 SELECT r.id FROM recordings r
 LEFT JOIN audio_files af ON af.recording_id = r.id
 WHERE af.id IS NULL;
@@ -1,50 +0,0 @@
 -- name: CreateReleaseGroupRecording :one
 INSERT INTO release_group_recordings (
  release_group_id, recording_id, track_number, disc_number, total_tracks
 )
 VALUES (?, ?, ?, ?, ?)
 RETURNING *;
 -- name: GetAlbumCompleteness :one
 WITH discs AS (
    SELECT
        COALESCE(rgr.disc_number, 1) AS disc,
        MAX(COALESCE(rgr.total_tracks, 0)) AS declared,
        COUNT(DISTINCT COALESCE(rgr.track_number, -rgr.recording_id)) AS owned
    FROM release_group_recordings rgr
    WHERE rgr.release_group_id = ?
    GROUP BY COALESCE(rgr.disc_number, 1)
 )
 SELECT
    CAST(COALESCE(SUM(owned), 0) AS INTEGER) AS owned,
    CAST(COALESCE(SUM(declared), 0) AS INTEGER) AS expected,
    CAST(COALESCE(SUM(CASE WHEN declared = 0 THEN 1 ELSE 0 END), 0) AS INTEGER) AS discs_untotalled
 FROM discs;
 -- name: GetReleaseGroupRecording :one
 SELECT * FROM release_group_recordings 
 WHERE id = ? LIMIT 1;
 -- name: GetReleaseGroupRecordings :many
 SELECT * FROM release_group_recordings
 WHERE release_group_id = ?
 ORDER BY disc_number, track_number;
 -- name: GetRecordingReleaseGroups :many
 SELECT * FROM release_group_recordings
 WHERE recording_id = ?;
 -- name: DeleteReleaseGroupRecording :exec
 DELETE FROM release_group_recordings 
 WHERE id = ?;
 -- name: DeleteReleaseGroupRecordingByFK :exec
 DELETE FROM release_group_recordings
 WHERE release_group_id = ? AND recording_id = ?;
 -- name: DeleteAllReleaseGroupRecordings :exec
 DELETE FROM release_group_recordings;
 -- name: DeleteReleaseGroupRecordingsByRecording :exec
 DELETE FROM release_group_recordings
 WHERE recording_id = ?;
@@ -1,216 +0,0 @@
 -- name: CreateReleaseGroup :one
 INSERT INTO release_groups (name) VALUES (?)
 RETURNING *;
 -- name: CreateReleaseGroupFull :one
 INSERT INTO release_groups (
  name, cover_art_id, album_artist_credit_id, year, total_tracks, total_discs
 ) VALUES (?, ?, ?, ?, ?, ?)
 RETURNING *;
 -- name: GetReleaseGroup :one
 SELECT * FROM release_groups 
 WHERE id = ? LIMIT 1;
 -- name: GetReleaseGroupByNameAndArtist :one
 SELECT * FROM release_groups
 WHERE name = ? AND album_artist_credit_id = ? LIMIT 1;
 -- name: UpsertReleaseGroup :one
 INSERT INTO release_groups (name, album_artist_credit_id, year)
 VALUES (?, ?, ?)
 ON CONFLICT(name, album_artist_credit_id) DO UPDATE SET
  album_artist_credit_id = COALESCE(excluded.album_artist_credit_id, release_groups.album_artist_credit_id),
  year = COALESCE(excluded.year, release_groups.year)
 RETURNING *;
 -- name: SetReleaseGroupOriginalYear :exec
 -- Set the release group's original-release-year (release-group's
 -- first-release-date from MusicBrainz).  Called from autotag apply
 -- when the user confirms a candidate; the file-tag year stays in
 -- the year column.
 UPDATE release_groups SET original_year = ? WHERE id = ?;
 -- name: UpdateReleaseGroup :exec
 UPDATE release_groups 
 SET name = ?
 WHERE id = ?;
 -- name: UpdateReleaseGroupCoverArt :exec
 UPDATE release_groups
 SET cover_art_id = ?
 WHERE id = ?;
 -- name: DeleteReleaseGroup :exec
 DELETE FROM release_groups 
 WHERE id = ?;
 -- name: DeleteAllReleaseGroups :exec
 DELETE FROM release_groups;
 -- name: GetAllReleaseGroups :many
 SELECT * FROM release_groups
 ORDER BY name;
 -- name: GetAllAlbumsWithDetails :many
 SELECT
    rg.id,
    rg.name,
    -- year prefers original release year (MB first-release-date)
    -- over the file-tag year so the UI surfaces the album's
    -- original year by default.  release_year keeps the file-tag
    -- year accessible.
    COALESCE(rg.original_year, rg.year) AS year,
    COALESCE(rg.year, 0) AS release_year,
    rg.mbid,
    COALESCE(ac.text, fallback_ac.text, '') as artist_name,
    -- primary (first-credited) album artist's MBID, for linking the
    -- artist name to its detail page.  Empty when the album has no
    -- MB-tagged album-artist credit.
    CAST(COALESCE((
        SELECT a.mbid
        FROM artist_credit_artist aca_p
        JOIN artists a ON a.id = aca_p.artist_id
        WHERE aca_p.credit_id = rg.album_artist_credit_id
        ORDER BY aca_p.id
        LIMIT 1
    ), '') AS TEXT) as artist_mbid,
    COALESCE(ca.file_path, '') as cover_art_path
 FROM release_groups rg
 LEFT JOIN artist_credit ac ON rg.album_artist_credit_id = ac.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 LEFT JOIN (
    SELECT rgr.release_group_id, ac2.text
    FROM release_group_recordings rgr
    JOIN recordings rec ON rec.id = rgr.recording_id
    JOIN artist_credit ac2 ON ac2.id = rec.artist_credit_id
    GROUP BY rgr.release_group_id
 ) fallback_ac ON fallback_ac.release_group_id = rg.id
 ORDER BY rg.name;
 -- name: GetAllAlbumsWithDetailsByLibrary :many
 SELECT
    rg.id,
    rg.name,
    -- year prefers original release year (MB first-release-date)
    -- over the file-tag year so the UI surfaces the album's
    -- original year by default.  release_year keeps the file-tag
    -- year accessible.
    COALESCE(rg.original_year, rg.year) AS year,
    COALESCE(rg.year, 0) AS release_year,
    rg.mbid,
    COALESCE(ac.text, fallback_ac.text, '') as artist_name,
    -- primary (first-credited) album artist's MBID, for linking the
    -- artist name to its detail page.  Empty when the album has no
    -- MB-tagged album-artist credit.
    CAST(COALESCE((
        SELECT a.mbid
        FROM artist_credit_artist aca_p
        JOIN artists a ON a.id = aca_p.artist_id
        WHERE aca_p.credit_id = rg.album_artist_credit_id
        ORDER BY aca_p.id
        LIMIT 1
    ), '') AS TEXT) as artist_mbid,
    COALESCE(ca.file_path, '') as cover_art_path
 FROM release_groups rg
 LEFT JOIN artist_credit ac ON rg.album_artist_credit_id = ac.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 LEFT JOIN (
    SELECT rgr.release_group_id, ac2.text
    FROM release_group_recordings rgr
    JOIN recordings rec ON rec.id = rgr.recording_id
    JOIN artist_credit ac2 ON ac2.id = rec.artist_credit_id
    GROUP BY rgr.release_group_id
 ) fallback_ac ON fallback_ac.release_group_id = rg.id
 WHERE rg.id IN (
    SELECT DISTINCT rgr2.release_group_id
    FROM release_group_recordings rgr2
    JOIN recordings r2 ON r2.id = rgr2.recording_id
    JOIN audio_files af2 ON af2.recording_id = r2.id
    WHERE af2.library_id = ?
 )
 ORDER BY rg.name;
 -- name: GetAlbumsByArtist :many
 SELECT
    rg.id,
    rg.name,
    COALESCE(rg.original_year, rg.year) AS year,
    COALESCE(rg.year, 0) AS release_year,
    COALESCE(ac.text, fallback_ac.text, '') as artist_name,
    -- primary (first-credited) album artist's MBID, for linking the
    -- artist name to its detail page.  Empty when the album has no
    -- MB-tagged album-artist credit.
    CAST(COALESCE((
        SELECT a.mbid
        FROM artist_credit_artist aca_p
        JOIN artists a ON a.id = aca_p.artist_id
        WHERE aca_p.credit_id = rg.album_artist_credit_id
        ORDER BY aca_p.id
        LIMIT 1
    ), '') AS TEXT) as artist_mbid,
    COALESCE(ca.file_path, '') as cover_art_path
 FROM release_groups rg
 JOIN artist_credit ac ON rg.album_artist_credit_id = ac.id
 JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 LEFT JOIN (
    SELECT rgr.release_group_id, ac2.text
    FROM release_group_recordings rgr
    JOIN recordings rec ON rec.id = rgr.recording_id
    JOIN artist_credit ac2 ON ac2.id = rec.artist_credit_id
    GROUP BY rgr.release_group_id
 ) fallback_ac ON fallback_ac.release_group_id = rg.id
 WHERE aca.artist_id = ?
 ORDER BY rg.name;
 -- name: CountReleaseGroupRecordings :one
 SELECT COUNT(*) FROM release_group_recordings WHERE release_group_id = ?;
 -- name: GetOrphanedReleaseGroupIDs :many
 -- Release groups with no recordings left in them - run after orphaned
 -- recordings (and their release_group_recordings rows) are deleted, so
 -- a release group whose last owned track was removed is cleaned up too.
 SELECT rg.id FROM release_groups rg
 LEFT JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
 WHERE rgr.id IS NULL;
 -- name: GetAlbumsByArtistByLibrary :many
 SELECT
    rg.id,
    rg.name,
    COALESCE(rg.original_year, rg.year) AS year,
    COALESCE(rg.year, 0) AS release_year,
    COALESCE(ac.text, fallback_ac.text, '') as artist_name,
    -- primary (first-credited) album artist's MBID, for linking the
    -- artist name to its detail page.  Empty when the album has no
    -- MB-tagged album-artist credit.
    CAST(COALESCE((
        SELECT a.mbid
        FROM artist_credit_artist aca_p
        JOIN artists a ON a.id = aca_p.artist_id
        WHERE aca_p.credit_id = rg.album_artist_credit_id
        ORDER BY aca_p.id
        LIMIT 1
    ), '') AS TEXT) as artist_mbid,
    COALESCE(ca.file_path, '') as cover_art_path
 FROM release_groups rg
 JOIN artist_credit ac ON rg.album_artist_credit_id = ac.id
 JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 LEFT JOIN (
    SELECT rgr.release_group_id, ac2.text
    FROM release_group_recordings rgr
    JOIN recordings rec ON rec.id = rgr.recording_id
    JOIN artist_credit ac2 ON ac2.id = rec.artist_credit_id
    GROUP BY rgr.release_group_id
 ) fallback_ac ON fallback_ac.release_group_id = rg.id
 WHERE aca.artist_id = ?
  AND rg.id IN (
    SELECT DISTINCT rgr2.release_group_id
    FROM release_group_recordings rgr2
    JOIN recordings r2 ON r2.id = rgr2.recording_id
    JOIN audio_files af2 ON af2.recording_id = r2.id
    WHERE af2.library_id = ?
 )
 ORDER BY rg.name;
@@ -87,10 +87,7 @@ LIMIT 1;
 SELECT ti.group_key
 FROM tagging_items ti
 JOIN audio_files af ON af.group_key = ti.group_key
-LEFT JOIN recordings r ON af.recording_id = r.id
+LEFT JOIN albums rg ON rg.id = af.album_id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN release_group_recordings rgr ON rgr.recording_id = r.id
 LEFT JOIN release_groups rg ON rg.id = rgr.release_group_id
 WHERE ti.synthetic = 0
  AND ti.track_count >= 4
  AND (
@@ -98,13 +95,23 @@ WHERE ti.synthetic = 0
    OR LOWER(TRIM(ti.album_artist)) IN ('various artists', 'various', 'va', 'v.a.', 'v a', 'unknown')
  )
 GROUP BY ti.group_key
-HAVING COUNT(DISTINCT CASE WHEN ac.text != '' THEN LOWER(TRIM(ac.text)) END) > 1
+HAVING COUNT(DISTINCT CASE WHEN af.artist_credit != '' THEN LOWER(TRIM(af.artist_credit)) END) > 1
   AND COUNT(DISTINCT CASE WHEN rg.name != '' THEN LOWER(TRIM(rg.name)) END) > 1;
 -- name: CountPendingTaggingItems :one
-SELECT COUNT(*) FROM tagging_items
+-- "Needs tagging" is a question about the files, not about the row:
-WHERE status = 'pending'
+-- every scanned folder gets a tagging_items row (see
-  AND (CAST(@library_id AS INTEGER) = 0 OR library_id = @library_id);
+-- UpsertTaggingItemOnTrackAdd), including one whose files all arrived
 -- carrying a recording MBID.  Without the EXISTS a fully MB-tagged
 -- library reports its entire album count as pending work.  See the
 -- same predicate on the three list queries below.
 SELECT COUNT(*) FROM tagging_items ti
 WHERE ti.status = 'pending'
  AND (CAST(@library_id AS INTEGER) = 0 OR ti.library_id = @library_id)
  AND EXISTS (
    SELECT 1 FROM audio_files af
    WHERE af.group_key = ti.group_key AND af.tag_status = 'untagged'
  );
 -- name: ListPendingTaggingItemsAlphabetical :many
 SELECT
@@ -125,6 +132,18 @@ LEFT JOIN libraries lb ON lb.id = ti.library_id
 WHERE (CAST(@library_id AS INTEGER) = 0 OR ti.library_id = @library_id)
  AND (CAST(@status_filter AS TEXT) = 'all' OR ti.status = @status_filter)
  AND ti.cleared_at IS NULL
  -- Actionable rows must have something to act on: see
  -- CountPendingTaggingItems.  Reviewed rows (confirmed/skipped) are
  -- exempt because they are history, not work -- an applied folder is
  -- fully tagged by definition and would otherwise vanish from the
  -- sidebar's Completed section the instant it succeeded.
  AND (
    ti.status IN ('confirmed', 'skipped')
    OR EXISTS (
      SELECT 1 FROM audio_files af
      WHERE af.group_key = ti.group_key AND af.tag_status = 'untagged'
    )
  )
 ORDER BY LOWER(ti.album_artist), LOWER(ti.album_name), ti.disc_number
 LIMIT @row_limit OFFSET @row_offset;
@@ -150,6 +169,14 @@ LEFT JOIN libraries lb ON lb.id = ti.library_id
 WHERE (CAST(@library_id AS INTEGER) = 0 OR ti.library_id = @library_id)
  AND (CAST(@status_filter AS TEXT) = 'all' OR ti.status = @status_filter)
  AND ti.cleared_at IS NULL
  -- See ListPendingTaggingItemsAlphabetical.
  AND (
    ti.status IN ('confirmed', 'skipped')
    OR EXISTS (
      SELECT 1 FROM audio_files af
      WHERE af.group_key = ti.group_key AND af.tag_status = 'untagged'
    )
  )
 ORDER BY ti.score IS NULL, ti.score DESC, LOWER(ti.album_artist), LOWER(ti.album_name)
 LIMIT @row_limit OFFSET @row_offset;
@@ -204,61 +231,53 @@ ORDER BY ti.created_at DESC, ti.group_key
 LIMIT @row_limit OFFSET @row_offset;
 -- name: ListAudioFilesInTaggingGroup :many
-- album_name/album_artist are the PER-TRACK tags (via each track's
+-- album_name/album_artist are the PER-TRACK tags (each file's own
-- own release_group link), not the folder-level tagging_items
+-- album link), not the folder-level tagging_items values.
-- values.  SplitMixedFolder clusters on these to find sub-albums
+-- SplitMixedFolder clusters on these to find sub-albums hiding inside
-- hiding inside a folder full of unrelated tracks.
+-- a folder full of unrelated tracks.
 SELECT
  af.id,
  af.file_path,
  af.basename,
  af.length_milliseconds,
  af.tag_status,
-  COALESCE(r.track_number, 0) AS track_number,
+  COALESCE(af.track_number, 0) AS track_number,
-  COALESCE(r.disc_number, 0) AS disc_number,
+  COALESCE(af.disc_number, 0) AS disc_number,
-  COALESCE(r.name, '') AS title,
+  af.title,
-  COALESCE(ac.text, '') AS artist_name,
+  af.artist_credit AS artist_name,
-  COALESCE(r.mbid, '') AS recording_mbid,
+  COALESCE(af.recording_mbid, '') AS recording_mbid,
-  COALESCE(rg.name, '') AS album_name,
+  COALESCE(al.name, '') AS album_name,
-  COALESCE(rgac.text, '') AS album_artist
+  COALESCE(al.artist_credit, '') AS album_artist
 FROM audio_files af
-LEFT JOIN recordings r ON af.recording_id = r.id
+LEFT JOIN albums al ON al.id = af.album_id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN release_group_recordings rgr ON rgr.recording_id = r.id
 LEFT JOIN release_groups rg ON rg.id = rgr.release_group_id
 LEFT JOIN artist_credit rgac ON rg.album_artist_credit_id = rgac.id
 WHERE af.group_key = ?
-ORDER BY COALESCE(r.disc_number, 0),
+ORDER BY COALESCE(af.disc_number, 0),
-         COALESCE(r.track_number, 0),
+         COALESCE(af.track_number, 0),
         af.file_path;
-- name: ListLocalReleaseGroupCandidates :many
+-- name: ListLocalAlbumCandidates :many
-- Returns one row per (release_group, track) combination for any
+-- One row per (album, track) for any local album carrying an MBID.
-- local release_group that has an MBID.  Callers group these in Go
+-- Callers group these in Go and filter by normalized album-name match;
-- and filter by normalized album-name match.  Joined case-insensitive
+-- the join is case-insensitive on name to pre-filter cheaply.
 -- on name to pre-filter cheaply; Go does the real normalization.
 SELECT
-  rg.id AS release_group_id,
+  al.id AS album_id,
-  rg.mbid AS release_group_mbid,
+  al.mbid AS album_mbid,
-  rg.name AS album_name,
+  al.name AS album_name,
-  COALESCE(rg.year, 0) AS year,
+  COALESCE(al.year, 0) AS year,
-  COALESCE(ac.text, '') AS artist_credit,
+  al.artist_credit,
-  COALESCE(rgr.track_number, 0) AS track_number,
+  COALESCE(af.track_number, 0) AS track_number,
-  COALESCE(rgr.disc_number, 0) AS disc_number,
+  COALESCE(af.disc_number, 0) AS disc_number,
-  COALESCE(r.name, '') AS track_title,
+  af.title AS track_title,
-  COALESCE(r.mbid, '') AS recording_mbid,
+  COALESCE(af.recording_mbid, '') AS recording_mbid,
-  COALESCE(local_af.length_milliseconds, 0) AS length_milliseconds
+  af.length_milliseconds
-FROM release_groups rg
+FROM albums al
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = al.id
-JOIN recordings r ON r.id = rgr.recording_id
+WHERE al.mbid IS NOT NULL
-LEFT JOIN artist_credit ac ON rg.album_artist_credit_id = ac.id
+  AND al.mbid != ''
-LEFT JOIN audio_files local_af ON local_af.recording_id = r.id
+  AND af.recording_mbid IS NOT NULL
-WHERE rg.mbid IS NOT NULL
+  AND af.recording_mbid != ''
-  AND rg.mbid != ''
+  AND al.name = ? COLLATE NOCASE
-  AND r.mbid IS NOT NULL
+ORDER BY al.id, af.disc_number, af.track_number;
  AND r.mbid != ''
  AND rg.name = ? COLLATE NOCASE
 ORDER BY rg.id, rgr.disc_number, rgr.track_number;
 -- name: SetTaggingItemBestMatch :exec
 UPDATE tagging_items
@@ -284,17 +303,16 @@ WHERE group_key = ?;
 -- name: SetAudioFileTagStatus :exec
 UPDATE audio_files SET tag_status = ? WHERE id = ?;
-- name: SetRecordingMBID :exec
+-- name: SetFileRecordingMBID :exec
-UPDATE recordings SET mbid = ? WHERE id = ?;
+UPDATE audio_files SET recording_mbid = ? WHERE id = ?;
-- name: SetReleaseGroupMBID :exec
+-- name: SetFileAlbumMBID :exec
-UPDATE release_groups SET mbid = ? WHERE id = ?;
+-- The album MBID for the album a file belongs to.  Keyed by file
-
+-- because that is what the autotag apply path holds; under the old
-- name: GetRecordingReleaseGroupID :one
+-- schema it had to look the release group up through two join tables
-SELECT COALESCE(rgr.release_group_id, 0) AS release_group_id
+-- first (GetRecordingReleaseGroupID), which is gone.
-FROM release_group_recordings rgr
+UPDATE albums SET mbid = ?
-WHERE rgr.recording_id = ?
+WHERE albums.id = (SELECT af.album_id FROM audio_files af WHERE af.id = ?);
 LIMIT 1;
 -- name: GetNextPendingTaggingItem :one
 SELECT
@@ -315,5 +333,12 @@ LEFT JOIN libraries lb ON lb.id = ti.library_id
 WHERE ti.status = 'pending'
  AND (CAST(@library_id AS INTEGER) = 0 OR ti.library_id = @library_id)
  AND ti.group_key > @after_group_key
  -- See CountPendingTaggingItems: the cursor must not stop on a
  -- folder the list query no longer shows, or "next" walks folders
  -- that are not in the sidebar.
  AND EXISTS (
    SELECT 1 FROM audio_files af
    WHERE af.group_key = ti.group_key AND af.tag_status = 'untagged'
  )
 ORDER BY ti.group_key
 LIMIT 1;
@@ -0,0 +1,43 @@
 -- One row per album in the library.
 --
 -- This is `release_groups` renamed, and the rename is the point: a
 -- release group is a *MusicBrainz* concept and the catalog still has
 -- them (`explore_index.entity_type = 'release_group'`).  What this
 -- table holds is the local thing — the album some files on disk belong
 -- to — which may or may not have a catalog counterpart.  Calling both
 -- of them "release group" is most of why "is this album mine" was a
 -- question three different subsystems answered three different ways.
 --
 -- `artist_credit` is the album artist as tagged ("Various Artists",
 -- "A & B"); `artist_id` is the primary artist it resolves to.  Album
 -- identity is (name, artist_credit), which is what the old
 -- UNIQUE(name, album_artist_credit_id) meant with a join in the way.
 CREATE TABLE IF NOT EXISTS albums (
  id                   INTEGER PRIMARY KEY,
  name                 TEXT NOT NULL,
  artist_credit        TEXT NOT NULL DEFAULT '',
  artist_id            INTEGER,
  mbid                 TEXT,
  -- year is the tagged year of the copy on disk; original_year is
  -- MusicBrainz's first-release date when known.  For a 2010 remaster
  -- of a 1973 album: original_year 1973, year 2010.
  year                 INTEGER,
  original_year        INTEGER,
  cover_art_id         INTEGER,
  -- Set when the files carried a release MBID but no release-group
  -- MBID; a background pass resolves it and clears this.
  pending_release_mbid TEXT,
  FOREIGN KEY(cover_art_id) REFERENCES cover_art(id),
  FOREIGN KEY(artist_id)    REFERENCES artists(id),
  UNIQUE(name, artist_credit)
 );
 CREATE INDEX IF NOT EXISTS idx_albums_artist_id
  ON albums(artist_id);
 CREATE INDEX IF NOT EXISTS idx_albums_cover_art_id
  ON albums(cover_art_id);
 CREATE INDEX IF NOT EXISTS idx_albums_mbid
  ON albums(mbid) WHERE mbid IS NOT NULL;
@@ -1,4 +0,0 @@
 CREATE TABLE IF NOT EXISTS artist_credit (
  id   INTEGER PRIMARY KEY,
  text TEXT NOT NULL UNIQUE
 );
@@ -1,16 +0,0 @@
 CREATE TABLE IF NOT EXISTS artist_credit_artist (
  id   integer PRIMARY KEY,
  artist_id int NOT NULL,
  credit_id int NOT NULL,
  FOREIGN KEY(artist_id) REFERENCES artists(id),
  FOREIGN KEY(credit_id) REFERENCES artist_credit(id)
 );
 CREATE INDEX IF NOT EXISTS idx_artist_credit_artist_artist_id
    ON artist_credit_artist(artist_id);
 CREATE INDEX IF NOT EXISTS idx_artist_credit_artist_credit_id
    ON artist_credit_artist(credit_id);
 CREATE UNIQUE INDEX IF NOT EXISTS idx_artist_credit_artist_unique
  ON artist_credit_artist(artist_id, credit_id);
@@ -12,8 +12,8 @@ CREATE TABLE IF NOT EXISTS artist_images (
  created_at  DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
    );
-CREATE INDEX IF NOT EXISTS idx_artist_images_mbid
+-- No index on artist_mbid alone: the UNIQUE index below has it as its
-    ON artist_images(artist_mbid);
+-- leftmost column.
 CREATE UNIQUE INDEX IF NOT EXISTS idx_artist_images_source
    ON artist_images(artist_mbid, source, source_url);
@@ -12,4 +12,6 @@ CREATE TABLE IF NOT EXISTS artist_metadata (
    PRIMARY KEY (mbid, source)
 );
-CREATE INDEX IF NOT EXISTS idx_artist_metadata_mbid ON artist_metadata(mbid);
+-- No index on mbid alone: PRIMARY KEY (mbid, source) already has it as
 -- its leftmost column, so a second one costs a write per row and serves
 -- no read.
@@ -1,34 +1,92 @@
 -- One row per audio file, and the file's tags live on it.
 --
 -- This table used to be a stub — path, format, a foreign key — with
 -- every tag-derived field one join away in `recordings`, which was in
 -- turn linked to an album through `release_group_recordings` and to an
 -- artist through `artist_credit` + `artist_credit_artist`.  That is
 -- MusicBrainz's data model, and it is the right model for MusicBrainz:
 -- a recording really can appear on many releases and a credit really
 -- can list many artists.
 --
 -- It was the wrong model here, and the library said so.  Measured on a
 -- real 25,966-file library: **no** recording had more than one file,
 -- **no** recording belonged to more than one release group, and 3 of
 -- 2,823 credits listed more than one artist.  Every many-to-many the
 -- schema modelled was 1:1 in the data, and the cost of modelling it
 -- anyway was a six-way join in every read, a `MIN(release_group_id)`
 -- subquery in eleven queries to collapse a fan-out that never happened,
 -- a first-credited-artist subquery in nine more to collapse the other
 -- one, and — the reason this changed — a whole class of bugs where a
 -- `recordings` row **outlived the file that created it**.  Retagging a
 -- file created a new recording and abandoned the old one, so the same
 -- library carried 812 recordings, 216 release groups and 260 artists
 -- with no file behind them, and everything that asked "do I own this"
 -- by looking for a metadata row got 129 confident yeses for tracks
 -- that could not be played.
 --
 -- With the tags on the file, ownership is not a rule anyone can forget:
 -- the row *is* the file.
 CREATE TABLE IF NOT EXISTS audio_files (
-  id integer PRIMARY KEY,
+  id                 INTEGER PRIMARY KEY,
-  file_path text NOT NULL UNIQUE,
+  file_path          TEXT NOT NULL UNIQUE,
-  length_milliseconds int NOT NULL,
+  library_id         INTEGER NOT NULL DEFAULT 0,
-  file_type_id int NOT NULL,
+  file_type_id       INTEGER NOT NULL,
-  recording_id int NOT NULL,
+
-  sample_rate int NOT NULL DEFAULT 0,
+  -- Audio properties, read from the file itself.
-  bit_depth int NOT NULL DEFAULT 0,
+  length_milliseconds INTEGER NOT NULL,
-  channels int NOT NULL DEFAULT 0,
+  sample_rate        INTEGER NOT NULL DEFAULT 0,
-  bitrate int NOT NULL DEFAULT 0,
+  bit_depth          INTEGER NOT NULL DEFAULT 0,
-  file_size int NOT NULL DEFAULT 0,
+  channels           INTEGER NOT NULL DEFAULT 0,
-  basename text NOT NULL DEFAULT '',
+  bitrate            INTEGER NOT NULL DEFAULT 0,
-  library_id int NOT NULL DEFAULT 0,
+  file_size          INTEGER NOT NULL DEFAULT 0,
-  play_count int NOT NULL DEFAULT 0,
+
-  last_played datetime,
+  -- Tags.  `artist_credit` is the credit as tagged ("A feat. B") and is
-  tag_status TEXT NOT NULL DEFAULT 'untagged'
+  -- for display; `artist_id` is the primary artist it resolves to, and
  -- is what grouping, browsing and the artist page use.  Keeping both
  -- is what makes the credit table unnecessary: the string is the only
  -- thing that was ever read off it.
  title              TEXT NOT NULL DEFAULT '',
  artist_credit      TEXT NOT NULL DEFAULT '',
  artist_id          INTEGER,
  album_id           INTEGER,
  track_number       INTEGER,
  disc_number        INTEGER,
  -- The denominator the tag declared: the 12 in "5/12", per disc.  It
  -- is what lets "do I have all of this album" be answered from disk
  -- instead of from MusicBrainz.  NULL means the tag did not say, which
  -- is a third state and not the same as zero.
  total_tracks       INTEGER,
  year               INTEGER,
  composer           TEXT NOT NULL DEFAULT '',
  comment            TEXT NOT NULL DEFAULT '',
  recording_mbid     TEXT,
  -- Library bookkeeping.
  basename           TEXT NOT NULL DEFAULT '',
  group_key          TEXT NOT NULL DEFAULT '',
  -- File mtime as a Unix timestamp in seconds, captured at import and
  -- compared against the on-disk mtime during a scan to detect files
  -- another application retagged in place.
  modified_at        INTEGER NOT NULL DEFAULT 0,
  play_count         INTEGER NOT NULL DEFAULT 0,
  last_played        DATETIME,
  tag_status         TEXT NOT NULL DEFAULT 'untagged'
    CHECK(tag_status IN (
      'untagged', 'auto_matched', 'user_confirmed', 'user_skipped_permanent'
    )),
-  group_key TEXT NOT NULL DEFAULT '',
+
  -- File mtime as a Unix timestamp in seconds, captured at import.
  -- Compared against the on-disk mtime during a scan to detect files
  -- another application retagged in place.  0 means "never recorded"
  -- (rows predating migration 47) and is treated as not-stale so an
  -- upgrade does not re-import the whole library.
  modified_at int NOT NULL DEFAULT 0,
  FOREIGN KEY(file_type_id) REFERENCES file_types(id),
-  FOREIGN KEY(recording_id) REFERENCES recordings(id),
+  FOREIGN KEY(library_id)   REFERENCES libraries(id),
-  FOREIGN KEY(library_id) REFERENCES libraries(id)
+  FOREIGN KEY(artist_id)    REFERENCES artists(id),
  FOREIGN KEY(album_id)     REFERENCES albums(id)
 );
 CREATE INDEX IF NOT EXISTS idx_audio_files_album_id
    ON audio_files(album_id);
 CREATE INDEX IF NOT EXISTS idx_audio_files_artist_id
    ON audio_files(artist_id);
 CREATE INDEX IF NOT EXISTS idx_audio_files_basename
    ON audio_files(basename);
@@ -36,10 +94,17 @@ CREATE INDEX IF NOT EXISTS idx_audio_files_group_key
    ON audio_files(group_key) WHERE group_key != '';
 CREATE INDEX IF NOT EXISTS idx_audio_files_library_id
-  ON audio_files(library_id);
+    ON audio_files(library_id);
-CREATE INDEX IF NOT EXISTS idx_audio_files_recording_id
+-- The ownership question, asked by MBID: "is there a *file* with this
-    ON audio_files(recording_id);
+-- recording MBID".  Nothing may answer it from a metadata table again.
 CREATE INDEX IF NOT EXISTS idx_audio_files_recording_mbid
    ON audio_files(recording_mbid) WHERE recording_mbid IS NOT NULL;
 -- Answers "does this tagging group still contain untagged files" in one
 -- seek per group.  The autotag queue asks it once per row.
 CREATE INDEX IF NOT EXISTS idx_audio_files_untagged_group_key
    ON audio_files(group_key) WHERE tag_status = 'untagged';
 CREATE INDEX IF NOT EXISTS idx_audio_files_tag_status_untagged
    ON audio_files(library_id) WHERE tag_status = 'untagged';
@@ -1,10 +1,25 @@
 -- The downloaded MusicBrainz/ListenBrainz catalog.
