Compare commits

..
Author SHA1 Message Date
yonlu 4e5c6b9f7a fix(maintenance): bound search_clicks and lyrics_index, clear stale queue source
CI / check (push) Skipped
CI / e2e (push) Skipped
CI / check (pull_request) Failing after 1m3s
CI / e2e (pull_request) Skipped
Three unbounded or stale surfaces, each small on its own:

- lyrics_index rows were never pruned on track removal, so the FTS index
  grew forever. Delete the entry where the library search FTS entry is
  already deleted, on the orphan and RemoveFromLibrary paths.
- search_clicks had no ceiling; age out ranking rows after a retention
  window via a daily janitor job.
- queue.source_* kept a "Playing from X" label after its playlist was
  deleted. Drop the source when the queue's own playlist goes, wired
  through a playlist-service hook like Library.SetRemovalHooks.

Closes #249
2026-09-09 10:22:57 -04:00
25 changed files with 288 additions and 2830 deletions
-1
View File
@@ -779,7 +779,6 @@ func (yj *YellowJacketApp) startJanitor() {
}
yj.janitor.Register(maintenance.ExpiredHTTPCacheJob(yj.database))
yj.janitor.Register(maintenance.StaleArtistMetadataJob(yj.database))
yj.janitor.Register(maintenance.StaleSearchClicksJob(yj.database))
yj.janitor.Register(maintenance.OrphanedCoverFilesJob(
yj.database, coversDir, library.CoverArtFileSet,
+8 -60
View File
@@ -5,7 +5,6 @@ import (
"database/sql"
"fmt"
"log/slog"
"strings"
)
// Preserving a playlist entry across the loss of its track is two
@@ -90,14 +89,12 @@ const (
// metadata on the entry itself, so the entry survives the rows being
// deleted underneath it.
//
// Every path that empties `audio_files` must call this (or the scoped
// variant below) first, inside the same transaction as the delete.
// These paths have drifted before: the full rescan in backend/library
// did this and the stale-shape retire in this package did not, so the
// *documented* repair ("delete and rescan") preserved playlists while
// the automatic one that exists to spare the user that work silently
// emptied them (#183). The incremental scan's orphan cleanup and
// RemoveFromLibrary drifted the same way and are #246.
// Every path that empties `audio_files` must call this first, inside
// the same transaction as the delete. There are two such paths and
// they had drifted: the full rescan in backend/library did this and the
// stale-shape retire in this package did not, so the *documented*
// repair ("delete and rescan") preserved playlists while the automatic
// one that exists to spare the user that work silently emptied them.
//
// The display half is skipped, with a warning, when `track_metadata`
// cannot answer -- see the note above. Skipping it costs a phantom
@@ -106,39 +103,13 @@ const (
func PreservePlaylistPhantoms(
ctx context.Context, tx *sql.Tx, logger *slog.Logger,
) error {
return preservePlaylistPhantoms(ctx, tx, nil, logger)
}
// PreservePlaylistPhantomsForFiles is PreservePlaylistPhantoms scoped to
// the given audio file ids, for the two removal paths that delete a
// known subset of the table rather than all of it: the incremental
// scan's orphan cleanup and RemoveFromLibrary. A bulk pass there would
// rewrite every linked playlist row on every scan for nothing.
func PreservePlaylistPhantomsForFiles(
ctx context.Context, tx *sql.Tx, ids []int64, logger *slog.Logger,
) error {
return preservePlaylistPhantoms(ctx, tx, ids, logger)
}
func preservePlaylistPhantoms(
ctx context.Context, tx *sql.Tx, ids []int64, logger *slog.Logger,
) error {
clause, args, skip := phantomIDFilter(ids)
if skip {
return nil
}
if _, err := tx.ExecContext(
ctx, preservePhantomPathSQL+clause, args...,
); err != nil {
if _, err := tx.ExecContext(ctx, preservePhantomPathSQL); err != nil {
return fmt.Errorf(
"could not preserve playlist track file paths: %w", err,
)
}
if _, err := tx.ExecContext(
ctx, preservePhantomDisplaySQL+clause, args...,
); err != nil {
if _, err := tx.ExecContext(ctx, preservePhantomDisplaySQL); err != nil {
// A failed statement does not roll back a SQLite transaction,
// so the path half above stands and the entries remain
// re-linkable.
@@ -152,26 +123,3 @@ func preservePlaylistPhantoms(
return nil
}
// phantomIDFilter builds the extra WHERE terms and arguments that scope
// a preservation pass to a set of audio file ids. A nil ids returns the
// empty clause (a bulk run over every linked entry); an empty slice
// reports skip, since there is nothing to preserve.
func phantomIDFilter(ids []int64) (clause string, args []any, skip bool) {
switch {
case ids == nil:
return "", nil, false
case len(ids) == 0:
return "", nil, true
}
clause = " AND audio_file_id IN (" +
strings.Repeat("?,", len(ids)-1) + "?)"
args = make([]any, len(ids))
for i, id := range ids {
args[i] = id
}
return clause, args, false
}
-94
View File
@@ -1,94 +0,0 @@
package database
import (
"context"
"database/sql"
"testing"
)
// TestPreservePlaylistPhantomsForFilesScopesToTheRequestedIDs is the
// scoping half of the scoped variant: a run over one file's id must
// fill that file's playlist entries and leave every other entry alone,
// because the incremental scan calls this once per orphan batch and a
// pass that rewrote the whole table would touch every playlist row on
// every scan.
func TestPreservePlaylistPhantomsForFilesScopesToTheRequestedIDs(t *testing.T) {
ctx := context.Background()
db := openRaw(t, t.TempDir())
if _, err := db.ExecContext(ctx, "PRAGMA foreign_keys = ON"); err != nil {
t.Fatalf("pragma: %v", err)
}
if err := applySchema(ctx, db); err != nil {
t.Fatalf("applySchema: %v", err)
}
if _, err := db.ExecContext(ctx, `
INSERT INTO playlists (id, name) VALUES (1, 'keepme');
INSERT INTO libraries (id, name, path) VALUES (0, 'test', '/music');
INSERT INTO artists (id, name) VALUES (3, 'Aurora Fields');
INSERT INTO cover_art (id, file_path, mime_type)
VALUES (9, 'covers/7.jpg', 'image/jpeg');
INSERT INTO albums (id, name, artist_id, cover_art_id)
VALUES (4, 'Tideline', 3, 9);
INSERT INTO audio_files
(id, file_path, file_type_id, length_milliseconds,
title, artist_credit, artist_id, album_id)
VALUES
(7, '/music/a.flac', 1, 1000,
'Slack Water', 'Aurora Fields', 3, 4),
(8, '/music/b.flac', 1, 2000,
'Second Tide', 'Aurora Fields', 3, 4);
INSERT INTO playlist_tracks (playlist_id, audio_file_id, position)
VALUES (1, 7, 0), (1, 8, 1);
`); err != nil {
t.Fatalf("seed: %v", err)
}
tx, err := db.BeginTx(ctx, nil)
if err != nil {
t.Fatalf("begin: %v", err)
}
defer func() { _ = tx.Rollback() }()
if err := PreservePlaylistPhantomsForFiles(
ctx, tx, []int64{7}, testLogger(),
); err != nil {
t.Fatalf("preserve: %v", err)
}
if err := tx.Commit(); err != nil {
t.Fatalf("commit: %v", err)
}
var filled, untouched sql.NullString
if err := db.QueryRowContext(ctx,
"SELECT phantom_file_path FROM playlist_tracks WHERE audio_file_id = 7",
).Scan(&filled); err != nil {
t.Fatalf("read the requested entry: %v", err)
}
if filled.String != "/music/a.flac" {
t.Errorf(
"requested entry phantom_file_path = %q, want %q",
filled.String, "/music/a.flac",
)
}
if err := db.QueryRowContext(ctx,
"SELECT phantom_file_path FROM playlist_tracks WHERE audio_file_id = 8",
).Scan(&untouched); err != nil {
t.Fatalf("read the untouched entry: %v", err)
}
if untouched.Valid {
t.Errorf(
"untouched entry got phantom_file_path = %q, want NULL "+
"(a scoped run must not rewrite the whole table)",
untouched.String,
)
}
}
-248
View File
@@ -1,248 +0,0 @@
package download
import (
"context"
"os"
"path/filepath"
"testing"
)
// Multi-disc rips, single-track results and coverage counted in tracks
// rather than files (#270).
func TestParsePathReadsTheDiscFromItsFolder(t *testing.T) {
t.Parallel()
cases := []struct {
path string
disc int
track int
folder string
}{
{`\share\Pink Floyd - The Wall (1979)\CD2\03 Hey You.flac`, 2, 3, "The Wall"},
{`\share\The Wall\Disc 1\01 In The Flesh.flac`, 1, 1, "The Wall"},
{`\share\The Wall\[Disk-2]\01 Hey You.flac`, 2, 1, "The Wall"},
{`\share\The Wall\CD1 - Live\04 Mother.flac`, 1, 4, "The Wall"},
// The filename's own disc number is more specific than the folder.
{`\share\The Wall\CD1\2-05 Comfortably Numb.flac`, 2, 5, "The Wall"},
// Not a disc folder: a number is required.
{`\share\CDs\The Wall\01 In The Flesh.flac`, 0, 1, "The Wall"},
}
for _, tc := range cases {
t.Run(tc.path, func(t *testing.T) {
t.Parallel()
got := ParsePath(tc.path)
if got.Disc != tc.disc || got.Track != tc.track || got.Folder != tc.folder {
t.Errorf(
"ParsePath = disc %d track %d folder %q, want %d %d %q",
got.Disc, got.Track, got.Folder, tc.disc, tc.track, tc.folder,
)
}
})
}
}
func TestAlbumDir(t *testing.T) {
t.Parallel()
cases := map[string]string{
`\share\Album\CD1\01 A.flac`: "/share/Album",
`\share\Album\01 A.flac`: "/share/Album",
`CD1\01 A.flac`: "CD1",
`\share\CD Collection\01.mp3`: "/share/CD Collection",
}
for in, want := range cases {
if got := AlbumDir(in); got != want {
t.Errorf("AlbumDir(%q) = %q, want %q", in, got, want)
}
}
}
// One album shared as CD1/CD2 is one candidate, named after the album.
func TestSlskdGroupsDiscFoldersIntoOneCandidate(t *testing.T) {
t.Parallel()
stub := newSlskdStub(t)
stub.responses = []slskdResponse{{
Username: "peer",
Files: []slskdFile{
{Filename: `\m\The Wall\CD1\01 In The Flesh.flac`, Size: 1},
{Filename: `\m\The Wall\CD1\02 The Thin Ice.flac`, Size: 1},
{Filename: `\m\The Wall\CD2\01 Hey You.flac`, Size: 1},
{Filename: `\m\The Wall\CD2\02 Is There Anybody Out There.flac`, Size: 1},
},
}}
s, _ := newStubSlskd(t, stub)
got, err := s.Search(context.Background(), Download{Query: "the wall"})
if err != nil {
t.Fatalf("Search: %v", err)
}
if len(got) != 1 {
t.Fatalf("got %d candidates, want the two discs as one", len(got))
}
if got[0].Title != "The Wall" || len(got[0].Files) != 4 {
t.Errorf(
"candidate = %q with %d files, want \"The Wall\" with 4",
got[0].Title, len(got[0].Files),
)
}
}
// A track search matches one file per folder, so a single-track request
// must accept a one-file folder that an album request rightly drops.
func TestSlskdKeepsASingleFileForATrackRequest(t *testing.T) {
t.Parallel()
stub := newSlskdStub(t)
stub.responses = []slskdResponse{{
Username: "peer",
Files: []slskdFile{
{Filename: `\m\OK Computer\02 Paranoid Android.flac`, Size: 1},
},
}}
s, _ := newStubSlskd(t, stub)
track, err := s.Search(context.Background(), Download{
RecordingMBID: "rec-1", Artist: "Radiohead", Album: "Paranoid Android",
})
if err != nil {
t.Fatalf("Search: %v", err)
}
if len(track) != 1 {
t.Errorf("track request: got %d candidates, want 1", len(track))
}
album, err := s.Search(context.Background(), Download{
ReleaseMBID: "rel-1", Artist: "Radiohead", Album: "OK Computer",
})
if err != nil {
t.Fatalf("Search: %v", err)
}
if len(album) != 0 {
t.Errorf("album request: got %d candidates, want the one-file folder dropped", len(album))
}
}
// Two discs with a file of the same name both reach staging, each under
// its disc folder, where the importer reads the disc number from.
func TestSlskdCollectKeepsDiscFolders(t *testing.T) {
t.Parallel()
stub := newSlskdStub(t)
s, downloads := newStubSlskd(t, stub)
for _, disc := range []string{"CD1", "CD2"} {
dir := filepath.Join(downloads, disc)
if err := os.MkdirAll(dir, 0o750); err != nil {
t.Fatalf("mkdir: %v", err)
}
if err := os.WriteFile(
filepath.Join(dir, "01 Intro.flac"), []byte(disc), 0o600,
); err != nil {
t.Fatalf("write: %v", err)
}
}
dst := t.TempDir()
got, err := s.collect(Candidate{Files: []CandidateFile{
{Path: `\m\Album\CD1\01 Intro.flac`, IsAudio: true},
{Path: `\m\Album\CD2\01 Intro.flac`, IsAudio: true},
}}, dst)
if err != nil {
t.Fatalf("collect: %v", err)
}
if len(got.Files) != 2 {
t.Fatalf("collected %d files, want 2", len(got.Files))
}
for _, disc := range []string{"CD1", "CD2"} {
data, err := os.ReadFile(filepath.Join(dst, disc, "01 Intro.flac"))
if err != nil || string(data) != disc {
t.Errorf("%s's file missing or overwritten: %q, %v", disc, data, err)
}
if hint := ParsePath(filepath.Join(dst, disc, "01 Intro.flac")); hint.Disc == 0 {
t.Errorf("staged %s file lost its disc number", disc)
}
}
}
// A two-disc release whose discs both number from 01 aligns completely
// once the disc comes from the folder; before, disc 2's 01 collided with
// disc 1's.
