Files
yellowjacket/backend/home/home_test.go
yonluandClaude Opus 5 e7748f1fd5
CI / check (push) Successful in 3m7s
CI / e2e (push) Canceled after 1m45s
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

317 lines
8.5 KiB
Go

package home_test
import (
"log/slog"
"testing"
"yellowjacket/backend/database"
"yellowjacket/backend/home"
"yellowjacket/backend/library"
)
// fakeLibrary answers the one question the home service asks of the
// library package, so these tests are about shelf selection rather than
// about album rendering.
type fakeLibrary struct {
albums []library.Album
err error
}
func (f fakeLibrary) GetAlbums(int64) ([]library.Album, error) {
return f.albums, f.err
}
// seed inserts one album with one played-or-not track, returning the
// release group id. Written with raw SQL rather than the library
// scanner because the shelves are queries, and a query is best tested
// against rows it can be given precisely.
func seed(
t *testing.T,
db *database.DB,
name, artist, genre string,
playCount int,
lastPlayed string,
) int64 {
t.Helper()
var genres []string
if genre != "" {
genres = []string{genre}
}
fileID := database.InsertTestTrack(t, db, database.TestTrack{
FilePath: "/music/" + name + ".mp3",
Title: name + " track",
Artist: artist,
Album: name,
Genres: genres,
LengthMs: 1000,
PlayCount: int64(playCount),
})
if lastPlayed != "" {
if _, err := db.ExecContext(
"UPDATE audio_files SET last_played = ? WHERE id = ?", lastPlayed, fileID,
); err != nil {
t.Fatalf("seed last_played: %v", err)
}
}
var id int64
if err := db.QueryRowWriter(
"SELECT album_id FROM audio_files WHERE id = ?", fileID,
).Scan(&id); err != nil {
t.Fatalf("seed: read album id: %v", err)
}
return id
}
func shelfKinds(shelves []home.Shelf) []home.Kind {
kinds := make([]home.Kind, 0, len(shelves))
for _, s := range shelves {
kinds = append(kinds, s.Kind)
}
return kinds
}
func hasKind(shelves []home.Shelf, kind home.Kind) bool {
for _, s := range shelves {
if s.Kind == kind {
return true
}
}
return false
}
func shelfFor(shelves []home.Shelf, kind home.Kind) home.Shelf {
for _, s := range shelves {
if s.Kind == kind {
return s
}
}
return home.Shelf{}
}
func TestGetShelvesEmptyLibraryHasNoShelves(t *testing.T) {
db := database.NewTestDB(t)
svc := home.NewService(slog.Default(), db, fakeLibrary{})
shelves, err := svc.GetShelves()
if err != nil {
t.Fatalf("GetShelves: %v", err)
}
if len(shelves) != 0 {
t.Fatalf("empty library produced shelves: %v", shelfKinds(shelves))
}
}
func TestGetShelvesOmitsShelvesWithNothingBehindThem(t *testing.T) {
// A library nothing has ever been played from must not claim to
// know what is on repeat: an empty row labelled with a reason is a
// worse answer than no row.
db := database.NewTestDB(t)
id := seed(t, db, "Quiet", "Nobody", "", 0, "")
svc := home.NewService(slog.Default(), db, fakeLibrary{
albums: []library.Album{{ID: id, Name: "Quiet", ArtistName: "Nobody"}},
})
shelves, err := svc.GetShelves()
if err != nil {
t.Fatalf("GetShelves: %v", err)
}
if hasKind(shelves, home.KindRecentlyPlayed) {
t.Error("recently-played shelf built from no plays")
}
if hasKind(shelves, home.KindMostPlayed) {
t.Error("most-played shelf built from no plays")
}
if !hasKind(shelves, home.KindUnplayed) {
t.Errorf("expected an unplayed shelf, got %v", shelfKinds(shelves))
}
if !hasKind(shelves, home.KindRecentlyAdded) {
t.Errorf("expected a recently-added shelf, got %v", shelfKinds(shelves))
}
}
func TestGetShelvesRanksPlayHistory(t *testing.T) {
db := database.NewTestDB(t)
old := seed(t, db, "Old Favourite", "A", "", 20, "2020-01-01 00:00:00")
recent := seed(t, db, "Last Night", "B", "", 3, "2999-01-01 00:00:00")
svc := home.NewService(slog.Default(), db, fakeLibrary{
albums: []library.Album{
{ID: old, Name: "Old Favourite", ArtistName: "A"},
{ID: recent, Name: "Last Night", ArtistName: "B"},
},
})
shelves, err := svc.GetShelves()
if err != nil {
t.Fatalf("GetShelves: %v", err)
}
played := shelfFor(shelves, home.KindRecentlyPlayed)
if len(played.Albums) == 0 || played.Albums[0].ID != recent {
t.Errorf("recently played led with %v, want the newest play", played.Albums)
}
most := shelfFor(shelves, home.KindMostPlayed)
if len(most.Albums) == 0 || most.Albums[0].ID != old {
t.Errorf("most played led with %v, want the highest play count", most.Albums)
}
// Played long ago is "forgotten"; the never-played fallback must
// not take over while there is real history to report.