 --
 -- MusicBrainz ids are stored as their 16 raw bytes and entity types as
 -- small integers, which is a size decision: on a real 2,052,200-row
 -- catalog those four columns were 220 MB of a 383 MB table and were
 -- carried again in every index keyed on them, and the conversion took
 -- the table and its four indexes from 677 MB to 389 MB.  See
 -- backend/explore/mbid.go, which is the only place that encoding is
 -- known -- everything above it speaks dashed strings and entity names.
 --
 -- The CHECK constraints are what make a mistake loud.  SQLite does not
 -- coerce between TEXT and BLOB, so a query comparing this column
 -- against a 36-character string returns no rows rather than an error;
 -- a *write* of one fails here instead, at the insert that made it.
 CREATE TABLE IF NOT EXISTS explore_index (
    id                       INTEGER PRIMARY KEY AUTOINCREMENT,
-    entity_type              TEXT NOT NULL,
+    entity_type              INTEGER NOT NULL,
-    mbid                     TEXT NOT NULL,
+    mbid                     BLOB NOT NULL CHECK(length(mbid) = 16),
    title                    TEXT NOT NULL,
    artist_name              TEXT NOT NULL,
-    artist_mbid              TEXT NOT NULL,
+    artist_mbid              BLOB NOT NULL
        CHECK(length(artist_mbid) IN (0, 16)),
    aliases                  TEXT NOT NULL DEFAULT '',
    -- Popularity signals, derived from the ListenBrainz listens dump.
@@ -13,7 +28,8 @@ CREATE TABLE IF NOT EXISTS explore_index (
    -- Recording-specific fields.
    duration                 INTEGER NOT NULL DEFAULT 0,
-    caa_release_mbid         TEXT NOT NULL DEFAULT '',
+    caa_release_mbid         BLOB NOT NULL DEFAULT x''
        CHECK(length(caa_release_mbid) IN (0, 16)),
    release_name             TEXT NOT NULL DEFAULT '',
    -- Release-group-specific fields.
@@ -21,6 +37,13 @@ CREATE TABLE IF NOT EXISTS explore_index (
    secondary_types          TEXT NOT NULL DEFAULT '',
    release_date             TEXT NOT NULL DEFAULT '',
    -- How many tracks the release group's canonical release has, so
    -- "do I have all of this" is answerable offline for an album the
    -- library holds no tags for.  Zero means the catalog does not say,
    -- which is the same third state the local answer has -- and is what
    -- every row carries until a central dump build fills it.
    total_tracks             INTEGER NOT NULL DEFAULT 0,
    -- Artist-specific fields.
    artist_type              TEXT NOT NULL DEFAULT '',
    country                  TEXT NOT NULL DEFAULT '',
@@ -46,13 +69,22 @@ CREATE TABLE IF NOT EXISTS explore_index (
    UNIQUE(mbid)
  );
 -- The exact-match tier's two indexes.
 --
 -- Their predicate is the champion set - the popular rows plus whatever
 -- the user owns - and matching it to `ExactMatches`' own WHERE clause is
 -- what makes them small.  They used to say `popularity > 0`, which on a
 -- real 2,052,200-row catalog covered 2,046,645 of them: a full index
 -- wearing a partial index's clothes, 101 MB for the pair.  Narrowed to
 -- the set the tier can actually return, they are 3 MB and the query
 -- plan is unchanged (measured, on that catalog).
 CREATE INDEX IF NOT EXISTS idx_explore_artist_lower
    ON explore_index(LOWER(artist_name))
-    WHERE popularity > 0;
+    WHERE popularity >= 10000 OR in_library = 1;
 CREATE INDEX IF NOT EXISTS idx_explore_caa_release
    ON explore_index(caa_release_mbid)
-    WHERE entity_type = 'release_group' AND caa_release_mbid != '';
+    WHERE entity_type = 2 AND caa_release_mbid != x'';
 CREATE INDEX IF NOT EXISTS idx_explore_index_artist_mbid
  ON explore_index(artist_mbid, entity_type, popularity DESC);
@@ -62,4 +94,4 @@ CREATE INDEX IF NOT EXISTS idx_explore_index_entity_pop
 CREATE INDEX IF NOT EXISTS idx_explore_title_lower
    ON explore_index(LOWER(title))
-    WHERE popularity > 0;
+    WHERE popularity >= 10000 OR in_library = 1;
@@ -0,0 +1,18 @@
 -- Genres per file.  This is `recording_genres` with the recording taken
 -- out of the middle: it is the one many-to-many in the local library
 -- that is actually many-to-many (a real library runs about four genre
 -- rows per file), which is why it stays a join table when the others
 -- did not.
 CREATE TABLE IF NOT EXISTS file_genres (
  audio_file_id INTEGER NOT NULL,
  genre_id      INTEGER NOT NULL,
  PRIMARY KEY (audio_file_id, genre_id),
  FOREIGN KEY(audio_file_id) REFERENCES audio_files(id) ON DELETE CASCADE,
  FOREIGN KEY(genre_id)      REFERENCES genres(id)
 ) WITHOUT ROWID;
 -- The reverse direction ("which files are in this genre").  The
 -- forward direction is served by the primary key, so — unlike the
 -- table this replaces — there is no third index restating it.
 CREATE INDEX IF NOT EXISTS idx_file_genres_genre_id
    ON file_genres(genre_id);
@@ -1,15 +1,11 @@
-- One row per "go find me this", from the moment the user asks until
+-- One row per folder the user has added as a music library.
 -- the files are in the library or the attempt is abandoned.
 --
-- release_mbid / release_group_mbid are the anchor: a request that
+-- Everything else keyed by library_id means "which of these folders did
-- carries one can be matched against a known tracklist at import time,
+-- this come from"; a library_id of 0 in a query means "all of them".
 -- which is what makes unattended completion safe.  Free-text requests
 -- (both NULL) are always presented to the user for confirmation.
 --
-- `expected` caches the anchor's tracklist as JSON so ranking and
+-- autotag_warning_acked records that the user has been told what
-- import do not have to re-resolve it, and so a request survives the
+-- autotagging will do to the files in this folder, which is a decision
-- explore index being rebuilt underneath it.
+-- they made and not something a rescan can rediscover.
 CREATE TABLE IF NOT EXISTS libraries (
    id                     INTEGER PRIMARY KEY,
@@ -0,0 +1,25 @@
 -- Lyrics for a file, and where they came from.
 --
 -- These used to be a column on `recordings`, in a table classified
 -- `Owned` — data a rescan can rebuild from the files.  That was true of
 -- lyrics read out of a USLT frame and false of lyrics fetched from
 -- LRCLIB, and nothing recorded which was which, so a library with
 -- 24,294 of them could not answer how many were free to rebuild and how
 -- many were network traffic waiting to happen.  `source` answers it.
 --
 -- `recording_mbid` is carried alongside the file id so a future
 -- re-import can re-adopt fetched lyrics without asking LRCLIB again;
 -- the file id is the key because untagged files have no MBID and are
 -- exactly the ones whose lyrics had to be fetched.
 CREATE TABLE IF NOT EXISTS lyrics (
  audio_file_id  INTEGER PRIMARY KEY,
  text           TEXT NOT NULL,
  source         TEXT NOT NULL DEFAULT 'tag'
    CHECK(source IN ('tag', 'lrclib')),
  recording_mbid TEXT,
  fetched_at     DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
  FOREIGN KEY(audio_file_id) REFERENCES audio_files(id) ON DELETE CASCADE
 );
 CREATE INDEX IF NOT EXISTS idx_lyrics_recording_mbid
    ON lyrics(recording_mbid) WHERE recording_mbid IS NOT NULL;
@@ -1,14 +0,0 @@
 CREATE TABLE IF NOT EXISTS recording_genres (
  id           INTEGER PRIMARY KEY,
  recording_id INTEGER NOT NULL,
  genre_id     INTEGER NOT NULL,
  FOREIGN KEY(recording_id) REFERENCES recordings(id),
  FOREIGN KEY(genre_id) REFERENCES genres(id),
  UNIQUE(recording_id, genre_id)
 );
 CREATE INDEX IF NOT EXISTS idx_recording_genres_genre_id
    ON recording_genres(genre_id);
 CREATE INDEX IF NOT EXISTS idx_recording_genres_recording_id
    ON recording_genres(recording_id);
@@ -1,19 +0,0 @@
 CREATE TABLE IF NOT EXISTS recordings (
  id               INTEGER PRIMARY KEY,
  name             TEXT NOT NULL,
  artist_credit_id INTEGER NOT NULL,
  track_number     INTEGER,
  disc_number      INTEGER,
  year             INTEGER,
  genre            TEXT,
  composer         TEXT,
  lyrics           TEXT,
  comment          TEXT,
  mbid             TEXT,
  FOREIGN KEY(artist_credit_id) REFERENCES artist_credit(id)
 );
 CREATE INDEX IF NOT EXISTS idx_recordings_artist_credit_id
    ON recordings(artist_credit_id);
 CREATE INDEX IF NOT EXISTS idx_recordings_mbid ON recordings(mbid) WHERE mbid IS NOT NULL;
@@ -1,21 +0,0 @@
 CREATE TABLE IF NOT EXISTS release_group_recordings (
  id               INTEGER PRIMARY KEY,
  release_group_id INTEGER NOT NULL,
  recording_id     INTEGER NOT NULL,
  track_number     INTEGER,
  disc_number      INTEGER,
  -- The denominator the file's own tag declared: the 12 in "5/12", per
  -- disc.  Read off every file at scan and, until now, discarded — so
  -- "do I have all of this album" had no local answer and the album
  -- page asked MusicBrainz.  NULL means the tag did not say, which is
  -- a third state and not the same as zero.
  total_tracks     INTEGER,
  FOREIGN KEY(release_group_id) REFERENCES release_groups(id),
  FOREIGN KEY(recording_id) REFERENCES recordings(id)
 );
 CREATE INDEX IF NOT EXISTS idx_release_group_recordings_recording_id
    ON release_group_recordings(recording_id);
 CREATE INDEX IF NOT EXISTS idx_release_group_recordings_release_group_id
    ON release_group_recordings(release_group_id);
@@ -1,20 +0,0 @@
 CREATE TABLE IF NOT EXISTS "release_groups" (
  id                     INTEGER PRIMARY KEY,
  name                   TEXT NOT NULL,
  cover_art_id           INTEGER,
  album_artist_credit_id INTEGER,
  year                   INTEGER,
  total_tracks           INTEGER,
  total_discs            INTEGER, mbid TEXT, original_year INTEGER, pending_release_mbid TEXT,
  FOREIGN KEY(cover_art_id) REFERENCES cover_art(id),
  FOREIGN KEY(album_artist_credit_id) REFERENCES artist_credit(id),
  UNIQUE(name, album_artist_credit_id)
    );
 CREATE INDEX IF NOT EXISTS idx_release_groups_album_artist_credit_id
  ON release_groups(album_artist_credit_id);
 CREATE INDEX IF NOT EXISTS idx_release_groups_cover_art_id
  ON release_groups(cover_art_id);
 CREATE INDEX IF NOT EXISTS idx_release_groups_mbid ON release_groups(mbid) WHERE mbid IS NOT NULL;
@@ -1,4 +1,11 @@
 -- Release MBID -> release-group MBID, captured during a dump import.
 --
 -- It is empty on an ordinary install and looks droppable for that
 -- reason: only a local dump build (`indexbuild`) fills it.  The daily
 -- incremental refresh reads it to roll per-release listen counts up to
 -- the release group they belong to, so an install that has built its
 -- own index does need it.
 CREATE TABLE IF NOT EXISTS release_to_rg (
  release_mbid TEXT PRIMARY KEY,
  rg_mbid      TEXT NOT NULL
-    ) WITHOUT ROWID;
+) WITHOUT ROWID;
@@ -6,5 +6,5 @@ CREATE TABLE IF NOT EXISTS similar_artist_map (
  PRIMARY KEY (source_artist_mbid, similar_artist_mbid)
    );
-CREATE INDEX IF NOT EXISTS idx_similar_artist_map_source
+-- No index on source_artist_mbid alone: the PRIMARY KEY has it as its
-    ON similar_artist_map(source_artist_mbid);
+-- leftmost column.
@@ -1,23 +1,42 @@
-CREATE VIEW IF NOT EXISTS track_metadata AS
+-- The one definition of "a track, with everything a list needs".
 --
 -- A view is a definition, not data, so it is dropped and recreated on
 -- every open rather than carrying a migration alongside it: CREATE VIEW
 -- IF NOT EXISTS silently keeps an older database on the old definition,
 -- and a migration file restating it would be the second description of
 -- the schema the migration rules exist to prevent.
 --
 -- This projection used to exist **nine times** — four copies in
 -- audio_files.sql, two in playlists.sql, two in genres.sql, one in
 -- queue.sql — plus this view, which only the raw-SQL search paths used.
 -- They had already drifted: this view preferred the album's
 -- original_year for `year` and GetAllTracksWithFullMetadata used the
 -- track's own, so the same library reported different years on
 -- different screens.  Every query that wants a track row now selects
 -- from here, which is also why there is one row type and one mapper on
 -- the Go side instead of nine and a twenty-two-argument function.
 DROP VIEW IF EXISTS track_metadata;
 CREATE VIEW track_metadata AS
    SELECT
        af.id,
        af.file_path,
        af.length_milliseconds,
-        COALESCE(r.name, '') AS title,
+        af.title,
-        COALESCE(ac.text, '') AS artist_name,
+        af.artist_credit AS artist_name,
-        r.track_number,
+        af.track_number,
-        r.disc_number,
+        af.disc_number,
-        COALESCE(rg.name, '') AS album,
+        COALESCE(al.name, '') AS album,
        CAST(COALESCE(
            (SELECT GROUP_CONCAT(g.name, '||')
-             FROM recording_genres rg_sub
+             FROM file_genres fg
-             JOIN genres g ON rg_sub.genre_id = g.id
+             JOIN genres g ON g.id = fg.genre_id
-             WHERE rg_sub.recording_id = r.id),
+             WHERE fg.audio_file_id = af.id),
            ''
        ) AS TEXT) AS genre,
-        COALESCE(rg.original_year, rg.year, r.year, 0) AS year,
+        COALESCE(al.original_year, al.year, af.year, 0) AS year,
-        COALESCE(rg.year, r.year, 0) AS release_year,
+        COALESCE(al.year, af.year, 0) AS release_year,
-        COALESCE(r.composer, '') AS composer,
+        af.composer,
        COALESCE(ft.extension, '') AS file_type,
        af.sample_rate,
        af.bit_depth,
@@ -28,20 +47,13 @@ CREATE VIEW IF NOT EXISTS track_metadata AS
        af.play_count,
        af.last_played,
        COALESCE(ca.file_path, '') AS cover_art_path,
-        COALESCE(a.mbid, '') AS artist_mbid,
+        COALESCE(ar.mbid, '') AS artist_mbid,
-        COALESCE(rg.mbid, '') AS release_group_mbid,
+        COALESCE(al.mbid, '') AS release_group_mbid,
-        COALESCE(r.mbid, '') AS recording_mbid
+        COALESCE(af.recording_mbid, '') AS recording_mbid,
        af.album_id,
        af.artist_id
    FROM audio_files af
-    LEFT JOIN recordings r ON af.recording_id = r.id
+    LEFT JOIN albums al     ON al.id = af.album_id
-    LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
+    LEFT JOIN artists ar    ON ar.id = af.artist_id
-    LEFT JOIN artist_credit_artist aca ON aca.credit_id = ac.id
+    LEFT JOIN cover_art ca  ON ca.id = al.cover_art_id
-    LEFT JOIN artists a ON a.id = aca.artist_id
+    LEFT JOIN file_types ft ON ft.id = af.file_type_id;
    LEFT JOIN (
        SELECT recording_id,
            MIN(release_group_id) AS release_group_id
        FROM release_group_recordings
        GROUP BY recording_id
    ) rgr ON r.id = rgr.recording_id
    LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
    LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
    LEFT JOIN file_types ft ON af.file_type_id = ft.id;
@@ -0,0 +1,426 @@
 // Code generated by sqlc. DO NOT EDIT.
 // versions:
 //   sqlc v1.30.0
 // source: albums.sql
 package sqlcgen
 import (
 	"context"
 	"database/sql"
 )
 const deleteAlbum = `-- name: DeleteAlbum :exec
 DELETE FROM albums WHERE id = ?
 `
 func (q *Queries) DeleteAlbum(ctx context.Context, id int64) error {
 	_, err := q.db.ExecContext(ctx, deleteAlbum, id)
 	return err
 }
 const deleteAllAlbums = `-- name: DeleteAllAlbums :exec
 DELETE FROM albums
 `
 func (q *Queries) DeleteAllAlbums(ctx context.Context) error {
 	_, err := q.db.ExecContext(ctx, deleteAllAlbums)
 	return err
 }
 const getAlbum = `-- name: GetAlbum :one
 SELECT id, name, artist_credit, artist_id, mbid, year, original_year, cover_art_id, pending_release_mbid FROM albums WHERE id = ? LIMIT 1
 `
 func (q *Queries) GetAlbum(ctx context.Context, id int64) (Album, error) {
 	row := q.db.QueryRowContext(ctx, getAlbum, id)
 	var i Album
 	err := row.Scan(
 		&i.ID,
 		&i.Name,
 		&i.ArtistCredit,
 		&i.ArtistID,
 		&i.Mbid,
 		&i.Year,
 		&i.OriginalYear,
 		&i.CoverArtID,
 		&i.PendingReleaseMbid,
 	)
 	return i, err
 }
 const getAlbumCompleteness = `-- name: GetAlbumCompleteness :one
 SELECT
    -- Distinct (disc, track) pairs: this app detects duplicates, and
    -- counting two files of track 3 twice would report a short album as
    -- complete.  A file with no track number falls back to its own id,
    -- because three untagged files are three tracks, not one.
    CAST(COUNT(DISTINCT CAST(COALESCE(a.disc_number, 1) AS TEXT) || ':' ||
                        COALESCE(CAST(a.track_number AS TEXT), 'f' || a.id)
    ) AS INTEGER) AS owned,
    CAST(COALESCE((
        SELECT SUM(per_disc.total)
        FROM (
            SELECT MAX(b.total_tracks) AS total
            FROM audio_files b
            WHERE b.album_id = ?1 AND b.total_tracks IS NOT NULL
            GROUP BY COALESCE(b.disc_number, 1)
        ) per_disc
    ), 0) AS INTEGER) AS expected,
    CAST((
        SELECT COUNT(*) = 0 FROM audio_files c
        WHERE c.album_id = ?1 AND c.total_tracks IS NULL
    ) AS INTEGER) AS known
 FROM audio_files a
 WHERE a.album_id = ?1
 `
 type GetAlbumCompletenessRow struct {
 	Owned    int64
 	Expected int64
 	Known    int64
 }
 // "Do I have all of this album", answered from the tags on disk.
 //
 // The expectation is a **sum over discs**, not one number: totals are
 // declared per disc ("5/12" on disc 2 means 12 tracks on disc 2), so a
 // multi-disc album's expectation is the sum of each disc's declared
 // total.  A disc whose files declared nothing leaves the whole album
 // unknowable rather than being covered by the discs that did -- which is
 // what `known` reports.
 //
 // Owned counts DISTINCT track numbers: this app detects duplicates, and
 // counting two files of track 3 twice would report a short album as
 // complete.
 func (q *Queries) GetAlbumCompleteness(ctx context.Context, albumID sql.NullInt64) (GetAlbumCompletenessRow, error) {
 	row := q.db.QueryRowContext(ctx, getAlbumCompleteness, albumID)
 	var i GetAlbumCompletenessRow
 	err := row.Scan(&i.Owned, &i.Expected, &i.Known)
 	return i, err
 }
 const getAlbums = `-- name: GetAlbums :many
 SELECT
    al.id,
    al.name,
    COALESCE(al.original_year, al.year) AS year,
    COALESCE(al.year, 0) AS release_year,
    al.mbid,
    al.artist_credit AS artist_name,
    CAST(COALESCE(ar.mbid, '') AS TEXT) AS artist_mbid,
    COALESCE(ca.file_path, '') AS cover_art_path
 FROM albums al
 LEFT JOIN artists ar   ON ar.id = al.artist_id
 LEFT JOIN cover_art ca ON ca.id = al.cover_art_id
 WHERE EXISTS (
    SELECT 1 FROM audio_files af
    WHERE af.album_id = al.id
      AND af.library_id = COALESCE(NULLIF(CAST(?1 AS INTEGER), 0), af.library_id)
 )
 ORDER BY al.name
 `
 type GetAlbumsRow struct {
 	ID           int64
 	Name         string
 	Year         sql.NullInt64
 	ReleaseYear  int64
 	Mbid         sql.NullString
 	ArtistName   string
 	ArtistMbid   string
 	CoverArtPath string
 }
 func (q *Queries) GetAlbums(ctx context.Context, libraryID int64) ([]GetAlbumsRow, error) {
 	rows, err := q.db.QueryContext(ctx, getAlbums, libraryID)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []GetAlbumsRow
 	for rows.Next() {
 		var i GetAlbumsRow
 		if err := rows.Scan(
 			&i.ID,
 			&i.Name,
 			&i.Year,
 			&i.ReleaseYear,
 			&i.Mbid,
 			&i.ArtistName,
 			&i.ArtistMbid,
 			&i.CoverArtPath,
 		); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getAlbumsByArtistName = `-- name: GetAlbumsByArtistName :many
 SELECT
    al.id,
    al.name,
    COALESCE(al.original_year, al.year) AS year,
    COALESCE(al.year, 0) AS release_year,
    al.mbid,
    al.artist_credit AS artist_name,
    CAST(COALESCE(ar.mbid, '') AS TEXT) AS artist_mbid,
    COALESCE(ca.file_path, '') AS cover_art_path
 FROM albums al
 LEFT JOIN artists ar   ON ar.id = al.artist_id
 LEFT JOIN cover_art ca ON ca.id = al.cover_art_id
 WHERE (al.artist_credit = ?1 OR ar.name = ?1)
  AND EXISTS (
      SELECT 1 FROM audio_files af
      WHERE af.album_id = al.id
        AND af.library_id = COALESCE(NULLIF(CAST(?2 AS INTEGER), 0), af.library_id)
  )
 ORDER BY year, al.name
 `
 type GetAlbumsByArtistNameParams struct {
 	Artist    string
 	LibraryID int64
 }
 type GetAlbumsByArtistNameRow struct {
 	ID           int64
 	Name         string
 	Year         sql.NullInt64
 	ReleaseYear  int64
 	Mbid         sql.NullString
 	ArtistName   string
 	ArtistMbid   string
 	CoverArtPath string
 }
 func (q *Queries) GetAlbumsByArtistName(ctx context.Context, arg GetAlbumsByArtistNameParams) ([]GetAlbumsByArtistNameRow, error) {
 	rows, err := q.db.QueryContext(ctx, getAlbumsByArtistName, arg.Artist, arg.LibraryID)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []GetAlbumsByArtistNameRow
 	for rows.Next() {
 		var i GetAlbumsByArtistNameRow
 		if err := rows.Scan(
 			&i.ID,
 			&i.Name,
 			&i.Year,
 			&i.ReleaseYear,
 			&i.Mbid,
 			&i.ArtistName,
 			&i.ArtistMbid,
 			&i.CoverArtPath,
 		); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getAlbumsWithPendingReleaseMBID = `-- name: GetAlbumsWithPendingReleaseMBID :many
 SELECT id, pending_release_mbid FROM albums
 WHERE pending_release_mbid IS NOT NULL AND pending_release_mbid != ''
  AND (mbid IS NULL OR mbid = '')
 `
 type GetAlbumsWithPendingReleaseMBIDRow struct {
 	ID                 int64
 	PendingReleaseMbid sql.NullString
 }
 func (q *Queries) GetAlbumsWithPendingReleaseMBID(ctx context.Context) ([]GetAlbumsWithPendingReleaseMBIDRow, error) {
 	rows, err := q.db.QueryContext(ctx, getAlbumsWithPendingReleaseMBID)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []GetAlbumsWithPendingReleaseMBIDRow
 	for rows.Next() {
 		var i GetAlbumsWithPendingReleaseMBIDRow
 		if err := rows.Scan(&i.ID, &i.PendingReleaseMbid); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getEmptyAlbumIDs = `-- name: GetEmptyAlbumIDs :many
 SELECT id FROM albums al
 WHERE NOT EXISTS (
    SELECT 1 FROM audio_files af WHERE af.album_id = al.id
 )
 `
 // Albums with no file left behind them.  Under the old schema this was
 // one of three orphan sweeps that had to run by hand and did not;
 // audio_files is the only thing that can leave an album empty now, so
 // this is the whole of it.
 func (q *Queries) GetEmptyAlbumIDs(ctx context.Context) ([]int64, error) {
 	rows, err := q.db.QueryContext(ctx, getEmptyAlbumIDs)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []int64
 	for rows.Next() {
 		var id int64
 		if err := rows.Scan(&id); err != nil {
 			return nil, err
 		}
 		items = append(items, id)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const resolveAlbumPendingReleaseMBID = `-- name: ResolveAlbumPendingReleaseMBID :exec
 UPDATE albums
 SET mbid = ?, pending_release_mbid = NULL
 WHERE id = ? AND (mbid IS NULL OR mbid = '')
 `
 type ResolveAlbumPendingReleaseMBIDParams struct {
 	Mbid sql.NullString
 	ID   int64
 }
 // Clears the pending marker once the release-group MBID it stood in for
 // has been resolved.  Guarded so a real MBID is never overwritten.
 func (q *Queries) ResolveAlbumPendingReleaseMBID(ctx context.Context, arg ResolveAlbumPendingReleaseMBIDParams) error {
 	_, err := q.db.ExecContext(ctx, resolveAlbumPendingReleaseMBID, arg.Mbid, arg.ID)
 	return err
 }
 const setAlbumCoverArt = `-- name: SetAlbumCoverArt :exec
 UPDATE albums SET cover_art_id = ? WHERE id = ?
 `
 type SetAlbumCoverArtParams struct {
 	CoverArtID sql.NullInt64
 	ID         int64
 }
 func (q *Queries) SetAlbumCoverArt(ctx context.Context, arg SetAlbumCoverArtParams) error {
 	_, err := q.db.ExecContext(ctx, setAlbumCoverArt, arg.CoverArtID, arg.ID)
 	return err
 }
 const setAlbumMBID = `-- name: SetAlbumMBID :exec
 UPDATE albums SET mbid = ? WHERE id = ?
 `
 type SetAlbumMBIDParams struct {
 	Mbid sql.NullString
 	ID   int64
 }
 func (q *Queries) SetAlbumMBID(ctx context.Context, arg SetAlbumMBIDParams) error {
 	_, err := q.db.ExecContext(ctx, setAlbumMBID, arg.Mbid, arg.ID)
 	return err
 }
 const setAlbumOriginalYear = `-- name: SetAlbumOriginalYear :exec
 UPDATE albums SET original_year = ? WHERE id = ?
 `
 type SetAlbumOriginalYearParams struct {
 	OriginalYear sql.NullInt64
 	ID           int64
 }
 func (q *Queries) SetAlbumOriginalYear(ctx context.Context, arg SetAlbumOriginalYearParams) error {
 	_, err := q.db.ExecContext(ctx, setAlbumOriginalYear, arg.OriginalYear, arg.ID)
 	return err
 }
 const setAlbumPendingReleaseMBID = `-- name: SetAlbumPendingReleaseMBID :exec
 UPDATE albums SET pending_release_mbid = ? WHERE id = ?
 `
 type SetAlbumPendingReleaseMBIDParams struct {
 	PendingReleaseMbid sql.NullString
 	ID                 int64
 }
 func (q *Queries) SetAlbumPendingReleaseMBID(ctx context.Context, arg SetAlbumPendingReleaseMBIDParams) error {
 	_, err := q.db.ExecContext(ctx, setAlbumPendingReleaseMBID, arg.PendingReleaseMbid, arg.ID)
 	return err
 }
 const upsertAlbum = `-- name: UpsertAlbum :one
 INSERT INTO albums (name, artist_credit, artist_id, year, cover_art_id)
 VALUES (?, ?, ?, ?, ?)
 ON CONFLICT(name, artist_credit) DO UPDATE SET
    artist_id    = COALESCE(excluded.artist_id, albums.artist_id),
    year         = COALESCE(excluded.year, albums.year),
    cover_art_id = COALESCE(excluded.cover_art_id, albums.cover_art_id)
 RETURNING id, name, artist_credit, artist_id, mbid, year, original_year, cover_art_id, pending_release_mbid
 `
 type UpsertAlbumParams struct {
 	Name         string
 	ArtistCredit string
 	ArtistID     sql.NullInt64
 	Year         sql.NullInt64
 	CoverArtID   sql.NullInt64
 }
 // Queries over albums (formerly release_groups).
 //
 // The two-copy pattern is gone here too: one query answers both the
 // whole-library and the single-library case.  The `fallback_ac`
 // subquery every album read used to carry -- "if the album has no album
 // artist credit, borrow one from any of its recordings" -- is gone with
 // it, because the album carries its own credit text now.
 func (q *Queries) UpsertAlbum(ctx context.Context, arg UpsertAlbumParams) (Album, error) {
 	row := q.db.QueryRowContext(ctx, upsertAlbum,
 		arg.Name,
 		arg.ArtistCredit,
 		arg.ArtistID,
 		arg.Year,
 		arg.CoverArtID,
 	)
 	var i Album
 	err := row.Scan(
 		&i.ID,
 		&i.Name,
 		&i.ArtistCredit,
 		&i.ArtistID,
 		&i.Mbid,
 		&i.Year,
 		&i.OriginalYear,
 		&i.CoverArtID,
 		&i.PendingReleaseMbid,
 	)
 	return i, err
 }
@@ -1,140 +0,0 @@
 // Code generated by sqlc. DO NOT EDIT.
 // versions:
 //   sqlc v1.30.0
 // source: artist_credit.sql
 package sqlcgen
 import (
 	"context"
 )
 const countArtistCreditReferences = `-- name: CountArtistCreditReferences :one
 SELECT
  (SELECT COUNT(*) FROM recordings WHERE artist_credit_id = ?1) +
  (SELECT COUNT(*) FROM release_groups WHERE album_artist_credit_id = ?1)
 AS total
 `
 func (q *Queries) CountArtistCreditReferences(ctx context.Context, artistCreditID int64) (int64, error) {
 	row := q.db.QueryRowContext(ctx, countArtistCreditReferences, artistCreditID)
 	var total int64
 	err := row.Scan(&total)
 	return total, err
 }
 const createArtistCredit = `-- name: CreateArtistCredit :one
 INSERT INTO artist_credit (text) VALUES (?)
 RETURNING id, text
 `
 func (q *Queries) CreateArtistCredit(ctx context.Context, text string) (ArtistCredit, error) {
 	row := q.db.QueryRowContext(ctx, createArtistCredit, text)
 	var i ArtistCredit
 	err := row.Scan(&i.ID, &i.Text)
 	return i, err
 }
 const deleteAllArtistCredits = `-- name: DeleteAllArtistCredits :exec
 DELETE FROM artist_credit
 `
 func (q *Queries) DeleteAllArtistCredits(ctx context.Context) error {
 	_, err := q.db.ExecContext(ctx, deleteAllArtistCredits)
 	return err
 }
 const deleteArtistCredit = `-- name: DeleteArtistCredit :exec
 DELETE FROM artist_credit 
 WHERE id = ?