func TestMultiDiscCandidateAlignsEveryTrack(t *testing.T) {
t.Parallel()
dl := Download{
ReleaseMBID: "the-wall",
Artist: "Pink Floyd",
Album: "The Wall",
Expected: []ExpectedTrack{
{DiscNumber: 1, Position: 1, Title: "In the Flesh?"},
{DiscNumber: 1, Position: 2, Title: "The Thin Ice"},
{DiscNumber: 2, Position: 1, Title: "Hey You"},
{DiscNumber: 2, Position: 2, Title: "Is There Anybody Out There?"},
},
}
c := Candidate{
Title: "The Wall",
Files: []CandidateFile{
{Path: `\m\Pink Floyd - The Wall\CD1\01 In the Flesh.flac`, Size: 1},
{Path: `\m\Pink Floyd - The Wall\CD1\02 The Thin Ice.flac`, Size: 1},
{Path: `\m\Pink Floyd - The Wall\CD2\01 Hey You.flac`, Size: 1},
{Path: `\m\Pink Floyd - The Wall\CD2\02 Is There Anybody Out There.flac`, Size: 1},
},
}
got := Score(dl, c, 50, AutoDownloadPrefs{})
if got.Match.Completeness != 1 {
t.Errorf("completeness = %f, want 1", got.Match.Completeness)
}
if got.Match.AlbumFit < 0.99 {
t.Errorf("album fit = %f, want the album's own name to match", got.Match.AlbumFit)
}
if got.Match.Overall < minMatch {
t.Errorf("match = %f, want it to clear the auto-pick bar %f", got.Match.Overall, minMatch)
}
}
// Ten files against a ten-track album is not a complete album when only
// three of them are its tracks.
func TestCompletenessCountsTracksNotFiles(t *testing.T) {
t.Parallel()
dl := okComputer()
c := Candidate{Title: "OK Computer", Files: []CandidateFile{
{Path: `\m\Radiohead - OK Computer\Airbag.flac`, Size: 1},
{Path: `\m\Radiohead - OK Computer\Paranoid Android.flac`, Size: 1},
{Path: `\m\Radiohead - OK Computer\Exit Music (For a Film).flac`, Size: 1},
{Path: `\m\Radiohead - OK Computer\Creep.flac`, Size: 1},
}}
got := Score(dl, c, 50, AutoDownloadPrefs{})
if got.Match.Completeness > 0.76 {
t.Errorf(
"completeness = %f with 3 of 4 tracks present, want at most 0.75",
got.Match.Completeness,
)
}
}
+1 -1
View File
@@ -47,7 +47,7 @@ func grabAll(
go func() {
defer wg.Done()
f.manager.grab(ctx, dl, candidate, nil, false)
f.manager.grab(ctx, dl, candidate, nil)
}()
}
-261
View File
@@ -1,261 +0,0 @@
package download
import (
"context"
"errors"
"os"
"testing"
)
// A transfer that fails on one copy of an album is not a failed
// download while another acceptable copy exists. On Soulseek the usual
// failure is one peer being offline, with several others offering the
// same folder.
var errPeerOffline = errors.New("peer went offline")
func TestManagerFallsBackToTheNextCandidate(t *testing.T) {
t.Parallel()
f := newManagerFixture(t)
// The failing source ranks first on priority, so the fallback is
// what reaches the one that works.
bad := fakeWithAlbum(1, "offline-peer", ".flac")
bad.GrabErr = errPeerOffline
good := fakeWithAlbum(2, "online-peer", ".flac")
f.manager.installProvider(Config{ID: 1, Priority: 90}, bad)
f.manager.installProvider(Config{ID: 2, Priority: 10}, good)
dl := fourTrackDownload()
if _, err := f.manager.Start(context.Background(), dl); err != nil {
t.Fatalf("Start: %v", err)
}
waitForDownloadState(t, f.store, dl.ID, StateComplete)
if bad.GrabCalls != 1 || good.GrabCalls != 1 {
t.Errorf(
"grabs: failing=%d working=%d, want 1 and 1",
bad.GrabCalls, good.GrabCalls,
)
}
// The abandoned attempt's staging goes with it; only a request that
// fails outright keeps its staging for inspection.
waitFor(t, func() bool {
entries, err := os.ReadDir(f.staging.Root())
return err == nil && len(entries) == 0
}, "the failed attempt's staging was never released")
}
// Falling back must not lower the bar. A second choice outside the
// user's guardrails is not a choice auto-pick may make, first or second.
func TestManagerFallbackRespectsTheGuardrails(t *testing.T) {
t.Parallel()
f := newManagerFixture(t)
f.manager.SetPreferences(AutoDownloadPrefs{MaxSizeMB: 50})
bad := fakeWithAlbum(1, "offline-peer", ".flac")
bad.GrabErr = errPeerOffline
bad.Candidates[0].TotalSize = 40 << 20
huge := fakeWithAlbum(2, "oversized", ".flac")
huge.Candidates[0].TotalSize = 900 << 20
f.manager.installProvider(Config{ID: 1, Priority: 90}, bad)
f.manager.installProvider(Config{ID: 2, Priority: 10}, huge)
dl := fourTrackDownload()
if _, err := f.manager.Start(context.Background(), dl); err != nil {
t.Fatalf("Start: %v", err)
}
waitForDownloadState(t, f.store, dl.ID, StateFailed)
if huge.GrabCalls != 0 {
t.Errorf("fell back to a candidate over the size ceiling")
}
}
// A copy the user picked by hand is the copy they asked for. Quietly
// substituting another is a decision they did not make.
func TestManagerPickDoesNotFallBack(t *testing.T) {
t.Parallel()
f := newManagerFixture(t)
bad := fakeWithAlbum(1, "offline-peer", ".flac")
bad.GrabErr = errPeerOffline
good := fakeWithAlbum(2, "online-peer", ".flac")
f.manager.installProvider(Config{ID: 1, Priority: 90}, bad)
f.manager.installProvider(Config{ID: 2, Priority: 10}, good)
// A ceiling below both copies parks the result set for the user.
f.manager.SetPreferences(AutoDownloadPrefs{MaxSizeMB: 1})
bad.Candidates[0].TotalSize = 30 << 20
good.Candidates[0].TotalSize = 30 << 20
dl := fourTrackDownload()
if _, err := f.manager.Start(context.Background(), dl); err != nil {
t.Fatalf("Start: %v", err)
}
if err := f.manager.Pick(
context.Background(), dl.ID, "offline-peer-cand",
); err != nil {
t.Fatalf("Pick: %v", err)
}
waitForDownloadState(t, f.store, dl.ID, StateFailed)
if good.GrabCalls != 0 {
t.Errorf("a hand-picked grab fell back to another candidate")
}
}
// Fallback is for surviving an offline peer or two, not for walking a
// forty-peer list for six hours.
func TestManagerFallbackIsBounded(t *testing.T) {
t.Parallel()
f := newManagerFixture(t)
var providers []*FakeProvider
for i := int64(1); i <= maxGrabAttempts+2; i++ {
p := fakeWithAlbum(i, "peer-"+itoa(int(i)), ".flac")
p.GrabErr = errPeerOffline
f.manager.installProvider(Config{ID: i, Priority: 50}, p)
providers = append(providers, p)
}
dl := fourTrackDownload()
if _, err := f.manager.Start(context.Background(), dl); err != nil {
t.Fatalf("Start: %v", err)
}
waitForDownloadState(t, f.store, dl.ID, StateFailed)
grabs := 0
for _, p := range providers {
grabs += p.GrabCalls
}
if grabs != maxGrabAttempts {
t.Errorf("grabs = %d, want %d", grabs, maxGrabAttempts)
}
}
// The veto judges the best candidate *inside* the guardrails, so the
// grab has to take that one — not the overall best, which may be the
// very copy the user said not to take unattended.
func TestManagerAutoPickTakesTheBestEligibleCandidate(t *testing.T) {
t.Parallel()
f := newManagerFixture(t)
f.manager.SetPreferences(AutoDownloadPrefs{MaxSizeMB: 50})
huge := fakeWithAlbum(1, "oversized", ".flac")
huge.Candidates[0].TotalSize = 900 << 20
fits := fakeWithAlbum(2, "fits", ".flac")
fits.Candidates[0].TotalSize = 40 << 20
f.manager.installProvider(Config{ID: 1, Priority: 90}, huge)
f.manager.installProvider(Config{ID: 2, Priority: 10}, fits)
dl := fourTrackDownload()
ranked, err := f.manager.Start(context.Background(), dl)
if err != nil {
t.Fatalf("Start: %v", err)
}
if ranked[0].ID != "oversized-cand" {
t.Fatalf("fixture: best overall is %s, want the oversized copy", ranked[0].ID)
}
waitForDownloadState(t, f.store, dl.ID, StateComplete)
if huge.GrabCalls != 0 || fits.GrabCalls != 1 {
t.Errorf(
"grabs: oversized=%d fits=%d, want 0 and 1",
huge.GrabCalls, fits.GrabCalls,
)
}
}
// On Soulseek a failure is the peer's, so every folder that peer offered
// goes with it. Elsewhere a failure is the release's, and one indexer's
// other releases are still worth trying.
func TestRuledOutBy(t *testing.T) {
t.Parallel()
failed := []Candidate{
{ID: "slskd:alice:Album", Kind: KindSlskd, ProviderID: 1, Origin: "alice"},
{ID: "tracker-1", Kind: KindProwlarr, ProviderID: 2, Origin: "indexer"},
}
cases := []struct {
name string
c Candidate
want bool
}{
{
name: "the same candidate",
c: Candidate{ID: "tracker-1", Kind: KindProwlarr, ProviderID: 2, Origin: "indexer"},
want: true,
},
{
name: "another folder from a failed peer",
c: Candidate{
ID: "slskd:alice:Album (2)",
Kind: KindSlskd,
ProviderID: 1,
Origin: "alice",
},
want: true,
},
{
name: "another peer",
c: Candidate{ID: "slskd:bob:Album", Kind: KindSlskd, ProviderID: 1, Origin: "bob"},
want: false,
},
{
name: "another release from the same indexer",
c: Candidate{ID: "tracker-2", Kind: KindProwlarr, ProviderID: 2, Origin: "indexer"},
want: false,
},
{
name: "a peer of the same name on a different daemon",
c: Candidate{
ID: "slskd:alice:Album",
Kind: KindSlskd,
ProviderID: 3,
Origin: "alice",
},
want: false,
},
}
for _, tc := range cases {
t.Run(tc.name, func(t *testing.T) {
t.Parallel()
if got := ruledOutBy(tc.c, failed); got != tc.want {
t.Errorf("ruledOutBy = %v, want %v", got, tc.want)
}
})
}
}
+59 -206
View File
@@ -577,12 +577,12 @@ func (m *Manager) Start(
))
}
if pick, ok := autoPick(dl, ranked, m.preferences()); ok {
if m.AutoPickable(dl, ranked) {
if job != nil {
job.Logf(jobs.LevelInfo, "Auto-selected best candidate")
}
go m.grab(context.WithoutCancel(ctx), dl, pick, job, true)
go m.grab(context.WithoutCancel(ctx), dl, ranked[0], job)
return ranked, nil
}
@@ -622,13 +622,6 @@ func (m *Manager) Attempt(
return false, veto, nil
}
pick, ok := autoPick(dl, ranked, m.preferences())
if !ok {
// Unreachable while autoPick and AutoPickVeto agree; kept so a
// future divergence refuses rather than grabbing blind.
return false, "no candidate clears the auto-download bar", nil
}
if err := m.store.CreateDownload(ctx, dl); err != nil {
return false, "", err
}
@@ -650,7 +643,7 @@ func (m *Manager) Attempt(
))
}
go m.grab(context.WithoutCancel(ctx), dl, pick, job, true)
go m.grab(context.WithoutCancel(ctx), dl, ranked[0], job)
return true, "", nil
}
@@ -685,7 +678,7 @@ func (m *Manager) Pick(
job := m.startJob(dl)
go m.grab(context.WithoutCancel(ctx), dl, *chosen, job, false)
go m.grab(context.WithoutCancel(ctx), dl, *chosen, job)
return nil
}
@@ -709,20 +702,13 @@ func (m *Manager) Cancel(ctx context.Context, downloadID string) error {
return nil
}
// grab drives one request all the way to the library. It runs on its
// grab drives one candidate all the way to the library. It runs on its
// own goroutine and owns the job from here on.