forgotten := shelfFor(shelves, home.KindStale)
if len(forgotten.Albums) == 0 || forgotten.Albums[0].ID != old {
t.Errorf("forgotten shelf = %v, want the album last played in 2020", forgotten.Albums)
}
}
func TestGetShelvesBuildsAGenreShelf(t *testing.T) {
db := database.NewTestDB(t)
albums := make([]library.Album, 0, 4)
for _, name := range []string{"One", "Two", "Three", "Four"} {
id := seed(t, db, name, "Various", "Doom Jazz", 1, "2024-01-01 00:00:00")
albums = append(albums, library.Album{
ID: id, Name: name, ArtistName: "Various",
})
}
svc := home.NewService(slog.Default(), db, fakeLibrary{albums: albums})
shelves, err := svc.GetShelves()
if err != nil {
t.Fatalf("GetShelves: %v", err)
}
genre := shelfFor(shelves, home.KindGenre)
if genre.Title != "Doom Jazz" {
t.Fatalf("genre shelf = %q, want the library's one genre", genre.Title)
}
if len(genre.Albums) != len(albums) {
t.Errorf("genre shelf had %d albums, want %d", len(genre.Albums), len(albums))
}
}
func TestGetShelvesSkipsAlbumsTheLibraryNoLongerHas(t *testing.T) {
// The id queries and the album list are two reads, and a scan can
// remove an album between them. A stale id must vanish from the
// shelf, not render as a blank card.
db := database.NewTestDB(t)
kept := seed(t, db, "Kept", "A", "", 5, "2024-01-01 00:00:00")
seed(t, db, "Removed", "B", "", 5, "2024-01-02 00:00:00")
svc := home.NewService(slog.Default(), db, fakeLibrary{
albums: []library.Album{{ID: kept, Name: "Kept", ArtistName: "A"}},
})
shelves, err := svc.GetShelves()
if err != nil {
t.Fatalf("GetShelves: %v", err)
}
for _, shelf := range shelves {
for _, album := range shelf.Albums {
if album.ID != kept {
t.Fatalf("shelf %q surfaced a removed album: %+v", shelf.ID, album)
}
}
}
}
// A shelf has to be worth its own row.
//
// A library with one signal answers several questions with the same
// albums, so shelves render the same covers in different orders under
// different reasons and the page reads as repeating itself (H-9).
// Confirmed in the running app on the fixture library before this
// existed: "On repeat" was "Pick up where you left off" reordered.
//
// The library has to be bigger than one shelf for the rule to apply at
// all, which is the point: a repeat is only a fault if a different row
// was possible.
func TestGetShelvesOmitsAShelfThatRepeatsTheOneAboveIt(t *testing.T) {
db := database.NewTestDB(t)
albums := make([]library.Album, 0, 16)
// Four albums carry the whole play history, so "what you played
// last" and "what you play most" can only answer with those four.
for i := range 4 {
name := "Played " + string(rune('A'+i))
id := seed(t, db, name, "Solo", "", 50+i, "2026-08-1"+string(rune('0'+i))+" 00:00:00")
albums = append(albums, library.Album{ID: id, Name: name, ArtistName: "Solo"})
}
// …and ten more the user has never touched, so the library is
// larger than a single shelf and a different row was possible.
for i := range 10 {
name := "Quiet " + string(rune('A'+i))
id := seed(t, db, name, "Nobody", "", 0, "")
albums = append(albums, library.Album{ID: id, Name: name, ArtistName: "Nobody"})
}
svc := home.NewService(slog.Default(), db, fakeLibrary{albums: albums})
shelves, err := svc.GetShelves()
if err != nil {
t.Fatalf("GetShelves: %v", err)
}
// The first of a duplicate pair survives: the reason the user sees
// is the one that came first, not the one built last.
if !hasKind(shelves, home.KindRecentlyPlayed) {
t.Fatalf("expected a recently-played shelf, got %v", shelfKinds(shelves))
}
// And the page still has somewhere to start.
if len(shelves) < 2 {
t.Fatalf("suppression left %d shelves: %v", len(shelves), shelfKinds(shelves))
}
for i := 1; i < len(shelves); i++ {
above := shelves[i-1]
shared := 0
for _, a := range shelves[i].Albums {
for _, b := range above.Albums {
if a.ID == b.ID {
shared++
break
}
}
}
if len(shelves[i].Albums) < 3 {
continue
}
if ratio := float64(shared) / float64(len(shelves[i].Albums)); ratio >= 2.0/3.0 {
t.Errorf(
"shelf %q repeats %q (%d of %d albums)",
shelves[i].Kind, above.Kind, shared, len(shelves[i].Albums),
)
}
}
}