 `
 func (q *Queries) DeleteArtistCredit(ctx context.Context, id int64) error {
 	_, err := q.db.ExecContext(ctx, deleteArtistCredit, id)
 	return err
 }
 const getArtistCredit = `-- name: GetArtistCredit :one
 SELECT id, text FROM artist_credit 
 WHERE id = ? LIMIT 1
 `
 func (q *Queries) GetArtistCredit(ctx context.Context, id int64) (ArtistCredit, error) {
 	row := q.db.QueryRowContext(ctx, getArtistCredit, id)
 	var i ArtistCredit
 	err := row.Scan(&i.ID, &i.Text)
 	return i, err
 }
 const getArtistCreditByText = `-- name: GetArtistCreditByText :one
 SELECT id, text FROM artist_credit
 WHERE text = ? LIMIT 1
 `
 func (q *Queries) GetArtistCreditByText(ctx context.Context, text string) (ArtistCredit, error) {
 	row := q.db.QueryRowContext(ctx, getArtistCreditByText, text)
 	var i ArtistCredit
 	err := row.Scan(&i.ID, &i.Text)
 	return i, err
 }
 const getOrphanedArtistCreditIDs = `-- name: GetOrphanedArtistCreditIDs :many
 SELECT ac.id FROM artist_credit ac
 WHERE NOT EXISTS (SELECT 1 FROM recordings r WHERE r.artist_credit_id = ac.id)
  AND NOT EXISTS (SELECT 1 FROM release_groups rg WHERE rg.album_artist_credit_id = ac.id)
 `
 // Artist credits no longer used by any recording or release group - run
 // after orphaned recordings/release groups are deleted, so a credit
 // that only existed for now-removed tracks is cleaned up too.
 func (q *Queries) GetOrphanedArtistCreditIDs(ctx context.Context) ([]int64, error) {
 	rows, err := q.db.QueryContext(ctx, getOrphanedArtistCreditIDs)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []int64
 	for rows.Next() {
 		var id int64
 		if err := rows.Scan(&id); err != nil {
 			return nil, err
 		}
 		items = append(items, id)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const updateArtistCredit = `-- name: UpdateArtistCredit :exec
 UPDATE artist_credit 
 SET text = ?
 WHERE id = ?
 `
 type UpdateArtistCreditParams struct {
 	Text string
 	ID   int64
 }
 func (q *Queries) UpdateArtistCredit(ctx context.Context, arg UpdateArtistCreditParams) error {
 	_, err := q.db.ExecContext(ctx, updateArtistCredit, arg.Text, arg.ID)
 	return err
 }
 const upsertArtistCredit = `-- name: UpsertArtistCredit :one
 INSERT INTO artist_credit (text) VALUES (?)
 ON CONFLICT(text) DO UPDATE SET text = excluded.text
 RETURNING id, text
 `
 func (q *Queries) UpsertArtistCredit(ctx context.Context, text string) (ArtistCredit, error) {
 	row := q.db.QueryRowContext(ctx, upsertArtistCredit, text)
 	var i ArtistCredit
 	err := row.Scan(&i.ID, &i.Text)
 	return i, err
 }
@@ -1,85 +0,0 @@
 // Code generated by sqlc. DO NOT EDIT.
 // versions:
 //   sqlc v1.30.0
 // source: artist_credit_artists.sql
 package sqlcgen
 import (
 	"context"
 )
 const createArtistCreditArtist = `-- name: CreateArtistCreditArtist :one
 INSERT INTO artist_credit_artist (artist_id, credit_id) VALUES (?, ?)
 RETURNING id, artist_id, credit_id
 `
 type CreateArtistCreditArtistParams struct {
 	ArtistID int64
 	CreditID int64
 }
 func (q *Queries) CreateArtistCreditArtist(ctx context.Context, arg CreateArtistCreditArtistParams) (ArtistCreditArtist, error) {
 	row := q.db.QueryRowContext(ctx, createArtistCreditArtist, arg.ArtistID, arg.CreditID)
 	var i ArtistCreditArtist
 	err := row.Scan(&i.ID, &i.ArtistID, &i.CreditID)
 	return i, err
 }
 const deleteAllArtistCreditArtists = `-- name: DeleteAllArtistCreditArtists :exec
 DELETE FROM artist_credit_artist
 `
 func (q *Queries) DeleteAllArtistCreditArtists(ctx context.Context) error {
 	_, err := q.db.ExecContext(ctx, deleteAllArtistCreditArtists)
 	return err
 }
 const deleteArtistCreditArtist = `-- name: DeleteArtistCreditArtist :exec
 DELETE FROM artist_credit_artist 
 WHERE id =?
 `
 func (q *Queries) DeleteArtistCreditArtist(ctx context.Context, id int64) error {
 	_, err := q.db.ExecContext(ctx, deleteArtistCreditArtist, id)
 	return err
 }
 const deleteArtistCreditArtistByCredit = `-- name: DeleteArtistCreditArtistByCredit :exec
 DELETE FROM artist_credit_artist
 WHERE credit_id = ?
 `
 func (q *Queries) DeleteArtistCreditArtistByCredit(ctx context.Context, creditID int64) error {
 	_, err := q.db.ExecContext(ctx, deleteArtistCreditArtistByCredit, creditID)
 	return err
 }
 const getArtistCreditArtist = `-- name: GetArtistCreditArtist :one
 SELECT id, artist_id, credit_id FROM artist_credit_artist 
 WHERE id = ? LIMIT 1
 `
 func (q *Queries) GetArtistCreditArtist(ctx context.Context, id int64) (ArtistCreditArtist, error) {
 	row := q.db.QueryRowContext(ctx, getArtistCreditArtist, id)
 	var i ArtistCreditArtist
 	err := row.Scan(&i.ID, &i.ArtistID, &i.CreditID)
 	return i, err
 }
 const updateArtistCreditArtist = `-- name: UpdateArtistCreditArtist :exec
 UPDATE artist_credit_artist 
 SET artist_id = ?, credit_id = ?
 WHERE id =?
 `
 type UpdateArtistCreditArtistParams struct {
 	ArtistID int64
 	CreditID int64
 	ID       int64
 }
 func (q *Queries) UpdateArtistCreditArtist(ctx context.Context, arg UpdateArtistCreditArtistParams) error {
 	_, err := q.db.ExecContext(ctx, updateArtistCreditArtist, arg.ArtistID, arg.CreditID, arg.ID)
 	return err
 }
@@ -7,20 +7,9 @@ package sqlcgen
 import (
 	"context"
 	"database/sql"
 )
 const createArtist = `-- name: CreateArtist :one
 INSERT INTO artists (name) VALUES (?)
 RETURNING id, name, mbid
 `
 func (q *Queries) CreateArtist(ctx context.Context, name string) (Artist, error) {
 	row := q.db.QueryRowContext(ctx, createArtist, name)
 	var i Artist
 	err := row.Scan(&i.ID, &i.Name, &i.Mbid)
 	return i, err
 }
 const deleteAllArtists = `-- name: DeleteAllArtists :exec
 DELETE FROM artists
 `
@@ -31,8 +20,7 @@ func (q *Queries) DeleteAllArtists(ctx context.Context) error {
 }
 const deleteArtist = `-- name: DeleteArtist :exec
-DELETE FROM artists 
+DELETE FROM artists WHERE id = ?
 WHERE id = ?
 `
 func (q *Queries) DeleteArtist(ctx context.Context, id int64) error {
@@ -43,56 +31,17 @@ func (q *Queries) DeleteArtist(ctx context.Context, id int64) error {
 const getAlbumArtists = `-- name: GetAlbumArtists :many
 SELECT DISTINCT a.id, a.name, a.mbid
 FROM artists a
-JOIN artist_credit_artist aca ON aca.artist_id = a.id
+JOIN albums al ON al.artist_id = a.id
-JOIN artist_credit ac ON ac.id = aca.credit_id
+WHERE EXISTS (
-JOIN release_groups rg ON rg.album_artist_credit_id = ac.id
+    SELECT 1 FROM audio_files af
-ORDER BY a.name
+    WHERE af.album_id = al.id
-`
+      AND af.library_id = COALESCE(NULLIF(CAST(?1 AS INTEGER), 0), af.library_id)
 func (q *Queries) GetAlbumArtists(ctx context.Context) ([]Artist, error) {
 	rows, err := q.db.QueryContext(ctx, getAlbumArtists)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []Artist
 	for rows.Next() {
 		var i Artist
 		if err := rows.Scan(&i.ID, &i.Name, &i.Mbid); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getAlbumArtistsByLibrary = `-- name: GetAlbumArtistsByLibrary :many
 SELECT DISTINCT a.id, a.name, a.mbid
 FROM artists a
 JOIN artist_credit_artist aca ON aca.artist_id = a.id
 JOIN artist_credit ac ON ac.id = aca.credit_id
 JOIN release_groups rg ON rg.album_artist_credit_id = ac.id
 WHERE a.id IN (
    SELECT DISTINCT aca2.artist_id
    FROM artist_credit_artist aca2
    JOIN artist_credit ac2 ON ac2.id = aca2.credit_id
    JOIN release_groups rg2 ON rg2.album_artist_credit_id = ac2.id
    JOIN release_group_recordings rgr2 ON rgr2.release_group_id = rg2.id
    JOIN recordings r2 ON r2.id = rgr2.recording_id
    JOIN audio_files af2 ON af2.recording_id = r2.id
    WHERE af2.library_id = ?
 )
 ORDER BY a.name
 `
-func (q *Queries) GetAlbumArtistsByLibrary(ctx context.Context, libraryID int64) ([]Artist, error) {
+func (q *Queries) GetAlbumArtists(ctx context.Context, libraryID int64) ([]Artist, error) {
-	rows, err := q.db.QueryContext(ctx, getAlbumArtistsByLibrary, libraryID)
+	rows, err := q.db.QueryContext(ctx, getAlbumArtists, libraryID)
 	if err != nil {
 		return nil, err
 	}
@@ -115,8 +64,7 @@ func (q *Queries) GetAlbumArtistsByLibrary(ctx context.Context, libraryID int64)
 }
 const getAllArtists = `-- name: GetAllArtists :many
-SELECT id, name, mbid FROM artists
+SELECT id, name, mbid FROM artists ORDER BY name
 ORDER BY name
 `
 func (q *Queries) GetAllArtists(ctx context.Context) ([]Artist, error) {
@@ -143,8 +91,7 @@ func (q *Queries) GetAllArtists(ctx context.Context) ([]Artist, error) {
 }
 const getArtist = `-- name: GetArtist :one
-SELECT id, name, mbid FROM artists 
+SELECT id, name, mbid FROM artists WHERE id = ? LIMIT 1
 WHERE id = ? LIMIT 1
 `
 func (q *Queries) GetArtist(ctx context.Context, id int64) (Artist, error) {
@@ -154,9 +101,28 @@ func (q *Queries) GetArtist(ctx context.Context, id int64) (Artist, error) {
 	return i, err
 }
 const getArtistByFilePath = `-- name: GetArtistByFilePath :one
 SELECT COALESCE(a.name, '') AS artist_name, COALESCE(a.mbid, '') AS artist_mbid
 FROM audio_files af
 LEFT JOIN artists a ON a.id = af.artist_id
 WHERE af.file_path = ?
 LIMIT 1
 `
 type GetArtistByFilePathRow struct {
 	ArtistName string
 	ArtistMbid string
 }
 func (q *Queries) GetArtistByFilePath(ctx context.Context, filePath string) (GetArtistByFilePathRow, error) {
 	row := q.db.QueryRowContext(ctx, getArtistByFilePath, filePath)
 	var i GetArtistByFilePathRow
 	err := row.Scan(&i.ArtistName, &i.ArtistMbid)
 	return i, err
 }
 const getArtistByName = `-- name: GetArtistByName :one
-SELECT id, name, mbid FROM artists
+SELECT id, name, mbid FROM artists WHERE name = ? LIMIT 1
 WHERE name = ? LIMIT 1
 `
 func (q *Queries) GetArtistByName(ctx context.Context, name string) (Artist, error) {
@@ -166,18 +132,15 @@ func (q *Queries) GetArtistByName(ctx context.Context, name string) (Artist, err
 	return i, err
 }
-const getOrphanedArtistIDs = `-- name: GetOrphanedArtistIDs :many
+const getUnreferencedArtistIDs = `-- name: GetUnreferencedArtistIDs :many
-SELECT a.id FROM artists a
+SELECT id FROM artists a
-WHERE NOT EXISTS (
+WHERE NOT EXISTS (SELECT 1 FROM audio_files af WHERE af.artist_id = a.id)
-    SELECT 1 FROM artist_credit_artist aca WHERE aca.artist_id = a.id
+  AND NOT EXISTS (SELECT 1 FROM albums al WHERE al.artist_id = a.id)
 )
 `
-// Artists no longer credited on any recording or release group - left
+// Artists no file and no album points at any more.
-// behind when a scan's orphan cleanup removes the audio_files that used
+func (q *Queries) GetUnreferencedArtistIDs(ctx context.Context) ([]int64, error) {
-// to justify them, since deleting an audio_files row doesn't cascade.
+	rows, err := q.db.QueryContext(ctx, getUnreferencedArtistIDs)
 func (q *Queries) GetOrphanedArtistIDs(ctx context.Context) ([]int64, error) {
 	rows, err := q.db.QueryContext(ctx, getOrphanedArtistIDs)
 	if err != nil {
 		return nil, err
 	}
@@ -199,30 +162,42 @@ func (q *Queries) GetOrphanedArtistIDs(ctx context.Context) ([]int64, error) {
 	return items, nil
 }
-const updateArtist = `-- name: UpdateArtist :exec
+const setArtistMBID = `-- name: SetArtistMBID :exec
-UPDATE artists 
+UPDATE artists SET mbid = ? WHERE id = ?
 SET name = ?
 WHERE id = ?
 `
-type UpdateArtistParams struct {
+type SetArtistMBIDParams struct {
-	Name string
+	Mbid sql.NullString
 	ID   int64
 }
-func (q *Queries) UpdateArtist(ctx context.Context, arg UpdateArtistParams) error {
+func (q *Queries) SetArtistMBID(ctx context.Context, arg SetArtistMBIDParams) error {
-	_, err := q.db.ExecContext(ctx, updateArtist, arg.Name, arg.ID)
+	_, err := q.db.ExecContext(ctx, setArtistMBID, arg.Mbid, arg.ID)
 	return err
 }
 const upsertArtist = `-- name: UpsertArtist :one
-INSERT INTO artists (name) VALUES (?)
+
-ON CONFLICT(name) DO UPDATE SET name = excluded.name
+INSERT INTO artists (name, mbid) VALUES (?, ?)
 ON CONFLICT(name) DO UPDATE SET
    mbid = COALESCE(excluded.mbid, artists.mbid)
 RETURNING id, name, mbid
 `
-func (q *Queries) UpsertArtist(ctx context.Context, name string) (Artist, error) {
+type UpsertArtistParams struct {
-	row := q.db.QueryRowContext(ctx, upsertArtist, name)
+	Name string
 	Mbid sql.NullString
 }
 // Queries over artists.
 //
 // An artist row is reachable two ways: as a file's primary artist
 // (audio_files.artist_id) and as an album's artist (albums.artist_id).
 // Both used to route through artist_credit + artist_credit_artist,
 // which is how "which album artists are in library 2" came to be a
 // five-join subquery inside a three-join query.
 func (q *Queries) UpsertArtist(ctx context.Context, arg UpsertArtistParams) (Artist, error) {
 	row := q.db.QueryRowContext(ctx, upsertArtist, arg.Name, arg.Mbid)
 	var i Artist
 	err := row.Scan(&i.ID, &i.Name, &i.Mbid)
 	return i, err
@@ -7,36 +7,9 @@ package sqlcgen
 import (
 	"context"
 	"database/sql"
 	"strings"
 )
 const countGenreReferences = `-- name: CountGenreReferences :one
 SELECT COUNT(*) FROM recording_genres WHERE genre_id = ?
 `
 func (q *Queries) CountGenreReferences(ctx context.Context, genreID int64) (int64, error) {
 	row := q.db.QueryRowContext(ctx, countGenreReferences, genreID)
 	var count int64
 	err := row.Scan(&count)
 	return count, err
 }
 const createRecordingGenre = `-- name: CreateRecordingGenre :exec
 INSERT OR IGNORE INTO recording_genres (recording_id, genre_id)
 VALUES (?, ?)
 `
 type CreateRecordingGenreParams struct {
 	RecordingID int64
 	GenreID     int64
 }
 func (q *Queries) CreateRecordingGenre(ctx context.Context, arg CreateRecordingGenreParams) error {
 	_, err := q.db.ExecContext(ctx, createRecordingGenre, arg.RecordingID, arg.GenreID)
 	return err
 }
 const deleteAllGenres = `-- name: DeleteAllGenres :exec
 DELETE FROM genres
 `
@@ -46,12 +19,12 @@ func (q *Queries) DeleteAllGenres(ctx context.Context) error {
 	return err
 }
-const deleteAllRecordingGenres = `-- name: DeleteAllRecordingGenres :exec
+const deleteFileGenres = `-- name: DeleteFileGenres :exec
-DELETE FROM recording_genres
+DELETE FROM file_genres WHERE audio_file_id = ?
 `
-func (q *Queries) DeleteAllRecordingGenres(ctx context.Context) error {
+func (q *Queries) DeleteFileGenres(ctx context.Context, audioFileID int64) error {
-	_, err := q.db.ExecContext(ctx, deleteAllRecordingGenres)
+	_, err := q.db.ExecContext(ctx, deleteFileGenres, audioFileID)
 	return err
 }
@@ -64,20 +37,12 @@ func (q *Queries) DeleteGenre(ctx context.Context, id int64) error {
 	return err
 }
 const deleteRecordingGenres = `-- name: DeleteRecordingGenres :exec
 DELETE FROM recording_genres
 WHERE recording_id = ?
 `
 func (q *Queries) DeleteRecordingGenres(ctx context.Context, recordingID int64) error {
 	_, err := q.db.ExecContext(ctx, deleteRecordingGenres, recordingID)
 	return err
 }
 const getAllGenresWithCounts = `-- name: GetAllGenresWithCounts :many
-SELECT g.name, COUNT(rg.recording_id) AS track_count
+SELECT g.name, COUNT(fg.audio_file_id) AS track_count
 FROM genres g
-JOIN recording_genres rg ON g.id = rg.genre_id
+JOIN file_genres fg ON fg.genre_id = g.id
 JOIN audio_files af ON af.id = fg.audio_file_id
 WHERE af.library_id = COALESCE(NULLIF(CAST(?1 AS INTEGER), 0), af.library_id)
 GROUP BY g.id, g.name
 ORDER BY g.name
 `
@@ -87,8 +52,8 @@ type GetAllGenresWithCountsRow struct {
 	TrackCount int64
 }
-func (q *Queries) GetAllGenresWithCounts(ctx context.Context) ([]GetAllGenresWithCountsRow, error) {
+func (q *Queries) GetAllGenresWithCounts(ctx context.Context, libraryID int64) ([]GetAllGenresWithCountsRow, error) {
-	rows, err := q.db.QueryContext(ctx, getAllGenresWithCounts)
+	rows, err := q.db.QueryContext(ctx, getAllGenresWithCounts, libraryID)
 	if err != nil {
 		return nil, err
 	}
@@ -110,35 +75,25 @@ func (q *Queries) GetAllGenresWithCounts(ctx context.Context) ([]GetAllGenresWit
 	return items, nil
 }
-const getAllGenresWithCountsByLibrary = `-- name: GetAllGenresWithCountsByLibrary :many
+const getGenreNamesByFile = `-- name: GetGenreNamesByFile :many
-SELECT g.name, COUNT(rg.recording_id) AS track_count
+SELECT g.name FROM genres g
-FROM genres g
+JOIN file_genres fg ON fg.genre_id = g.id
-JOIN recording_genres rg ON g.id = rg.genre_id
+WHERE fg.audio_file_id = ?
 JOIN recordings r ON rg.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 WHERE af.library_id = ?
 GROUP BY g.id, g.name
 ORDER BY g.name
 `
-type GetAllGenresWithCountsByLibraryRow struct {
+func (q *Queries) GetGenreNamesByFile(ctx context.Context, audioFileID int64) ([]string, error) {
-	Name       string
+	rows, err := q.db.QueryContext(ctx, getGenreNamesByFile, audioFileID)
 	TrackCount int64
 }
 func (q *Queries) GetAllGenresWithCountsByLibrary(ctx context.Context, libraryID int64) ([]GetAllGenresWithCountsByLibraryRow, error) {
 	rows, err := q.db.QueryContext(ctx, getAllGenresWithCountsByLibrary, libraryID)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
-	var items []GetAllGenresWithCountsByLibraryRow
+	var items []string
 	for rows.Next() {
-		var i GetAllGenresWithCountsByLibraryRow
+		var name string
-		if err := rows.Scan(&i.Name, &i.TrackCount); err != nil {
+		if err := rows.Scan(&name); err != nil {
 			return nil, err
 		}
-		items = append(items, i)
+		items = append(items, name)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
@@ -149,45 +104,42 @@ func (q *Queries) GetAllGenresWithCountsByLibrary(ctx context.Context, libraryID
 	return items, nil
 }
-const getFilePathsByGenres = `-- name: GetFilePathsByGenres :many
+const getGenreNamesByFilePaths = `-- name: GetGenreNamesByFilePaths :many
-
+SELECT af.file_path, g.name
-SELECT g.name AS genre_name, af.file_path
+FROM audio_files af
-FROM genres g
+JOIN file_genres fg ON fg.audio_file_id = af.id
-JOIN recording_genres rg ON g.id = rg.genre_id
+JOIN genres g ON g.id = fg.genre_id
-JOIN recordings r ON rg.recording_id = r.id
+WHERE af.file_path IN (/*SLICE:paths*/?)
 JOIN audio_files af ON af.recording_id = r.id
 WHERE g.name IN (/*SLICE:genre_names*/?)
 ORDER BY r.name
 `
-type GetFilePathsByGenresRow struct {
+type GetGenreNamesByFilePathsRow struct {
-	GenreName string
+	FilePath string
-	FilePath  string
+	Name     string
 }
-// Same as GetFilePathsByReleaseGroups, for "play these genres" (perf.m2):
+// Genres for many files at once.  The mix builder asked this one file
-// one query instead of one per genre, and file paths instead of whole
+// at a time, inside two nested loops -- twelve thousand single-row
-// track rows, which was 6 MB over the IPC for five genres.
+// queries to assemble one mix.
-func (q *Queries) GetFilePathsByGenres(ctx context.Context, genreNames []string) ([]GetFilePathsByGenresRow, error) {
+func (q *Queries) GetGenreNamesByFilePaths(ctx context.Context, paths []string) ([]GetGenreNamesByFilePathsRow, error) {
-	query := getFilePathsByGenres
+	query := getGenreNamesByFilePaths
 	var queryParams []interface{}
-	if len(genreNames) > 0 {
+	if len(paths) > 0 {
-		for _, v := range genreNames {
+		for _, v := range paths {
 			queryParams = append(queryParams, v)
 		}
-		query = strings.Replace(query, "/*SLICE:genre_names*/?", strings.Repeat(",?", len(genreNames))[1:], 1)
+		query = strings.Replace(query, "/*SLICE:paths*/?", strings.Repeat(",?", len(paths))[1:], 1)
 	} else {
-		query = strings.Replace(query, "/*SLICE:genre_names*/?", "NULL", 1)
+		query = strings.Replace(query, "/*SLICE:paths*/?", "NULL", 1)
 	}
 	rows, err := q.db.QueryContext(ctx, query, queryParams...)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
-	var items []GetFilePathsByGenresRow
+	var items []GetGenreNamesByFilePathsRow
 	for rows.Next() {
-		var i GetFilePathsByGenresRow
+		var i GetGenreNamesByFilePathsRow
-		if err := rows.Scan(&i.GenreName, &i.FilePath); err != nil {
+		if err := rows.Scan(&i.FilePath, &i.Name); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
@@ -201,51 +153,24 @@ func (q *Queries) GetFilePathsByGenres(ctx context.Context, genreNames []string)
 	return items, nil
 }
-const getFilePathsByGenresByLibrary = `-- name: GetFilePathsByGenresByLibrary :many
+const getUnusedGenreIDs = `-- name: GetUnusedGenreIDs :many
-SELECT g.name AS genre_name, af.file_path
+SELECT id FROM genres g
-FROM genres g
+WHERE NOT EXISTS (SELECT 1 FROM file_genres fg WHERE fg.genre_id = g.id)
 JOIN recording_genres rg ON g.id = rg.genre_id
 JOIN recordings r ON rg.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 WHERE g.name IN (/*SLICE:genre_names*/?)
  AND af.library_id = ?
 ORDER BY r.name
 `
-type GetFilePathsByGenresByLibraryParams struct {
+func (q *Queries) GetUnusedGenreIDs(ctx context.Context) ([]int64, error) {
-	GenreNames []string
+	rows, err := q.db.QueryContext(ctx, getUnusedGenreIDs)
 	LibraryID  int64
 }
 type GetFilePathsByGenresByLibraryRow struct {
 	GenreName string
 	FilePath  string
 }
 func (q *Queries) GetFilePathsByGenresByLibrary(ctx context.Context, arg GetFilePathsByGenresByLibraryParams) ([]GetFilePathsByGenresByLibraryRow, error) {
 	query := getFilePathsByGenresByLibrary
 	var queryParams []interface{}
 	if len(arg.GenreNames) > 0 {
 		for _, v := range arg.GenreNames {
 			queryParams = append(queryParams, v)
 		}
 		query = strings.Replace(query, "/*SLICE:genre_names*/?", strings.Repeat(",?", len(arg.GenreNames))[1:], 1)
 	} else {
 		query = strings.Replace(query, "/*SLICE:genre_names*/?", "NULL", 1)
 	}
 	queryParams = append(queryParams, arg.LibraryID)
 	rows, err := q.db.QueryContext(ctx, query, queryParams...)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
-	var items []GetFilePathsByGenresByLibraryRow
+	var items []int64
 	for rows.Next() {
-		var i GetFilePathsByGenresByLibraryRow
+		var id int64
-		if err := rows.Scan(&i.GenreName, &i.FilePath); err != nil {
+		if err := rows.Scan(&id); err != nil {
 			return nil, err
 		}
-		items = append(items, i)
+		items = append(items, id)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
@@ -256,247 +181,32 @@ func (q *Queries) GetFilePathsByGenresByLibrary(ctx context.Context, arg GetFile
 	return items, nil
 }
-const getGenresByRecordingID = `-- name: GetGenresByRecordingID :many
+const linkFileGenre = `-- name: LinkFileGenre :exec
-SELECT g.id, g.name
+INSERT OR IGNORE INTO file_genres (audio_file_id, genre_id) VALUES (?, ?)
 FROM genres g
 JOIN recording_genres rg ON g.id = rg.genre_id
 WHERE rg.recording_id = ?
 `
-func (q *Queries) GetGenresByRecordingID(ctx context.Context, recordingID int64) ([]Genre, error) {
+type LinkFileGenreParams struct {
-	rows, err := q.db.QueryContext(ctx, getGenresByRecordingID, recordingID)
+	AudioFileID int64
-	if err != nil {
+	GenreID     int64
 		return nil, err
 	}
 	defer rows.Close()
 	var items []Genre
 	for rows.Next() {
 		var i Genre
 		if err := rows.Scan(&i.ID, &i.Name); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
-const getTracksByGenre = `-- name: GetTracksByGenre :many
+func (q *Queries) LinkFileGenre(ctx context.Context, arg LinkFileGenreParams) error {
-SELECT
+	_, err := q.db.ExecContext(ctx, linkFileGenre, arg.AudioFileID, arg.GenreID)
-    af.file_path,
+	return err
    af.length_milliseconds,
    COALESCE(r.name, '') AS title,
    COALESCE(ac.text, '') AS artist_name,
    r.track_number,
    r.disc_number,
    COALESCE(rlg.name, '') AS album,
    CAST(COALESCE(
        (SELECT GROUP_CONCAT(g2.name, '||')
         FROM recording_genres rg2
         JOIN genres g2 ON rg2.genre_id = g2.id
         WHERE rg2.recording_id = r.id),
        ''
    ) AS TEXT) AS genre,
    COALESCE(r.year, 0) AS year,
    COALESCE(r.composer, '') AS composer,
    COALESCE(ft.extension, '') AS file_type,
    af.sample_rate,
    af.bit_depth,
    af.channels,
    af.bitrate,
    af.file_size
 FROM genres g
 JOIN recording_genres rg ON g.id = rg.genre_id
 JOIN recordings r ON rg.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN (
    SELECT recording_id,
        MIN(release_group_id) AS release_group_id
    FROM release_group_recordings
    GROUP BY recording_id
 ) rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rlg ON rgr.release_group_id = rlg.id
 LEFT JOIN file_types ft ON af.file_type_id = ft.id
 WHERE g.name = ?
 ORDER BY r.name
 `
 type GetTracksByGenreRow struct {
 	FilePath           string
 	LengthMilliseconds int64
 	Title              string
 	ArtistName         string
 	TrackNumber        sql.NullInt64
 	DiscNumber         sql.NullInt64
 	Album              string
 	Genre              string
 	Year               int64
 	Composer           string
 	FileType           string
 	SampleRate         int64
 	BitDepth           int64
 	Channels           int64
 	Bitrate            int64
 	FileSize           int64
 }
 func (q *Queries) GetTracksByGenre(ctx context.Context, name string) ([]GetTracksByGenreRow, error) {
 	rows, err := q.db.QueryContext(ctx, getTracksByGenre, name)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []GetTracksByGenreRow
 	for rows.Next() {
 		var i GetTracksByGenreRow
 		if err := rows.Scan(
 			&i.FilePath,
 			&i.LengthMilliseconds,
 			&i.Title,
 			&i.ArtistName,
 			&i.TrackNumber,
 			&i.DiscNumber,
 			&i.Album,
 			&i.Genre,
 			&i.Year,
 			&i.Composer,
 			&i.FileType,
 			&i.SampleRate,
 			&i.BitDepth,
 			&i.Channels,
 			&i.Bitrate,
 			&i.FileSize,
 		); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getTracksByGenreByLibrary = `-- name: GetTracksByGenreByLibrary :many
 SELECT
    af.file_path,
    af.length_milliseconds,
    COALESCE(r.name, '') AS title,
    COALESCE(ac.text, '') AS artist_name,
    r.track_number,
    r.disc_number,
    COALESCE(rlg.name, '') AS album,
    CAST(COALESCE(
        (SELECT GROUP_CONCAT(g2.name, '||')
         FROM recording_genres rg2
         JOIN genres g2 ON rg2.genre_id = g2.id
         WHERE rg2.recording_id = r.id),
        ''
    ) AS TEXT) AS genre,
    COALESCE(r.year, 0) AS year,
    COALESCE(r.composer, '') AS composer,
    COALESCE(ft.extension, '') AS file_type,
    af.sample_rate,
    af.bit_depth,
    af.channels,
    af.bitrate,
    af.file_size
 FROM genres g
 JOIN recording_genres rg ON g.id = rg.genre_id
 JOIN recordings r ON rg.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN (
    SELECT recording_id,
        MIN(release_group_id) AS release_group_id
    FROM release_group_recordings
    GROUP BY recording_id
 ) rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rlg ON rgr.release_group_id = rlg.id
 LEFT JOIN file_types ft ON af.file_type_id = ft.id
 WHERE g.name = ? AND af.library_id = ?
 ORDER BY r.name
 `
 type GetTracksByGenreByLibraryParams struct {
 	Name      string
 	LibraryID int64
 }
 type GetTracksByGenreByLibraryRow struct {
 	FilePath           string
 	LengthMilliseconds int64
 	Title              string
 	ArtistName         string
 	TrackNumber        sql.NullInt64
 	DiscNumber         sql.NullInt64
 	Album              string
 	Genre              string
 	Year               int64
 	Composer           string
 	FileType           string
 	SampleRate         int64
 	BitDepth           int64
 	Channels           int64
 	Bitrate            int64
 	FileSize           int64
 }
 func (q *Queries) GetTracksByGenreByLibrary(ctx context.Context, arg GetTracksByGenreByLibraryParams) ([]GetTracksByGenreByLibraryRow, error) {
 	rows, err := q.db.QueryContext(ctx, getTracksByGenreByLibrary, arg.Name, arg.LibraryID)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []GetTracksByGenreByLibraryRow
 	for rows.Next() {
 		var i GetTracksByGenreByLibraryRow
 		if err := rows.Scan(
 			&i.FilePath,
 			&i.LengthMilliseconds,
 			&i.Title,
 			&i.ArtistName,
 			&i.TrackNumber,
 			&i.DiscNumber,
 			&i.Album,
 			&i.Genre,
 			&i.Year,
 			&i.Composer,
 			&i.FileType,
 			&i.SampleRate,
 			&i.BitDepth,
 			&i.Channels,
 			&i.Bitrate,
 			&i.FileSize,
 		); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const upsertGenre = `-- name: UpsertGenre :one
 INSERT INTO genres (name) VALUES (?)
-ON CONFLICT(name) DO UPDATE SET name = name
+ON CONFLICT(name) DO UPDATE SET name = excluded.name
 RETURNING id, name
 `
 // Queries over genres and file_genres.