//
// When fallback is set and a candidate's transfer fails, the next
// candidate that auto-pick would itself have accepted is tried in its
// place (see nextCandidate). It is set for the two unattended routes
// and not for a candidate the user picked by hand: they chose that copy,
// and quietly substituting another is a decision they did not make.
func (m *Manager) grab(
ctx context.Context,
dl Download,
c Candidate,
job *jobs.Handle,
fallback bool,
) {
ctx, cancel := context.WithTimeout(ctx, grabTimeout)
defer cancel()
@@ -737,82 +723,6 @@ func (m *Manager) grab(
m.actMu.Unlock()
}()
var failed []Candidate
for {
out := m.attemptGrab(ctx, dl, c, job)
if out.err == nil {
m.finishGrab(ctx, dl, out.item, out.imported, job)
return
}
failed = append(failed, c)
next, ok := m.nextCandidate(ctx, dl, failed, out, fallback)
if !ok {
m.failDownload(ctx, job, dl.ID, out.err)
return
}
m.logger.Info(
"download candidate failed; trying the next",
"download", dl.ID,
"failed", c.ID,
"next", next.ID,
"error", out.err,
)
if job != nil {
job.Logf(jobs.LevelWarn, fmt.Sprintf(
"%s failed (%v); trying %s instead",
describeCandidate(c), out.err, describeCandidate(next),
))
}
// The failed attempt's staging holds at most a partial folder
// nobody is going to import, and the next attempt reserves its
// own. Only the final failure keeps its staging for inspection.
if out.item.StagingDir != "" {
if err := m.staging.Release(out.item.StagingDir); err != nil {
m.logger.Warn("could not release staging dir", "error", err)
}
}
c = next
}
}
// maxGrabAttempts bounds how many candidates one request will try. A
// popular album can have dozens of peers; the point of falling back is
// to survive the ordinary one or two that are offline, not to walk the
// whole list for six hours.
const maxGrabAttempts = 3
// grabOutcome is how one candidate's attempt ended.
type grabOutcome struct {
item DownloadItem
imported ImportResult
err error
// retryable reports whether another candidate might succeed where
// this one failed: the transfer failed, or delivered too little of
// the album. Anything else — no staging space, no library root, a
// tag write failing — would fail the next candidate identically.
retryable bool
}
// attemptGrab takes one candidate through transfer and import. It
// records the item's own failure, but not the download's: whether the
// download has failed is the caller's decision, since another candidate
// may yet succeed.
func (m *Manager) attemptGrab(
ctx context.Context,
dl Download,
c Candidate,
job *jobs.Handle,
) grabOutcome {
// Who will move the bytes is decided before any slot is taken, so
// the transfer waits in its own provider's queue rather than in a
// global one. A delegate takes no slot at all: the transfer is
@@ -821,7 +731,9 @@ func (m *Manager) attemptGrab(
// work against our budget.
plan, err := m.planTransfer(dl, c)
if err != nil {
return grabOutcome{err: err}
m.failDownload(ctx, job, dl.ID, err)
return
}
if !plan.delegated() {
@@ -831,7 +743,9 @@ func (m *Manager) attemptGrab(
case provSem <- struct{}{}:
defer func() { <-provSem }()
case <-ctx.Done():
return grabOutcome{err: ctx.Err()}
m.failDownload(ctx, job, dl.ID, ctx.Err())
return
}
globalSem := m.globalSem()
@@ -840,7 +754,9 @@ func (m *Manager) attemptGrab(
case globalSem <- struct{}{}:
defer func() { <-globalSem }()
case <-ctx.Done():
return grabOutcome{err: ctx.Err()}
m.failDownload(ctx, job, dl.ID, ctx.Err())
return
}
}
@@ -855,30 +771,24 @@ func (m *Manager) attemptGrab(
dir, err := m.staging.Reserve(item.ID)
if err != nil {
return grabOutcome{err: err}
m.failDownload(ctx, job, dl.ID, err)
return
}
item.StagingDir = dir
if err := m.store.CreateItem(ctx, item); err != nil {
return grabOutcome{item: item, err: err}
}
m.failDownload(ctx, job, dl.ID, err)
fail := func(err error, retryable bool) grabOutcome {
if serr := m.store.SetItemState(
ctx, item.ID, StateFailed, err.Error(),
); serr != nil {
m.logger.Warn("could not record item failure", "error", serr)
}
return grabOutcome{item: item, err: err, retryable: retryable}
return
}
result, err := m.transfer(ctx, dl, item, plan, job)
if err != nil {
// A delegate's failure is the external manager's verdict on the
// whole request, not on one copy of it.
return fail(err, !plan.delegated())
m.failItem(ctx, job, item, dl.ID, err)
return
}
m.setStates(ctx, dl.ID, item.ID, StateImporting)
@@ -888,116 +798,42 @@ func (m *Manager) attemptGrab(
job.SetStages(importStages(2))
}
var imported ImportResult
if result.Delegated {
// The external manager already placed and tagged these files in
// its own library. Moving them out from under a system that is
// still managing them would be worse than useless, so the files
// are recorded where they are and the library scan picks them
// up in place.
imported = ImportResult{Paths: result.Files}
if job != nil {
job.Logf(jobs.LevelInfo, fmt.Sprintf(
"External manager imported %d files; recording them in place",
len(result.Files),
))
}
} else {
opts := m.importOptions()
opts.WriteTags = true
return grabOutcome{
item: item,
imported: ImportResult{Paths: result.Files},
opts.LibraryRoot, err = m.library.LibraryPath(dl.LibraryID)
if err != nil {
m.failItem(ctx, job, item, dl.ID,
fmt.Errorf("resolve library root: %w", err))
return
}
imported, err = m.importer.Import(ctx, dl, result, opts)
if err != nil {
m.failItem(ctx, job, item, dl.ID, err)
return
}
}
opts := m.importOptions()
opts.WriteTags = true
opts.LibraryRoot, err = m.library.LibraryPath(dl.LibraryID)
if err != nil {
return fail(fmt.Errorf("resolve library root: %w", err), false)
}
imported, err := m.importer.Import(ctx, dl, result, opts)
if err != nil {
return fail(err, errors.Is(err, ErrTooIncomplete))
}
return grabOutcome{item: item, imported: imported}
}
// nextCandidate picks the candidate to try after the ones in failed.
//
// It only ever offers a candidate auto-pick would have taken on its own
// (autoAcceptable), so falling back cannot lower the bar an unattended
// download is held to: the second choice has to clear the same gates
// the first did.
//
// On Soulseek a failure belongs to the *peer* — offline, refusing, or
// holding us in a queue — so every folder that peer offered is skipped
// with it. Elsewhere a failure belongs to the release, and only that
// candidate is.
func (m *Manager) nextCandidate(
ctx context.Context,
dl Download,
failed []Candidate,
out grabOutcome,
fallback bool,
) (Candidate, bool) {
if !fallback || !out.retryable || ctx.Err() != nil ||
len(failed) >= maxGrabAttempts {
return Candidate{}, false
}
m.resMu.RLock()
ranked := m.results[dl.ID]
m.resMu.RUnlock()
prefs := m.preferences()
for _, c := range ranked {
if ruledOutBy(c, failed) || !autoAcceptable(dl, c, prefs) {
continue
}
return c, true
}
return Candidate{}, false
}
// ruledOutBy reports whether a failure among failed also rules out c.
func ruledOutBy(c Candidate, failed []Candidate) bool {
for _, f := range failed {
if c.ID == f.ID && c.ProviderID == f.ProviderID {
return true
}
if c.Kind == KindSlskd && f.Kind == KindSlskd &&
c.ProviderID == f.ProviderID && c.Origin != "" &&
c.Origin == f.Origin {
return true
}
}
return false
}
// describeCandidate names a candidate for the job log.
func describeCandidate(c Candidate) string {
if c.Origin != "" {
return fmt.Sprintf("%q from %s", c.Title, c.Origin)
}
return fmt.Sprintf("%q", c.Title)
}
// finishGrab records a successful import and retires what the request
// was holding.
func (m *Manager) finishGrab(
ctx context.Context,
dl Download,
item DownloadItem,
imported ImportResult,
job *jobs.Handle,
) {
if err := m.store.SetItemImported(
ctx, item.ID, imported.Paths,
); err != nil {
@@ -1341,6 +1177,23 @@ func (m *Manager) failDownload(
}
}
// failItem records an item-level failure and fails its download.
func (m *Manager) failItem(
ctx context.Context,
job *jobs.Handle,
item DownloadItem,
downloadID string,
err error,
) {
if serr := m.store.SetItemState(
ctx, item.ID, StateFailed, err.Error(),
); serr != nil {
m.logger.Warn("could not record item failure", "error", serr)
}
m.failDownload(ctx, job, downloadID, err)
}
// startJob registers the request in the background jobs panel.
func (m *Manager) startJob(dl Download) *jobs.Handle {
if m.jobsReg == nil {
+8 -140
View File
@@ -66,14 +66,6 @@ var (
// separatorPattern splits "Artist - Album" style folder names.
separatorPattern = regexp.MustCompile(`\s+[-–—]\s+`)
// discFolderPattern matches a directory that holds one disc of an
// album rather than the album: "CD1", "CD 2", "Disc 3", "Disk-1",
// "[Disc 2]", "CD1 - The Early Years". A number is required, so a
// folder merely called "CDs" is not one.
discFolderPattern = regexp.MustCompile(
`(?i)^\s*[\[(]?\s*(?:cd|disc|disk)\s*[-_.#]?\s*(\d{1,2})\b`,
)
)
// FormatForPath returns the audio format implied by a path's extension,
@@ -102,63 +94,16 @@ type TrackHint struct {
Folder string
}
// discFolder reports whether a directory name is one disc of an album,
// and which.
func discFolder(name string) (int, bool) {
m := discFolderPattern.FindStringSubmatch(name)
if m == nil {
return 0, false
}
n, err := strconv.Atoi(m[1])
if err != nil || n == 0 {
return 0, false
}
return n, true
}
// AlbumDir is the directory that holds a file's *album*: its parent,
// or its grandparent when the parent is a disc folder.
//
// Multi-disc rips are shared as `Album/CD1/…` and `Album/CD2/…`, and
// grouping candidates by the immediate parent split one album into two
// half-albums, each titled "CD1". Neither could clear the completeness
// or album-title bars, so a multi-disc release could not be auto-picked
// at all. A disc folder at the root has no album above it and is
// returned as it is.
func AlbumDir(p string) string {
dir := path.Dir(strings.ReplaceAll(p, `\`, "/"))
if _, ok := discFolder(path.Base(dir)); !ok {
return dir
}
parent := path.Dir(dir)
if parent == "." || parent == "/" || parent == "" {
return dir
}
return parent
}
// ParsePath extracts what it can from one candidate file path.
func ParsePath(p string) TrackHint {
// Soulseek paths are Windows-style; normalize before splitting.
norm := strings.ReplaceAll(p, `\`, "/")
base := path.Base(norm)
folder := path.Base(path.Dir(norm))
name := strings.TrimSuffix(base, path.Ext(base))
// The album's name is the album directory's, not a disc folder's,
// and the disc folder is where a multi-disc rip says which disc a
// file is on. A disc number in the filename ("2-01 …") is more
// specific and overrides it below.
hint := TrackHint{Folder: cleanAlbumName(path.Base(AlbumDir(norm)))}
if disc, ok := discFolder(path.Base(path.Dir(norm))); ok {
hint.Disc = disc
}
hint := TrackHint{Folder: cleanAlbumName(folder)}
if m := trackNumPattern.FindStringSubmatch(name); m != nil {
if m[1] != "" {
@@ -258,72 +203,21 @@ func AnnotateFiles(files []CandidateFile) []CandidateFile {
// matchFiles aligns a candidate's audio files to the expected tracklist
// and returns the per-file assignment plus the mean title similarity of
// the aligned pairs. alignFiles is the same alignment with the
// duration evidence as well.
func matchFiles(
files []CandidateFile,
expected []ExpectedTrack,
) ([]CandidateFile, float64) {
a := alignFiles(files, expected)
return a.files, a.titleFit
}
// alignment is what aligning a candidate to a tracklist found.
type alignment struct {
files []CandidateFile
// titleFit is the mean title similarity over aligned pairs.
titleFit float64
// durationFit is the mean duration agreement over aligned pairs
// where both sides state a length, and timedPairs is how many such
// pairs there were.
durationFit float64
timedPairs int
aligned int
}
// durationAgreement scores how well a file's length matches the
// expected track's, in 0..1. Rips of the same master differ by a
// second or two of silence; a different edit, a live take or a
// truncated file differs by tens of seconds.
func durationAgreement(got, want int64) float64 {
const (
exactMillis = 3_000
wrongMillis = 30_000
)
d := got - want
if d < 0 {
d = -d
}
switch {
case d <= exactMillis:
return 1
case d >= wrongMillis:
return 0
default:
return 1 - float64(d-exactMillis)/float64(wrongMillis-exactMillis)
}
}
// alignFiles aligns a candidate's audio files to the expected tracklist.
// the aligned pairs.
//
// Alignment is greedy by score rather than optimal: candidate folders
// are small (a few dozen files at most) and the common cases — correct
// track numbers, or clean "NN Title" names — are unambiguous, so the
// extra machinery of Hungarian assignment buys nothing here.
func alignFiles(
func matchFiles(
files []CandidateFile,
expected []ExpectedTrack,
) alignment {
) ([]CandidateFile, float64) {
annotated := make([]CandidateFile, len(files))
copy(annotated, files)
if len(expected) == 0 {
return alignment{files: annotated}
return annotated, 0
}
hints := make([]TrackHint, len(annotated))
@@ -336,19 +230,8 @@ func alignFiles(
var (
total float64
matched int
durTotal float64
timed int
)
// timing adds a pair's duration evidence when both sides state one.
timing := func(f CandidateFile, e ExpectedTrack) {
if f.LengthMillis > 0 && e.LengthMillis > 0 {
durTotal += durationAgreement(f.LengthMillis, e.LengthMillis)
timed++
}
}
// Pass 1: trust explicit track numbers when they are unique and in
// range. A folder that numbers its files correctly is the strong
// case, and title comparison only adds noise there.
@@ -367,8 +250,6 @@ func alignFiles(
total += autotag.TitleSimilarity(hints[i].Title, expected[idx].Title)
matched++
timing(annotated[i], expected[idx])
}
// Pass 2: title similarity for whatever is left.
@@ -403,26 +284,13 @@ func alignFiles(
total += bestSim
matched++
timing(annotated[i], expected[bestIdx])
}
if matched == 0 {
return alignment{files: annotated}
return annotated, 0
}
a := alignment{
files: annotated,
titleFit: total / float64(matched),
timedPairs: timed,
aligned: matched,
}
if timed > 0 {
a.durationFit = durTotal / float64(timed)
}
return a
return annotated, total / float64(matched)
}
// indexForPosition finds the expected track at a disc/track position.
+86 -538
View File
@@ -5,11 +5,9 @@ import (
"errors"
"fmt"
"log/slog"
"net/url"
"os"
"path"
"path/filepath"
"regexp"
"strings"
"time"
@@ -71,34 +69,12 @@ const (
slskdTransferPoll = 3 * time.Second
// slskdMinFiles is the fewest audio files a folder needs before it
// is offered as a candidate for an album. Soulseek returns a lot of
// one-file noise for common queries. A single-track request takes
// one (see minFilesFor).
// is offered as a candidate. Soulseek returns a lot of one-file
// noise for common queries.
slskdMinFiles = 2
// slskdHTTPTimeout bounds one API call.
slskdHTTPTimeout = 20 * time.Second
// millisPerSecond converts slskd's whole-second file lengths.
millisPerSecond = 1000
// slskdStallAfter is how long a grab may go without a byte arriving
// before the peer is given up on. It is measured from enqueue, so
// it covers a peer that queues us and never starts as well as one
// that starts and stops. Ten minutes is long enough for a short
// queue ahead of us to clear and short enough that one unresponsive
// peer does not hold slskd's single transfer slot for an evening.
slskdStallAfter = 10 * time.Minute
// slskdAbsentGrace is how long a requested file may be missing from
// slskd's transfer list before it is counted as failed. slskd lists
// a transfer as soon as it accepts it, so a file still absent after
// a few polls was refused.
slskdAbsentGrace = 30 * time.Second
// slskdCancelTimeout bounds the cleanup that cancels abandoned
// transfers.
slskdCancelTimeout = 15 * time.Second
)
func init() {
@@ -158,8 +134,6 @@ type slskd struct {
searchPoll time.Duration
searchWait time.Duration
transferPoll time.Duration
stallAfter time.Duration
absentGrace time.Duration
}
// newSlskd builds the provider from config.
@@ -214,8 +188,6 @@ func newSlskd(
searchPoll: slskdSearchPoll,
searchWait: slskdSearchWait,
transferPoll: slskdTransferPoll,
stallAfter: slskdStallAfter,
absentGrace: slskdAbsentGrace,
}, nil
}
@@ -259,13 +231,14 @@ type slskdSearch struct {
// slskdResponse is one peer's answer to a search.
type slskdResponse struct {
Username string `json:"username"`
HasFreeUploadSlot bool `json:"hasFreeUploadSlot"`
QueueLength int `json:"queueLength"`
UploadSpeed int64 `json:"uploadSpeed"`
Files []slskdFile `json:"files"`
LockedFileCount int `json:"lockedFileCount"`
FileCount int `json:"fileCount"`
Username string `json:"username"`
HasFreeUploadSlot bool `json:"hasFreeUploadSlot"`
QueueLength int `json:"queueLength"`
UploadSpeed int64 `json:"uploadSpeed"`
Files []slskdFile `json:"files"`
LockedFileCount int `json:"lockedFileCount"`
FileCount int `json:"fileCount"`
FreeUploadSlotFlag bool `json:"freeUploadSlots"`
}
// slskdFile is one file a peer is offering.
@@ -273,9 +246,7 @@ type slskdFile struct {
Filename string `json:"filename"`
Size int64 `json:"size"`
BitRate int `json:"bitRate"`
// Length is the duration in whole seconds.
Length int `json:"length"`
Length int `json:"length"`
}
// slskdTransfer is one download's state.
@@ -307,89 +278,24 @@ func (t slskdTransfer) done() (finished, ok bool) {
// per-folder candidates. A folder from one peer is the unit a user
// actually wants: Soulseek has no album concept, but people organise
// their shares by album directory.
//
// Up to two queries run at once — the request as written and a
// normalised form of it (see slskdQueries) — and their candidates are
// merged. They run concurrently rather than as a fallback because the
// manager gives a provider one search budget, and a Soulseek search
// spends most of it waiting for peers to answer; a second query after
// the first would not fit.
func (s *slskd) Search(ctx context.Context, dl Download) ([]Candidate, error) {
queries := slskdQueries(dl)
if len(queries) == 0 {
return nil, nil
}
type found struct {
candidates []Candidate
err error
}
results := make(chan found, len(queries))
for _, q := range queries {
go func(q string) {
c, err := s.searchOnce(ctx, q, minFilesFor(dl))
results <- found{candidates: c, err: err}
}(q)
}
var (
out []Candidate
seen = map[string]bool{}
firstErr error
answered int
)
for range queries {
r := <-results
if r.err != nil {
s.logger.Debug("slskd search failed", "error", r.err)
if firstErr == nil {
firstErr = r.err
}
continue
}
answered++
// The same peer's folder turns up under both queries; the ID is
// peer and folder, so it is the same candidate.
for _, c := range r.candidates {
if seen[c.ID] {
continue
}
seen[c.ID] = true
out = append(out, c)
}
}
if answered == 0 {
return nil, firstErr
}
return out, nil
}
// searchOnce runs one query to completion and returns its candidates.
func (s *slskd) searchOnce(
ctx context.Context,
text string,
minFiles int,
) ([]Candidate, error) {
// slskd's search endpoint deserializes id as a .NET Guid server-side,
// so it must be a dashed UUID — the app's own newID() (a plain hex
// string, used for request/item IDs elsewhere) is rejected with an
// HTTP 400 before any search happens.
searchID := uuid.NewString()
if err := s.client.post(
ctx, "/api/v0/searches", s.searchRequest(searchID, text, minFiles), nil,
); err != nil {
body := map[string]any{
"id": searchID,
"searchText": dl.SearchText(),
}
if err := s.client.post(ctx, "/api/v0/searches", body, nil); err != nil {
return nil, err
}
search, err := s.awaitSearch(ctx, searchID)
if err != nil {
return nil, err
}
@@ -401,222 +307,52 @@ func (s *slskd) searchOnce(
)
}()
if err := s.awaitSearch(ctx, searchID); err != nil {
return nil, err
}
responses, err := s.searchResponses(ctx, searchID)
if err != nil {
return nil, err
}
return s.candidatesFrom(responses, minFiles), nil
}
// searchRequest is the body that starts a search.
//
// Every option is stated rather than left to the daemon, because
// slskd's defaults are its own and not ours. Its search timeout in
// particular has to finish inside our wait: a search that slskd is still
// running when we stop polling is results we asked for and discarded.
// The response and file limits are raised well above what a popular
// album produces, and the peer filters let slskd drop answers this
// provider would only score down to nothing — a folder too small to be
// a candidate, a peer with a queue it will not reach today.
func (s *slskd) searchRequest(id, text string, minFiles int) map[string]any {
const (
responseLimit = 500
fileLimit = 20_000
maximumPeerQueueLength = 100
)
// A tenth of the wait is left for the last poll and the responses
// fetch.
timeout := s.searchWait - s.searchWait/10
return map[string]any{
"id": id,
"searchText": text,
"searchTimeout": timeout.Milliseconds(),
"responseLimit": responseLimit,
"fileLimit": fileLimit,
"filterResponses": true,
"minimumResponseFileCount": minFiles,
"maximumPeerQueueLength": maximumPeerQueueLength,
}
}
// slskdQueries is what is searched for a request: the request's own
// search text, and a normalised form of it when that differs.
//
// Soulseek matches every term against the file's full path, so each
// extra word is a filter, and some words filter wrongly:
//
// - edition qualifiers — "(Deluxe Edition)", "[2011 Remaster]" — are
// in the catalog's title and rarely in anyone's folder name;
// - punctuation splits a term oddly, and a term that starts with "-"
// is an *exclusion*, so an album called "-ism" searches for
// everything without it;
// - "Various Artists" is in no one's path for a compilation.
//
// A query the user typed is theirs and is searched exactly as written.
func slskdQueries(dl Download) []string {
primary := strings.TrimSpace(dl.SearchText())
if primary == "" {
return nil
}
out := []string{primary}
if dl.Query != "" {
return out
}
artist := dl.Artist
if isVariousArtists(artist) {
artist = ""
}
normal := Download{
Artist: normalizeSearchTerms(artist),
Album: normalizeSearchTerms(editionPattern.ReplaceAllString(dl.Album, " ")),
}
if alt := strings.TrimSpace(normal.SearchText()); alt != "" &&
!strings.EqualFold(alt, primary) {
out = append(out, alt)
}
return out
}
var (
// editionPattern finds an edition qualifier: a bracketed group that
// names an edition, or a trailing " - 2011 Remaster".
editionPattern = regexp.MustCompile(
`(?i)\s*[(\[][^)\]]*\b(?:deluxe|edition|remaster(?:ed)?|expanded|` +
`anniversary|bonus|explicit|reissue|special|collector'?s?|` +
`version|mono|stereo)\b[^)\]]*[)\]]` +
`|\s+-\s+(?:\d{4}\s+)?remaster(?:ed)?\b.*$`,
)
// nonWordPattern is everything that is not a letter or a digit.
nonWordPattern = regexp.MustCompile(`[^\p{L}\p{N}]+`)
)
// normalizeSearchTerms reduces text to plain words.
func normalizeSearchTerms(s string) string {
return strings.Join(strings.Fields(nonWordPattern.ReplaceAllString(s, " ")), " ")
}
// isVariousArtists reports whether an artist credit is a compilation's
// placeholder rather than an artist.
func isVariousArtists(artist string) bool {
switch strings.ToLower(strings.TrimSpace(artist)) {
case "various artists", "various", "va":
return true
default:
return false
}
}
// minFilesFor is the fewest audio files a folder must offer to be a
// candidate for this request.
//
// Soulseek answers a search with the files that match it, not with the
// folders they sit in. An album query matches every file in the album's
// folder, because the folder name carries the terms; a *track* query
// usually matches one file per folder. The two-file floor that filters
// out one-file noise for an album therefore filtered out every result
// for a track, and a single-track request could never be served here.
func minFilesFor(dl Download) int {
if dl.RecordingMBID != "" {
return 1
}
return slskdMinFiles
return s.candidatesFrom(search), nil
}
// awaitSearch polls until the search completes or the budget runs out.
// A timeout is not an error: partial Soulseek results are normal and
// often good enough.
//
// The poll asks for the search's state only. It used to ask for every
// response on every one-second tick, which for a popular album is the
// same few thousand file entries serialised twenty times to be read
// once; searchResponses fetches them once at the end.
func (s *slskd) awaitSearch(ctx context.Context, searchID string) error {
func (s *slskd) awaitSearch(
ctx context.Context,
searchID string,
) (slskdSearch, error) {
deadline := time.Now().Add(s.searchWait)
var last slskdSearch
for time.Now().Before(deadline) {
select {
case <-ctx.Done():
return fmt.Errorf("%w: search cancelled", ErrSlskdTimeout)
return last, fmt.Errorf("%w: search cancelled", ErrSlskdTimeout)
case <-time.After(s.searchPoll):
}
var search slskdSearch
if err := s.client.get(
ctx, "/api/v0/searches/"+searchID, &search,
ctx,
"/api/v0/searches/"+searchID+"?includeResponses=true",
&search,
); err != nil {
return err
return last, err
}
last = search
if search.IsComplete {
return nil
return search, nil
}
}
return nil
return last, nil
}
// searchResponses fetches a search's responses once.
//
// `/searches/{id}/responses` is the endpoint for that; a daemon that
// does not answer it is asked the older way, with the search itself
// carrying its responses, so an older slskd degrades to the previous
// behaviour rather than to no results at all.
func (s *slskd) searchResponses(
ctx context.Context,
searchID string,
) ([]slskdResponse, error) {
var responses []slskdResponse
// candidatesFrom groups a search's responses into candidates.
func (s *slskd) candidatesFrom(search slskdSearch) []Candidate {
out := make([]Candidate, 0, len(search.Responses))
err := s.client.get(
ctx, "/api/v0/searches/"+searchID+"/responses", &responses,
)
if err == nil {
return responses, nil
}
s.logger.Debug(
"slskd responses endpoint failed; asking with the search",
"error", err,
)
var search slskdSearch
if err := s.client.get(
ctx,
"/api/v0/searches/"+searchID+"?includeResponses=true",
&search,
); err != nil {
return nil, err
}
return search.Responses, nil
}
// candidatesFrom groups a search's responses into candidates, dropping
// folders with fewer than minFiles audio files.
func (s *slskd) candidatesFrom(
responses []slskdResponse,
minFiles int,
) []Candidate {
out := make([]Candidate, 0, len(responses))
for _, resp := range responses {
for _, resp := range search.Responses {
for folder, files := range groupByFolder(resp.Files) {
audio := 0
@@ -636,14 +372,12 @@ func (s *slskd) candidatesFrom(
Format: format,
Bitrate: f.BitRate,
IsAudio: isAudio,
LengthMillis: int64(f.Length) * millisPerSecond,
})
total += f.Size
}
if audio < minFiles {
if audio < slskdMinFiles {
continue
}
@@ -664,15 +398,13 @@ func (s *slskd) candidatesFrom(
return out
}
// groupByFolder buckets a peer's files by the album directory they sit
// in — the containing directory, or the one above it for a disc folder
// (see AlbumDir), so a multi-disc rip is one candidate and not two.