 //
 // The track-returning ones live in audio_files.sql with the rest of the
 // track_metadata reads; what is left here is the genre list itself and
 // the link table's writes.
 func (q *Queries) UpsertGenre(ctx context.Context, name string) (Genre, error) {
 	row := q.db.QueryRowContext(ctx, upsertGenre, name)
 	var i Genre
@@ -12,10 +12,10 @@ import (
 const homeAlbumsByGenre = `-- name: HomeAlbumsByGenre :many
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
-JOIN recording_genres rgen ON rgen.recording_id = rgr.recording_id
+JOIN file_genres fg ON fg.audio_file_id = af.id
-JOIN genres g ON g.id = rgen.genre_id
+JOIN genres g ON g.id = fg.genre_id
 WHERE g.name = ?
 GROUP BY rg.id
 ORDER BY RANDOM()
@@ -54,9 +54,8 @@ func (q *Queries) HomeAlbumsByGenre(ctx context.Context, arg HomeAlbumsByGenrePa
 const homeMostPlayedAlbums = `-- name: HomeMostPlayedAlbums :many
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
 JOIN audio_files af ON af.recording_id = rgr.recording_id
 GROUP BY rg.id
 HAVING SUM(af.play_count) > 0
 ORDER BY SUM(af.play_count) DESC
@@ -89,9 +88,8 @@ func (q *Queries) HomeMostPlayedAlbums(ctx context.Context, limit int64) ([]int6
 const homeRandomAlbums = `-- name: HomeRandomAlbums :many
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
 JOIN audio_files af ON af.recording_id = rgr.recording_id
 GROUP BY rg.id
 ORDER BY RANDOM()
 LIMIT ?
@@ -122,9 +120,8 @@ func (q *Queries) HomeRandomAlbums(ctx context.Context, limit int64) ([]int64, e
 const homeRecentlyAddedAlbums = `-- name: HomeRecentlyAddedAlbums :many
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
 JOIN audio_files af ON af.recording_id = rgr.recording_id
 GROUP BY rg.id
 ORDER BY MAX(af.id) DESC
 LIMIT ?
@@ -159,9 +156,8 @@ func (q *Queries) HomeRecentlyAddedAlbums(ctx context.Context, limit int64) ([]i
 const homeRecentlyPlayedAlbums = `-- name: HomeRecentlyPlayedAlbums :many
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
 JOIN audio_files af ON af.recording_id = rgr.recording_id
 WHERE af.last_played IS NOT NULL
 GROUP BY rg.id
 ORDER BY MAX(af.last_played) DESC
@@ -172,7 +168,7 @@ LIMIT ?
 //
 // Every one of these returns album ids and nothing else.  The display
 // columns (cover art, artist credit, year) already have exactly one
-// correct expression of them, in GetAllAlbumsWithDetails, and a second
+// correct expression of them, in GetAlbums, and a second
 // copy per shelf would be six more places for that to drift.  The home
 // service joins the ids back to that one album list in Go.
 // Albums with the most recent play, newest first.
@@ -201,9 +197,8 @@ func (q *Queries) HomeRecentlyPlayedAlbums(ctx context.Context, limit int64) ([]
 const homeStaleAlbums = `-- name: HomeStaleAlbums :many
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
 JOIN audio_files af ON af.recording_id = rgr.recording_id
 WHERE af.last_played IS NOT NULL
 GROUP BY rg.id
 HAVING MAX(af.last_played) < datetime('now', ?)
@@ -242,14 +237,12 @@ func (q *Queries) HomeStaleAlbums(ctx context.Context, arg HomeStaleAlbumsParams
 const homeTopArtists = `-- name: HomeTopArtists :many
 SELECT
-    COALESCE(ac.text, '') AS artist_name,
+    rg.artist_credit AS artist_name,
    SUM(af.play_count) AS plays
-FROM release_groups rg
+FROM albums rg
-JOIN artist_credit ac ON ac.id = rg.album_artist_credit_id
+JOIN audio_files af ON af.album_id = rg.id
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+WHERE rg.artist_credit <> ''
-JOIN audio_files af ON af.recording_id = rgr.recording_id
+GROUP BY rg.artist_credit
 WHERE ac.text <> ''
 GROUP BY ac.text
 HAVING plays > 0
 ORDER BY plays DESC
 LIMIT ?
@@ -288,10 +281,10 @@ func (q *Queries) HomeTopArtists(ctx context.Context, limit int64) ([]HomeTopArt
 const homeTopGenres = `-- name: HomeTopGenres :many
 SELECT
    g.name AS genre,
-    COUNT(DISTINCT rgr.release_group_id) AS album_count
+    COUNT(DISTINCT af.album_id) AS album_count
 FROM genres g
-JOIN recording_genres rgen ON rgen.genre_id = g.id
+JOIN file_genres fg ON fg.genre_id = g.id
-JOIN release_group_recordings rgr ON rgr.recording_id = rgen.recording_id
+JOIN audio_files af ON af.id = fg.audio_file_id
 GROUP BY g.id
 HAVING album_count >= 3
 ORDER BY album_count DESC
@@ -331,9 +324,8 @@ func (q *Queries) HomeTopGenres(ctx context.Context, limit int64) ([]HomeTopGenr
 const homeUnplayedAlbums = `-- name: HomeUnplayedAlbums :many
 SELECT rg.id AS album_id
-FROM release_groups rg
+FROM albums rg
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = rg.id
 JOIN audio_files af ON af.recording_id = rgr.recording_id
 GROUP BY rg.id
 HAVING SUM(af.play_count) = 0
 ORDER BY RANDOM()
@@ -1,103 +0,0 @@
 // Code generated by sqlc. DO NOT EDIT.
 // versions:
 //   sqlc v1.30.0
 // source: mix.sql
 package sqlcgen
 import (
 	"context"
 	"database/sql"
 )
 const getArtistByFilePath = `-- name: GetArtistByFilePath :one
 SELECT COALESCE(a.name, '') AS artist_name, COALESCE(a.mbid, '') AS artist_mbid
 FROM audio_files af
 JOIN recordings r ON af.recording_id = r.id
 JOIN artist_credit ac ON r.artist_credit_id = ac.id
 JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 JOIN artists a ON a.id = aca.artist_id
 WHERE af.file_path = ?
 LIMIT 1
 `
 type GetArtistByFilePathRow struct {
 	ArtistName string
 	ArtistMbid string
 }
 func (q *Queries) GetArtistByFilePath(ctx context.Context, filePath string) (GetArtistByFilePathRow, error) {
 	row := q.db.QueryRowContext(ctx, getArtistByFilePath, filePath)
 	var i GetArtistByFilePathRow
 	err := row.Scan(&i.ArtistName, &i.ArtistMbid)
 	return i, err
 }
 const getFilePathsByArtistMBID = `-- name: GetFilePathsByArtistMBID :many
 SELECT DISTINCT af.file_path
 FROM audio_files af
 JOIN recordings r ON af.recording_id = r.id
 JOIN artist_credit ac ON r.artist_credit_id = ac.id
 JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 JOIN artists a ON a.id = aca.artist_id
 WHERE a.mbid = ?
 `
 // Queries backing the dynamic-mix queue fallback (backend/explore/mix.go):
 // expanding a seed selection into a candidate pool by artist similarity
 // and genre overlap, restricted to what is actually in the library.
 func (q *Queries) GetFilePathsByArtistMBID(ctx context.Context, mbid sql.NullString) ([]string, error) {
 	rows, err := q.db.QueryContext(ctx, getFilePathsByArtistMBID, mbid)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []string
 	for rows.Next() {
 		var file_path string
 		if err := rows.Scan(&file_path); err != nil {
 			return nil, err
 		}
 		items = append(items, file_path)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getGenreNamesByFilePath = `-- name: GetGenreNamesByFilePath :many
 SELECT DISTINCT g.name
 FROM genres g
 JOIN recording_genres rg ON g.id = rg.genre_id
 JOIN recordings r ON rg.recording_id = r.id
 JOIN audio_files af ON af.recording_id = r.id
 WHERE af.file_path = ?
 `
 func (q *Queries) GetGenreNamesByFilePath(ctx context.Context, filePath string) ([]string, error) {
 	rows, err := q.db.QueryContext(ctx, getGenreNamesByFilePath, filePath)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []string
 	for rows.Next() {
 		var name string
 		if err := rows.Scan(&name); err != nil {
 			return nil, err
 		}
 		items = append(items, name)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
@@ -9,23 +9,24 @@ import (
 	"time"
 )
 type Album struct {
 	ID                 int64
 	Name               string
 	ArtistCredit       string
 	ArtistID           sql.NullInt64
 	Mbid               sql.NullString
 	Year               sql.NullInt64
 	OriginalYear       sql.NullInt64
 	CoverArtID         sql.NullInt64
 	PendingReleaseMbid sql.NullString
 }
 type Artist struct {
 	ID   int64
 	Name string
 	Mbid sql.NullString
 }
 type ArtistCredit struct {
 	ID   int64
 	Text string
 }
 type ArtistCreditArtist struct {
 	ID       int64
 	ArtistID int64
 	CreditID int64
 }
 type ArtistEnrichment struct {
 	ArtistMbid string
 	BrowsedAt  sql.NullTime
@@ -56,21 +57,31 @@ type ArtistMetadatum struct {
 type AudioFile struct {
 	ID                 int64
 	FilePath           string
-	LengthMilliseconds int64
+	LibraryID          int64
 	FileTypeID         int64
-	RecordingID        int64
+	LengthMilliseconds int64
 	SampleRate         int64
 	BitDepth           int64
 	Channels           int64
 	Bitrate            int64
 	FileSize           int64
 	Title              string
 	ArtistCredit       string
 	ArtistID           sql.NullInt64
 	AlbumID            sql.NullInt64
 	TrackNumber        sql.NullInt64
 	DiscNumber         sql.NullInt64
 	TotalTracks        sql.NullInt64
 	Year               sql.NullInt64
 	Composer           string
 	Comment            string
 	RecordingMbid      sql.NullString
 	Basename           string
-	LibraryID          int64
+	GroupKey           string
 	ModifiedAt         int64
 	PlayCount          int64
 	LastPlayed         sql.NullTime
 	TagStatus          string
 	GroupKey           string
 	ModifiedAt         int64
 }
 type CoverArt struct {
@@ -160,20 +171,21 @@ type ExploreChampionFt struct {
 type ExploreIndex struct {
 	ID                  int64
-	EntityType          string
+	EntityType          int64
-	Mbid                string
+	Mbid                []byte
 	Title               string
 	ArtistName          string
-	ArtistMbid          string
+	ArtistMbid          []byte
 	Aliases             string
 	Popularity          int64
 	ListenerCount       int64
 	Duration            int64
-	CaaReleaseMbid      string
+	CaaReleaseMbid      []byte
 	ReleaseName         string
 	PrimaryType         string
 	SecondaryTypes      string
 	ReleaseDate         string
 	TotalTracks         int64
 	ArtistType          string
 	Country             string
 	Disambiguation      string
@@ -197,6 +209,11 @@ type ExploreIndexMetum struct {
 	Value string
 }
 type FileGenre struct {
 	AudioFileID int64
 	GenreID     int64
 }
 type FileType struct {
 	ID        int64
 	Extension string
@@ -231,6 +248,14 @@ type Library struct {
 	AutotagWarningAcked int64
 }
 type Lyric struct {
 	AudioFileID   int64
 	Text          string
 	Source        string
 	RecordingMbid sql.NullString
 	FetchedAt     time.Time
 }
 type LyricsIndex struct {
 	Lyrics string
 }
@@ -291,48 +316,6 @@ type QueueTrack struct {
 	Position    int64
 }
 type Recording struct {
 	ID             int64
 	Name           string
 	ArtistCreditID int64
 	TrackNumber    sql.NullInt64
 	DiscNumber     sql.NullInt64
 	Year           sql.NullInt64
 	Genre          sql.NullString
 	Composer       sql.NullString
 	Lyrics         sql.NullString
 	Comment        sql.NullString
 	Mbid           sql.NullString
 }
 type RecordingGenre struct {
 	ID          int64
 	RecordingID int64
 	GenreID     int64
 }
 type ReleaseGroup struct {
 	ID                  int64
 	Name                string
 	CoverArtID          sql.NullInt64
 	AlbumArtistCreditID sql.NullInt64
 	Year                sql.NullInt64
 	TotalTracks         sql.NullInt64
 	TotalDiscs          sql.NullInt64
 	Mbid                sql.NullString
 	OriginalYear        sql.NullInt64
 	PendingReleaseMbid  sql.NullString
 }
 type ReleaseGroupRecording struct {
 	ID             int64
 	ReleaseGroupID int64
 	RecordingID    int64
 	TrackNumber    sql.NullInt64
 	DiscNumber     sql.NullInt64
 	TotalTracks    sql.NullInt64
 }
 type ReleaseToRg struct {
 	ReleaseMbid string
 	RgMbid      string
@@ -410,4 +393,6 @@ type TrackMetadatum struct {
 	ArtistMbid         string
 	ReleaseGroupMbid   string
 	RecordingMbid      string
 	AlbumID            sql.NullInt64
 	ArtistID           sql.NullInt64
 }
@@ -124,29 +124,18 @@ SELECT
    pt.playlist_id,
    pt.audio_file_id,
    pt.position,
-    COALESCE(af.file_path, '') AS file_path,
+    COALESCE(tm.file_path, '') AS file_path,
-    COALESCE(af.length_milliseconds, 0) AS length_milliseconds,
+    COALESCE(tm.length_milliseconds, 0) AS length_milliseconds,
-    COALESCE(r.name, pt.phantom_title, '') AS title,
+    COALESCE(tm.title, pt.phantom_title, '') AS title,
-    COALESCE(ac.text, pt.phantom_artist, '') AS artist,
+    COALESCE(tm.artist_name, pt.phantom_artist, '') AS artist,
-    COALESCE(rg.name, pt.phantom_album, '') AS album,
+    COALESCE(tm.album, pt.phantom_album, '') AS album,
-    COALESCE(ca.file_path, pt.phantom_cover_art_path, '') AS cover_art_path,
+    COALESCE(NULLIF(tm.cover_art_path, ''), pt.phantom_cover_art_path, '') AS cover_art_path,
    CASE WHEN pt.audio_file_id IS NULL THEN 1 ELSE 0 END AS is_phantom,
-    COALESCE(a.mbid, '') AS artist_mbid,
+    CAST(COALESCE(tm.artist_mbid, '') AS TEXT) AS artist_mbid,
-    COALESCE(rg.mbid, '') AS release_group_mbid,
+    COALESCE(tm.release_group_mbid, '') AS release_group_mbid,
-    COALESCE(r.mbid, '') AS recording_mbid
+    COALESCE(tm.recording_mbid, '') AS recording_mbid
 FROM playlist_tracks pt
-LEFT JOIN audio_files af ON pt.audio_file_id = af.id
+LEFT JOIN track_metadata tm ON tm.id = pt.audio_file_id
 LEFT JOIN recordings r ON af.recording_id = r.id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN artists a ON a.id = aca.artist_id
 LEFT JOIN (
    SELECT recording_id, MIN(release_group_id) AS release_group_id
    FROM release_group_recordings
    GROUP BY recording_id
 ) rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 ORDER BY pt.playlist_id, pt.position
 `
@@ -368,29 +357,18 @@ SELECT
    pt.playlist_id,
    pt.audio_file_id,
    pt.position,
-    COALESCE(af.file_path, '') AS file_path,
+    COALESCE(tm.file_path, '') AS file_path,
-    COALESCE(af.length_milliseconds, 0) AS length_milliseconds,
+    COALESCE(tm.length_milliseconds, 0) AS length_milliseconds,
-    COALESCE(r.name, pt.phantom_title, '') AS title,
+    COALESCE(tm.title, pt.phantom_title, '') AS title,
-    COALESCE(ac.text, pt.phantom_artist, '') AS artist,
+    COALESCE(tm.artist_name, pt.phantom_artist, '') AS artist,
-    COALESCE(rg.name, pt.phantom_album, '') AS album,
+    COALESCE(tm.album, pt.phantom_album, '') AS album,
-    COALESCE(ca.file_path, pt.phantom_cover_art_path, '') AS cover_art_path,
+    COALESCE(NULLIF(tm.cover_art_path, ''), pt.phantom_cover_art_path, '') AS cover_art_path,
    CASE WHEN pt.audio_file_id IS NULL THEN 1 ELSE 0 END AS is_phantom,
-    COALESCE(a.mbid, '') AS artist_mbid,
+    CAST(COALESCE(tm.artist_mbid, '') AS TEXT) AS artist_mbid,
-    COALESCE(rg.mbid, '') AS release_group_mbid,
+    COALESCE(tm.release_group_mbid, '') AS release_group_mbid,
-    COALESCE(r.mbid, '') AS recording_mbid
+    COALESCE(tm.recording_mbid, '') AS recording_mbid
 FROM playlist_tracks pt
-LEFT JOIN audio_files af ON pt.audio_file_id = af.id
+LEFT JOIN track_metadata tm ON tm.id = pt.audio_file_id
 LEFT JOIN recordings r ON af.recording_id = r.id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN artists a ON a.id = aca.artist_id
 LEFT JOIN (
    SELECT recording_id, MIN(release_group_id) AS release_group_id
    FROM release_group_recordings
    GROUP BY recording_id
 ) rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 WHERE pt.playlist_id = ?
 ORDER BY pt.position
 `
@@ -451,30 +429,10 @@ func (q *Queries) GetPlaylistTracksWithMetadata(ctx context.Context, playlistID
 }
 const getTrackPhantomMetadata = `-- name: GetTrackPhantomMetadata :one
-SELECT
+SELECT title, artist_name AS artist, album,
-    COALESCE(r.name, '') AS title,
+       length_milliseconds AS duration_ms, genre, cover_art_path
-    COALESCE(ac.text, '') AS artist,
+FROM track_metadata
-    COALESCE(rg.name, '') AS album,
+WHERE id = ?
    af.length_milliseconds AS duration_ms,
    CAST(COALESCE(
        (SELECT GROUP_CONCAT(g.name, '||')
         FROM recording_genres rg_sub
         JOIN genres g ON rg_sub.genre_id = g.id
         WHERE rg_sub.recording_id = r.id),
        ''
    ) AS TEXT) AS genre,
    COALESCE(ca.file_path, '') AS cover_art_path
 FROM audio_files af
 LEFT JOIN recordings r ON af.recording_id = r.id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN (
    SELECT recording_id, MIN(release_group_id) AS release_group_id
    FROM release_group_recordings
    GROUP BY recording_id
 ) rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 WHERE af.id = ?
 `
 type GetTrackPhantomMetadataRow struct {
@@ -486,6 +444,7 @@ type GetTrackPhantomMetadataRow struct {
 	CoverArtPath string
 }
 // The display fields a playlist row keeps after its file goes away.
 func (q *Queries) GetTrackPhantomMetadata(ctx context.Context, id int64) (GetTrackPhantomMetadataRow, error) {
 	row := q.db.QueryRowContext(ctx, getTrackPhantomMetadata, id)
 	var i GetTrackPhantomMetadataRow
@@ -61,27 +61,11 @@ func (q *Queries) GetQueueTrackCount(ctx context.Context) (int64, error) {
 }
 const getQueueTracks = `-- name: GetQueueTracks :many
-SELECT qt.id, qt.audio_file_id, qt.position, af.file_path,
+SELECT qt.id, qt.audio_file_id, qt.position, tm.file_path,
-    COALESCE(r.name, '') AS title,
+    tm.title, tm.artist_name AS artist, tm.album, tm.cover_art_path,
-    COALESCE(ac.text, '') AS artist,
+    tm.artist_mbid, tm.release_group_mbid, tm.recording_mbid
    COALESCE(rg.name, '') AS album,
    COALESCE(ca.file_path, '') AS cover_art_path,
    COALESCE(a.mbid, '') AS artist_mbid,
    COALESCE(rg.mbid, '') AS release_group_mbid,
    COALESCE(r.mbid, '') AS recording_mbid
 FROM queue_tracks qt
-JOIN audio_files af ON qt.audio_file_id = af.id
+JOIN track_metadata tm ON tm.id = qt.audio_file_id
 LEFT JOIN recordings r ON af.recording_id = r.id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN artists a ON a.id = aca.artist_id
 LEFT JOIN (
    SELECT recording_id, MIN(release_group_id) AS release_group_id
    FROM release_group_recordings
    GROUP BY recording_id
 ) rgr ON r.id = rgr.recording_id
 LEFT JOIN release_groups rg ON rgr.release_group_id = rg.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 ORDER BY qt.position
 `
@@ -99,6 +83,7 @@ type GetQueueTracksRow struct {
 	RecordingMbid    string
 }
 // The queue's rows, joined to the one track projection.
 func (q *Queries) GetQueueTracks(ctx context.Context) ([]GetQueueTracksRow, error) {
 	rows, err := q.db.QueryContext(ctx, getQueueTracks)
 	if err != nil {
@@ -1,268 +0,0 @@
 // Code generated by sqlc. DO NOT EDIT.
 // versions:
 //   sqlc v1.30.0
 // source: recordings.sql
 package sqlcgen
 import (
 	"context"
 	"database/sql"
 )
 const countRecordingsByArtistCredit = `-- name: CountRecordingsByArtistCredit :one
 SELECT COUNT(*) FROM recordings WHERE artist_credit_id = ?
 `
 func (q *Queries) CountRecordingsByArtistCredit(ctx context.Context, artistCreditID int64) (int64, error) {
 	row := q.db.QueryRowContext(ctx, countRecordingsByArtistCredit, artistCreditID)
 	var count int64
 	err := row.Scan(&count)
 	return count, err
 }
 const createRecording = `-- name: CreateRecording :one
 INSERT INTO recordings (name, artist_credit_id) VALUES (?, ?)
 RETURNING id, name, artist_credit_id, track_number, disc_number, year, genre, composer, lyrics, comment, mbid
 `
 type CreateRecordingParams struct {
 	Name           string
 	ArtistCreditID int64
 }
 func (q *Queries) CreateRecording(ctx context.Context, arg CreateRecordingParams) (Recording, error) {
 	row := q.db.QueryRowContext(ctx, createRecording, arg.Name, arg.ArtistCreditID)
 	var i Recording
 	err := row.Scan(
 		&i.ID,
 		&i.Name,
 		&i.ArtistCreditID,
 		&i.TrackNumber,
 		&i.DiscNumber,
 		&i.Year,
 		&i.Genre,
 		&i.Composer,
 		&i.Lyrics,
 		&i.Comment,
 		&i.Mbid,
 	)
 	return i, err
 }
 const createRecordingFull = `-- name: CreateRecordingFull :one
 INSERT INTO recordings (
  name, artist_credit_id, track_number, disc_number,
  year, genre, composer, lyrics, comment
 ) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)
 RETURNING id, name, artist_credit_id, track_number, disc_number, year, genre, composer, lyrics, comment, mbid
 `
 type CreateRecordingFullParams struct {
 	Name           string
 	ArtistCreditID int64
 	TrackNumber    sql.NullInt64
 	DiscNumber     sql.NullInt64
 	Year           sql.NullInt64
 	Genre          sql.NullString
 	Composer       sql.NullString
 	Lyrics         sql.NullString
 	Comment        sql.NullString
 }
 func (q *Queries) CreateRecordingFull(ctx context.Context, arg CreateRecordingFullParams) (Recording, error) {
 	row := q.db.QueryRowContext(ctx, createRecordingFull,
 		arg.Name,
 		arg.ArtistCreditID,
 		arg.TrackNumber,
 		arg.DiscNumber,
 		arg.Year,
 		arg.Genre,
 		arg.Composer,
 		arg.Lyrics,
 		arg.Comment,
 	)
 	var i Recording
 	err := row.Scan(
 		&i.ID,
 		&i.Name,
 		&i.ArtistCreditID,
 		&i.TrackNumber,
 		&i.DiscNumber,
 		&i.Year,
 		&i.Genre,
 		&i.Composer,
 		&i.Lyrics,
 		&i.Comment,
 		&i.Mbid,
 	)
 	return i, err
 }
 const deleteAllRecordings = `-- name: DeleteAllRecordings :exec
 DELETE FROM recordings
 `
 func (q *Queries) DeleteAllRecordings(ctx context.Context) error {
 	_, err := q.db.ExecContext(ctx, deleteAllRecordings)
 	return err
 }
 const deleteRecording = `-- name: DeleteRecording :exec
 DELETE FROM recordings 
 WHERE id = ?
 `
 func (q *Queries) DeleteRecording(ctx context.Context, id int64) error {
 	_, err := q.db.ExecContext(ctx, deleteRecording, id)
 	return err
 }
 const getAllRecordings = `-- name: GetAllRecordings :many
 SELECT id, name, artist_credit_id, track_number, disc_number, year, genre, composer, lyrics, comment, mbid FROM recordings
 ORDER BY name
 `
 func (q *Queries) GetAllRecordings(ctx context.Context) ([]Recording, error) {
 	rows, err := q.db.QueryContext(ctx, getAllRecordings)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []Recording
 	for rows.Next() {
 		var i Recording
 		if err := rows.Scan(
 			&i.ID,
 			&i.Name,
 			&i.ArtistCreditID,
 			&i.TrackNumber,
 			&i.DiscNumber,
 			&i.Year,
 			&i.Genre,
 			&i.Composer,
 			&i.Lyrics,
 			&i.Comment,
 			&i.Mbid,
 		); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getOrphanedRecordingIDs = `-- name: GetOrphanedRecordingIDs :many
 SELECT r.id FROM recordings r
 LEFT JOIN audio_files af ON af.recording_id = r.id
 WHERE af.id IS NULL
 `
 // Recordings no longer backed by any audio_files row - left behind
 // when a scan's orphan cleanup deletes the file that used to own them,
 // since deleting audio_files doesn't cascade to recordings.
 func (q *Queries) GetOrphanedRecordingIDs(ctx context.Context) ([]int64, error) {
 	rows, err := q.db.QueryContext(ctx, getOrphanedRecordingIDs)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []int64
 	for rows.Next() {
 		var id int64
 		if err := rows.Scan(&id); err != nil {
 			return nil, err
 		}
 		items = append(items, id)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getRecording = `-- name: GetRecording :one
 SELECT id, name, artist_credit_id, track_number, disc_number, year, genre, composer, lyrics, comment, mbid FROM recordings 
 WHERE id = ? LIMIT 1
 `
 func (q *Queries) GetRecording(ctx context.Context, id int64) (Recording, error) {
 	row := q.db.QueryRowContext(ctx, getRecording, id)
 	var i Recording
 	err := row.Scan(
 		&i.ID,
 		&i.Name,
 		&i.ArtistCreditID,
 		&i.TrackNumber,
 		&i.DiscNumber,
 		&i.Year,
 		&i.Genre,
 		&i.Composer,
 		&i.Lyrics,
 		&i.Comment,
 		&i.Mbid,
 	)
 	return i, err
 }
 const updateRecording = `-- name: UpdateRecording :exec
 UPDATE recordings 
 SET name = ?, artist_credit_id = ?
 WHERE id = ?
 `
 type UpdateRecordingParams struct {
 	Name           string
 	ArtistCreditID int64
 	ID             int64
 }
 func (q *Queries) UpdateRecording(ctx context.Context, arg UpdateRecordingParams) error {
 	_, err := q.db.ExecContext(ctx, updateRecording, arg.Name, arg.ArtistCreditID, arg.ID)
 	return err
 }
 const updateRecordingFull = `-- name: UpdateRecordingFull :exec
 UPDATE recordings
 SET name = ?, artist_credit_id = ?, track_number = ?, disc_number = ?,
    year = ?, genre = ?, composer = ?, lyrics = ?, comment = ?
 WHERE id = ?
 `
 type UpdateRecordingFullParams struct {
 	Name           string
 	ArtistCreditID int64
 	TrackNumber    sql.NullInt64
 	DiscNumber     sql.NullInt64
 	Year           sql.NullInt64
 	Genre          sql.NullString
 	Composer       sql.NullString
 	Lyrics         sql.NullString
 	Comment        sql.NullString
 	ID             int64
 }
 func (q *Queries) UpdateRecordingFull(ctx context.Context, arg UpdateRecordingFullParams) error {
 	_, err := q.db.ExecContext(ctx, updateRecordingFull,
 		arg.Name,
 		arg.ArtistCreditID,
 		arg.TrackNumber,
 		arg.DiscNumber,
 		arg.Year,
 		arg.Genre,
 		arg.Composer,
 		arg.Lyrics,
 		arg.Comment,
 		arg.ID,
 	)
 	return err
 }
@@ -1,211 +0,0 @@
 // Code generated by sqlc. DO NOT EDIT.
 // versions:
 //   sqlc v1.30.0
 // source: release_group_recordings.sql
 package sqlcgen
 import (
 	"context"
 	"database/sql"
 )
 const createReleaseGroupRecording = `-- name: CreateReleaseGroupRecording :one
 INSERT INTO release_group_recordings (
  release_group_id, recording_id, track_number, disc_number, total_tracks
 )
 VALUES (?, ?, ?, ?, ?)
 RETURNING id, release_group_id, recording_id, track_number, disc_number, total_tracks
 `
 type CreateReleaseGroupRecordingParams struct {
 	ReleaseGroupID int64
 	RecordingID    int64
 	TrackNumber    sql.NullInt64
 	DiscNumber     sql.NullInt64
 	TotalTracks    sql.NullInt64
 }
 func (q *Queries) CreateReleaseGroupRecording(ctx context.Context, arg CreateReleaseGroupRecordingParams) (ReleaseGroupRecording, error) {
 	row := q.db.QueryRowContext(ctx, createReleaseGroupRecording,
 		arg.ReleaseGroupID,
 		arg.RecordingID,
 		arg.TrackNumber,
 		arg.DiscNumber,
 		arg.TotalTracks,
 	)
 	var i ReleaseGroupRecording
 	err := row.Scan(
 		&i.ID,
 		&i.ReleaseGroupID,
 		&i.RecordingID,
 		&i.TrackNumber,
 		&i.DiscNumber,
 		&i.TotalTracks,
 	)
 	return i, err
 }
 const deleteAllReleaseGroupRecordings = `-- name: DeleteAllReleaseGroupRecordings :exec
 DELETE FROM release_group_recordings
 `
 func (q *Queries) DeleteAllReleaseGroupRecordings(ctx context.Context) error {
 	_, err := q.db.ExecContext(ctx, deleteAllReleaseGroupRecordings)
 	return err
 }
 const deleteReleaseGroupRecording = `-- name: DeleteReleaseGroupRecording :exec
 DELETE FROM release_group_recordings 
 WHERE id = ?
 `
 func (q *Queries) DeleteReleaseGroupRecording(ctx context.Context, id int64) error {
 	_, err := q.db.ExecContext(ctx, deleteReleaseGroupRecording, id)
 	return err
 }
 const deleteReleaseGroupRecordingByFK = `-- name: DeleteReleaseGroupRecordingByFK :exec
 DELETE FROM release_group_recordings
 WHERE release_group_id = ? AND recording_id = ?
 `
 type DeleteReleaseGroupRecordingByFKParams struct {
 	ReleaseGroupID int64
 	RecordingID    int64
 }
 func (q *Queries) DeleteReleaseGroupRecordingByFK(ctx context.Context, arg DeleteReleaseGroupRecordingByFKParams) error {
 	_, err := q.db.ExecContext(ctx, deleteReleaseGroupRecordingByFK, arg.ReleaseGroupID, arg.RecordingID)
 	return err
 }
 const deleteReleaseGroupRecordingsByRecording = `-- name: DeleteReleaseGroupRecordingsByRecording :exec
 DELETE FROM release_group_recordings
 WHERE recording_id = ?
 `
 func (q *Queries) DeleteReleaseGroupRecordingsByRecording(ctx context.Context, recordingID int64) error {
 	_, err := q.db.ExecContext(ctx, deleteReleaseGroupRecordingsByRecording, recordingID)
 	return err
 }
 const getAlbumCompleteness = `-- name: GetAlbumCompleteness :one
 WITH discs AS (
    SELECT
        COALESCE(rgr.disc_number, 1) AS disc,
        MAX(COALESCE(rgr.total_tracks, 0)) AS declared,
        COUNT(DISTINCT COALESCE(rgr.track_number, -rgr.recording_id)) AS owned
    FROM release_group_recordings rgr
    WHERE rgr.release_group_id = ?
    GROUP BY COALESCE(rgr.disc_number, 1)
 )
 SELECT
    CAST(COALESCE(SUM(owned), 0) AS INTEGER) AS owned,
    CAST(COALESCE(SUM(declared), 0) AS INTEGER) AS expected,
    CAST(COALESCE(SUM(CASE WHEN declared = 0 THEN 1 ELSE 0 END), 0) AS INTEGER) AS discs_untotalled
 FROM discs
 `
 type GetAlbumCompletenessRow struct {
 	Owned           int64
 	Expected        int64
 	DiscsUntotalled int64
 }
 func (q *Queries) GetAlbumCompleteness(ctx context.Context, releaseGroupID int64) (GetAlbumCompletenessRow, error) {
 	row := q.db.QueryRowContext(ctx, getAlbumCompleteness, releaseGroupID)
 	var i GetAlbumCompletenessRow
 	err := row.Scan(&i.Owned, &i.Expected, &i.DiscsUntotalled)
 	return i, err
 }
 const getRecordingReleaseGroups = `-- name: GetRecordingReleaseGroups :many
 SELECT id, release_group_id, recording_id, track_number, disc_number, total_tracks FROM release_group_recordings
 WHERE recording_id = ?
 `
 func (q *Queries) GetRecordingReleaseGroups(ctx context.Context, recordingID int64) ([]ReleaseGroupRecording, error) {
 	rows, err := q.db.QueryContext(ctx, getRecordingReleaseGroups, recordingID)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []ReleaseGroupRecording
 	for rows.Next() {
 		var i ReleaseGroupRecording
 		if err := rows.Scan(
 			&i.ID,
 			&i.ReleaseGroupID,
 			&i.RecordingID,
 			&i.TrackNumber,
 			&i.DiscNumber,
 			&i.TotalTracks,
 		); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getReleaseGroupRecording = `-- name: GetReleaseGroupRecording :one
 SELECT id, release_group_id, recording_id, track_number, disc_number, total_tracks FROM release_group_recordings 
 WHERE id = ? LIMIT 1
 `
 func (q *Queries) GetReleaseGroupRecording(ctx context.Context, id int64) (ReleaseGroupRecording, error) {
 	row := q.db.QueryRowContext(ctx, getReleaseGroupRecording, id)
 	var i ReleaseGroupRecording
 	err := row.Scan(
 		&i.ID,
 		&i.ReleaseGroupID,
 		&i.RecordingID,
 		&i.TrackNumber,
 		&i.DiscNumber,
 		&i.TotalTracks,
 	)
 	return i, err
 }
 const getReleaseGroupRecordings = `-- name: GetReleaseGroupRecordings :many
 SELECT id, release_group_id, recording_id, track_number, disc_number, total_tracks FROM release_group_recordings
 WHERE release_group_id = ?
 ORDER BY disc_number, track_number
 `
 func (q *Queries) GetReleaseGroupRecordings(ctx context.Context, releaseGroupID int64) ([]ReleaseGroupRecording, error) {
 	rows, err := q.db.QueryContext(ctx, getReleaseGroupRecordings, releaseGroupID)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []ReleaseGroupRecording
 	for rows.Next() {
 		var i ReleaseGroupRecording
 		if err := rows.Scan(
 			&i.ID,
 			&i.ReleaseGroupID,
 			&i.RecordingID,
 			&i.TrackNumber,
 			&i.DiscNumber,
 			&i.TotalTracks,
 		); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
@@ -1,639 +0,0 @@
 // Code generated by sqlc. DO NOT EDIT.
 // versions:
 //   sqlc v1.30.0
 // source: release_groups.sql
 package sqlcgen
 import (
 	"context"
 	"database/sql"
 )
 const countReleaseGroupRecordings = `-- name: CountReleaseGroupRecordings :one
 SELECT COUNT(*) FROM release_group_recordings WHERE release_group_id = ?
 `
 func (q *Queries) CountReleaseGroupRecordings(ctx context.Context, releaseGroupID int64) (int64, error) {
 	row := q.db.QueryRowContext(ctx, countReleaseGroupRecordings, releaseGroupID)
 	var count int64
 	err := row.Scan(&count)
 	return count, err
 }
 const createReleaseGroup = `-- name: CreateReleaseGroup :one
 INSERT INTO release_groups (name) VALUES (?)
 RETURNING id, name, cover_art_id, album_artist_credit_id, year, total_tracks, total_discs, mbid, original_year, pending_release_mbid
 `
 func (q *Queries) CreateReleaseGroup(ctx context.Context, name string) (ReleaseGroup, error) {
 	row := q.db.QueryRowContext(ctx, createReleaseGroup, name)
 	var i ReleaseGroup
 	err := row.Scan(
 		&i.ID,
 		&i.Name,
 		&i.CoverArtID,
 		&i.AlbumArtistCreditID,
 		&i.Year,
 		&i.TotalTracks,
 		&i.TotalDiscs,
 		&i.Mbid,
 		&i.OriginalYear,
 		&i.PendingReleaseMbid,
 	)
 	return i, err
 }
 const createReleaseGroupFull = `-- name: CreateReleaseGroupFull :one
 INSERT INTO release_groups (
  name, cover_art_id, album_artist_credit_id, year, total_tracks, total_discs
 ) VALUES (?, ?, ?, ?, ?, ?)
 RETURNING id, name, cover_art_id, album_artist_credit_id, year, total_tracks, total_discs, mbid, original_year, pending_release_mbid
 `
 type CreateReleaseGroupFullParams struct {
 	Name                string
 	CoverArtID          sql.NullInt64
 	AlbumArtistCreditID sql.NullInt64
 	Year                sql.NullInt64
 	TotalTracks         sql.NullInt64
 	TotalDiscs          sql.NullInt64
 }
 func (q *Queries) CreateReleaseGroupFull(ctx context.Context, arg CreateReleaseGroupFullParams) (ReleaseGroup, error) {
 	row := q.db.QueryRowContext(ctx, createReleaseGroupFull,
 		arg.Name,
 		arg.CoverArtID,
 		arg.AlbumArtistCreditID,
 		arg.Year,
 		arg.TotalTracks,
 		arg.TotalDiscs,
 	)
 	var i ReleaseGroup
 	err := row.Scan(
 		&i.ID,
 		&i.Name,
 		&i.CoverArtID,
 		&i.AlbumArtistCreditID,
 		&i.Year,
 		&i.TotalTracks,
 		&i.TotalDiscs,
 		&i.Mbid,
 		&i.OriginalYear,
 		&i.PendingReleaseMbid,
 	)
 	return i, err
 }
 const deleteAllReleaseGroups = `-- name: DeleteAllReleaseGroups :exec
 DELETE FROM release_groups
 `
 func (q *Queries) DeleteAllReleaseGroups(ctx context.Context) error {
 	_, err := q.db.ExecContext(ctx, deleteAllReleaseGroups)
 	return err
 }
 const deleteReleaseGroup = `-- name: DeleteReleaseGroup :exec
 DELETE FROM release_groups 
 WHERE id = ?
 `
 func (q *Queries) DeleteReleaseGroup(ctx context.Context, id int64) error {
 	_, err := q.db.ExecContext(ctx, deleteReleaseGroup, id)
 	return err
 }
 const getAlbumsByArtist = `-- name: GetAlbumsByArtist :many
 SELECT
    rg.id,
    rg.name,
    COALESCE(rg.original_year, rg.year) AS year,
    COALESCE(rg.year, 0) AS release_year,
    COALESCE(ac.text, fallback_ac.text, '') as artist_name,
    -- primary (first-credited) album artist's MBID, for linking the
    -- artist name to its detail page.  Empty when the album has no
    -- MB-tagged album-artist credit.
    CAST(COALESCE((
        SELECT a.mbid
        FROM artist_credit_artist aca_p
        JOIN artists a ON a.id = aca_p.artist_id
        WHERE aca_p.credit_id = rg.album_artist_credit_id
        ORDER BY aca_p.id
        LIMIT 1
    ), '') AS TEXT) as artist_mbid,
    COALESCE(ca.file_path, '') as cover_art_path
 FROM release_groups rg
 JOIN artist_credit ac ON rg.album_artist_credit_id = ac.id
 JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 LEFT JOIN (
    SELECT rgr.release_group_id, ac2.text
    FROM release_group_recordings rgr
    JOIN recordings rec ON rec.id = rgr.recording_id
    JOIN artist_credit ac2 ON ac2.id = rec.artist_credit_id
    GROUP BY rgr.release_group_id
 ) fallback_ac ON fallback_ac.release_group_id = rg.id
 WHERE aca.artist_id = ?
 ORDER BY rg.name
 `
 type GetAlbumsByArtistRow struct {
 	ID           int64
 	Name         string
 	Year         sql.NullInt64
 	ReleaseYear  int64
 	ArtistName   string
 	ArtistMbid   string
 	CoverArtPath string
 }
 func (q *Queries) GetAlbumsByArtist(ctx context.Context, artistID int64) ([]GetAlbumsByArtistRow, error) {
 	rows, err := q.db.QueryContext(ctx, getAlbumsByArtist, artistID)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []GetAlbumsByArtistRow
 	for rows.Next() {
 		var i GetAlbumsByArtistRow
 		if err := rows.Scan(
 			&i.ID,
 			&i.Name,
 			&i.Year,
 			&i.ReleaseYear,
 			&i.ArtistName,
 			&i.ArtistMbid,
 			&i.CoverArtPath,
 		); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getAlbumsByArtistByLibrary = `-- name: GetAlbumsByArtistByLibrary :many
 SELECT
    rg.id,
    rg.name,
    COALESCE(rg.original_year, rg.year) AS year,
    COALESCE(rg.year, 0) AS release_year,
    COALESCE(ac.text, fallback_ac.text, '') as artist_name,
    -- primary (first-credited) album artist's MBID, for linking the
    -- artist name to its detail page.  Empty when the album has no
    -- MB-tagged album-artist credit.
    CAST(COALESCE((
        SELECT a.mbid
        FROM artist_credit_artist aca_p
        JOIN artists a ON a.id = aca_p.artist_id
        WHERE aca_p.credit_id = rg.album_artist_credit_id
        ORDER BY aca_p.id
        LIMIT 1
    ), '') AS TEXT) as artist_mbid,
    COALESCE(ca.file_path, '') as cover_art_path
 FROM release_groups rg
 JOIN artist_credit ac ON rg.album_artist_credit_id = ac.id
 JOIN artist_credit_artist aca ON aca.credit_id = ac.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 LEFT JOIN (
    SELECT rgr.release_group_id, ac2.text
    FROM release_group_recordings rgr
    JOIN recordings rec ON rec.id = rgr.recording_id
    JOIN artist_credit ac2 ON ac2.id = rec.artist_credit_id
    GROUP BY rgr.release_group_id
 ) fallback_ac ON fallback_ac.release_group_id = rg.id
 WHERE aca.artist_id = ?
  AND rg.id IN (
    SELECT DISTINCT rgr2.release_group_id
    FROM release_group_recordings rgr2
    JOIN recordings r2 ON r2.id = rgr2.recording_id
    JOIN audio_files af2 ON af2.recording_id = r2.id
    WHERE af2.library_id = ?
 )
 ORDER BY rg.name
 `
 type GetAlbumsByArtistByLibraryParams struct {
 	ArtistID  int64
 	LibraryID int64
 }
 type GetAlbumsByArtistByLibraryRow struct {
 	ID           int64
 	Name         string
 	Year         sql.NullInt64
 	ReleaseYear  int64
 	ArtistName   string
 	ArtistMbid   string
 	CoverArtPath string
 }
 func (q *Queries) GetAlbumsByArtistByLibrary(ctx context.Context, arg GetAlbumsByArtistByLibraryParams) ([]GetAlbumsByArtistByLibraryRow, error) {
 	rows, err := q.db.QueryContext(ctx, getAlbumsByArtistByLibrary, arg.ArtistID, arg.LibraryID)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []GetAlbumsByArtistByLibraryRow
 	for rows.Next() {
 		var i GetAlbumsByArtistByLibraryRow
 		if err := rows.Scan(
 			&i.ID,
 			&i.Name,
 			&i.Year,
 			&i.ReleaseYear,
 			&i.ArtistName,
 			&i.ArtistMbid,
 			&i.CoverArtPath,
 		); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getAllAlbumsWithDetails = `-- name: GetAllAlbumsWithDetails :many
 SELECT
    rg.id,
    rg.name,
    -- year prefers original release year (MB first-release-date)
    -- over the file-tag year so the UI surfaces the album's
    -- original year by default.  release_year keeps the file-tag
    -- year accessible.
    COALESCE(rg.original_year, rg.year) AS year,
    COALESCE(rg.year, 0) AS release_year,
    rg.mbid,
    COALESCE(ac.text, fallback_ac.text, '') as artist_name,
    -- primary (first-credited) album artist's MBID, for linking the
    -- artist name to its detail page.  Empty when the album has no
    -- MB-tagged album-artist credit.
    CAST(COALESCE((
        SELECT a.mbid
        FROM artist_credit_artist aca_p
        JOIN artists a ON a.id = aca_p.artist_id
        WHERE aca_p.credit_id = rg.album_artist_credit_id
        ORDER BY aca_p.id
        LIMIT 1
    ), '') AS TEXT) as artist_mbid,
    COALESCE(ca.file_path, '') as cover_art_path
 FROM release_groups rg
 LEFT JOIN artist_credit ac ON rg.album_artist_credit_id = ac.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 LEFT JOIN (
    SELECT rgr.release_group_id, ac2.text
    FROM release_group_recordings rgr
    JOIN recordings rec ON rec.id = rgr.recording_id
    JOIN artist_credit ac2 ON ac2.id = rec.artist_credit_id
    GROUP BY rgr.release_group_id
 ) fallback_ac ON fallback_ac.release_group_id = rg.id
 ORDER BY rg.name
 `
 type GetAllAlbumsWithDetailsRow struct {
 	ID           int64
 	Name         string
 	Year         sql.NullInt64
 	ReleaseYear  int64
 	Mbid         sql.NullString
 	ArtistName   string
 	ArtistMbid   string
 	CoverArtPath string
 }
 func (q *Queries) GetAllAlbumsWithDetails(ctx context.Context) ([]GetAllAlbumsWithDetailsRow, error) {
 	rows, err := q.db.QueryContext(ctx, getAllAlbumsWithDetails)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []GetAllAlbumsWithDetailsRow
 	for rows.Next() {
 		var i GetAllAlbumsWithDetailsRow
 		if err := rows.Scan(
 			&i.ID,
 			&i.Name,
 			&i.Year,
 			&i.ReleaseYear,
 			&i.Mbid,
 			&i.ArtistName,
 			&i.ArtistMbid,
 			&i.CoverArtPath,
 		); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getAllAlbumsWithDetailsByLibrary = `-- name: GetAllAlbumsWithDetailsByLibrary :many
 SELECT
    rg.id,
    rg.name,
    -- year prefers original release year (MB first-release-date)
    -- over the file-tag year so the UI surfaces the album's
    -- original year by default.  release_year keeps the file-tag
    -- year accessible.
    COALESCE(rg.original_year, rg.year) AS year,
    COALESCE(rg.year, 0) AS release_year,
    rg.mbid,
    COALESCE(ac.text, fallback_ac.text, '') as artist_name,
    -- primary (first-credited) album artist's MBID, for linking the
    -- artist name to its detail page.  Empty when the album has no
    -- MB-tagged album-artist credit.
    CAST(COALESCE((
        SELECT a.mbid
        FROM artist_credit_artist aca_p
        JOIN artists a ON a.id = aca_p.artist_id
        WHERE aca_p.credit_id = rg.album_artist_credit_id
        ORDER BY aca_p.id
        LIMIT 1
    ), '') AS TEXT) as artist_mbid,
    COALESCE(ca.file_path, '') as cover_art_path
 FROM release_groups rg
 LEFT JOIN artist_credit ac ON rg.album_artist_credit_id = ac.id
 LEFT JOIN cover_art ca ON rg.cover_art_id = ca.id
 LEFT JOIN (
    SELECT rgr.release_group_id, ac2.text
    FROM release_group_recordings rgr
    JOIN recordings rec ON rec.id = rgr.recording_id
    JOIN artist_credit ac2 ON ac2.id = rec.artist_credit_id
    GROUP BY rgr.release_group_id
 ) fallback_ac ON fallback_ac.release_group_id = rg.id
 WHERE rg.id IN (
    SELECT DISTINCT rgr2.release_group_id
    FROM release_group_recordings rgr2
    JOIN recordings r2 ON r2.id = rgr2.recording_id
    JOIN audio_files af2 ON af2.recording_id = r2.id
    WHERE af2.library_id = ?
 )
 ORDER BY rg.name
 `
 type GetAllAlbumsWithDetailsByLibraryRow struct {
 	ID           int64
 	Name         string
 	Year         sql.NullInt64
 	ReleaseYear  int64
 	Mbid         sql.NullString
 	ArtistName   string
 	ArtistMbid   string
 	CoverArtPath string
 }
 func (q *Queries) GetAllAlbumsWithDetailsByLibrary(ctx context.Context, libraryID int64) ([]GetAllAlbumsWithDetailsByLibraryRow, error) {
 	rows, err := q.db.QueryContext(ctx, getAllAlbumsWithDetailsByLibrary, libraryID)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []GetAllAlbumsWithDetailsByLibraryRow
 	for rows.Next() {
 		var i GetAllAlbumsWithDetailsByLibraryRow
 		if err := rows.Scan(
 			&i.ID,
 			&i.Name,
 			&i.Year,
 			&i.ReleaseYear,
 			&i.Mbid,
 			&i.ArtistName,
 			&i.ArtistMbid,
 			&i.CoverArtPath,
 		); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getAllReleaseGroups = `-- name: GetAllReleaseGroups :many
 SELECT id, name, cover_art_id, album_artist_credit_id, year, total_tracks, total_discs, mbid, original_year, pending_release_mbid FROM release_groups
 ORDER BY name
 `
 func (q *Queries) GetAllReleaseGroups(ctx context.Context) ([]ReleaseGroup, error) {
 	rows, err := q.db.QueryContext(ctx, getAllReleaseGroups)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []ReleaseGroup
 	for rows.Next() {
 		var i ReleaseGroup
 		if err := rows.Scan(
 			&i.ID,
 			&i.Name,
 			&i.CoverArtID,
 			&i.AlbumArtistCreditID,
 			&i.Year,
 			&i.TotalTracks,
 			&i.TotalDiscs,
 			&i.Mbid,
 			&i.OriginalYear,
 			&i.PendingReleaseMbid,
 		); err != nil {
 			return nil, err
 		}
 		items = append(items, i)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getOrphanedReleaseGroupIDs = `-- name: GetOrphanedReleaseGroupIDs :many
 SELECT rg.id FROM release_groups rg
 LEFT JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
 WHERE rgr.id IS NULL
 `
 // Release groups with no recordings left in them - run after orphaned
 // recordings (and their release_group_recordings rows) are deleted, so
 // a release group whose last owned track was removed is cleaned up too.
 func (q *Queries) GetOrphanedReleaseGroupIDs(ctx context.Context) ([]int64, error) {
 	rows, err := q.db.QueryContext(ctx, getOrphanedReleaseGroupIDs)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
 	var items []int64
 	for rows.Next() {
 		var id int64
 		if err := rows.Scan(&id); err != nil {
 			return nil, err
 		}
 		items = append(items, id)
 	}
 	if err := rows.Close(); err != nil {
 		return nil, err
 	}
 	if err := rows.Err(); err != nil {
 		return nil, err
 	}
 	return items, nil
 }
 const getReleaseGroup = `-- name: GetReleaseGroup :one
 SELECT id, name, cover_art_id, album_artist_credit_id, year, total_tracks, total_discs, mbid, original_year, pending_release_mbid FROM release_groups 
 WHERE id = ? LIMIT 1
 `
 func (q *Queries) GetReleaseGroup(ctx context.Context, id int64) (ReleaseGroup, error) {
 	row := q.db.QueryRowContext(ctx, getReleaseGroup, id)
 	var i ReleaseGroup
 	err := row.Scan(
 		&i.ID,
 		&i.Name,
 		&i.CoverArtID,
 		&i.AlbumArtistCreditID,
 		&i.Year,
 		&i.TotalTracks,
 		&i.TotalDiscs,
 		&i.Mbid,
 		&i.OriginalYear,
 		&i.PendingReleaseMbid,
 	)
 	return i, err
 }
 const getReleaseGroupByNameAndArtist = `-- name: GetReleaseGroupByNameAndArtist :one
 SELECT id, name, cover_art_id, album_artist_credit_id, year, total_tracks, total_discs, mbid, original_year, pending_release_mbid FROM release_groups
 WHERE name = ? AND album_artist_credit_id = ? LIMIT 1
 `
 type GetReleaseGroupByNameAndArtistParams struct {
 	Name                string
 	AlbumArtistCreditID sql.NullInt64
 }
 func (q *Queries) GetReleaseGroupByNameAndArtist(ctx context.Context, arg GetReleaseGroupByNameAndArtistParams) (ReleaseGroup, error) {
 	row := q.db.QueryRowContext(ctx, getReleaseGroupByNameAndArtist, arg.Name, arg.AlbumArtistCreditID)
 	var i ReleaseGroup
 	err := row.Scan(
 		&i.ID,
 		&i.Name,
 		&i.CoverArtID,
 		&i.AlbumArtistCreditID,
 		&i.Year,
 		&i.TotalTracks,
 		&i.TotalDiscs,
 		&i.Mbid,
 		&i.OriginalYear,
 		&i.PendingReleaseMbid,
 	)
 	return i, err
 }
 const setReleaseGroupOriginalYear = `-- name: SetReleaseGroupOriginalYear :exec
 UPDATE release_groups SET original_year = ? WHERE id = ?
 `
 type SetReleaseGroupOriginalYearParams struct {
 	OriginalYear sql.NullInt64
 	ID           int64
 }
 // Set the release group's original-release-year (release-group's
 // first-release-date from MusicBrainz).  Called from autotag apply
 // when the user confirms a candidate; the file-tag year stays in
 // the year column.
 func (q *Queries) SetReleaseGroupOriginalYear(ctx context.Context, arg SetReleaseGroupOriginalYearParams) error {
 	_, err := q.db.ExecContext(ctx, setReleaseGroupOriginalYear, arg.OriginalYear, arg.ID)
 	return err
 }
 const updateReleaseGroup = `-- name: UpdateReleaseGroup :exec
 UPDATE release_groups 
 SET name = ?
 WHERE id = ?
 `
 type UpdateReleaseGroupParams struct {
 	Name string
 	ID   int64
 }
 func (q *Queries) UpdateReleaseGroup(ctx context.Context, arg UpdateReleaseGroupParams) error {
 	_, err := q.db.ExecContext(ctx, updateReleaseGroup, arg.Name, arg.ID)
 	return err
 }
 const updateReleaseGroupCoverArt = `-- name: UpdateReleaseGroupCoverArt :exec
 UPDATE release_groups
 SET cover_art_id = ?
 WHERE id = ?
 `
 type UpdateReleaseGroupCoverArtParams struct {
 	CoverArtID sql.NullInt64
 	ID         int64
 }
 func (q *Queries) UpdateReleaseGroupCoverArt(ctx context.Context, arg UpdateReleaseGroupCoverArtParams) error {
 	_, err := q.db.ExecContext(ctx, updateReleaseGroupCoverArt, arg.CoverArtID, arg.ID)
 	return err
 }
 const upsertReleaseGroup = `-- name: UpsertReleaseGroup :one
 INSERT INTO release_groups (name, album_artist_credit_id, year)
 VALUES (?, ?, ?)
 ON CONFLICT(name, album_artist_credit_id) DO UPDATE SET
  album_artist_credit_id = COALESCE(excluded.album_artist_credit_id, release_groups.album_artist_credit_id),
  year = COALESCE(excluded.year, release_groups.year)
 RETURNING id, name, cover_art_id, album_artist_credit_id, year, total_tracks, total_discs, mbid, original_year, pending_release_mbid
 `
 type UpsertReleaseGroupParams struct {
 	Name                string
 	AlbumArtistCreditID sql.NullInt64
 	Year                sql.NullInt64
 }
 func (q *Queries) UpsertReleaseGroup(ctx context.Context, arg UpsertReleaseGroupParams) (ReleaseGroup, error) {
 	row := q.db.QueryRowContext(ctx, upsertReleaseGroup, arg.Name, arg.AlbumArtistCreditID, arg.Year)
 	var i ReleaseGroup
 	err := row.Scan(
 		&i.ID,
 		&i.Name,
 		&i.CoverArtID,
 		&i.AlbumArtistCreditID,
 		&i.Year,
 		&i.TotalTracks,
 		&i.TotalDiscs,
 		&i.Mbid,
 		&i.OriginalYear,
 		&i.PendingReleaseMbid,
 	)
 	return i, err
 }
@@ -25,11 +25,21 @@ func (q *Queries) ClearCompletedTaggingItems(ctx context.Context, libraryID int6
 }
 const countPendingTaggingItems = `-- name: CountPendingTaggingItems :one
-SELECT COUNT(*) FROM tagging_items
+SELECT COUNT(*) FROM tagging_items ti
-WHERE status = 'pending'
+WHERE ti.status = 'pending'
-  AND (CAST(?1 AS INTEGER) = 0 OR library_id = ?1)
+  AND (CAST(?1 AS INTEGER) = 0 OR ti.library_id = ?1)
  AND EXISTS (
    SELECT 1 FROM audio_files af
    WHERE af.group_key = ti.group_key AND af.tag_status = 'untagged'
  )
 `
 // "Needs tagging" is a question about the files, not about the row:
 // every scanned folder gets a tagging_items row (see
 // UpsertTaggingItemOnTrackAdd), including one whose files all arrived
 // carrying a recording MBID.  Without the EXISTS a fully MB-tagged
 // library reports its entire album count as pending work.  See the
 // same predicate on the three list queries below.
 func (q *Queries) CountPendingTaggingItems(ctx context.Context, libraryID int64) (int64, error) {
 	row := q.db.QueryRowContext(ctx, countPendingTaggingItems, libraryID)
 	var count int64
@@ -77,6 +87,13 @@ LEFT JOIN libraries lb ON lb.id = ti.library_id
 WHERE ti.status = 'pending'
  AND (CAST(?1 AS INTEGER) = 0 OR ti.library_id = ?1)
  AND ti.group_key > ?2
  -- See CountPendingTaggingItems: the cursor must not stop on a
  -- folder the list query no longer shows, or "next" walks folders
  -- that are not in the sidebar.
  AND EXISTS (
    SELECT 1 FROM audio_files af
    WHERE af.group_key = ti.group_key AND af.tag_status = 'untagged'
  )
 ORDER BY ti.group_key
 LIMIT 1
 `
@@ -185,20 +202,6 @@ func (q *Queries) GetPendingFolderDetail(ctx context.Context, groupKey string) (
 	return i, err
 }
 const getRecordingReleaseGroupID = `-- name: GetRecordingReleaseGroupID :one
 SELECT COALESCE(rgr.release_group_id, 0) AS release_group_id
 FROM release_group_recordings rgr
 WHERE rgr.recording_id = ?
 LIMIT 1
 `
 func (q *Queries) GetRecordingReleaseGroupID(ctx context.Context, recordingID int64) (int64, error) {
 	row := q.db.QueryRowContext(ctx, getRecordingReleaseGroupID, recordingID)
 	var release_group_id int64
 	err := row.Scan(&release_group_id)
 	return release_group_id, err
 }
 const getTaggingItem = `-- name: GetTaggingItem :one
 SELECT group_key, library_id, track_count, album_name, album_artist, disc_number, best_match_release_mbid, score, last_checked_at, status, cleared_at, created_at, synthetic, parent_group_key, album_artist_conflict FROM tagging_items
 WHERE group_key = ?
@@ -235,22 +238,18 @@ SELECT
  af.basename,
  af.length_milliseconds,
  af.tag_status,
-  COALESCE(r.track_number, 0) AS track_number,
+  COALESCE(af.track_number, 0) AS track_number,
-  COALESCE(r.disc_number, 0) AS disc_number,
+  COALESCE(af.disc_number, 0) AS disc_number,
-  COALESCE(r.name, '') AS title,
+  af.title,
-  COALESCE(ac.text, '') AS artist_name,
+  af.artist_credit AS artist_name,
-  COALESCE(r.mbid, '') AS recording_mbid,
+  COALESCE(af.recording_mbid, '') AS recording_mbid,
-  COALESCE(rg.name, '') AS album_name,
+  COALESCE(al.name, '') AS album_name,
-  COALESCE(rgac.text, '') AS album_artist
+  COALESCE(al.artist_credit, '') AS album_artist
 FROM audio_files af
-LEFT JOIN recordings r ON af.recording_id = r.id
+LEFT JOIN albums al ON al.id = af.album_id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN release_group_recordings rgr ON rgr.recording_id = r.id
 LEFT JOIN release_groups rg ON rg.id = rgr.release_group_id
 LEFT JOIN artist_credit rgac ON rg.album_artist_credit_id = rgac.id
 WHERE af.group_key = ?
-ORDER BY COALESCE(r.disc_number, 0),
+ORDER BY COALESCE(af.disc_number, 0),
-         COALESCE(r.track_number, 0),
+         COALESCE(af.track_number, 0),
         af.file_path
 `
@@ -269,10 +268,10 @@ type ListAudioFilesInTaggingGroupRow struct {
 	AlbumArtist        string
 }
-// album_name/album_artist are the PER-TRACK tags (via each track's
+// album_name/album_artist are the PER-TRACK tags (each file's own
-// own release_group link), not the folder-level tagging_items
+// album link), not the folder-level tagging_items values.
-// values.  SplitMixedFolder clusters on these to find sub-albums
+// SplitMixedFolder clusters on these to find sub-albums hiding inside
-// hiding inside a folder full of unrelated tracks.
+// a folder full of unrelated tracks.
 func (q *Queries) ListAudioFilesInTaggingGroup(ctx context.Context, groupKey string) ([]ListAudioFilesInTaggingGroupRow, error) {
 	rows, err := q.db.QueryContext(ctx, listAudioFilesInTaggingGroup, groupKey)
 	if err != nil {
@@ -313,10 +312,7 @@ const listLikelyMixedBagGroupKeys = `-- name: ListLikelyMixedBagGroupKeys :many
 SELECT ti.group_key
 FROM tagging_items ti
 JOIN audio_files af ON af.group_key = ti.group_key
-LEFT JOIN recordings r ON af.recording_id = r.id
+LEFT JOIN albums rg ON rg.id = af.album_id
 LEFT JOIN artist_credit ac ON r.artist_credit_id = ac.id
 LEFT JOIN release_group_recordings rgr ON rgr.recording_id = r.id
 LEFT JOIN release_groups rg ON rg.id = rgr.release_group_id
 WHERE ti.synthetic = 0
  AND ti.track_count >= 4
  AND (
@@ -324,7 +320,7 @@ WHERE ti.synthetic = 0
    OR LOWER(TRIM(ti.album_artist)) IN ('various artists', 'various', 'va', 'v.a.', 'v a', 'unknown')
  )
 GROUP BY ti.group_key
-HAVING COUNT(DISTINCT CASE WHEN ac.text != '' THEN LOWER(TRIM(ac.text)) END) > 1
+HAVING COUNT(DISTINCT CASE WHEN af.artist_credit != '' THEN LOWER(TRIM(af.artist_credit)) END) > 1
   AND COUNT(DISTINCT CASE WHEN rg.name != '' THEN LOWER(TRIM(rg.name)) END) > 1
 `
@@ -359,34 +355,31 @@ func (q *Queries) ListLikelyMixedBagGroupKeys(ctx context.Context) ([]string, er
 	return items, nil
 }
-const listLocalReleaseGroupCandidates = `-- name: ListLocalReleaseGroupCandidates :many
+const listLocalAlbumCandidates = `-- name: ListLocalAlbumCandidates :many
 SELECT
-  rg.id AS release_group_id,
+  al.id AS album_id,
-  rg.mbid AS release_group_mbid,
+  al.mbid AS album_mbid,
-  rg.name AS album_name,
+  al.name AS album_name,
-  COALESCE(rg.year, 0) AS year,
+  COALESCE(al.year, 0) AS year,
-  COALESCE(ac.text, '') AS artist_credit,
+  al.artist_credit,
-  COALESCE(rgr.track_number, 0) AS track_number,
+  COALESCE(af.track_number, 0) AS track_number,
-  COALESCE(rgr.disc_number, 0) AS disc_number,
+  COALESCE(af.disc_number, 0) AS disc_number,
-  COALESCE(r.name, '') AS track_title,
+  af.title AS track_title,
-  COALESCE(r.mbid, '') AS recording_mbid,
+  COALESCE(af.recording_mbid, '') AS recording_mbid,
-  COALESCE(local_af.length_milliseconds, 0) AS length_milliseconds
+  af.length_milliseconds
-FROM release_groups rg
+FROM albums al
-JOIN release_group_recordings rgr ON rgr.release_group_id = rg.id
+JOIN audio_files af ON af.album_id = al.id
-JOIN recordings r ON r.id = rgr.recording_id
+WHERE al.mbid IS NOT NULL
-LEFT JOIN artist_credit ac ON rg.album_artist_credit_id = ac.id
+  AND al.mbid != ''
-LEFT JOIN audio_files local_af ON local_af.recording_id = r.id
+  AND af.recording_mbid IS NOT NULL
-WHERE rg.mbid IS NOT NULL
+  AND af.recording_mbid != ''
-  AND rg.mbid != ''
+  AND al.name = ? COLLATE NOCASE
-  AND r.mbid IS NOT NULL
+ORDER BY al.id, af.disc_number, af.track_number
  AND r.mbid != ''
  AND rg.name = ? COLLATE NOCASE
 ORDER BY rg.id, rgr.disc_number, rgr.track_number
 `
-type ListLocalReleaseGroupCandidatesRow struct {
+type ListLocalAlbumCandidatesRow struct {
-	ReleaseGroupID     int64
+	AlbumID            int64
-	ReleaseGroupMbid   sql.NullString
+	AlbumMbid          sql.NullString
 	AlbumName          string
 	Year               int64
 	ArtistCredit       string
@@ -397,22 +390,21 @@ type ListLocalReleaseGroupCandidatesRow struct {
 	LengthMilliseconds int64
 }
-// Returns one row per (release_group, track) combination for any
+// One row per (album, track) for any local album carrying an MBID.