// groupByFolder buckets a peer's files by their containing directory.
func groupByFolder(files []slskdFile) map[string][]slskdFile {
out := map[string][]slskdFile{}
for _, f := range files {
dir := AlbumDir(f.Filename)
out[dir] = append(out[dir], f)
norm := strings.ReplaceAll(f.Filename, `\`, "/")
out[path.Dir(norm)] = append(out[path.Dir(norm)], f)
}
return out
@@ -686,7 +418,7 @@ func groupByFolder(files []slskdFile) map[string][]slskdFile {
func peerHealth(r slskdResponse) float64 {
score := 0.35
if r.HasFreeUploadSlot {
if r.HasFreeUploadSlot || r.FreeUploadSlotFlag {
score += 0.4
}
@@ -735,15 +467,6 @@ func (s *slskd) Grab(
)
}
// slskd keeps finished transfers listed until someone removes them,
// and a transfer is matched to the request by filename. A record
// left by an earlier attempt at the same file from the same peer
// would otherwise be read as this attempt's answer the moment the
// first poll came back — an old failure failing a transfer that has
// not started. So what is already terminal is noted before enqueueing
// and ignored after.
stale := s.terminalTransferIDs(ctx, username)
wanted := make([]map[string]any, 0, len(c.Files))
for _, f := range c.Files {
wanted = append(wanted, map[string]any{
@@ -753,69 +476,24 @@ func (s *slskd) Grab(
}
if err := s.client.post(
ctx, slskdDownloadsPath(username), wanted, nil,
ctx, "/api/v0/transfers/downloads/"+username, wanted, nil,
); err != nil {
return Result{}, err
}
if err := s.awaitTransfers(
ctx, username, stale, c, onProgress,
); err != nil {
if err := s.awaitTransfers(ctx, username, c, onProgress); err != nil {
return Result{}, err
}
return s.collect(c, dst)
}
// slskdDownloadsPath is the transfers endpoint for one peer. Soulseek
// usernames may contain spaces and punctuation, so the name is escaped
// rather than spliced into the path.
func slskdDownloadsPath(username string) string {
return "/api/v0/transfers/downloads/" + url.PathEscape(username)
}
// terminalTransferIDs returns the ids of this peer's transfers that are
// already finished. Best effort: slskd answers 404 for a peer it has no
// transfers with, and any failure here means only that there is nothing
// to ignore.
func (s *slskd) terminalTransferIDs(
ctx context.Context,
username string,
) map[string]bool {
transfers, err := s.transfersFor(ctx, username)
if err != nil {
return nil
}
out := make(map[string]bool, len(transfers))
for _, t := range transfers {
if finished, _ := t.done(); finished && t.ID != "" {
out[t.ID] = true
}
}
return out
}
// awaitTransfers polls until every requested file reaches a terminal
// state, the transfer stalls, or the caller gives up.
//
// Soulseek queues are measured in hours, so there is no deadline on the
// transfer as a whole — but there is one on *progress*. slskd's
// transfer limit is one, so a peer that holds us in its queue without
// sending a byte is not only failing this download, it is holding every
// other Soulseek download behind it. After stallAfter with nothing
// moving the peer is given up on, and the manager tries another.
//
// Whatever way this ends short of every file finishing, the transfers
// still live in slskd are cancelled there. Returning without doing so
// leaves the daemon downloading into its own folder for a request
// nobody is waiting on any more.
// state. Soulseek queues are measured in hours, so the only deadline
// is the caller's context.
func (s *slskd) awaitTransfers(
ctx context.Context,
username string,
stale map[string]bool,
c Candidate,
onProgress ProgressFunc,
) error {
@@ -824,18 +502,9 @@ func (s *slskd) awaitTransfers(
wanted[f.Path] = true
}
var (
started = time.Now()
lastProgress = started
lastBytes int64
live []slskdTransfer
)
for {
select {
case <-ctx.Done():
s.cancelTransfers(username, live)
return fmt.Errorf("%w: transfer cancelled", ErrSlskdTimeout)
case <-time.After(s.transferPoll):
}
@@ -843,177 +512,60 @@ func (s *slskd) awaitTransfers(
transfers, err := s.transfersFor(ctx, username)
if err != nil {
// A blip talking to the daemon should not abandon a
// transfer that may be hours in — but a daemon that stays
// away is a stall like any other.
// transfer that may be hours in.
s.logger.Debug("slskd transfer poll failed", "error", err)
if time.Since(lastProgress) >= s.stallAfter {
s.cancelTransfers(username, live)
return fmt.Errorf(
"%w: slskd has not answered for %s: %w",
ErrSlskdTimeout, s.stallAfter, err,
)
}
continue
}
tally := tallyTransfers(
transfers, wanted, stale,
time.Since(started) >= s.absentGrace,
var (
done, failed int
current int64
)
live = tally.live
if tally.bytes > lastBytes {
lastBytes = tally.bytes
lastProgress = time.Now()
for _, t := range transfers {
if !wanted[t.Filename] {
continue
}
current += t.BytesTransferred
finished, ok := t.done()
if !finished {
continue
}
if ok {
done++
} else {
failed++
}
}
if onProgress != nil {
onProgress(Progress{
Current: tally.bytes,
Current: current,
Total: c.TotalSize,
Phase: fmt.Sprintf(
"Transferring from %s (%d/%d)",
username, tally.done, len(wanted),
"Transferring from %s (%d/%d)", username, done, len(wanted),
),
})
}
if tally.done+tally.failed >= len(wanted) {
// Some files failing is normal — a peer goes offline
// mid-folder. Let the importer's completeness check decide
// whether what arrived is enough, rather than discarding it
// here.
if tally.done == 0 {
return fmt.Errorf(
"%w: all %d files failed",
ErrSlskdTransferFailed, tally.failed,
)
}
return nil
}
if time.Since(lastProgress) < s.stallAfter {
if done+failed < len(wanted) {
continue
}
s.cancelTransfers(username, live)
// A folder that stalls on its last track is the same shape as
// one whose last track failed, and goes forward the same way.
if tally.done > 0 {
s.logger.Info(
"slskd transfer stalled; keeping what arrived",
"peer", username,
"done", tally.done,
"wanted", len(wanted),
)
return nil
}
return fmt.Errorf(
"%w: %s sent nothing in %s",
ErrSlskdTimeout, username, s.stallAfter,
)
}
}
// transferTally is one poll's reading of the files a grab asked for.
type transferTally struct {
done, failed int
bytes int64
// live are the requested transfers slskd is still working on,
// which are what has to be cancelled if the grab is abandoned.
live []slskdTransfer
}
// tallyTransfers reads a peer's transfer list against the files a grab
// asked for.
//
// A requested file slskd does not list at all is one it never accepted
// — refused at enqueue, or dropped — and it will never reach a terminal
// state to be counted by. Once absentExpired, such a file counts as
// failed, or the grab would wait on it until the six-hour ceiling.
func tallyTransfers(
transfers []slskdTransfer,
wanted map[string]bool,
stale map[string]bool,
absentExpired bool,
) transferTally {
seen := make(map[string]slskdTransfer, len(wanted))
for _, t := range transfers {
if !wanted[t.Filename] || stale[t.ID] {
continue
}
seen[t.Filename] = t
}
var out transferTally
for name := range wanted {
t, ok := seen[name]
if !ok {
if absentExpired {
out.failed++
}
continue
}
out.bytes += t.BytesTransferred
finished, succeeded := t.done()
switch {
case !finished:
out.live = append(out.live, t)
case succeeded:
out.done++
default:
out.failed++
}
}
return out
}
// cancelTransfers asks slskd to cancel and forget transfers this grab
// is abandoning. It runs on a context of its own: the usual reason to
// be here is that the caller's context has just been cancelled, and a
// cleanup that inherited it would never be sent.
func (s *slskd) cancelTransfers(username string, live []slskdTransfer) {
if len(live) == 0 {
return
}
ctx, cancel := context.WithTimeout(
context.Background(), slskdCancelTimeout,
)
defer cancel()
for _, t := range live {
if t.ID == "" {
continue
}
endpoint := slskdDownloadsPath(username) + "/" +
url.PathEscape(t.ID) + "?remove=true"
if err := s.client.delete(ctx, endpoint); err != nil {
s.logger.Warn(
"could not cancel slskd transfer",
"peer", username,
"file", t.Filename,
"error", err,
// Some files failing is normal — a peer goes offline mid-folder.
// Let the importer's completeness check decide whether what
// arrived is enough, rather than discarding it here.
if done == 0 {
return fmt.Errorf(
"%w: all %d files failed", ErrSlskdTransferFailed, failed,
)
}
return nil
}
}
@@ -1029,7 +581,9 @@ func (s *slskd) transfersFor(
} `json:"directories"`
}
if err := s.client.get(ctx, slskdDownloadsPath(username), &raw); err != nil {
if err := s.client.get(
ctx, "/api/v0/transfers/downloads/"+username, &raw,
); err != nil {
return nil, err
}
@@ -1062,13 +616,7 @@ func (s *slskd) collect(c Candidate, dst string) (Result, error) {
continue
}
// A multi-disc candidate keeps its disc folders in staging.
// Flattened, disc 2's "01 Intro.flac" overwrites disc 1's, and
// the importer loses the folder it reads the disc number from.
target := filepath.Join(dst, base)
if _, ok := discFolder(folder); ok {
target = filepath.Join(dst, folder, base)
}
if err := movePath(src, target); err != nil {
return Result{}, fmt.Errorf("collect %s: %w", base, err)
+1 -368
View File
@@ -31,30 +31,11 @@ type slskdStub struct {
transfers [][]slskdTransfer
pollCount int
// before is what the downloads endpoint reports until something is
// enqueued: records slskd already held from earlier attempts.
before []slskdTransfer
// enqueued records what was requested for download.
enqueued []map[string]any
posted bool
// paths records the escaped path of every transfers call, and
// cancelled the escaped request URI of every DELETE.
paths []string
cancelled []string
// unauthorized makes every call return 401.
unauthorized bool
// searches records every search request body, and searchGets the
// request URI of every search GET.
searches []map[string]any
searchGets []string
// noResponsesEndpoint makes /searches/{id}/responses 404, as an
// older daemon would.
noResponsesEndpoint bool
}
func newSlskdStub(t *testing.T) *slskdStub {
@@ -76,16 +57,6 @@ func newSlskdStub(t *testing.T) *slskdStub {
return
}
var body map[string]any
if err := json.NewDecoder(r.Body).Decode(&body); err != nil {
t.Errorf("decode search body: %v", err)
}
s.mu.Lock()
s.searches = append(s.searches, body)
s.mu.Unlock()
w.WriteHeader(http.StatusCreated)
})
@@ -102,26 +73,8 @@ func newSlskdStub(t *testing.T) *slskdStub {
s.mu.Lock()
responses := s.responses
noEndpoint := s.noResponsesEndpoint
s.searchGets = append(s.searchGets, r.URL.RequestURI())
s.mu.Unlock()
if strings.HasSuffix(r.URL.Path, "/responses") {
if noEndpoint {
w.WriteHeader(http.StatusNotFound)
return
}
writeJSON(t, w, responses)
return
}
if r.URL.Query().Get("includeResponses") != "true" {
responses = nil
}
writeJSON(t, w, slskdSearch{
ID: "search-1",
IsComplete: true,
@@ -134,12 +87,7 @@ func newSlskdStub(t *testing.T) *slskdStub {
return
}
s.mu.Lock()
s.paths = append(s.paths, r.URL.EscapedPath())
s.mu.Unlock()
switch r.Method {
case http.MethodPost:
if r.Method == http.MethodPost {
var body []map[string]any
if err := json.NewDecoder(r.Body).Decode(&body); err != nil {
@@ -148,35 +96,15 @@ func newSlskdStub(t *testing.T) *slskdStub {
s.mu.Lock()
s.enqueued = body
s.posted = true
s.mu.Unlock()
w.WriteHeader(http.StatusCreated)
return
case http.MethodDelete:
s.mu.Lock()
s.cancelled = append(s.cancelled, r.URL.RequestURI())
s.mu.Unlock()
w.WriteHeader(http.StatusNoContent)
return
}
s.mu.Lock()
if !s.posted {
before := s.before
s.mu.Unlock()
writeJSON(t, w, map[string]any{
"directories": []map[string]any{{"files": before}},
})
return
}
idx := s.pollCount
if idx >= len(s.transfers) {
idx = len(s.transfers) - 1
@@ -263,11 +191,6 @@ func newStubSlskd(t *testing.T, stub *slskdStub) (*slskd, string) {
s.searchWait = 200 * time.Millisecond
s.transferPoll = time.Millisecond
// Long enough that no existing test trips them by accident; the
// tests about stalls and absences set their own.
s.stallAfter = time.Minute
s.absentGrace = time.Minute
return s, downloads
}
@@ -642,293 +565,3 @@ func TestSlskdRequiresConfiguration(t *testing.T) {
})
}
}
// slskdAlbum is a two-file candidate from peer, with the files slskd
// would have written already in place under downloads.
func slskdAlbum(t *testing.T, downloads, peer string, arrived ...string) Candidate {
t.Helper()
folder := filepath.Join(downloads, "Album")
if err := os.MkdirAll(folder, 0o750); err != nil {
t.Fatalf("mkdir: %v", err)
}
for _, name := range arrived {
if err := os.WriteFile(
filepath.Join(folder, name), []byte("audio"), 0o600,
); err != nil {
t.Fatalf("write: %v", err)
}
}
return Candidate{
Files: []CandidateFile{
{Path: `\s\Album\01 A.flac`, Size: 500, IsAudio: true},
{Path: `\s\Album\02 B.flac`, Size: 500, IsAudio: true},
},
TotalSize: 1000,
Payload: map[string]string{"username": peer},
}
}
// cancelledURIs returns what the stub was asked to cancel.
func (s *slskdStub) cancelledURIs() []string {
s.mu.Lock()
defer s.mu.Unlock()
return append([]string(nil), s.cancelled...)
}
// A peer that queues us and never sends a byte is given up on, and the
// queued transfers are cancelled in slskd rather than left to start
// hours later for a request nobody is waiting on.
func TestSlskdGrabGivesUpOnAStalledPeer(t *testing.T) {
t.Parallel()
stub := newSlskdStub(t)
stub.transfers = [][]slskdTransfer{{
{ID: "t1", Filename: `\s\Album\01 A.flac`, State: "Queued, Remotely"},
{ID: "t2", Filename: `\s\Album\02 B.flac`, State: "Queued, Remotely"},
}}
s, downloads := newStubSlskd(t, stub)
s.stallAfter = 30 * time.Millisecond
_, err := s.Grab(
context.Background(), slskdAlbum(t, downloads, "peer"), t.TempDir(), nil,
)
if !errors.Is(err, ErrSlskdTimeout) {
t.Fatalf("error = %v, want ErrSlskdTimeout", err)
}
got := stub.cancelledURIs()
if len(got) != 2 {
t.Fatalf("cancelled %v, want both queued transfers", got)
}
for _, uri := range got {
if !strings.HasSuffix(uri, "?remove=true") {
t.Errorf("cancel %s does not remove the record", uri)
}
}
}
// A folder that stalls on its last track goes forward with what
// arrived, the same as one whose last track failed; the importer's
// completeness check decides whether that is enough.
func TestSlskdGrabKeepsWhatArrivedBeforeAStall(t *testing.T) {
t.Parallel()
stub := newSlskdStub(t)
stub.transfers = [][]slskdTransfer{{
{
ID: "t1", Filename: `\s\Album\01 A.flac`,
State: "Completed, Succeeded", BytesTransferred: 500,
},
{ID: "t2", Filename: `\s\Album\02 B.flac`, State: "Queued, Remotely"},
}}
s, downloads := newStubSlskd(t, stub)
s.stallAfter = 30 * time.Millisecond
got, err := s.Grab(
context.Background(),
slskdAlbum(t, downloads, "peer", "01 A.flac"),
t.TempDir(), nil,
)
if err != nil {
t.Fatalf("Grab: %v", err)
}
if len(got.Files) != 1 {
t.Errorf("collected %d files, want the 1 that arrived", len(got.Files))
}
if cancelled := stub.cancelledURIs(); len(cancelled) != 1 ||
!strings.Contains(cancelled[0], "/t2") {
t.Errorf("cancelled %v, want only the stalled t2", cancelled)
}
}
// Progress is what holds the stall timer off. A transfer that keeps
// moving bytes is never abandoned, however long it takes.
func TestSlskdGrabWaitsOnATransferThatIsMoving(t *testing.T) {
t.Parallel()
stub := newSlskdStub(t)
for b := int64(1); b <= 100; b++ {
stub.transfers = append(stub.transfers, []slskdTransfer{
{ID: "t1", Filename: `\s\Album\01 A.flac`, State: "InProgress", BytesTransferred: b},
{ID: "t2", Filename: `\s\Album\02 B.flac`, State: "Queued, Remotely"},
})
}
stub.transfers = append(stub.transfers, []slskdTransfer{
{
ID: "t1",
Filename: `\s\Album\01 A.flac`,
State: "Completed, Succeeded",
BytesTransferred: 500,
},
{
ID: "t2",
Filename: `\s\Album\02 B.flac`,
State: "Completed, Succeeded",
BytesTransferred: 500,
},
})
s, downloads := newStubSlskd(t, stub)
// A hundred polls take several times the stall window; each one
// moves a byte. The window is kept well above one poll so a
// descheduled test runner does not read as a stall.
s.transferPoll = 5 * time.Millisecond
s.stallAfter = 150 * time.Millisecond
got, err := s.Grab(
context.Background(),
slskdAlbum(t, downloads, "peer", "01 A.flac", "02 B.flac"),
t.TempDir(), nil,
)
if err != nil {
t.Fatalf("Grab: %v", err)
}
if len(got.Files) != 2 {
t.Errorf("collected %d files, want 2", len(got.Files))
}
}
// A file slskd never lists was refused at enqueue and will never reach
// a terminal state. It counts as failed once the grace period is up,
// rather than being waited on until the six-hour ceiling.
func TestSlskdGrabCountsAnUnlistedFileAsFailed(t *testing.T) {
t.Parallel()
stub := newSlskdStub(t)
stub.transfers = [][]slskdTransfer{{
{
ID: "t1", Filename: `\s\Album\01 A.flac`,
State: "Completed, Succeeded", BytesTransferred: 500,
},
}}
s, downloads := newStubSlskd(t, stub)
s.absentGrace = 20 * time.Millisecond
got, err := s.Grab(
context.Background(),
slskdAlbum(t, downloads, "peer", "01 A.flac"),
t.TempDir(), nil,
)
if err != nil {
t.Fatalf("Grab: %v", err)
}
if len(got.Files) != 1 {
t.Errorf("collected %d files, want 1", len(got.Files))
}
}
// A finished record left by an earlier attempt at the same file is not
// this attempt's answer. Without the snapshot it would fail the grab on
// the first poll, before the new transfer had started.
func TestSlskdGrabIgnoresAnEarlierAttemptsRecord(t *testing.T) {
t.Parallel()
stale := slskdTransfer{
ID: "old", Filename: `\s\Album\01 A.flac`, State: "Completed, Errored",
}
stub := newSlskdStub(t)
stub.before = []slskdTransfer{stale}
stub.transfers = [][]slskdTransfer{
{stale},
{
stale,
{
ID: "new", Filename: `\s\Album\01 A.flac`,
State: "Completed, Succeeded", BytesTransferred: 500,
},
},
}
s, downloads := newStubSlskd(t, stub)
c := slskdAlbum(t, downloads, "peer", "01 A.flac")
c.Files = c.Files[:1]
got, err := s.Grab(context.Background(), c, t.TempDir(), nil)
if err != nil {
t.Fatalf("Grab: %v", err)
}
if len(got.Files) != 1 {
t.Errorf("collected %d files, want 1", len(got.Files))
}
}
// Cancelling the download cancels the transfer in slskd too. The
// cleanup must not inherit the cancelled context, or it is never sent.
func TestSlskdGrabCancelsTransfersWhenTheCallerGivesUp(t *testing.T) {
t.Parallel()
stub := newSlskdStub(t)
stub.transfers = [][]slskdTransfer{{
{ID: "t1", Filename: `\s\Album\01 A.flac`, State: "InProgress", BytesTransferred: 10},
{ID: "t2", Filename: `\s\Album\02 B.flac`, State: "Queued, Remotely"},
}}
s, downloads := newStubSlskd(t, stub)
ctx, cancel := context.WithTimeout(context.Background(), 50*time.Millisecond)
defer cancel()
_, err := s.Grab(ctx, slskdAlbum(t, downloads, "peer"), t.TempDir(), nil)
if !errors.Is(err, ErrSlskdTimeout) {
t.Fatalf("error = %v, want ErrSlskdTimeout", err)
}
if got := stub.cancelledURIs(); len(got) != 2 {
t.Errorf("cancelled %v, want both live transfers", got)
}
}
// Soulseek usernames carry spaces and punctuation; spliced raw into the
// path, a name with a slash addresses a different endpoint entirely.
func TestSlskdEscapesTheUsername(t *testing.T) {
t.Parallel()
stub := newSlskdStub(t)
stub.transfers = [][]slskdTransfer{{
{
ID: "t1", Filename: `\s\Album\01 A.flac`,
State: "Completed, Succeeded", BytesTransferred: 500,
},
{
ID: "t2", Filename: `\s\Album\02 B.flac`,
State: "Completed, Succeeded", BytesTransferred: 500,
},
}}
s, downloads := newStubSlskd(t, stub)
if _, err := s.Grab(
context.Background(),
slskdAlbum(t, downloads, "dj a/b", "01 A.flac", "02 B.flac"),
t.TempDir(), nil,
); err != nil {
t.Fatalf("Grab: %v", err)
}
stub.mu.Lock()
paths := append([]string(nil), stub.paths...)
stub.mu.Unlock()
for _, p := range paths {
if p != "/api/v0/transfers/downloads/dj%20a%2Fb" {
t.Errorf("transfers call went to %s", p)
}
}
}
+18 -104
View File
@@ -35,15 +35,6 @@ const (
weightArtistFit = 0.12
)
// Match sub-weights when the candidate's durations are known. Duration
// takes its weight from title fit, the signal it corroborates: a title
// says which song a file claims to be, a length says whether it is that
// recording — the right edit, the whole file, not the live take.
const (
timedWeightTitleFit = 0.25
timedWeightDurationFit = 0.15
)
// Quality sub-weights. Each set sums to 1.0.
//
// There are two of them because a stated preference changes what the
@@ -328,13 +319,13 @@ func Score(dl Download, c Candidate, priority int, prefs AutoDownloadPrefs) Cand
audio := c.AudioFiles()
a := alignFiles(audio, dl.Expected)
matched, titleFit := matchFiles(audio, dl.Expected)
// Write the alignment back so the picker can show which file maps
// to which track.
c.Files = mergeMatched(c.Files, a.files)
c.Files = mergeMatched(c.Files, matched)
c.Match = scoreMatch(dl, c, audio, a)
c.Match = scoreMatch(dl, c, audio, titleFit)
c.Quality = scoreQuality(
c, audio, priority, prefs, dl.runtimeMillis(),
)
@@ -349,21 +340,14 @@ func scoreMatch(
dl Download,
c Candidate,
audio []CandidateFile,
a alignment,
titleFit float64,
) MatchScore {
m := MatchScore{
Anchored: dl.Anchored(),
TitleFit: a.titleFit,
DurationFit: a.durationFit,
// Durations count once at least half the aligned pairs state
// one; a single timed pair would be a coin toss carrying 15%.
DurationKnown: a.timedPairs > 0 && a.timedPairs*2 >= a.aligned,
Anchored: dl.Anchored(),
TitleFit: titleFit,
}
m.Completeness = completeness(
alignedCount(c.Files), len(audio), len(dl.Expected),
)
m.Completeness = completeness(len(audio), len(dl.Expected))
// The candidate's own title, and the folder its files sit in, are
// two independent guesses at the album name. Take the better one:
@@ -383,16 +367,9 @@ func scoreMatch(
// With no expected tracklist there is no title signal at all, so
// redistribute its weight onto the album/artist evidence rather
// than scoring every free-text result as half-wrong.
switch {
case len(dl.Expected) == 0:
if len(dl.Expected) == 0 {
m.Overall = 0.55*m.AlbumFit + 0.45*m.ArtistFit
case m.DurationKnown:
m.Overall = timedWeightTitleFit*m.TitleFit +
timedWeightDurationFit*m.DurationFit +
weightCompleteness*m.Completeness +
weightAlbumFit*m.AlbumFit +
weightArtistFit*m.ArtistFit
default:
} else {
m.Overall = weightTitleFit*m.TitleFit +
weightCompleteness*m.Completeness +
weightAlbumFit*m.AlbumFit +
@@ -438,54 +415,30 @@ func artistFit(want string, c Candidate) float64 {
return best
}
// completeness scores how much of the expected tracklist a candidate
// covers. Extra files are penalized far more gently than missing ones:
// completeness scores audio file count against the expected track
// count. Extra files are penalized far more gently than missing ones:
// a folder with bonus tracks or a stray intro is still the album, while
// a folder missing half the tracks is not.
//
// **Coverage is counted in aligned tracks, not in files.** It used to
// be the audio file count, so any ten files scored full marks against
// a ten-track album whether or not they were its tracks — and since
// title fit is the mean over the files that *did* align, a folder where
// three titles matched read as a near-perfect candidate on both counts.
// `aligned` is how many files matchFiles assigned to an expected track;
// `audio` still sets the penalty for extras, because a folder of thirty
// files holding the ten wanted is a worse copy than one holding ten.
func completeness(aligned, audio, want int) float64 {
func completeness(got, want int) float64 {
if want == 0 {
if audio > 0 {
if got > 0 {
return 0.5
}
return 0
}
if aligned == 0 {
if got == 0 {
return 0
}
cover := float64(min(aligned, want)) / float64(want)
if got >= want {
extra := float64(got-want) / float64(want)
if audio > want {
extra := float64(audio-want) / float64(want)
cover *= math.Max(0.75, 1.0-0.25*extra)
return math.Max(0.75, 1.0-0.25*extra)
}
return cover
}
// alignedCount is how many audio files were assigned to an expected
// track.
func alignedCount(files []CandidateFile) int {
n := 0
for _, f := range files {
if f.IsAudio && f.MatchedTo != 0 {
n++
}
}
return n
return float64(got) / float64(want)
}
// scoreQuality answers whether this is a good copy.
@@ -782,45 +735,6 @@ func AutoPickVeto(
return ""
}
// autoAcceptable reports whether auto-pick may take this one candidate
// without asking: the request is anchored to a tracklist, and the
// candidate is inside the user's guardrails and clears the match and
// quality bars. It is AutoPickVeto's test applied to a single
// candidate, which is what falling back to a second choice needs.
func autoAcceptable(dl Download, c Candidate, prefs AutoDownloadPrefs) bool {
return dl.Anchored() &&
len(dl.Expected) > 0 &&
prefs.eligible(c, dl.runtimeMillis()) &&
c.Match.Overall >= minMatch &&
c.Quality.Overall >= minQuality
}
// autoPick returns the candidate auto-pick takes: the best-ranked one
// it may take at all.
//
// That is not `ranked[0]`. AutoPickVeto judges the best candidate
// *inside* the guardrails, so when the overall best is outside them —
// over the size ceiling, say — the veto passes on the strength of the
// second, and grabbing the first would download exactly the copy the
// user said not to take unattended.
func autoPick(
dl Download,
ranked []Candidate,
prefs AutoDownloadPrefs,
) (Candidate, bool) {
if AutoPickVeto(dl, ranked, prefs) != "" {
return Candidate{}, false
}
for _, c := range ranked {
if autoAcceptable(dl, c, prefs) {
return c, true
}
}
return Candidate{}, false
}
// mergeMatched copies MatchedTo assignments from the audio-only slice
// back onto the full file list.
func mergeMatched(all, matched []CandidateFile) []CandidateFile {
+10 -14
View File
@@ -282,32 +282,28 @@ func TestCompleteness(t *testing.T) {
tests := []struct {
name string
aligned int
audio int
got int
want int
minScore float64
maxScore float64
}{
{"exact", 10, 10, 10, 1.0, 1.0},
{"half missing", 5, 5, 10, 0.49, 0.51},
{"one bonus track", 10, 11, 10, 0.95, 1.0},
{"double", 10, 20, 10, 0.74, 0.76},
{"nothing", 0, 0, 10, 0, 0},
{"no expectation", 0, 5, 0, 0.5, 0.5},
// Ten files are not ten tracks: three that align are three.