-// local release_group that has an MBID.  Callers group these in Go
+// Callers group these in Go and filter by normalized album-name match;
-// and filter by normalized album-name match.  Joined case-insensitive
+// the join is case-insensitive on name to pre-filter cheaply.
-// on name to pre-filter cheaply; Go does the real normalization.
+func (q *Queries) ListLocalAlbumCandidates(ctx context.Context, name string) ([]ListLocalAlbumCandidatesRow, error) {
-func (q *Queries) ListLocalReleaseGroupCandidates(ctx context.Context, name string) ([]ListLocalReleaseGroupCandidatesRow, error) {
+	rows, err := q.db.QueryContext(ctx, listLocalAlbumCandidates, name)
 	rows, err := q.db.QueryContext(ctx, listLocalReleaseGroupCandidates, name)
 	if err != nil {
 		return nil, err
 	}
 	defer rows.Close()
-	var items []ListLocalReleaseGroupCandidatesRow
+	var items []ListLocalAlbumCandidatesRow
 	for rows.Next() {
-		var i ListLocalReleaseGroupCandidatesRow
+		var i ListLocalAlbumCandidatesRow
 		if err := rows.Scan(
-			&i.ReleaseGroupID,
+			&i.AlbumID,
-			&i.ReleaseGroupMbid,
+			&i.AlbumMbid,
 			&i.AlbumName,
 			&i.Year,
 			&i.ArtistCredit,
@@ -454,6 +446,18 @@ LEFT JOIN libraries lb ON lb.id = ti.library_id
 WHERE (CAST(?1 AS INTEGER) = 0 OR ti.library_id = ?1)
  AND (CAST(?2 AS TEXT) = 'all' OR ti.status = ?2)
  AND ti.cleared_at IS NULL
  -- Actionable rows must have something to act on: see
  -- CountPendingTaggingItems.  Reviewed rows (confirmed/skipped) are
  -- exempt because they are history, not work -- an applied folder is
  -- fully tagged by definition and would otherwise vanish from the
  -- sidebar's Completed section the instant it succeeded.
  AND (
    ti.status IN ('confirmed', 'skipped')
    OR EXISTS (
      SELECT 1 FROM audio_files af
      WHERE af.group_key = ti.group_key AND af.tag_status = 'untagged'
    )
  )
 ORDER BY LOWER(ti.album_artist), LOWER(ti.album_name), ti.disc_number
 LIMIT ?4 OFFSET ?3
 `
@@ -628,6 +632,14 @@ LEFT JOIN libraries lb ON lb.id = ti.library_id
 WHERE (CAST(?1 AS INTEGER) = 0 OR ti.library_id = ?1)
  AND (CAST(?2 AS TEXT) = 'all' OR ti.status = ?2)
  AND ti.cleared_at IS NULL
  -- See ListPendingTaggingItemsAlphabetical.
  AND (
    ti.status IN ('confirmed', 'skipped')
    OR EXISTS (
      SELECT 1 FROM audio_files af
      WHERE af.group_key = ti.group_key AND af.tag_status = 'untagged'
    )
  )
 ORDER BY ti.score IS NULL, ti.score DESC, LOWER(ti.album_artist), LOWER(ti.album_name)
 LIMIT ?4 OFFSET ?3
 `
@@ -758,31 +770,36 @@ func (q *Queries) SetAudioFileTagStatus(ctx context.Context, arg SetAudioFileTag
 	return err
 }
-const setRecordingMBID = `-- name: SetRecordingMBID :exec
+const setFileAlbumMBID = `-- name: SetFileAlbumMBID :exec
-UPDATE recordings SET mbid = ? WHERE id = ?
+UPDATE albums SET mbid = ?
 WHERE albums.id = (SELECT af.album_id FROM audio_files af WHERE af.id = ?)
 `
-type SetRecordingMBIDParams struct {
+type SetFileAlbumMBIDParams struct {
 	Mbid sql.NullString
 	ID   int64
 }
-func (q *Queries) SetRecordingMBID(ctx context.Context, arg SetRecordingMBIDParams) error {
+// The album MBID for the album a file belongs to.  Keyed by file
-	_, err := q.db.ExecContext(ctx, setRecordingMBID, arg.Mbid, arg.ID)
+// because that is what the autotag apply path holds; under the old
 // schema it had to look the release group up through two join tables
 // first (GetRecordingReleaseGroupID), which is gone.
 func (q *Queries) SetFileAlbumMBID(ctx context.Context, arg SetFileAlbumMBIDParams) error {
 	_, err := q.db.ExecContext(ctx, setFileAlbumMBID, arg.Mbid, arg.ID)
 	return err
 }
-const setReleaseGroupMBID = `-- name: SetReleaseGroupMBID :exec
+const setFileRecordingMBID = `-- name: SetFileRecordingMBID :exec
-UPDATE release_groups SET mbid = ? WHERE id = ?
+UPDATE audio_files SET recording_mbid = ? WHERE id = ?
 `
-type SetReleaseGroupMBIDParams struct {
+type SetFileRecordingMBIDParams struct {
-	Mbid sql.NullString
+	RecordingMbid sql.NullString
-	ID   int64
+	ID            int64
 }
-func (q *Queries) SetReleaseGroupMBID(ctx context.Context, arg SetReleaseGroupMBIDParams) error {
+func (q *Queries) SetFileRecordingMBID(ctx context.Context, arg SetFileRecordingMBIDParams) error {
-	_, err := q.db.ExecContext(ctx, setReleaseGroupMBID, arg.Mbid, arg.ID)
+	_, err := q.db.ExecContext(ctx, setFileRecordingMBID, arg.RecordingMbid, arg.ID)
 	return err
 }
@@ -69,10 +69,7 @@ func TestTagStatusBackfillFromRecordingMBID(t *testing.T) {
 		UPDATE audio_files
 		SET tag_status = 'user_confirmed'
 		WHERE tag_status = 'untagged'
-		  AND recording_id IN (
+		  AND recording_mbid IS NOT NULL AND recording_mbid != ''
 		    SELECT id FROM recordings
 		    WHERE mbid IS NOT NULL AND mbid != ''
 		  )
 	`); err != nil {
 		t.Fatalf("backfill: %v", err)
 	}
@@ -205,6 +202,12 @@ func TestTaggingItems_ListPendingAndCount(t *testing.T) {
 	seedTaggingItem(t, db, "g2", 0, "Album B", "Artist B", 1, "pending")
 	seedTaggingItem(t, db, "g3", 0, "Album C", "Artist C", 3, "confirmed")
 	// A pending group is only listed while it still holds untagged
 	// files — see TestTaggingItems_FullyTaggedGroupIsNotPending.
 	seedGroupFile(t, db, "g1", "/music/a1.mp3", "untagged")
 	seedGroupFile(t, db, "g2", "/music/b1.mp3", "untagged")
 	seedGroupFile(t, db, "g3", "/music/c1.mp3", "user_confirmed")
 	count, err := db.Queries.CountPendingTaggingItems(db.Ctx, 0)
 	if err != nil {
 		t.Fatalf("count: %v", err)
@@ -239,6 +242,62 @@ func TestTaggingItems_ListPendingAndCount(t *testing.T) {
 	}
 }
 // TestClearUnreviewedConfirmedTaggingItems mirrors migration 0007 the
 // way TestTagStatusBackfillFromRecordingMBID mirrors the tag_status
 // backfill: NewTestDB applies migrations to an empty database, so the
 // only way to exercise one that rewrites existing rows is to seed the
 // shapes and re-issue its statement.  Keep the two in step.
 func TestClearUnreviewedConfirmedTaggingItems(t *testing.T) {
 	t.Parallel()
 	db := database.NewTestDB(t)
 	// Stamped 'confirmed' by the old backfill: never scored, never
 	// checked, never matched — the app has not touched it.
 	seedTaggingItem(t, db, "g-backfill", 0, "Bulk", "Artist A", 2, "confirmed")
 	// Confirmed by a real apply, which stamps last_checked_at.
 	seedTaggingItem(t, db, "g-applied", 0, "Applied", "Artist B", 2, "confirmed")
 	if _, err := db.ExecContext(
 		`UPDATE tagging_items SET last_checked_at = CURRENT_TIMESTAMP, score = 0.98
 		 WHERE group_key = 'g-applied'`,
 	); err != nil {
 		t.Fatalf("mark applied: %v", err)
 	}
 	// A skipped row is not confirmed and must be left alone.
 	seedTaggingItem(t, db, "g-skipped", 0, "Skipped", "Artist C", 1, "skipped")
 	if _, err := db.ExecContext(`
 		UPDATE tagging_items
 		SET cleared_at = CURRENT_TIMESTAMP
 		WHERE status = 'confirmed'
 		  AND cleared_at IS NULL
 		  AND last_checked_at IS NULL
 		  AND score IS NULL
 		  AND (best_match_release_mbid IS NULL OR best_match_release_mbid = '')
 	`); err != nil {
 		t.Fatalf("migration: %v", err)
 	}
 	cases := map[string]bool{
 		"g-backfill": true,
 		"g-applied":  false,
 		"g-skipped":  false,
 	}
 	for key, wantCleared := range cases {
 		got := scalarInt(t, db,
 			`SELECT cleared_at IS NOT NULL FROM tagging_items WHERE group_key = ?`,
 			key,
 		)
 		if (got == 1) != wantCleared {
 			t.Errorf("%s cleared = %v, want %v", key, got == 1, wantCleared)
 		}
 	}
 }
 func TestCountPendingTaggingItems_UsesPartialIndex(t *testing.T) {
 	t.Parallel()
@@ -248,9 +307,13 @@ func TestCountPendingTaggingItems_UsesPartialIndex(t *testing.T) {
 	rows, err := db.QueryContext(`
 		EXPLAIN QUERY PLAN
-		SELECT COUNT(*) FROM tagging_items
+		SELECT COUNT(*) FROM tagging_items ti
-		WHERE status = 'pending'
+		WHERE ti.status = 'pending'
-		  AND (CAST(0 AS INTEGER) = 0 OR library_id = 0)
+		  AND (CAST(0 AS INTEGER) = 0 OR ti.library_id = 0)
 		  AND EXISTS (
 		    SELECT 1 FROM audio_files af
 		    WHERE af.group_key = ti.group_key AND af.tag_status = 'untagged'
 		  )
 	`)
 	if err != nil {
 		t.Fatalf("explain: %v", err)
@@ -281,6 +344,97 @@ func TestCountPendingTaggingItems_UsesPartialIndex(t *testing.T) {
 			plan.String(),
 		)
 	}
 	// The untagged-files existence check is asked once per candidate
 	// row, so it has to be a seek.  idx_audio_files_tag_status_untagged
 	// is keyed on library_id and cannot serve it; the group_key one
 	// can, and covers the query outright.
 	if !strings.Contains(plan.String(), "idx_audio_files_untagged_group_key") {
 		t.Errorf(
 			"untagged-files check does not use idx_audio_files_untagged_group_key:\n%s",
 			plan.String(),
 		)
 	}
 }
 // TestTaggingItems_FullyTaggedGroupIsNotPending pins the rule that
 // decides what the autotag review page shows: every scanned folder
 // gets a tagging_items row, so "pending" has to mean "still holds
 // untagged files" rather than "has a row".  Without it a fully
 // MB-tagged library queues its entire album count for review — and
 // the background prefetch scores every one of them against
 // MusicBrainz.
 func TestTaggingItems_FullyTaggedGroupIsNotPending(t *testing.T) {
 	t.Parallel()
 	db := database.NewTestDB(t)
 	seedTaggingItem(t, db, "g-partial", 0, "Half Tagged", "Artist A", 2, "pending")
 	seedGroupFile(t, db, "g-partial", "/music/partial-1.mp3", "user_confirmed")
 	seedGroupFile(t, db, "g-partial", "/music/partial-2.mp3", "untagged")
 	seedTaggingItem(t, db, "g-done", 0, "Already Tagged", "Artist B", 2, "pending")
 	seedGroupFile(t, db, "g-done", "/music/done-1.mp3", "user_confirmed")
 	seedGroupFile(t, db, "g-done", "/music/done-2.mp3", "user_confirmed")
 	// Reviewed rows are history, not work: an applied folder is fully
 	// tagged by definition and must stay in the Completed section.
 	seedTaggingItem(t, db, "g-applied", 0, "Applied Here", "Artist C", 1, "confirmed")
 	seedGroupFile(t, db, "g-applied", "/music/applied-1.mp3", "user_confirmed")
 	count, err := db.Queries.CountPendingTaggingItems(db.Ctx, 0)
 	if err != nil {
 		t.Fatalf("count: %v", err)
 	}
 	if count != 1 {
 		t.Errorf("pending count = %d, want 1 (only the part-tagged folder)", count)
 	}
 	items, err := db.Queries.ListPendingTaggingItemsByScore(
 		db.Ctx,
 		sqlcgen.ListPendingTaggingItemsByScoreParams{
 			LibraryID:    0,
 			StatusFilter: "all",
 			RowLimit:     50,
 			RowOffset:    0,
 		},
 	)
 	if err != nil {
 		t.Fatalf("list by score: %v", err)
 	}
 	listed := make(map[string]bool, len(items))
 	for _, it := range items {
 		listed[it.GroupKey] = true
 	}
 	if !listed["g-partial"] {
 		t.Error("expected g-partial (one untagged file left) to be listed")
 	}
 	if listed["g-done"] {
 		t.Error("expected g-done (nothing left to tag) to be filtered out")
 	}
 	if !listed["g-applied"] {
 		t.Error("expected g-applied (confirmed by a real apply) to stay listed")
 	}
 	next, err := db.Queries.GetNextPendingTaggingItem(
 		db.Ctx,
 		sqlcgen.GetNextPendingTaggingItemParams{LibraryID: 0, AfterGroupKey: ""},
 	)
 	if err != nil {
 		t.Fatalf("next pending: %v", err)
 	}
 	// The cursor must not stop on a folder the sidebar no longer
 	// shows: alphabetically g-done sorts before g-partial, so a
 	// missing predicate here surfaces as "next" landing on nothing.
 	if next.GroupKey != "g-partial" {
 		t.Errorf("next pending = %q, want %q", next.GroupKey, "g-partial")
 	}
 }
 // TestListPendingFolders_SampleFilePathUsesIndex guards the folder-list
@@ -450,9 +604,7 @@ func TestGetTaggingItemAndListAudioFilesInGroup(t *testing.T) {
 // helpers
 // ---------------------------------------------------------------------------
-// seedAF inserts a minimal recording + audio_files pair and returns
+// seedAF inserts one file and returns its audio_files id.
 // the new audio_files id.  All FK-satisfying rows (artist_credit,
 // recordings, file_types[0]) are created inline.
 func seedAF(
 	t *testing.T,
 	db *database.DB,
@@ -462,49 +614,32 @@ func seedAF(
 ) int64 {
 	t.Helper()
-	ac, err := db.Queries.UpsertArtistCredit(db.Ctx, "Test Artist")
+	return database.InsertTestTrack(t, db, database.TestTrack{
-	if err != nil {
+		FilePath:      filePath,
-		t.Fatalf("upsert artist credit: %v", err)
+		Title:         recordingName,
-	}
+		RecordingMBID: recordingMBID,
 		DiscNumber:    discNumber,
 		LibraryID:     libraryID,
 		LengthMs:      1000,
 	})
 }
-	rec, err := db.Queries.CreateRecordingFull(
+// seedGroupFile attaches one file to a tagging group with an explicit
-		db.Ctx,
+// tag_status - the thing the queue's "is there anything left to tag
-		sqlcgen.CreateRecordingFullParams{
+// here" predicate reads.
-			Name:           recordingName,
+func seedGroupFile(
-			ArtistCreditID: ac.ID,
+	t *testing.T,
-		},
+	db *database.DB,
-	)
+	groupKey, filePath, tagStatus string,
-	if err != nil {
+) {
-		t.Fatalf("create recording: %v", err)
+	t.Helper()
 	}
-	if recordingMBID != "" {
+	database.InsertTestTrack(t, db, database.TestTrack{
-		if _, err := db.ExecContext(
+		FilePath:  filePath,
-			`UPDATE recordings SET mbid = ? WHERE id = ?`,
+		Title:     filePath,
-			recordingMBID, rec.ID,
+		GroupKey:  groupKey,
-		); err != nil {
+		TagStatus: tagStatus,
-			t.Fatalf("set mbid: %v", err)
+	})
 		}
 	}
 	af, err := db.Queries.CreateAudioFile(
 		db.Ctx,
 		sqlcgen.CreateAudioFileParams{
 			FilePath:           filePath,
 			LengthMilliseconds: 1000,
 			FileTypeID:         0,
 			RecordingID:        rec.ID,
 			Basename:           filePath,
 			LibraryID:          libraryID,
 		},
 	)
 	if err != nil {
 		t.Fatalf("create audio file: %v", err)
 	}
 	_ = discNumber // reserved for callers that want specific disc values
 	return af.ID
 }
 func seedTaggingItem(
@@ -2,7 +2,10 @@ package database
 import (
 	"database/sql"
 	"fmt"
 	"log/slog"
 	"path/filepath"
 	"sync/atomic"
 	"testing"
 	_ "modernc.org/sqlite" // Register sqlite driver.
@@ -10,16 +13,30 @@ import (
 	"yellowjacket/backend/database/sql/sqlcgen"
 )
-// NewTestDB returns an in-memory SQLite database that mirrors the
+// NewTestDB returns an in-memory SQLite database shaped like the real
-// production setup (PRAGMAs + all migrations). The database is
+// one: a single-writer handle and a separate query-only read pool over
-// automatically closed when the test completes via t.Cleanup.
+// the same database, built by the same applySchema production uses.
 //
 // The two handles matter. The test DB used to be one shared connection
 // with readDB nil, so `reader()` returned the *writer* — which is how a
 // query-shaped write (`INSERT ... RETURNING` through QueryContext)
 // passed every test and then failed for a user with "attempt to write a
 // readonly database". `TestNoWritesOnTheReadPool` had to walk the source
 // tree to catch what a test could not.
 //
 // A shared-cache in-memory database is what lets two handles see one
 // database; the connections are capped the way production caps them.
 // It is closed when the test completes via t.Cleanup.
 func NewTestDB(t *testing.T) *DB {
 	t.Helper()
-	db, err := sql.Open(
+	// A per-test name, so parallel tests do not share a database.
-		"sqlite",
+	dsn := fmt.Sprintf(
-		":memory:?_busy_timeout=5000&_journal_mode=WAL",
+		"file:testdb%d?mode=memory&cache=shared&_pragma=busy_timeout(5000)",
 		testDBSeq.Add(1),
 	)
 	db, err := sql.Open("sqlite", dsn)
 	if err != nil {
 		t.Fatalf("could not open test database: %v", err)
 	}
@@ -47,21 +64,32 @@ func NewTestDB(t *testing.T) *DB {
 		t.Fatalf("could not insert test library: %v", err)
 	}
-	queries := sqlcgen.New(db)
+	readDB, err := sql.Open("sqlite", dsn+"&_pragma=query_only(true)")
 	if err != nil {
 		t.Fatalf("could not open test read pool: %v", err)
 	}
-	t.Cleanup(func() { _ = db.Close() })
+	readDB.SetMaxOpenConns(readPoolConns)
 	t.Cleanup(func() {
 		_ = readDB.Close()
 		_ = db.Close()
 	})
 	return &DB{
-		db:  db,
+		db:          db,
-		Ctx: ctx,
+		readDB:      readDB,
-		// The in-memory test DB shares one connection, so reads and
+		Ctx:         ctx,
-		// writes use the same handle; ReadQueries aliases Queries.
+		Queries:     sqlcgen.New(db),
-		Queries:     queries,
+		ReadQueries: sqlcgen.New(readDB),
 		ReadQueries: queries,
 		logger:      slog.Default(),
 	}
 }
 // testDBSeq names each test database uniquely, so parallel tests do not
 // share one through the shared cache.
 var testDBSeq atomic.Int64
 // NewTestDBWithLibrary returns a test DB with a library row
 // pre-inserted. Returns the DB and the library ID.
 func NewTestDBWithLibrary(
@@ -85,3 +113,162 @@ func NewTestDBWithLibrary(
 	return db, lib.ID
 }
 // TestTrack describes one file to seed into a test database.  Zero
 // values are fine: only FilePath is required.
 type TestTrack struct {
 	FilePath      string
 	Title         string
 	Artist        string
 	ArtistMBID    string
 	Album         string
 	AlbumArtist   string
 	AlbumMBID     string
 	RecordingMBID string
 	Genres        []string
 	TrackNumber   int64
 	DiscNumber    int64
 	TotalTracks   int64
 	Year          int64
 	LengthMs      int64
 	LibraryID     int64
 	PlayCount     int64
 	TagStatus     string
 	GroupKey      string
 	// SkipSearchIndex leaves the file out of the FTS index, for the
 	// tests that assert on a rebuild putting it there.
 	SkipSearchIndex bool
 }
 // InsertTestTrack seeds one file, with the artist and album its tags
 // name, and returns the audio_files id.
 //
 // There is one of these because there is one shape.  Twenty test files
 // used to carry their own seeder, each inserting a recording, an artist
 // credit, a credit-artist link and a release-group link in the right
 // order - which is exactly the ceremony the schema change removed, and
 // exactly why every one of those seeders was subtly different.
 func InsertTestTrack(t *testing.T, db *DB, tr TestTrack) int64 {
 	t.Helper()
 	if tr.Title == "" {
 		tr.Title = "Test Track"
 	}
 	if tr.Artist == "" {
 		tr.Artist = "Test Artist"
 	}
 	if tr.TagStatus == "" {
 		tr.TagStatus = "untagged"
 	}
 	artist, err := db.Queries.UpsertArtist(db.Ctx, sqlcgen.UpsertArtistParams{
 		Name: tr.Artist,
 		Mbid: nullString(tr.ArtistMBID),
 	})
 	if err != nil {
 		t.Fatalf("seed artist: %v", err)
 	}
 	artistID := sql.NullInt64{Int64: artist.ID, Valid: true}
 	albumID := sql.NullInt64{}
 	if tr.Album != "" {
 		credit := tr.AlbumArtist
 		if credit == "" {
 			credit = tr.Artist
 		}
 		album, albErr := db.Queries.UpsertAlbum(db.Ctx, sqlcgen.UpsertAlbumParams{
 			Name:         tr.Album,
 			ArtistCredit: credit,
 			ArtistID:     artistID,
 			Year:         nullInt64(tr.Year),
 		})
 		if albErr != nil {
 			t.Fatalf("seed album: %v", albErr)
 		}
 		if tr.AlbumMBID != "" {
 			if err := db.Queries.SetAlbumMBID(db.Ctx, sqlcgen.SetAlbumMBIDParams{
 				Mbid: nullString(tr.AlbumMBID),
 				ID:   album.ID,
 			}); err != nil {
 				t.Fatalf("seed album mbid: %v", err)
 			}
 		}
 		albumID = sql.NullInt64{Int64: album.ID, Valid: true}
 	}
 	af, err := db.Queries.CreateAudioFile(db.Ctx, sqlcgen.CreateAudioFileParams{
 		FilePath:           tr.FilePath,
 		LibraryID:          tr.LibraryID,
 		LengthMilliseconds: tr.LengthMs,
 		Title:              tr.Title,
 		ArtistCredit:       tr.Artist,
 		ArtistID:           artistID,
 		AlbumID:            albumID,
 		TrackNumber:        nullInt64(tr.TrackNumber),
 		DiscNumber:         nullInt64(tr.DiscNumber),
 		TotalTracks:        nullInt64(tr.TotalTracks),
 		Year:               nullInt64(tr.Year),
 		RecordingMbid:      nullString(tr.RecordingMBID),
 		Basename:           filepath.Base(tr.FilePath),
 		GroupKey:           tr.GroupKey,
 		TagStatus:          tr.TagStatus,
 	})
 	if err != nil {
 		t.Fatalf("seed audio file %q: %v", tr.FilePath, err)
 	}
 	for _, name := range tr.Genres {
 		g, gErr := db.Queries.UpsertGenre(db.Ctx, name)
 		if gErr != nil {
 			t.Fatalf("seed genre %q: %v", name, gErr)
 		}
 		if err := db.Queries.LinkFileGenre(db.Ctx, sqlcgen.LinkFileGenreParams{
 			AudioFileID: af.ID,
 			GenreID:     g.ID,
 		}); err != nil {
 			t.Fatalf("seed file genre: %v", err)
 		}
 	}
 	if tr.PlayCount > 0 {
 		if _, err := db.db.ExecContext(db.Ctx,
 			"UPDATE audio_files SET play_count = ? WHERE id = ?",
 			tr.PlayCount, af.ID,
 		); err != nil {
 			t.Fatalf("seed play count: %v", err)
 		}
 	}
 	if !tr.SkipSearchIndex {
 		if err := db.InsertSearchIndex(
 			af.ID, tr.FilePath, tr.Title, tr.Artist, tr.Album,
 		); err != nil {
 			t.Fatalf("seed search index: %v", err)
 		}
 	}
 	return af.ID
 }
 func nullString(v string) sql.NullString {
 	if v == "" {
 		return sql.NullString{}
 	}
 	return sql.NullString{String: v, Valid: true}
 }
 func nullInt64(v int64) sql.NullInt64 {
 	if v == 0 {
 		return sql.NullInt64{}
 	}
 	return sql.NullInt64{Int64: v, Valid: true}
 }
@@ -105,13 +105,10 @@ var internalTables = map[string]bool{
 // tables is the catalog.  Keep it alphabetical.
 var tables = []Table{
 	{
-		Name: "artist_credit", Kind: Owned, Lifetime: Swept,
+		Name: "albums", Kind: Owned, Lifetime: Swept,
-		Note: "Credit strings parsed from file tags. Orphan-swept when " +
+		Note: "Albums as named by file tags — the local counterpart of a " +
-			"no recording references them.",
+			"catalog release group, which is a different thing and lives " +
-	},
+			"in explore_index. Swept when the last file on one goes.",
 	{
 		Name: "artist_credit_artist", Kind: Owned, Lifetime: Swept,
 		Note: "Join table between credits and artists.",
 	},
 	{
 		Name: "artist_enrichment", Kind: Derived, Lifetime: Retained,
@@ -138,11 +135,12 @@ var tables = []Table{
 	},
 	{
 		Name: "audio_files", Kind: Owned, Lifetime: Swept,
-		Note: "MIXED KIND. Mostly an owned projection of files on disk, " +
+		Note: "MIXED KIND. One row per file, carrying its tags: mostly an " +
-			"but play_count, last_played and tag_status are authored and " +
+			"owned projection of what is on disk, but play_count, " +
-			"exist nowhere else. Deleting a row to rebuild it destroys " +
+			"last_played and tag_status are authored and exist nowhere " +
-			"that authored state — which is why a file rename currently " +
+			"else. Deleting a row to rebuild it destroys that authored " +
-			"loses play counts. See the data architecture plan.",
+			"state — which is why a file rename currently loses play " +
 			"counts. See the data architecture plan.",
 	},
 	{
 		Name: "cover_art", Kind: Owned, Lifetime: Swept,
@@ -209,6 +207,15 @@ var tables = []Table{
 			"rescan clears it \u2014 the only way back for a path removed by " +
 			"mistake.",
 	},
 	{
 		Name: "file_genres", Kind: Owned, Lifetime: Swept,
 		Note: "Genres per file. The one many-to-many in the local library " +
 			"that really is one. It cascades with the *file*, but the " +
 			"genre_id key is NO ACTION, so a genre cannot be deleted " +
 			"while a link survives — the genre sweep runs after the file " +
 			"sweep for that reason, and only deletes genres nothing " +
 			"references.",
 	},
 	{
 		Name: "file_types", Kind: Derived, Lifetime: Retained,
 		Note: "Static lookup rows seeded from code, not user data.",
@@ -232,6 +239,15 @@ var tables = []Table{
 		Note: "The directories the user chose. Removed only by explicit " +
 			"user action via RemoveLibrary.",
 	},
 	{
 		Name: "lyrics", Kind: Cache, Lifetime: Cascade,
 		Note: "MIXED KIND, and it says which: source='tag' is Owned (any " +
 			"rescan reads it back off the file) and source='lrclib' is " +
 			"Cache (re-fetching it is network traffic and someone else's " +
 			"rate limit). These used to be one untyped column on " +
 			"recordings, in a table classified Owned, so 24,294 rows in a " +
 			"real library could not say which of the two they were.",
 	},
 	{
 		Name: "lyrics_index", Kind: Derived, Lifetime: Retained, FTS: true,
 		Note: "Full-text index over embedded and fetched lyrics. Rebuilt " +
@@ -266,32 +282,11 @@ var tables = []Table{
 		Note: "Queue entries. Cascade with their track; the queue is " +
 			"compacted afterwards.",
 	},
 	{
 		Name: "recording_genres", Kind: Owned, Lifetime: Swept,
 		Note: "Join table between recordings and genres.",
 	},
 	{
 		Name: "recordings", Kind: Owned, Lifetime: Swept,
 		Note: "Tracks as parsed from file tags. Orphan-swept.",
 	},
 	{
 		Name: "release_group_recordings", Kind: Owned, Lifetime: Swept,
 		Note: "Join table between release groups and recordings.",
 	},
 	{
 		Name: "release_groups", Kind: Owned, Lifetime: Swept,
 		Note: "Albums as parsed from file tags. Orphan-swept.",
 	},
 	{
 		Name: "release_to_rg", Kind: Cache, Lifetime: Retained,
-		Note: "Release to release-group mapping from the dump.",
+		Note: "Release to release-group mapping, captured during a local " +
-	},
+			"dump import and read by the incremental listen-count " +
-	{
+			"refresh. Empty unless this install built its own index.",
 		Name: "schema_migrations", Kind: Derived, Lifetime: Retained,
 		Note: "Bookkeeping for sql/migrations: which numbered files have " +
 			"run. Safe to lose — replaying an already-applied migration " +
 			"tolerates its ALTER TABLE ADD COLUMN as a no-op and just " +
 			"re-records it.",
 	},
 	{
 		Name: "search_clicks", Kind: Authored, Lifetime: Retained,
@@ -486,6 +486,8 @@ func (s *Service) ClearFinished() error {
 // SetReconciler wires the request-list loop.  Optional: without it the
 // request list still stores and lists requests, it just never acts on
 // them.
 //
 //wails:ignore // internal wiring, not part of the app's IPC surface.
 func (s *Service) SetReconciler(r *Reconciler) {
 	s.reconciler = r
 }
@@ -7,6 +7,7 @@ import (
 	"fmt"
 	"os"
 	"strconv"
 	"strings"
 	"time"
 	_ "modernc.org/sqlite" // SQLite driver for reading the artifact file.
@@ -91,9 +92,90 @@ type artifactInfo struct {
 // TestArtifactColumnsMatchExporter.
 const artifactCatalogColumns = `entity_type, mbid, title, artist_name, artist_mbid,
 	aliases, popularity, listener_count, duration, caa_release_mbid,
-	release_name, primary_type, secondary_types, release_date,
+	release_name, primary_type, secondary_types, release_date, total_tracks,
 	artist_type, country, disambiguation, sort_name, discog_fetched`
 // artifactSelectColumns is the same list as read *from an artifact*,
 // converting the two columns whose storage this app changed.
 //
 // The local table stores an MBID as 16 raw bytes and an entity type as a
 // small integer, which took the catalog and its indexes from 677 MB to
 // 389 MB.  A published artifact still carries the text form, and there
 // is no reason it should not: converting on the way in costs one
 // `unhex` per row on a once-a-month import, and it means a new build
 // reads the artifact that is already out there rather than requiring
 // one to be rebuilt and re-downloaded first.
 //
 // An artifact that already carries the compact form is copied straight
 // through - `artifactStoresText` decides which, by asking the artifact
 // rather than by trusting a version number.
 func artifactSelectColumns(text, totals bool) string {
 	totalTracks := "total_tracks"
 	if !totals {
 		// An artifact built before the column existed. Zero is what the
 		// column means by "the catalog does not say", so an older
 		// artifact imports as one that declines to answer rather than
 		// failing to import at all.
 		totalTracks = "0"
 	}
 	if !text {
 		return strings.Replace(
 			artifactCatalogColumns, "total_tracks", totalTracks, 1,
 		)
 	}
 	return `CASE entity_type
 			WHEN 'artist' THEN 1
 			WHEN 'release_group' THEN 2
 			WHEN 'recording' THEN 3
 			ELSE 0 END,
 		unhex(replace(mbid, '-', '')),
 		title, artist_name,
 		CASE WHEN artist_mbid = '' THEN x''
 			ELSE unhex(replace(artist_mbid, '-', '')) END,
 		aliases, popularity, listener_count, duration,
 		CASE WHEN caa_release_mbid = '' THEN x''
 			ELSE unhex(replace(caa_release_mbid, '-', '')) END,
 		release_name, primary_type, secondary_types, release_date, ` +
 		totalTracks + `,
 		artist_type, country, disambiguation, sort_name, discog_fetched`
 }
 // artifactHasTotals reports whether the attached artifact carries the
 // per-release-group track denominator.  An artifact published before
 // that column existed is still a perfectly good catalog, so it is asked
 // rather than assumed - the same rule, and the same handle, as
 // artifactStoresText below.
 func (si *SearchIndex) artifactHasTotals() bool {
 	var n int
 	err := si.db.QueryRowWriter(
 		`SELECT COUNT(*) FROM pragma_table_info('explore_index', 'core')
 		 WHERE name = 'total_tracks'`,
 	).Scan(&n)
 	return err == nil && n > 0
 }
 // artifactStoresText reports whether the attached artifact carries the
 // old text encoding.
 func (si *SearchIndex) artifactStoresText() bool {
 	// The writer, not QueryContext: "core" is attached to that one
 	// connection and does not exist on the read pool.  Asking the wrong
 	// handle errors, and the fallback would then convert an artifact
 	// that needs no conversion.
 	var kind string
 	if err := si.db.QueryRowWriter(
 		"SELECT typeof(mbid) FROM core.explore_index LIMIT 1",
 	).Scan(&kind); err != nil {
 		return true
 	}
 	return kind == "text"
 }
 // inspectArtifact opens the artifact read-only and reports what it
 // declares, without touching the live index.  Validation happens here so
 // a bad download is rejected before anything is attached.
@@ -263,9 +345,13 @@ func (si *SearchIndex) analyzeIndex() {
 // is an index range scan and a cancelled import leaves committed work
 // behind rather than rolling it all back.
 func (si *SearchIndex) mergeArtifactRows(ctx context.Context, total int) (int, error) {
 	selectColumns := artifactSelectColumns(
 		si.artifactStoresText(), si.artifactHasTotals(),
 	)
 	insertSQL := `
 		INSERT INTO explore_index (` + artifactCatalogColumns + `)
-		SELECT ` + artifactCatalogColumns + `
+		SELECT ` + selectColumns + `
 		FROM core.explore_index
 		WHERE mbid > ?` + upsertIndexConflictSQL
@@ -273,7 +359,7 @@ func (si *SearchIndex) mergeArtifactRows(ctx context.Context, total int) (int, e
 	// appended only while one exists.
 	insertRangeSQL := `
 		INSERT INTO explore_index (` + artifactCatalogColumns + `)
-		SELECT ` + artifactCatalogColumns + `
+		SELECT ` + selectColumns + `
 		FROM core.explore_index
 		WHERE mbid > ? AND mbid <= ?` + upsertIndexConflictSQL
@@ -202,7 +202,7 @@ func TestImportCoreArtifactPreservesLocalData(t *testing.T) {
 	if err := db.QueryRowWriter(`
 		SELECT popularity, duration, in_library, discog_fetched
-		FROM explore_index WHERE mbid = ?`, recA,
+		FROM explore_index WHERE mbid = ?`, dbMBID(recA),
 	).Scan(&popularity, &duration, &inLibrary, &discogFetched); err != nil {
 		t.Fatalf("read merged row: %v", err)
 	}
@@ -376,7 +376,7 @@ func TestAddFromCacheNeverStoresMBIDAsName(t *testing.T) {
 	var artistName string
 	if err := db.QueryRowWriter(
-		`SELECT artist_name FROM explore_index WHERE mbid = ?`, rgA,
+		`SELECT artist_name FROM explore_index WHERE mbid = ?`, dbMBID(rgA),
 	).Scan(&artistName); err != nil {
 		t.Fatalf("read release group: %v", err)
 	}
@@ -391,7 +391,7 @@ func TestAddFromCacheNeverStoresMBIDAsName(t *testing.T) {
 	})
 	if err := db.QueryRowWriter(
-		`SELECT artist_name FROM explore_index WHERE mbid = ?`, rgA,
+		`SELECT artist_name FROM explore_index WHERE mbid = ?`, dbMBID(rgA),
 	).Scan(&artistName); err != nil {
 		t.Fatalf("re-read release group: %v", err)
 	}
@@ -442,3 +442,158 @@ func TestArtifactColumnsMatchExporter(t *testing.T) {
 			exporter, importer)
 	}
 }
 // TestImportCoreArtifactAcceptsBothEncodings is the compatibility half
 // of the storage change.
 //
 // The catalog stores an MBID as 16 raw bytes and an entity type as a
 // code, which took the table and its indexes from 677 MB to 389 MB. A
 // published artifact carries whichever form the exporter that built it
 // used, and there is one already out there in the older text form — so
 // the importer decides by asking the artifact, not by trusting a
 // version number, and both must land identically.
 func TestImportCoreArtifactAcceptsBothEncodings(t *testing.T) {
 	compact := filepath.Join(t.TempDir(), "core-index.db")
 	db, err := sql.Open("sqlite", "file:"+compact)
 	if err != nil {
 		t.Fatalf("open artifact: %v", err)
 	}
 	if _, err := db.Exec(`CREATE TABLE explore_index (
 		entity_type      INTEGER NOT NULL,
 		mbid             BLOB NOT NULL,
 		title            TEXT NOT NULL,
 		artist_name      TEXT NOT NULL,
 		artist_mbid      BLOB NOT NULL,
 		aliases          TEXT NOT NULL DEFAULT '',
 		popularity       INTEGER NOT NULL DEFAULT 0,
 		listener_count   INTEGER NOT NULL DEFAULT 0,
 		duration         INTEGER NOT NULL DEFAULT 0,
 		caa_release_mbid BLOB NOT NULL DEFAULT x'',
 		release_name     TEXT NOT NULL DEFAULT '',
 		primary_type     TEXT NOT NULL DEFAULT '',
 		secondary_types  TEXT NOT NULL DEFAULT '',
 		release_date     TEXT NOT NULL DEFAULT '',
 		artist_type      TEXT NOT NULL DEFAULT '',
 		country          TEXT NOT NULL DEFAULT '',
 		disambiguation   TEXT NOT NULL DEFAULT '',
 		sort_name        TEXT NOT NULL DEFAULT '',
 		discog_fetched   INTEGER NOT NULL DEFAULT 0,
 		PRIMARY KEY (mbid)
 	)`); err != nil {
 		t.Fatalf("create artifact table: %v", err)
 	}
 	if _, err := db.Exec(
 		`CREATE TABLE artifact_meta (key TEXT PRIMARY KEY, value TEXT NOT NULL)`,
 	); err != nil {
 		t.Fatalf("create artifact meta: %v", err)
 	}
 	for k, v := range validMeta() {
 		if _, err := db.Exec(
 			"INSERT INTO artifact_meta (key, value) VALUES (?, ?)", k, v,
 		); err != nil {
 			t.Fatalf("write artifact meta: %v", err)
 		}
 	}
 	if _, err := db.Exec(`
 		INSERT INTO explore_index (entity_type, mbid, title, artist_name, artist_mbid, popularity)
 		VALUES (1, ?, 'Artist A', 'Artist A', ?, 5000)`,
 		mbidBytes(artA), mbidBytes(artA),
 	); err != nil {
 		t.Fatalf("write artifact row: %v", err)
 	}
 	_ = db.Close()
 	live := database.NewTestDB(t)
 	si := NewSearchIndex(live, nil, nil, testLogger())
 	if err := si.importCoreArtifact(context.Background(), compact); err != nil {
 		t.Fatalf("importCoreArtifact (compact): %v", err)
 	}
 	got := si.LookupArtistByMBID(artA)
 	if got == nil {
 		t.Fatal("artist from a compact artifact was not imported")
 	}
 	if got.Title != "Artist A" || got.MBID != artA {
 		t.Errorf("imported %+v, want Artist A / %s", got, artA)
 	}
 }
 // TestImportCoreArtifactReadsTotalsWhenPresent covers both halves of the
 // denominator's arrival: an artifact that carries total_tracks imports
 // it, and one built before the column existed still imports at all.
 //
 // The second half is the one worth a test.  Adding a column to the
 // importer's SELECT list is how you break every artifact already
 // published - "no such column: total_tracks", on a file nobody can
 // re-cut retroactively - so the importer asks the artifact what it has,
 // the same way it asks which encoding it uses.
 func TestImportCoreArtifactReadsTotalsWhenPresent(t *testing.T) {
 	path := writeTestArtifact(t, validMeta(), []artifactRow{
 		{"release_group", rgA, "Big Album", "Solo Star", artA, 5000},
 	})
 	// The exporter writes the column; writeTestArtifact builds the older
 	// shape, so add it here rather than changing every other test's
 	// fixture to carry a value they do not use.
 	artifact, err := sql.Open("sqlite", "file:"+path)
 	if err != nil {
 		t.Fatalf("reopen artifact: %v", err)
 	}
 	for _, stmt := range []string{
 		`ALTER TABLE explore_index ADD COLUMN total_tracks INTEGER NOT NULL DEFAULT 0`,
 		`UPDATE explore_index SET total_tracks = 12`,
 	} {
 		if _, err := artifact.Exec(stmt); err != nil {
 			t.Fatalf("add total_tracks: %v", err)
 		}
 	}
 	_ = artifact.Close()
 	live := database.NewTestDB(t)
 	si := NewSearchIndex(live, nil, nil, testLogger())
 	if err := si.importCoreArtifact(context.Background(), path); err != nil {
 		t.Fatalf("importCoreArtifact: %v", err)
 	}
 	rg := si.LookupReleaseGroupByMBID(rgA)
 	if rg == nil {
 		t.Fatal("release group was not imported")
 	}
 	if rg.TotalTracks != 12 {
 		t.Errorf("TotalTracks = %d, want 12", rg.TotalTracks)
 	}
 	// And the shape that predates the column: the same import, from an
 	// artifact that has no total_tracks at all.
 	older := writeTestArtifact(t, validMeta(), []artifactRow{
 		{"release_group", rgB, "Duet Album", "Solo Star", artA, 4000},
 	})
 	live2 := database.NewTestDB(t)
 	si2 := NewSearchIndex(live2, nil, nil, testLogger())
 	if err := si2.importCoreArtifact(context.Background(), older); err != nil {
 		t.Fatalf("importCoreArtifact (no total_tracks column): %v", err)
 	}
 	old := si2.LookupReleaseGroupByMBID(rgB)
 	if old == nil {
 		t.Fatal("release group from a column-less artifact was not imported")
 	}
 	if old.TotalTracks != 0 {
 		t.Errorf("TotalTracks = %d, want 0 (the catalog does not say)", old.TotalTracks)
 	}
 }
@@ -12,19 +12,15 @@ import (
 func seedIndexArtist(t *testing.T, db *database.DB, mbid string, discogFetched int) {
 	t.Helper()
-	if _, err := db.ExecContext(
+	seedIndexResult(t, db, SearchIndexResult{
-		upsertIndexSQL,
+		EntityType:    EntityArtist,
-		"artist", mbid, "Seeded Artist", "Seeded Artist", mbid, "",
+		MBID:          testMBID(mbid),
-		0, 0,
+		Title:         "Seeded Artist",
-		0, "", "",
+		ArtistName:    "Seeded Artist",
-		"", "", "",
+		ArtistMBID:    testMBID(mbid),
-		"", "", "", "",
+		InLibrary:     true,
-		1, 0,
+		DiscogFetched: discogFetched == 1,
-		0, 0, 0,
+	})
 		discogFetched,
 	); err != nil {
 		t.Fatalf("seed explore_index row for %q: %v", mbid, err)
 	}
 }
 // TestArtistEnrichmentMarksAreIndependent is the reason these are two
@@ -0,0 +1,113 @@
 package explore
 import (
 	"log/slog"
 	"net/http"
 	"net/http/httptest"
 	"testing"
 	"yellowjacket/backend/database"
 )
 // lbPopularityServer serves both top-for-artist popularity endpoints
 // with a fixed status and body, and counts what it was asked.
 func lbPopularityServer(
 	t *testing.T, status int, body string,
 ) (*ListenBrainzClient, *int) {
 	t.Helper()
 	requests := 0
 	srv := httptest.NewServer(http.HandlerFunc(
 		func(w http.ResponseWriter, _ *http.Request) {
 			requests++
 			w.Header().Set("Content-Type", "application/json")
 			w.WriteHeader(status)
 			_, _ = w.Write([]byte(body))
 		},
 	))
 	t.Cleanup(srv.Close)
 	lb := NewListenBrainzClient(NewRateLimiterN(1000), nil, slog.Default())
 	lb.SetBaseURL(srv.URL)
 	return lb, &requests
 }
 // TestTopFetchesSeparateEmptyFromFailed is the distinction the
 // owned-artist backfill's mark rests on.  ListenBrainz answers 200 with
 // `[]` for an artist it has no popularity data for — which is most of a
 // long-tail library — and an empty answer is a *complete* one.  When
 // discog_fetched was keyed on "did rows come back", those artists were
 // never marked, stayed in unenrichedLibraryArtistMBIDs forever, and
 // "Filling in artist details" re-ran for them on every single launch.
 func TestTopFetchesSeparateEmptyFromFailed(t *testing.T) {
 	t.Parallel()
 	si := NewSearchIndex(database.NewTestDB(t), nil, nil, slog.Default())
 	artist := lbSitewideArtist{
 		ArtistMBID: "44444444-4444-4444-4444-444444444444",
 		ArtistName: "Nobody Has Listened",
 	}
 	t.Run("empty is success", func(t *testing.T) {
 		t.Parallel()
 		lb, _ := lbPopularityServer(t, http.StatusOK, `[]`)
 		rgs, err := si.fetchTopReleaseGroups(t.Context(), lb, artist, 50)
 		if err != nil || len(rgs) != 0 {
 			t.Errorf("top RGs: got %d rows, err %v; want 0 rows and no error", len(rgs), err)
 		}
 		recs, err := si.fetchTopRecordings(t.Context(), lb, artist, 200)
 		if err != nil || len(recs) != 0 {
 			t.Errorf("top recordings: got %d rows, err %v; want 0 rows and no error",
 				len(recs), err,
 			)
 		}
 	})
 	// Below indexMinPopularity is the same shape one step in: the
 	// endpoint answered, we simply keep none of it.
 	t.Run("everything below the popularity floor is success", func(t *testing.T) {
 		t.Parallel()
 		lb, _ := lbPopularityServer(t, http.StatusOK,
 			`[{"release_group_mbid":"rg-1","total_listen_count":3,`+
 				`"release_group":{"name":"Obscure"}}]`)
 		rgs, err := si.fetchTopReleaseGroups(t.Context(), lb, artist, 50)
 		if err != nil || len(rgs) != 0 {
 			t.Errorf("top RGs: got %d rows, err %v; want 0 rows and no error", len(rgs), err)
 		}
 	})
 	// A failure must stay a failure, or the retry this is built on goes
 	// away and a throttled run marks artists it never fetched.
 	t.Run("HTTP error is a failure", func(t *testing.T) {
 		t.Parallel()
 		lb, _ := lbPopularityServer(t, http.StatusServiceUnavailable, `nope`)
 		if _, err := si.fetchTopReleaseGroups(t.Context(), lb, artist, 50); err == nil {
 			t.Error("top RGs reported success on a 503")
 		}
 		if _, err := si.fetchTopRecordings(t.Context(), lb, artist, 200); err == nil {
 			t.Error("top recordings reported success on a 503")
 		}
 	})
 	t.Run("unparseable body is a failure", func(t *testing.T) {
 		t.Parallel()
 		lb, _ := lbPopularityServer(t, http.StatusOK, `{"not":"an array"}`)
 		if _, err := si.fetchTopReleaseGroups(t.Context(), lb, artist, 50); err == nil {
 			t.Error("top RGs reported success on a body it could not read")
 		}
 	})
 }
@@ -47,6 +47,13 @@ const (
 	// canonicalProgressRows controls progress reporting during the
 	// canonical CSV scan (~30M rows total).
 	canonicalProgressRows = 2_000_000
 	// maxReleaseTracks caps the per-release track count, so a malformed
 	// row cannot turn the denominator into nonsense.  Well above the
 	// longest real release, and it is a cap rather than a rejection
 	// because a box set reporting 999 is still a better answer than one
 	// reporting nothing.
 	maxReleaseTracks = 999
 )
 // Per-artist discography coverage (S2).  The global budgets above keep
@@ -329,6 +336,18 @@ type canonicalScan struct {
 	releaseToRG  map[uuid16]rgTarget
 	artistNames  map[uuid16]string
 	// releaseTracks counts the canonical dump's rows per kept release,
 	// which is that release's track count: the dump carries one row per
 	// recording per canonical release.
 	//
 	// It is counted *before* the popularity filter below, unlike almost
 	// everything else here, because a denominator built from the kept
 	// recordings would say "9" about a twelve-track album whose other
 	// three are unpopular - which is worse than saying nothing, and is
 	// exactly the confident lie that kept whole tracklists out of the
 	// artifact.  Bounded by the kept release set, not by MusicBrainz.
 	releaseTracks map[uuid16]uint16
 	// artistTracks accumulates each target artist's top recordings for
 	// S2 coverage; merged into recordings before assembly.
 	artistTracks *perArtistTracks
@@ -454,10 +473,11 @@ func (imp *dumpImporter) scanCanonicalDump(
 	defer zr.Close()
 	scan := &canonicalScan{
-		releaseInfos: make(map[uuid16]releaseInfo, len(ks.releases)),
+		releaseInfos:  make(map[uuid16]releaseInfo, len(ks.releases)),
-		releaseToRG:  make(map[uuid16]rgTarget, len(ks.releases)),
+		releaseToRG:   make(map[uuid16]rgTarget, len(ks.releases)),
-		artistNames:  make(map[uuid16]string, len(ks.artists)),
+		artistNames:   make(map[uuid16]string, len(ks.artists)),
-		artistTracks: newPerArtistTracks(),
+		releaseTracks: make(map[uuid16]uint16, len(ks.releases)),
 		artistTracks:  newPerArtistTracks(),
 	}
 	sawData, sawRedirect := false, false
@@ -614,9 +634,17 @@ func (imp *dumpImporter) scanCanonicalData(
 			}
 		}
-		// Release display info for release-group titling.
+		// Release display info for release-group titling, and the
 		// release's track count.  Both are for kept releases only, and
 		// the count is taken here rather than below because every row
 		// of this dump is one track of its release regardless of how
 		// often anyone played it.
 		if relOK {
 			if _, kept := ks.releases[relMBID]; kept {
 				if n := scan.releaseTracks[relMBID]; n < maxReleaseTracks {
 					scan.releaseTracks[relMBID] = n + 1
 				}
 				if _, seen := scan.releaseInfos[relMBID]; !seen {
 					firstArtist := ""
 					if len(artistMBIDs) > 0 {
@@ -1004,6 +1032,7 @@ func (imp *dumpImporter) assembleIndex(
 			ArtistName:     info.artistName,
 			ArtistMBID:     info.artistMBID,
 			Popularity:     int(agg.listens),
 			TotalTracks:    int(scan.releaseTracks[agg.bestRel]),
 			CAAReleaseMBID: formatUUID(agg.canonical[:]),
 		})
@@ -277,6 +277,12 @@ func canonicalDataCSV(t *testing.T) []byte {
 			"3", "11", "{" + artA + "," + artB + "}", "Solo Star feat. Guest",
 			relB, "Duet Album", recC, "Duet Song", "x", "1",
 		},
 		// Unplayed, so it survives no popularity floor and is not
 		// indexed -- but it is still a track on Big Album.
 		{
 			"4", "10", "{" + artA + "}", "Solo Star",
 			relA, "Big Album", recD, "Album Filler", "x", "1",
 		},
 	}
 	return csvBytes(t, rows)
@@ -594,7 +600,7 @@ func TestDumpImportEndToEnd(t *testing.T) {
 		rows, err := db.QueryContext(
 			"SELECT title, popularity FROM explore_index WHERE mbid = ? AND entity_type = ?",
-			mbid, entityType,
+			dbMBID(mbid), dbEntityType(entityType),
 		)
 		if err != nil {
 			t.Fatalf("query: %v", err)
@@ -627,11 +633,51 @@ func TestDumpImportEndToEnd(t *testing.T) {
 	assertRow(rgB, "release_group", "Duet Album", 12)
 	assertRow(artA, "artist", "Solo Star", 35)
 	// The per-release-group denominator: how many tracks the release
 	// has, counted from the canonical dump *before* the popularity
 	// filter.  Big Album has three, one of which nobody has played and
 	// which is therefore not indexed as a recording at all -- a
 	// denominator built from the kept recordings would say two, and
 	// "you have 2 of 2" about a three-track album is worse than saying
 	// nothing.
 	assertTotalTracks := func(mbid string, want int) {
 		t.Helper()
 		var got int
 		if err := db.QueryRowWriter(
 			"SELECT total_tracks FROM explore_index WHERE mbid = ? AND entity_type = 2",
 			dbMBID(mbid),
 		).Scan(&got); err != nil {
 			t.Fatalf("read total_tracks for %s: %v", mbid, err)
 		}
 		if got != want {
 			t.Errorf("total_tracks for %s = %d, want %d", mbid, got, want)
 		}
 	}
 	assertTotalTracks(rgA, 3)
 	assertTotalTracks(rgB, 1)
 	// And the unplayed track is still not an indexed recording.
 	var fillerRows int
 	if err := db.QueryRowWriter(
 		"SELECT COUNT(*) FROM explore_index WHERE mbid = ?", dbMBID(recD),
 	).Scan(&fillerRows); err != nil {
 		t.Fatalf("count recD rows: %v", err)
 	}
 	if fillerRows != 0 {
 		t.Errorf("unplayed track was indexed as a recording (%d rows)", fillerRows)
 	}
 	// artB only ever appears in a multi-artist credit: no name is
 	// derivable from the dump, so it must be queued for the API
 	// metadata patch instead of being written nameless.
 	rows, err := db.QueryContext(
-		"SELECT COUNT(*) FROM explore_index WHERE mbid = ?", artB,
+		"SELECT COUNT(*) FROM explore_index WHERE mbid = ?", dbMBID(artB),
 	)
 	if err != nil {
 		t.Fatalf("query artB: %v", err)
@@ -1006,7 +1052,7 @@ func TestListenerCountUpdateDoesNotTouchPopularity(t *testing.T) {
 	}
 	rows, err := db.QueryContext(
-		"SELECT popularity, listener_count FROM explore_index WHERE mbid = ?", recA,
+		"SELECT popularity, listener_count FROM explore_index WHERE mbid = ?", dbMBID(recA),
 	)
 	if err != nil {
 		t.Fatalf("query: %v", err)
@@ -277,7 +277,10 @@ func (si *SearchIndex) commitListenDeltas(
 	defer func() { _ = tx.Rollback() }()
 	if _, err := tx.Exec(
-		"CREATE TEMP TABLE IF NOT EXISTS incr_delta (mbid TEXT, kind TEXT, delta INTEGER)",
+		// mbid and kind are stored the way explore_index stores them,
 		// so the join below is a plain equality rather than a
 		// conversion per row.
 		"CREATE TEMP TABLE IF NOT EXISTS incr_delta (mbid BLOB, kind INTEGER, delta INTEGER)",
 	); err != nil {
 		return fmt.Errorf("incremental temp table: %w", err)
 	}
@@ -357,8 +360,10 @@ func insertDeltas(tx *sql.Tx, kind string, deltas map[string]uint32) error {
 		return nil
 	}
 	code := entityCode(kind)
 	for mbid, d := range deltas {
-		rowArgs = append(rowArgs, mbid, kind, int64(d))
+		rowArgs = append(rowArgs, mbidBytes(mbid), code, int64(d))
 		pending++
 		if pending >= deltaInsertBatch {
@@ -36,7 +36,7 @@ func popularityOf(t *testing.T, db *database.DB, mbid string) (int, bool) {
 	t.Helper()
 	rows, err := db.QueryContext(
-		"SELECT popularity FROM explore_index WHERE mbid = ?", mbid,
+		"SELECT popularity FROM explore_index WHERE mbid = ?", dbMBID(mbid),
 	)
 	if err != nil {
 		t.Fatalf("query popularity: %v", err)
@@ -54,7 +54,7 @@ func (imp *dumpImporter) runPatchPasses(ctx context.Context) {
 func (imp *dumpImporter) patchArtistMetadata(ctx context.Context) {
 	rows, err := imp.si.db.QueryContext(`
 		SELECT mbid FROM explore_index
-		WHERE entity_type = 'artist'
+		WHERE entity_type = 1 /* artist */
 		  AND (artist_type = '' OR country = '' OR title = '' OR title = mbid)
 	`)
 	if err != nil {
@@ -152,7 +152,7 @@ func (imp *dumpImporter) patchSimilarArtists(ctx context.Context) {
 			for _, s := range similar {
 				_, _ = imp.si.db.ExecContext(
 					"UPDATE explore_index SET is_similar = 1 WHERE artist_mbid = ?",
-					s.ArtistMBID,
+					dbMBID(s.ArtistMBID),
 				)
 			}
 		}
@@ -219,7 +219,7 @@ func (imp *dumpImporter) topMBIDs(entityType string, limit int) []string {
 		WHERE entity_type = ? AND listener_count = 0
 		ORDER BY popularity DESC
 		LIMIT ?
-	`, entityType, limit)
+	`, dbEntityType(entityType), limit)
 	if err != nil {
 		return nil
 	}
@@ -229,9 +229,9 @@ func (imp *dumpImporter) topMBIDs(entityType string, limit int) []string {
 	var mbids []string
 	for rows.Next() {
-		var m string
+		var m dbMBID
 		if err := rows.Scan(&m); err == nil {
-			mbids = append(mbids, m)
+			mbids = append(mbids, string(m))
 		}
 	}
@@ -264,7 +264,7 @@ func (si *SearchIndex) updateListenerCounts(updates map[string]PopularityData) i
 			`UPDATE explore_index
 			 SET listener_count = ?
 			 WHERE mbid = ? AND listener_count < ?`,
-			data.ListenerCount, strings.ToLower(mbid), data.ListenerCount,
+			data.ListenerCount, dbMBID(strings.ToLower(mbid)), data.ListenerCount,
 		)
 		if err != nil {
 			continue
@@ -0,0 +1,108 @@
 package explore
 import (
 	"testing"
 	"yellowjacket/backend/database"
 )
 // TestStoredEncodingRoundTrips walks every read path in the package
 // against a row written by the real upsert.
 //
 // It exists because of how this storage change fails when it fails.
 // MBIDs are stored as 16 raw bytes and entity types as codes - which
 // took the catalog and its indexes from 780 MB to 405 MB, measured on a
 // real 2,052,200-row catalog - and SQLite does not coerce between TEXT
 // and BLOB. A query that still compares against a 36-character string
 // therefore returns *no rows* rather than an error, and a scan into a
 // plain string yields sixteen bytes of mojibake. Neither shows up as a
 // failure anywhere except in a result that is quietly empty.
 //
 // So this is not a unit test of the encoding (that is TestMBIDRoundTrip)
 // but a sweep: every query that reads the catalog, asserted to return
 // something and to hand back canonical dashed ids.
 func TestStoredEncodingRoundTrips(t *testing.T) {
 	db := database.NewTestDB(t)
 	si := NewSearchIndex(db, nil, nil, testLogger())
 	const (
 		artist = "c0b2500e-0cef-4130-9b13-1b9d9a2f2c07"
 		album  = "11111111-2222-3333-4444-555555555555"
 	)
 	si.upsertBatch([]SearchIndexResult{
 		{
 			EntityType: EntityArtist,
 			MBID:       artist,
 			Title:      "Radiohead",
 			ArtistName: "Radiohead",
 			ArtistMBID: artist,
 			Popularity: 90000,
 		},
 		{
 			EntityType:     EntityReleaseGroup,
 			MBID:           album,
 			Title:          "Kid A",
 			ArtistName:     "Radiohead",
 			ArtistMBID:     artist,
 			Popularity:     80000,
 			CAAReleaseMBID: album,
 		},
 	})
 	// The stored form is the compact one, not the strings above.
 	var typeMBID, typeEntity string
 	if err := db.QueryRowWriter(
 		"SELECT typeof(mbid), typeof(entity_type) FROM explore_index LIMIT 1",
 	).Scan(&typeMBID, &typeEntity); err != nil {
 		t.Fatalf("typeof: %v", err)
 	}
 	if typeMBID != "blob" || typeEntity != "integer" {
 		t.Errorf("stored as mbid=%s entity_type=%s, want blob/integer", typeMBID, typeEntity)
 	}
 	si.MarkReadyIfPopulated()
 	// Read paths.
 	if got := si.LookupArtistByMBID(artist); got == nil {
 		t.Error("LookupArtistByMBID found nothing")
 	} else if got.MBID != artist || got.ArtistMBID != artist || got.EntityType != EntityArtist {
 		t.Errorf("lookup artist = %+v, want dashed ids and the artist type", got)
 	}
 	if got := si.LookupReleaseGroupByMBID(album); got == nil {
 		t.Error("LookupReleaseGroupByMBID found nothing")
 	} else if got.MBID != album || got.EntityType != EntityReleaseGroup {
 		t.Errorf("lookup album = %+v, want the dashed id and the release-group type", got)
 	}
 	if rgs := si.TopReleaseGroupsByArtist(artist, 5); len(rgs) == 0 {
 		t.Error("TopReleaseGroupsByArtist found nothing")
 	} else if rgs[0].MBID != album || rgs[0].ArtistMBID != artist {
 		t.Errorf("top release groups[0] = %+v, want %s by %s", rgs[0], album, artist)
 	}
 	// The exact-match tier reads the two partial LOWER() indexes, whose
 	// predicate has to agree with its WHERE clause or the seek silently
 	// becomes a scan.
 	if m := si.ExactMatches("radiohead", 3); len(m) == 0 {
 		t.Error("ExactMatches found nothing")
 	} else if m[0].MBID != artist || m[0].EntityType != EntityArtist {
 		t.Errorf("exact match[0] = %+v, want the artist", m[0])
 	}
 	if hits := si.Search(t.Context(), "radiohead", 5); len(hits) == 0 {
 		t.Error("Search found nothing")
 	} else if hits[0].MBID != artist || hits[0].EntityType != EntityArtist {
 		t.Errorf("search hit[0] = %+v, want the artist", hits[0])
 	}
 	if b := si.GetPopularityBatch([]string{artist, album}); b == nil || len(b.Popularity) != 2 {
 		t.Errorf("GetPopularityBatch = %+v, want two entries", b)
 	}
 	if m := si.ReleaseGroupMBIDsForCAAReleaseMBIDs([]string{album}); len(m) != 1 {
 		t.Errorf("CAA lookup = %v, want one entry", m)
 	}
 }
@@ -22,8 +22,8 @@ func seedIndexRow(
 	_, err := db.ExecContext(`
 		INSERT INTO explore_index
 			(entity_type, mbid, title, artist_name, artist_mbid, popularity, listener_count)
-		VALUES (?, ?, ?, ?, '', ?, ?)