{"right count, wrong tracks", 3, 10, 10, 0.29, 0.31},
{"files that align to nothing", 0, 10, 10, 0, 0},
{"exact", 10, 10, 1.0, 1.0},
{"half missing", 5, 10, 0.49, 0.51},
{"one bonus track", 11, 10, 0.95, 1.0},
{"double", 20, 10, 0.74, 0.76},
{"nothing", 0, 10, 0, 0},
{"no expectation", 5, 0, 0.5, 0.5},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
t.Parallel()
got := completeness(tt.aligned, tt.audio, tt.want)
got := completeness(tt.got, tt.want)
if got < tt.minScore || got > tt.maxScore {
t.Errorf(
"completeness(%d, %d, %d) = %f, want in [%f, %f]",
tt.aligned, tt.audio, tt.want, got, tt.minScore, tt.maxScore,
"completeness(%d, %d) = %f, want in [%f, %f]",
tt.got, tt.want, got, tt.minScore, tt.maxScore,
)
}
})
-238
View File
@@ -1,238 +0,0 @@
package download
import (
"context"
"strings"
"testing"
)
// What Soulseek is asked, how, and what is kept from the answer (#271).
func TestSlskdQueries(t *testing.T) {
t.Parallel()
cases := []struct {
name string
dl Download
want []string
}{
{
name: "a plain request is searched once",
dl: Download{Artist: "Radiohead", Album: "OK Computer"},
want: []string{"Radiohead OK Computer"},
},
{
name: "an edition qualifier gets a second query without it",
dl: Download{Artist: "Radiohead", Album: "OK Computer (Collector's Edition)"},
want: []string{
"Radiohead OK Computer (Collector's Edition)",
"Radiohead OK Computer",
},
},
{
name: "a trailing remaster note",
dl: Download{Artist: "Pink Floyd", Album: "Animals - 2018 Remaster"},
want: []string{
"Pink Floyd Animals - 2018 Remaster",
"Pink Floyd Animals",
},
},
{
name: "a leading dash would be an exclusion",
dl: Download{Artist: "Mocky", Album: "-ism"},
want: []string{"Mocky -ism", "Mocky ism"},
},
{
name: "a compilation is not searched by its placeholder artist",
dl: Download{Artist: "Various Artists", Album: "Pulp Fiction"},
want: []string{"Various Artists Pulp Fiction", "Pulp Fiction"},
},
{
name: "what the user typed is searched as written",
dl: Download{Query: "ok computer (deluxe)", Album: "OK Computer (Deluxe)"},
want: []string{"ok computer (deluxe)"},
},
}
for _, tc := range cases {
t.Run(tc.name, func(t *testing.T) {
t.Parallel()
got := slskdQueries(tc.dl)
if strings.Join(got, "|") != strings.Join(tc.want, "|") {
t.Errorf("slskdQueries = %q, want %q", got, tc.want)
}
})
}
}
// Both queries run, the options are stated rather than left to the
// daemon's defaults, and a folder both queries found is one candidate.
func TestSlskdSearchRunsBothQueriesAndMerges(t *testing.T) {
t.Parallel()
stub := newSlskdStub(t)
stub.responses = []slskdResponse{{
Username: "peer",
Files: []slskdFile{
{Filename: `\m\Radiohead - OK Computer\01 Airbag.flac`, Size: 1, Length: 284},
{Filename: `\m\Radiohead - OK Computer\02 Paranoid Android.flac`, Size: 1, Length: 383},
},
}}
s, _ := newStubSlskd(t, stub)
got, err := s.Search(context.Background(), Download{
ReleaseMBID: "rel", Artist: "Radiohead", Album: "OK Computer (Deluxe Edition)",
})
if err != nil {
t.Fatalf("Search: %v", err)
}
if len(got) != 1 {
t.Fatalf("got %d candidates, want the one folder once", len(got))
}
if got[0].Files[0].LengthMillis != 284_000 {
t.Errorf("length = %d ms, want 284000 from slskd's seconds", got[0].Files[0].LengthMillis)
}
stub.mu.Lock()
searches := append([]map[string]any(nil), stub.searches...)
gets := append([]string(nil), stub.searchGets...)
stub.mu.Unlock()
if len(searches) != 2 {
t.Fatalf("ran %d searches, want 2", len(searches))
}
for _, body := range searches {
for _, key := range []string{
"searchTimeout", "responseLimit", "fileLimit",
"minimumResponseFileCount", "maximumPeerQueueLength",
} {
if _, ok := body[key]; !ok {
t.Errorf("search %q does not state %s", body["searchText"], key)
}
}
}
// The responses are fetched once at the end, not with every poll.
for _, uri := range gets {
if strings.Contains(uri, "includeResponses") {
t.Errorf("poll %s asked for every response", uri)
}
}
}
// A daemon without the responses endpoint still returns results.
func TestSlskdSearchFallsBackForAnOlderDaemon(t *testing.T) {
t.Parallel()
stub := newSlskdStub(t)
stub.noResponsesEndpoint = true
stub.responses = []slskdResponse{{
Username: "peer",
Files: []slskdFile{
{Filename: `\m\Album\01 A.flac`, Size: 1},
{Filename: `\m\Album\02 B.flac`, Size: 1},
},
}}
s, _ := newStubSlskd(t, stub)
got, err := s.Search(context.Background(), Download{Query: "album"})
if err != nil {
t.Fatalf("Search: %v", err)
}
if len(got) != 1 {
t.Errorf("got %d candidates, want 1 through the fallback", len(got))
}
}
func TestDurationAgreement(t *testing.T) {
t.Parallel()
cases := []struct {
got, want int64
score float64
}{
{300_000, 300_000, 1},
{301_500, 300_000, 1}, // a second of silence
{300_000, 316_500, 0.5},
{300_000, 345_000, 0}, // a different edit
}
for _, tc := range cases {
if got := durationAgreement(tc.got, tc.want); got < tc.score-0.01 || got > tc.score+0.01 {
t.Errorf("durationAgreement(%d, %d) = %f, want %f", tc.got, tc.want, got, tc.score)
}
}
}
// Two folders with the same track names are told apart by their
// lengths: one is the album, the other a live record of the same songs.
func TestDurationsSeparateTheRightRecording(t *testing.T) {
t.Parallel()
dl := okComputer()
timed := func(id string, lengths ...int64) Candidate {
c := candidateFor(id, allTitles(), ".flac", 30_000_000)
for i := range c.Files {
c.Files[i].LengthMillis = lengths[i]
}
return c
}
studio := timed("studio", trackMillis, trackMillis+1_000, trackMillis, trackMillis-500)
live := timed(
"live",
trackMillis+60_000,
trackMillis+75_000,
trackMillis+50_000,
trackMillis+90_000,
)
ranked := Rank(dl, []Candidate{live, studio}, nil, AutoDownloadPrefs{})
if ranked[0].ID != "studio" {
t.Fatalf("winner = %s, want the recording whose lengths match", ranked[0].ID)
}
if !ranked[0].Match.DurationKnown || ranked[0].Match.DurationFit < 0.99 {
t.Errorf(
"studio duration fit = %f known=%v",
ranked[0].Match.DurationFit,
ranked[0].Match.DurationKnown,
)
}
if ranked[1].Match.DurationFit != 0 {
t.Errorf("live duration fit = %f, want 0", ranked[1].Match.DurationFit)
}
}
// Without lengths the score is exactly what it was before durations
// were read, so a provider that reports none is not penalised.
func TestUnknownDurationsLeaveTheScoreAlone(t *testing.T) {
t.Parallel()
dl := okComputer()
c := Score(dl, candidateFor("c", allTitles(), ".flac", 30_000_000), 50, AutoDownloadPrefs{})
if c.Match.DurationKnown {
t.Fatal("no file states a length, yet durations are known")
}
want := weightTitleFit*c.Match.TitleFit +
weightCompleteness*c.Match.Completeness +
weightAlbumFit*c.Match.AlbumFit +
weightArtistFit*c.Match.ArtistFit
if c.Match.Overall != want {
t.Errorf("match = %f, want the untimed formula's %f", c.Match.Overall, want)
}
}
+7 -19
View File
@@ -232,17 +232,12 @@ type Candidate struct {
// results give paths and sizes but no tags, so Format and duration are
// inferred from the path and size where possible.
type CandidateFile struct {
Path string `json:"path"`
Size int64 `json:"size"`
Format Format `json:"format"`
Bitrate int `json:"bitrate,omitempty"` // kbps, 0 when unknown
IsAudio bool `json:"isAudio"`
// LengthMillis is the file's duration as the source reports it, or
// 0 when it does not. Soulseek reports it for most audio files.
LengthMillis int64 `json:"lengthMillis,omitempty"`
MatchedTo int `json:"matchedTo,omitempty"` // expected track position
Path string `json:"path"`
Size int64 `json:"size"`
Format Format `json:"format"`
Bitrate int `json:"bitrate,omitempty"` // kbps, 0 when unknown
IsAudio bool `json:"isAudio"`
MatchedTo int `json:"matchedTo,omitempty"` // expected track position
}
// Format is a normalized audio container/codec name.
@@ -291,14 +286,7 @@ type MatchScore struct {
TitleFit float64 `json:"titleFit"` // filenames vs expected titles
ArtistFit float64 `json:"artistFit"` // path/origin vs expected artist
AlbumFit float64 `json:"albumFit"` // folder name vs album title
Completeness float64 `json:"completeness"` // aligned tracks vs expected count
// DurationFit is how well the aligned files' lengths agree with the
// expected tracks', and DurationKnown whether enough of them stated
// a length for that to count. When it does not, the score is the
// four text signals alone, exactly as before durations were read.
DurationFit float64 `json:"durationFit"`
DurationKnown bool `json:"durationKnown"`
Completeness float64 `json:"completeness"` // audio files vs expected count
// Anchored records whether an MBID drove this score. Unanchored
// matches are capped, because there is nothing to be right about.
-72
View File
@@ -3,7 +3,6 @@ package library
import (
"bytes"
"crypto/sha256"
"database/sql"
"encoding/hex"
"fmt"
"image"
@@ -70,77 +69,6 @@ func CoverArtFileSet(coverPath string) []string {
return paths
}
// sweepOrphanedCoverArt deletes the cover_art rows no album references
// and returns their file paths, for the caller to remove from disk
// after the transaction commits. Cover art is referenced only by
// albums.cover_art_id, so an orphan is a cover whose album is gone —
// which is every album the caller just swept.
//
// One implementation because the scan path, RemoveFromLibrary and
// RemoveLibrary all reach this state, and the scan side used to skip it
// entirely while RemoveLibrary did it inline (#247).
func (l *Library) sweepOrphanedCoverArt(tx *sql.Tx) ([]string, error) {
const orphanSQL = `
SELECT file_path FROM cover_art WHERE id NOT IN (
SELECT DISTINCT cover_art_id FROM albums
WHERE cover_art_id IS NOT NULL
)`
rows, err := tx.QueryContext(l.ctx, orphanSQL)
if err != nil {
return nil, fmt.Errorf("could not query orphaned cover art: %w", err)
}
var paths []string
for rows.Next() {
var filePath string
if err := rows.Scan(&filePath); err != nil {
l.logger.Warn("could not scan cover art path", "err", err)
continue
}
paths = append(paths, filePath)
}
// Close before the DELETE: the two run on the one writer connection.
if err := rows.Close(); err != nil {
l.logger.Warn("could not close cover art rows", "err", err)
}
if len(paths) > 0 {
if _, err := tx.ExecContext(l.ctx, `
DELETE FROM cover_art WHERE id NOT IN (
SELECT DISTINCT cover_art_id FROM albums
WHERE cover_art_id IS NOT NULL
)`); err != nil {
return nil, fmt.Errorf("could not delete orphaned cover_art: %w", err)
}
}
return paths, nil
}
// removeCoverArtFiles removes a cover original and its derived size
// variants. Only the original is stored in cover_art.file_path; the
// _sm/_md/_lg tiers are derived filenames beside it, so they have to be
// removed by name or they accumulate forever.
func (l *Library) removeCoverArtFiles(coverPaths []string) {
for _, coverPath := range coverPaths {
for _, path := range CoverArtFileSet(coverPath) {
if err := os.Remove(path); err != nil && !os.IsNotExist(err) {
l.logger.Warn(
"could not remove orphaned cover art file",
"path", path,
"err", err,
)
}
}
}
}
// saveCoverArt saves embedded cover art to the cache directory.
// Returns the file path where the art was saved, or empty string
// if no picture data. Timing is recorded in the provided metrics.