+		VALUES (?, ?, ?, ?, x'', ?, ?)
-	`, entityType, mbid, title, artist, popularity, popularity/10)
+	`, dbEntityType(entityType), dbMBID(testMBID(mbid)), title, artist, popularity, popularity/10)
 	if err != nil {
 		t.Fatalf("seed %s/%s: %v", entityType, mbid, err)
 	}
@@ -73,12 +73,12 @@ func TestEvalHarnessIndexRanking(t *testing.T) {
 		{
 			Query:  "radiohead",
 			Note:   "popular exact artist match",
-			Expect: []eval.Expected{{Type: "artist", MBID: "rh"}},
+			Expect: []eval.Expected{{Type: "artist", MBID: testMBID("rh")}},
 		},
 		{
 			Query:  "the teenagers",
 			Note:   "low-popularity exact match must beat high-popularity article match",
-			Expect: []eval.Expected{{Type: "artist", MBID: "teenagers"}},
+			Expect: []eval.Expected{{Type: "artist", MBID: testMBID("teenagers")}},
 		},
 	}
@@ -123,8 +123,13 @@ func TestExploreFTSDiacriticFolding(t *testing.T) {
 				t.Fatalf("query %q returned no hits", tc.query)
 			}
-			if hits[0].MBID != tc.wantMBID {
+			if hits[0].MBID != testMBID(tc.wantMBID) {
-				t.Errorf("query %q: top hit = %q, want %q", tc.query, hits[0].MBID, tc.wantMBID)
+				t.Errorf(
 					"query %q: top hit = %q, want %q",
 					tc.query,
 					hits[0].MBID,
 					testMBID(tc.wantMBID),
 				)
 			}
 		})
 	}
@@ -148,3 +153,30 @@ func TestEvalFixtureFileParses(t *testing.T) {
 		}
 	}
 }
 // seedIndexResult writes one explore_index row through the same
 // upsertBatch every other writer uses.
 //
 // The three seeders that used to bind upsertIndexSQL's parameters by
 // hand said, in a comment, that this was on purpose: a schema change
 // should break the tests where it breaks the app.  It did not - it
 // broke them at "missing argument with index 25", one file at a time,
 // for a column none of them cares about.  Going through the one writer
 // keeps the property they wanted (a field written to the wrong column
 // still fails here) without three copies of a 24-argument list.
 func seedIndexResult(t *testing.T, db *database.DB, r SearchIndexResult) {
 	t.Helper()
 	NewSearchIndex(db, nil, nil, slog.Default()).
 		upsertBatch([]SearchIndexResult{r})
 	// upsertBatch logs and swallows, which is right for a background
 	// merge and useless for a fixture, so the row is checked for.
 	var n int
 	if err := db.QueryRowWriter(
 		"SELECT COUNT(*) FROM explore_index WHERE mbid = ?", dbMBID(r.MBID),
 	).Scan(&n); err != nil || n == 0 {
 		t.Fatalf("seed explore_index row %q: not written (%v)", r.MBID, err)
 	}
 }
@@ -130,12 +130,16 @@ func NewExploreService(logger *slog.Logger, db *database.DB) *Service {
 // MusicBrainz returns the shared cached MB client so other services
 // (e.g. autotag) can reuse it without spinning up a second limiter.
 //
 //wails:ignore // internal wiring, not part of the app's IPC surface.
 func (e *Service) MusicBrainz() *MusicBrainzClient {
 	return e.mb
 }
 // CAALimiter returns the shared Cover Art Archive rate limiter.
 // Consumers must respect it for any fresh CAA HTTP GETs.
 //
 //wails:ignore // internal wiring, not part of the app's IPC surface.
 func (e *Service) CAALimiter() *RateLimiter {
 	return e.caaLimiter
 }
@@ -185,6 +189,8 @@ func (e *Service) CoreCatalogImported() bool {
 // SetJobRegistry wires the background job registry into the search
 // index so its build reports progress and controls to the frontend.
 //
 //wails:ignore // internal wiring, not part of the app's IPC surface.
 func (e *Service) SetJobRegistry(reg *jobs.Registry) {
 	e.index.SetJobRegistry(reg)
 }
@@ -561,6 +567,7 @@ func (e *Service) LookupReleaseGroup(mbid string) (*MBReleaseGroup, error) {
 			FirstReleaseDate: indexed.ReleaseDate,
 			InLibrary:        indexed.InLibrary || indexed.LocalReleaseGroupID > 0,
 			LocalID:          indexed.LocalReleaseGroupID,
 			TotalTracks:      indexed.TotalTracks,
 		}
 		// Background: fetch full MB data once per RG to fill in fields
@@ -18,9 +18,22 @@ func NewLibraryMBIDIndex(db *database.DB) *LibraryMBIDIndex {
 	return &LibraryMBIDIndex{db: db}
 }
-// CheckMBIDs returns which of the given MBIDs exist in the local
+// CheckMBIDs returns which of the given MBIDs the library actually
-// library.  The returned map has MBID → entity type ("artist",
+// has a file for.  The returned map is MBID -> entity type ("artist",
 // "release_group", or "recording").
 //
 // The "has a file" part is the whole point and is what this used to get
 // wrong.  It was three `SELECT mbid FROM <metadata table>` queries, and
 // a metadata row could outlive the file that created it - retagging a
 // file abandoned its old recording row, which kept the old MBID
 // forever.  Measured on a real library: 812 orphaned recordings, of
 // which 218 carried MBIDs, and 129 catalog rows that this function
 // therefore reported as owned.  Every one of them rendered as a track
 // you have, with a play button that could not work, because playback
 // resolves files and this resolved metadata.
 //
 // Each branch now joins audio_files.  An entity is in your library if
 // and only if a file says so.
 func (idx *LibraryMBIDIndex) CheckMBIDs(mbids []string) map[string]string {
 	if len(mbids) == 0 {
 		return nil
@@ -28,16 +41,26 @@ func (idx *LibraryMBIDIndex) CheckMBIDs(mbids []string) map[string]string {
 	result := make(map[string]string, len(mbids))
-	// Batch check all MBIDs against each table with a single IN query.
+	type entityQuery struct {
 	type tableEntity struct {
 		table      string
 		entityType string
 		query      string
 	}
-	tables := []tableEntity{
+	queries := []entityQuery{
-		{"artists", "artist"},
+		{"recording", `SELECT DISTINCT recording_mbid FROM audio_files
-		{"release_groups", "release_group"},
+			WHERE recording_mbid IN (%s)`},
-		{"recordings", "recording"},
+		{"release_group", `SELECT DISTINCT al.mbid FROM albums al
 			JOIN audio_files af ON af.album_id = al.id
 			WHERE al.mbid IN (%s)`},
 		{"artist", `SELECT DISTINCT a.mbid FROM artists a
 			WHERE a.mbid IN (%s) AND (
 				EXISTS (SELECT 1 FROM audio_files af WHERE af.artist_id = a.id)
 				OR EXISTS (
 					SELECT 1 FROM albums al
 					JOIN audio_files af2 ON af2.album_id = al.id
 					WHERE al.artist_id = a.id
 				)
 			)`},
 	}
 	// Build a set of MBIDs still unresolved.
@@ -48,12 +71,11 @@ func (idx *LibraryMBIDIndex) CheckMBIDs(mbids []string) map[string]string {
 		}
 	}
-	for _, te := range tables {
+	for _, eq := range queries {
 		if len(remaining) == 0 {
 			break
 		}
 		// Build IN clause from remaining MBIDs.
 		placeholders := make([]string, 0, len(remaining))
 		args := make([]any, 0, len(remaining))
@@ -62,9 +84,10 @@ func (idx *LibraryMBIDIndex) CheckMBIDs(mbids []string) map[string]string {
 			args = append(args, m)
 		}
-		//nolint:gosec // table name is hardcoded from the tables slice above
+		//nolint:gosec // the query text is a constant from the slice above
-		query := "SELECT mbid FROM " + te.table + " WHERE mbid IN (" +
+		query := strings.Replace(
-			strings.Join(placeholders, ",") + ")"
+			eq.query, "%s", strings.Join(placeholders, ","), 1,
 		)
 		rows, err := idx.db.QueryContext(query, args...)
 		if err != nil {
@@ -74,7 +97,7 @@ func (idx *LibraryMBIDIndex) CheckMBIDs(mbids []string) map[string]string {
 		for rows.Next() {
 			var mbid string
 			if err := rows.Scan(&mbid); err == nil {
-				result[mbid] = te.entityType
+				result[mbid] = eq.entityType
 				delete(remaining, mbid)
 			}
 		}
@@ -33,6 +33,12 @@ type ListenBrainzClient struct {
 	limiter *RateLimiter
 	cache   *Cache
 	logger  *slog.Logger
 	// baseURL is listenBrainzBaseURL unless a test redirects it.  It is
 	// per-client rather than a package variable — the MB client's
 	// SetBaseURL shape — so a test that points one client at an
 	// httptest server does not stop being parallel-safe.
 	baseURL string
 }
 // NewListenBrainzClient creates a ListenBrainz API client.
@@ -46,9 +52,15 @@ func NewListenBrainzClient(
 		limiter: limiter,
 		cache:   cache,
 		logger:  logger,
 		baseURL: listenBrainzBaseURL,
 	}
 }
 // SetBaseURL redirects this client at another host.  Tests only.
 func (c *ListenBrainzClient) SetBaseURL(url string) {
 	c.baseURL = strings.TrimSuffix(url, "/")
 }
 // TopRecordingsForArtist returns the most-listened recordings for
 // the artist identified by artistMBID.
 func (c *ListenBrainzClient) TopRecordingsForArtist(
@@ -56,7 +68,7 @@ func (c *ListenBrainzClient) TopRecordingsForArtist(
 ) ([]LBTopRecording, error) {
 	url := fmt.Sprintf(
 		"%s/1/popularity/top-recordings-for-artist/%s",
-		listenBrainzBaseURL,
+		c.baseURL,
 		artistMBID,
 	)
 	cacheKey := "lb:top-recordings:" + artistMBID
@@ -104,7 +116,7 @@ func (c *ListenBrainzClient) TopReleaseGroupsForArtist(
 ) ([]LBTopReleaseGroup, error) {
 	url := fmt.Sprintf(
 		"%s/1/popularity/top-release-groups-for-artist/%s",
-		listenBrainzBaseURL,
+		c.baseURL,
 		artistMBID,
 	)
 	cacheKey := "lb:top-release-groups:" + artistMBID
@@ -234,7 +246,7 @@ func (c *ListenBrainzClient) ArtistPopularity(
 		return nil, nil //nolint:nilnil
 	}
-	url := listenBrainzBaseURL + "/1/popularity/artist"
+	url := c.baseURL + "/1/popularity/artist"
 	cacheKey := "lb:pop:artist:" + hashMBIDs(mbids)
 	if data, ok := c.cache.Get(cacheKey); ok {
@@ -265,7 +277,7 @@ func (c *ListenBrainzClient) RecordingPopularity(
 		return nil, nil //nolint:nilnil
 	}
-	url := listenBrainzBaseURL + "/1/popularity/recording"
+	url := c.baseURL + "/1/popularity/recording"
 	cacheKey := "lb:pop:recording:" + hashMBIDs(mbids)
 	if data, ok := c.cache.Get(cacheKey); ok {
@@ -296,7 +308,7 @@ func (c *ListenBrainzClient) ReleaseGroupPopularity(
 		return nil, nil //nolint:nilnil
 	}
-	url := listenBrainzBaseURL + "/1/popularity/release-group"
+	url := c.baseURL + "/1/popularity/release-group"
 	cacheKey := "lb:pop:release-group:" + hashMBIDs(mbids)
 	if data, ok := c.cache.Get(cacheKey); ok {
@@ -342,7 +354,7 @@ func (c *ListenBrainzClient) BatchArtistMetadata(
 		return nil, nil //nolint:nilnil
 	}
-	url := listenBrainzBaseURL + "/1/metadata/artist/?artist_mbids=" + strings.Join(mbids, ",")
+	url := c.baseURL + "/1/metadata/artist/?artist_mbids=" + strings.Join(mbids, ",")
 	cacheKey := "lb:meta:artist:" + hashMBIDs(mbids)
 	if data, ok := c.cache.Get(cacheKey); ok {
@@ -10,7 +10,7 @@ import (
 // LyricsResult is a single lyric-search hit, mapped from the DB layer
 // into the camelCase shape the frontend consumes.
 type LyricsResult struct {
-	RecordingID int64  `json:"recordingId"`
+	AudioFileID int64  `json:"audioFileId"`
 	FilePath    string `json:"filePath"`
 	LengthMs    int64  `json:"lengthMs"`
 	Title       string `json:"title"`
@@ -55,7 +55,7 @@ func (e *Service) SearchLyrics(query string) []LyricsResult {
 	out := make([]LyricsResult, 0, len(hits))
 	for _, h := range hits {
 		out = append(out, LyricsResult{
-			RecordingID: h.RecordingID,
+			AudioFileID: h.AudioFileID,
 			FilePath:    h.FilePath,
 			LengthMs:    h.LengthMilliseconds,
 			Title:       h.Title,
@@ -67,18 +67,18 @@ func (e *Service) SearchLyrics(query string) []LyricsResult {
 	return out
 }
-// GetTrackLyrics returns lyrics for a recording.  If the library
+// GetTrackLyrics returns lyrics for a file.  If the library
 // already has them (from embedded tags) they're returned as-is;
 // otherwise it fetches from LRCLIB, persists them (updating the FTS
 // index), and returns them.  Never returns an error to the frontend —
 // a miss just yields an empty result.
-func (e *Service) GetTrackLyrics(recordingID int64) TrackLyrics {
+func (e *Service) GetTrackLyrics(audioFileID int64) TrackLyrics {
-	stored, err := e.db.GetRecordingLyrics(recordingID)
+	stored, err := e.db.GetLyrics(audioFileID)
 	if err == nil && stored != "" {
 		return TrackLyrics{Plain: stored, Source: "embedded"}
 	}
-	lookup, err := e.db.RecordingLyricLookup(recordingID)
+	lookup, err := e.db.FileLyricLookup(audioFileID)
 	if err != nil || lookup == nil {
 		return TrackLyrics{}
 	}
@@ -150,7 +150,7 @@ func (e *Service) backfillLibraryLyrics(ctx context.Context) {
 			return
 		}
-		candidates, err := e.db.RecordingsMissingLyrics(lyricsBackfillBatch)
+		candidates, err := e.db.FilesMissingLyrics(lyricsBackfillBatch)
 		if err != nil {
 			e.logger.Warn("lyrics backfill: query failed", "err", err)
@@ -223,8 +223,12 @@ func (e *Service) fetchAndStoreLyrics(
 		return nil
 	}
-	if err := e.db.SetRecordingLyrics(c.RecordingID, lyrics.Plain); err != nil {
+	// Marked `lrclib` rather than `tag`: these came off the network and
-		e.logger.Warn("lyrics store failed", "recordingId", c.RecordingID, "err", err)
+	// a rebuild that discards them pays for them again.
 	if err := e.db.SetLyrics(
 		c.AudioFileID, lyrics.Plain, "lrclib", c.RecordingMBID,
 	); err != nil {
 		e.logger.Warn("lyrics store failed", "audioFileId", c.AudioFileID, "err", err)
 		return nil
 	}
@@ -0,0 +1,203 @@
 package explore
 import (
 	"database/sql/driver"
 	"encoding/hex"
 	"errors"
 	"fmt"
 	"strings"
 )
 // The catalog stores a MusicBrainz id as its 16 raw bytes and an entity
 // type as a small integer, rather than as the 36-character text and the
 // words the rest of the app uses.
 //
 // This is a size decision and it is a large one. Measured on a real
 // 2,052,200-row catalog: the three MBID columns and `entity_type` are
 // 220 MB of a 383 MB table, and they are carried again in every index
 // that keys on them. Converting the table and its four indexes took
 // **677 MB to 389 MB** with the same row count.
 //
 // Everything above this file still speaks strings: `SearchIndexResult`
 // carries `"artist"` and a dashed MBID, the frontend receives them, and
 // the conversion happens only where a value crosses into SQL. The
 // alternative - blobs and codes reaching the rest of the app - would
 // trade 288 MB for a type that nothing else wants.
 //
 // Two things make a mistake here loud rather than silent, which matters
 // because SQLite does not coerce between TEXT and BLOB: a query
 // comparing a blob column against a string returns *no rows* rather
 // than an error.
 //
 //   - Writes are guarded by a CHECK on the column (16 bytes, or empty),
 //     so a stringly write fails at the insert rather than sitting in
 //     the table looking fine.
 //   - Reads scan into `dbMBID`, whose Scan rejects anything that is not
 //     16 bytes or empty. A column that somehow holds text produces an
 //     error instead of a garbled id.
 type dbMBID string
 // mbidLen is the byte length of a raw MusicBrainz id.
 const mbidLen = 16
 // Value encodes the id for storage: 16 raw bytes, or empty for "none".
 func (m dbMBID) Value() (driver.Value, error) {
 	return mbidBytes(string(m)), nil
 }
 // Scan decodes a stored id back to its canonical dashed form.
 func (m *dbMBID) Scan(src any) error {
 	switch v := src.(type) {
 	case nil:
 		*m = ""
 		return nil
 	case []byte:
 		s, err := mbidFromBytes(v)
 		if err != nil {
 			return err
 		}
 		*m = dbMBID(s)
 		return nil
 	case string:
 		// Tolerated for the one caller that reads through a view or a
 		// literal: a canonical id is already the right answer.
 		*m = dbMBID(v)
 		return nil
 	default:
 		return fmt.Errorf("%w: %T", errMBIDType, src)
 	}
 }
 // mbidBytes encodes a dashed MusicBrainz id as its 16 raw bytes.  An
 // id that is not one - including the empty string, which is how "no
 // MBID" is spelled throughout - encodes as empty.
 func mbidBytes(s string) []byte {
 	if s == "" {
 		return []byte{}
 	}
 	raw, err := hex.DecodeString(strings.ReplaceAll(s, "-", ""))
 	if err != nil || len(raw) != mbidLen {
 		return []byte{}
 	}
 	return raw
 }
 // mbidFromBytes decodes stored bytes back to the canonical dashed form.
 func mbidFromBytes(b []byte) (string, error) {
 	if len(b) == 0 {
 		return "", nil
 	}
 	if len(b) != mbidLen {
 		return "", fmt.Errorf("%w: %d bytes", errMBIDLength, len(b))
 	}
 	h := hex.EncodeToString(b)
 	return h[0:8] + "-" + h[8:12] + "-" + h[12:16] + "-" + h[16:20] + "-" + h[20:], nil
 }
 // Entity types are stored as codes.  The names are the app's, the codes
 // are the table's, and nothing outside this file should see a code.
 const (
 	entityCodeArtist       = 1
 	entityCodeReleaseGroup = 2
 	entityCodeRecording    = 3
 )
 // Entity type names, as everything above the SQL boundary spells them.
 const (
 	EntityArtist       = "artist"
 	EntityReleaseGroup = "release_group"
 	EntityRecording    = "recording"
 )
 // entityCode maps a name to its stored code.  An unknown name yields 0,
 // which matches no row - the same answer the old string comparison gave
 // and the reason this is not an error.
 func entityCode(name string) int {
 	switch name {
 	case EntityArtist:
 		return entityCodeArtist
 	case EntityReleaseGroup:
 		return entityCodeReleaseGroup
 	case EntityRecording:
 		return entityCodeRecording
 	default:
 		return 0
 	}
 }
 // entityName maps a stored code back to its name.
 func entityName(code int) string {
 	switch code {
 	case entityCodeArtist:
 		return EntityArtist
 	case entityCodeReleaseGroup:
 		return EntityReleaseGroup
 	case entityCodeRecording:
 		return EntityRecording
 	default:
 		return ""
 	}
 }
 // dbEntityType scans a stored entity-type code as its name.
 //
 // The db prefix keeps it out of the way of the many `mbid string` and
 // `entityType string` locals in this package: these two types exist
 // only at the SQL boundary, and a name that shadowed one of those
 // would turn a conversion into a confusing compile error rather than
 // an obvious one.
 type dbEntityType string
 // Value encodes the name for storage.
 func (e dbEntityType) Value() (driver.Value, error) {
 	return int64(entityCode(string(e))), nil
 }
 // Scan decodes a stored code back to its name.
 func (e *dbEntityType) Scan(src any) error {
 	switch v := src.(type) {
 	case nil:
 		*e = ""
 		return nil
 	case int64:
 		*e = dbEntityType(entityName(int(v)))
 		return nil
 	case string:
 		*e = dbEntityType(v)
 		return nil
 	case []byte:
 		*e = dbEntityType(v)
 		return nil
 	default:
 		return fmt.Errorf("%w: %T", errEntityTypeType, src)
 	}
 }
 // Errors from the encoding boundary.  They exist so a type confusion
 // here is reported rather than silently producing an id that matches
 // nothing.
 var (
 	errMBIDType       = errors.New("cannot scan MusicBrainz id from")
 	errMBIDLength     = errors.New("stored MusicBrainz id has the wrong length")
 	errEntityTypeType = errors.New("cannot scan entity type from")
 )
 // A query that names an entity type inline writes the code with the
 // name beside it - `entity_type = 1 /* artist */`.  Splicing a Go
 // constant into the SQL would keep them in step automatically but makes
 // every such query a concatenation; the codes are pinned by
 // TestEntityCodesAreStable instead, because they are a storage format
 // and changing one is not a refactor.
@@ -0,0 +1,105 @@
 package explore
 import (
 	"crypto/sha256"
 	"encoding/hex"
 	"testing"
 )
 // testMBID turns a short fixture label into a well-formed MusicBrainz
 // id, and passes a real one through unchanged.
 //
 // The catalog stores an id as its 16 raw bytes and the column says so
 // (`CHECK(length(mbid) = 16)`), so a fixture can no longer call itself
 // "rh". Deriving one from the label keeps the fixtures readable — the
 // same label is the same id in a seed and in the assertion that reads
 // it back — without letting a test write something the app could not.
 func testMBID(label string) string {
 	if len(mbidBytes(label)) == mbidLen {
 		return label
 	}
 	sum := sha256.Sum256([]byte(label))
 	h := hex.EncodeToString(sum[:mbidLen])
 	return h[0:8] + "-" + h[8:12] + "-" + h[12:16] + "-" + h[16:20] + "-" + h[20:]
 }
 // TestEntityCodesAreStable pins the stored entity-type codes.
 //
 // They are a storage format, not an enum: the queries that name a type
 // inline write the number with the name beside it
 // (`entity_type = 1 /* artist */`), so changing one here without
 // changing those would leave the catalog answering the wrong questions
 // silently.
 func TestEntityCodesAreStable(t *testing.T) {
 	t.Parallel()
 	for name, want := range map[string]int{
 		EntityArtist:       1,
 		EntityReleaseGroup: 2,
 		EntityRecording:    3,
 	} {
 		if got := entityCode(name); got != want {
 			t.Errorf("entityCode(%q) = %d, want %d", name, got, want)
 		}
 		if back := entityName(want); back != name {
 			t.Errorf("entityName(%d) = %q, want %q", want, back, name)
 		}
 	}
 	if got := entityCode("nonsense"); got != 0 {
 		t.Errorf("entityCode of an unknown name = %d, want 0 (matches nothing)", got)
 	}
 }
 // TestMBIDRoundTrip pins the encoding both ways, including the two
 // values that are not ids: empty, which is how "no MBID" is spelled
 // everywhere, and rubbish, which must not become a valid-looking id.
 func TestMBIDRoundTrip(t *testing.T) {
 	t.Parallel()
 	const canonical = "c0b2500e-0cef-4130-9b13-1b9d9a2f2c07"
 	encoded := mbidBytes(canonical)
 	if len(encoded) != mbidLen {
 		t.Fatalf("encoded length = %d, want %d", len(encoded), mbidLen)
 	}
 	back, err := mbidFromBytes(encoded)
 	if err != nil {
 		t.Fatalf("decode: %v", err)
 	}
 	if back != canonical {
 		t.Errorf("round trip = %q, want %q", back, canonical)
 	}
 	if got := mbidBytes(""); len(got) != 0 {
 		t.Errorf("empty encoded to %d bytes, want 0", len(got))
 	}
 	if got := mbidBytes("not-an-mbid"); len(got) != 0 {
 		t.Errorf("rubbish encoded to %d bytes, want 0", len(got))
 	}
 	// A stored value of the wrong length is an error, not a guess.
 	if _, err := mbidFromBytes([]byte{1, 2, 3}); err == nil {
 		t.Error("decoding three bytes should fail")
 	}
 }
 // TestMBIDScanRejectsText is the guard for the failure this encoding
 // could otherwise hide: SQLite does not coerce TEXT to BLOB, so a
 // column holding the old 36-character form would silently compare equal
 // to nothing. Scanning it must say so.
 func TestMBIDScanRejectsText(t *testing.T) {
 	t.Parallel()
 	var m dbMBID
 	if err := m.Scan([]byte("c0b2500e-0cef-4130-9b13-1b9d9a2f2c07")); err == nil {
 		t.Error("scanning 36 bytes as an MBID should fail")
 	}
 }
@@ -107,12 +107,11 @@ func (e *Service) mixSeedProfile(
 			artistCounts[artist.ArtistMbid]++
 			artistNames[artist.ArtistMbid] = artist.ArtistName
 		}
 	}
-		names, err := e.db.ReadQueries.GetGenreNamesByFilePath(ctx, p)
+	for _, names := range e.genresByPath(ctx, seedPaths) {
-		if err == nil {
+		for _, g := range names {
-			for _, g := range names {
+			genres[g] = true
 				genres[g] = true
 			}
 		}
 	}
@@ -148,6 +147,12 @@ func (e *Service) mixCandidates(
 	candidates := map[string]float64{}
 	// Every candidate path's genres, in one query.  This used to be a
 	// single-row lookup *per candidate* inside two nested loops -
 	// twenty seed artists by twenty similar artists by thirty paths is
 	// twelve thousand queries to assemble one mix.
 	pathGenres := e.genresByPath(ctx, e.similarArtistPaths(ctx, artistCounts))
 	for seedArtistMBID, count := range artistCounts {
 		similar, err := e.SimilarArtists(seedArtistMBID)
 		if err != nil {
@@ -178,13 +183,11 @@ func (e *Service) mixCandidates(
 					continue
 				}
-				if names, err := e.db.ReadQueries.GetGenreNamesByFilePath(ctx, p); err == nil {
+				for _, g := range pathGenres[p] {
-					for _, g := range names {
+					if seedGenres[g] {
-						if seedGenres[g] {
+						weight += mixGenreBoost
 							weight += mixGenreBoost
-							break
+						break
 						}
 					}
 				}
@@ -196,6 +199,64 @@ func (e *Service) mixCandidates(
 	return candidates
 }
 // genresByPath returns the genres of many files in one query.
 func (e *Service) genresByPath(
 	ctx context.Context, paths []string,
 ) map[string][]string {
 	out := make(map[string][]string, len(paths))
 	if len(paths) == 0 {
 		return out
 	}
 	rows, err := e.db.ReadQueries.GetGenreNamesByFilePaths(ctx, paths)
 	if err != nil {
 		return out
 	}
 	for _, row := range rows {
 		out[row.FilePath] = append(out[row.FilePath], row.Name)
 	}
 	return out
 }
 // similarArtistPaths collects every owned file by an artist similar to
 // one of the seeds, so their genres can be fetched in one go.
 func (e *Service) similarArtistPaths(
 	ctx context.Context, artistCounts map[string]int,
 ) []string {
 	var paths []string
 	for seedArtistMBID := range artistCounts {
 		similar, err := e.SimilarArtists(seedArtistMBID)
 		if err != nil {
 			continue
 		}
 		if len(similar) > mixSimilarArtistsPerSeed {
 			similar = similar[:mixSimilarArtistsPerSeed]
 		}
 		for _, sim := range similar {
 			if sim.ArtistMBID == "" {
 				continue
 			}
 			p, err := e.db.ReadQueries.GetFilePathsByArtistMBID(
 				ctx, sql.NullString{String: sim.ArtistMBID, Valid: true},
 			)
 			if err != nil {
 				continue
 			}
 			paths = append(paths, p...)
 		}
 	}
 	return paths
 }
 // weightedSample picks up to n distinct keys from weights without
 // replacement, biased toward higher weight (roulette-wheel selection).
 // A key with zero or negative weight is never picked.
@@ -23,69 +23,14 @@ func seedMixTrack(
 	fp := fmt.Sprintf("/music/%s/track%d.mp3", artistName, id)
-	_, err := db.ExecContext(
+	database.InsertTestTrack(t, db, database.TestTrack{
-		"INSERT INTO artists (id, name, mbid) VALUES (?, ?, ?) "+
+		FilePath:   fp,
-			"ON CONFLICT(name) DO NOTHING",
+		Title:      fmt.Sprintf("Track %d", id),
-		id, artistName, artistMBID,
+		Artist:     artistName,
-	)
+		ArtistMBID: artistMBID,
-	if err != nil {
+		Genres:     genreNames,
-		t.Fatalf("insert artist: %v", err)
+		LengthMs:   180000,
-	}
+	})
 	_, err = db.ExecContext(
 		"INSERT INTO artist_credit (id, text) VALUES (?, ?) "+
 			"ON CONFLICT(text) DO NOTHING",
 		id, artistName,
 	)
 	if err != nil {
 		t.Fatalf("insert artist_credit: %v", err)
 	}
 	_, err = db.ExecContext(
 		"INSERT OR IGNORE INTO artist_credit_artist (artist_id, credit_id) VALUES (?, ?)",
 		id, id,
 	)
 	if err != nil {
 		t.Fatalf("insert artist_credit_artist: %v", err)
 	}
 	_, err = db.ExecContext(
 		"INSERT INTO recordings (id, name, artist_credit_id) VALUES (?, ?, ?)",
 		id, fmt.Sprintf("Track %d", id), id,
 	)
 	if err != nil {
 		t.Fatalf("insert recording: %v", err)
 	}
 	_, err = db.ExecContext(
 		"INSERT INTO audio_files (id, file_path, length_milliseconds, file_type_id, recording_id) "+
 			"VALUES (?, ?, 180000, 0, ?)",
 		id, fp, id,
 	)
 	if err != nil {
 		t.Fatalf("insert audio_file: %v", err)
 	}
 	for _, g := range genreNames {
 		var genreID int64
 		row := db.QueryRowWriter(
 			"INSERT INTO genres (name) VALUES (?) "+
 				"ON CONFLICT(name) DO UPDATE SET name = name RETURNING id",
 			g,
 		)
 		if err := row.Scan(&genreID); err != nil {
 			t.Fatalf("upsert genre %q: %v", g, err)
 		}
 		_, err = db.ExecContext(
 			"INSERT OR IGNORE INTO recording_genres (recording_id, genre_id) VALUES (?, ?)",
 			id, genreID,
 		)
 		if err != nil {
 			t.Fatalf("insert recording_genre: %v", err)
 		}
 	}
 	return fp
 }
--- a/Show More
+++ b/Show More
		`@@ -1 +0,0 @@`
			`ALTER TABLE tagging_items ADD COLUMN album_artist_conflict INTEGER NOT NULL DEFAULT 0;`
		`@@ -1 +0,0 @@`
			`ALTER TABLE release_groups ADD COLUMN pending_release_mbid TEXT;`
		`@@ -1 +0,0 @@`
			`ALTER TABLE release_group_recordings ADD COLUMN total_tracks INTEGER;`