+4 -4
View File
@@ -8,7 +8,6 @@ import (
"yellowjacket/backend/coverart"
"yellowjacket/backend/database/sql/sqlcgen"
"yellowjacket/internal/testfixtures"
)
// TestScan_StoresOnlyCoverTiers pins the size decision: a scan writes
@@ -27,9 +26,10 @@ func TestScan_StoresOnlyCoverTiers(t *testing.T) {
lib, db := setupTestLibrary(t)
// Load skips when the fixture library has not been generated, as
// every other fixture test does.
root := testfixtures.Load(t).Root()
root, err := filepath.Abs("../../test_data/music_library_test")
if err != nil {
t.Fatalf("resolve fixture path: %v", err)
}
library, err := db.Queries.CreateLibrary(lib.ctx, sqlcgen.CreateLibraryParams{
Name: "Fixtures",
+50 -8
View File
@@ -320,12 +320,43 @@ func (l *Library) RemoveLibrary(id int64) (*RemovalSummary, error) {
genresRemoved, _ := result.RowsAffected()
// Collect and delete orphaned cover_art rows before the commit. The
// shared helper is the one place this sweep lives, so the scan path,
// RemoveFromLibrary and this removal cannot drift (#247).
orphanedCoverArtPaths, err := l.sweepOrphanedCoverArt(tx)
// 15. Collect orphaned cover_art file paths for post-commit cleanup.
// SAFETY: Hand-crafted SELECT for orphaned cover art identification.
// Parameterless.
rows, err := tx.QueryContext(l.ctx,
`SELECT file_path FROM cover_art WHERE id NOT IN (
SELECT DISTINCT cover_art_id FROM albums
WHERE cover_art_id IS NOT NULL
)`)
if err != nil {
return nil, err
return nil, fmt.Errorf("could not query orphaned cover art: %w", err)
}
var orphanedCoverArtPaths []string
for rows.Next() {
var filePath string
if err := rows.Scan(&filePath); err != nil {
l.logger.Warn("could not scan cover art path", "err", err)
continue
}
orphanedCoverArtPaths = append(orphanedCoverArtPaths, filePath)
}
if err := rows.Close(); err != nil {
l.logger.Warn("could not close cover art rows", "err", err)
}
// 16. Delete orphaned cover_art rows.
// SAFETY: Hand-crafted orphan cleanup SQL. Parameterless.
if _, err := tx.ExecContext(l.ctx,
`DELETE FROM cover_art WHERE id NOT IN (
SELECT DISTINCT cover_art_id FROM albums
WHERE cover_art_id IS NOT NULL
)`); err != nil {
return nil, fmt.Errorf("could not delete orphaned cover_art: %w", err)
}
// 17. Delete the library's tagging queue. tagging_items holds a
@@ -361,9 +392,20 @@ func (l *Library) RemoveLibrary(id int64) (*RemovalSummary, error) {
// avoids a costly full re-index of all remaining tracks (~10s for
// 25K tracks).
// Post-commit: remove the orphaned cover art files and their sized
// variants.
l.removeCoverArtFiles(orphanedCoverArtPaths)
// 21. Post-commit: Delete orphaned cover art files and their sized
// variants. Only the original is stored in cover_art.file_path; the
// _sm/_md/_lg thumbnails are derived filenames beside it, so they
// have to be removed by name or they accumulate forever.
for _, coverPath := range orphanedCoverArtPaths {
for _, path := range CoverArtFileSet(coverPath) {
if err := os.Remove(path); err != nil && !os.IsNotExist(err) {
l.logger.Warn("could not remove orphaned cover art file",
"path", path,
"err", err,
)
}
}
}
// 22. Post-commit: Compact queue.
if l.removalHooks.CompactQueue != nil {
+35 -110
View File
@@ -911,96 +911,30 @@ func (l *Library) scanInternal(
orphanStart := time.Now()
// Snapshot the orphan set first. The playlist-phantom
// preservation and the deletes are one transaction (the
// preservation has to land before the ON DELETE SET NULL, and
// both have to succeed or neither does), and the ids are what
// scope that preservation to just these files instead of
// rewriting every playlist row on a routine scan.
var (
orphans []sqlcgen.AudioFile
orphanPaths []string
)
existingPaths.Range(func(key, value any) bool {
orphanPaths = append(orphanPaths, key.(string))
orphans = append(orphans, value.(sqlcgen.AudioFile))
path := key.(string)
audioFile := value.(sqlcgen.AudioFile)
return true
})
l.logger.Debug(
"removing orphaned database entry",
"path", path, "id", audioFile.ID,
)
deleted := make([]bool, len(orphans))
if err := l.db.Queries.DeleteAudioFile(
l.ctx, audioFile.ID,
); err != nil {
l.logger.Warn(
"failed to delete orphaned audio file",
"path", path,
"id", audioFile.ID,
"err", err,
)
if len(orphans) > 0 {
orphanIDs := make([]int64, len(orphans))
for i, f := range orphans {
orphanIDs[i] = f.ID
metrics.addWarning(path, "orphan", err)
return true
}
tx, beginErr := l.db.BeginTx()
if beginErr != nil {
metrics.addWarning("", "orphan", beginErr)
} else {
defer func() { _ = tx.Rollback() }() // no-op after commit
if err := database.PreservePlaylistPhantomsForFiles(
l.ctx, tx, orphanIDs, l.logger,
); err != nil {
// Deleting without the phantoms is exactly the
// playlist-emptying bug the preservation exists to
// prevent, so leave the rows for the next scan
// rather than empty the playlists now.
l.logger.Error(
"skipping orphan deletion: could not preserve "+
"playlist entries",
"err", err,
)
metrics.addWarning("", "orphan", err)
_ = tx.Rollback()
} else {
txq := l.db.Queries.WithTx(tx)
for i, f := range orphans {
if err := txq.DeleteAudioFile(
l.ctx, f.ID,
); err != nil {
l.logger.Warn(
"failed to delete orphaned audio file",
"path", orphanPaths[i],
"id", f.ID,
"err", err,
)
metrics.addWarning(orphanPaths[i], "orphan", err)
continue
}
deleted[i] = true
}
if err := tx.Commit(); err != nil {
l.logger.Error(
"could not commit orphan deletion",
"err", err,
)
metrics.addWarning("", "orphan", err)
}
}
}
}
// Post-commit bookkeeping for the files that actually went.
for i, f := range orphans {
if !deleted[i] {
continue
}
path := orphanPaths[i]
// Keep the file's tagging group in sync: drop the group's
// track count and clear it out once empty, mirroring the
// bookkeeping maybeRebindTaggingGroup does for a group_key
@@ -1008,25 +942,25 @@ func (l *Library) scanInternal(
// and replaced leaves a stale tagging_items row behind —
// its track_count still counts the deleted files, and it
// never clears from the autotag queue.
if f.GroupKey != "" {
if audioFile.GroupKey != "" {
if err := l.db.Queries.DecrementTaggingItemTrackCount(
l.ctx, f.GroupKey,
l.ctx, audioFile.GroupKey,
); err != nil {
l.logger.Warn(
"failed to decrement tagging group for orphan",
"path", path,
"group_key", f.GroupKey,
"group_key", audioFile.GroupKey,
"err", err,
)
metrics.addWarning(path, "orphan", err)
} else if err := l.db.Queries.DeleteTaggingItemIfEmpty(
l.ctx, f.GroupKey,
l.ctx, audioFile.GroupKey,
); err != nil {
l.logger.Warn(
"failed to clean up emptied tagging group for orphan",
"path", path,
"group_key", f.GroupKey,
"group_key", audioFile.GroupKey,
"err", err,
)
@@ -1035,20 +969,24 @@ func (l *Library) scanInternal(
}
// Remove from FTS5 search index and the lyrics index.
if err := l.db.DeleteSearchIndex(f.ID); err != nil {
if err := l.db.DeleteSearchIndex(
audioFile.ID,
); err != nil {
l.logger.Warn(
"failed to delete FTS entry for orphan",
"id", f.ID,
"id", audioFile.ID,
"err", err,
)
metrics.addWarning(path, "orphan", err)
}
if err := l.db.DeleteLyricsIndex(f.ID); err != nil {
if err := l.db.DeleteLyricsIndex(
audioFile.ID,
); err != nil {
l.logger.Warn(
"failed to delete lyrics index entry for orphan",
"id", f.ID,
"id", audioFile.ID,
"err", err,
)
@@ -1056,7 +994,9 @@ func (l *Library) scanInternal(
}
removed.Add(1)
}
return true
})
metrics.OrphanCleanup = time.Since(orphanStart)
@@ -1265,32 +1205,17 @@ func (l *Library) pruneEmptyEntities() {
}
}
// Cover art after albums: a cover whose album just went is
// unreferenced, and leaving the row behind keeps its files exempt
// from the janitor's covers sweep forever (#247).
orphanedCovers, err := l.sweepOrphanedCoverArt(tx)
if err != nil {
l.logger.Warn("could not sweep orphaned cover art", "err", err)
return
}
if err := tx.Commit(); err != nil {
l.logger.Warn("could not commit entity cleanup", "err", err)
return
}
// Post-commit: the rows are gone, so their files can go too.
l.removeCoverArtFiles(orphanedCovers)
if len(albumIDs) > 0 || len(artistIDs) > 0 || len(genreIDs) > 0 ||
len(orphanedCovers) > 0 {
if len(albumIDs) > 0 || len(artistIDs) > 0 || len(genreIDs) > 0 {
l.logger.Info("pruned empty library entities",
"albums", len(albumIDs),
"artists", len(artistIDs),
"genres", len(genreIDs),
"covers", len(orphanedCovers),
)
}
}
-101
View File
@@ -94,57 +94,6 @@ func countRows(
return n
}
// RemoveFromLibrary is the one path that empties a track *deliberately*:
// the file stays on disk but is excluded, so nothing re-imports it. The
// playlist entry must still survive as a re-linkable phantom rather than
// an empty row, because a later full rescan clears the exclusion and is
// what re-links the entry then (#246).
func TestRemoveFromLibrary_PreservesPlaylistPhantoms(t *testing.T) {
t.Parallel()
lib, db := setupTestLibrary(t)
seedRemovableLibrary(t, lib, "/nonexistent/cover.jpg")
if _, err := lib.db.ExecContext(
`INSERT INTO playlists (name) VALUES ('keepme')`,
); err != nil {
t.Fatalf("seed playlist: %v", err)
}
playlistID := queryInt(
t, db, `SELECT id FROM playlists WHERE name = 'keepme'`,
)
if _, err := lib.db.ExecContext(
`INSERT INTO playlist_tracks (playlist_id, audio_file_id, position)
SELECT ?, id, 0 FROM audio_files
WHERE file_path = '/music/song.mp3'`,
playlistID,
); err != nil {
t.Fatalf("seed playlist_tracks: %v", err)
}
if _, err := lib.RemoveFromLibrary([]string{"/music/song.mp3"}); err != nil {
t.Fatalf("RemoveFromLibrary: %v", err)
}
phantomPath := queryString(
t, db,
`SELECT phantom_file_path FROM playlist_tracks
WHERE playlist_id = ?`,
playlistID,
)
if phantomPath != "/music/song.mp3" {
t.Fatalf(
"phantom_file_path is %q, want %q -- the entry cannot be "+
"re-linked after a later full rescan",
phantomPath, "/music/song.mp3",
)
}
}
// A library with tagging_items must still be removable. tagging_items
// FK-references libraries with no ON DELETE clause, so leaving those
// rows behind fails the DELETE and rolls back the entire removal.
@@ -249,53 +198,3 @@ func TestCoverArtFileSet(t *testing.T) {
}
}
}
// Removing the last track of an album must take the album's cover art
// with it — both the row and every derived file — or the row keeps its
// files exempt from the janitor's covers sweep forever (#247).
func TestRemoveFromLibrary_DeletesOrphanedCoverArt(t *testing.T) {
t.Parallel()
lib, _ := setupTestLibrary(t)
dir := t.TempDir()
// The largest tier is what cover_art.file_path names; write every
// variant so the sweep has a real set to remove.
for _, tier := range thumbnailTiers {
p := filepath.Join(dir, coverart.SizedFilename("abc123.jpg", tier.Suffix))
if err := os.WriteFile(p, []byte("img"), 0o600); err != nil {
t.Fatalf("write %s: %v", p, err)
}
}
cover := filepath.Join(dir, coverart.SizedFilename("abc123.jpg", "_lg"))
seedRemovableLibrary(t, lib, cover)
// Link the album to the cover so it is not orphaned until the track
// (and with it the album) goes.
if _, err := lib.db.ExecContext(
`UPDATE albums SET cover_art_id =
(SELECT id FROM cover_art WHERE file_path = ?)
WHERE name = 'Test Album'`,
cover,
); err != nil {
t.Fatalf("link cover art: %v", err)
}
if _, err := lib.RemoveFromLibrary([]string{"/music/song.mp3"}); err != nil {
t.Fatalf("RemoveFromLibrary: %v", err)
}
if n := countRows(t, lib, "cover_art"); n != 0 {
t.Errorf("cover_art has %d rows after removal, want 0", n)
}
for _, tier := range thumbnailTiers {
p := filepath.Join(dir, coverart.SizedFilename("abc123.jpg", tier.Suffix))
if _, err := os.Stat(p); !os.IsNotExist(err) {
t.Errorf("cover art file still present: %s", filepath.Base(p))
}
}
}
-20
View File
@@ -4,7 +4,6 @@ import (
"errors"
"fmt"
"yellowjacket/backend/database"
"yellowjacket/backend/events"
)
@@ -58,25 +57,6 @@ func (l *Library) RemoveFromLibrary(filePaths []string) (*RemovalResult, error)
var result RemovalResult
// Preserve the playlist entries before the rows go, so they survive
// as re-linkable phantoms rather than empty rows. The track is
// excluded and will not be re-imported on its own, but a later full
// rescan clears the exclusion and this is what lets the entry
// re-link then — the same preservation every other path that empties
// audio_files performs (#246).
rowIDs := make([]int64, len(rows))
for i, row := range rows {
rowIDs[i] = row.ID
}
if err := database.PreservePlaylistPhantomsForFiles(
l.ctx, tx, rowIDs, l.logger,
); err != nil {
return nil, fmt.Errorf(
"could not preserve playlist entries for removal: %w", err,
)
}
// Exclude every path the caller named, including one whose row has
// already gone: the user asked for that file to stay out, and a row
// that disappeared between the click and the commit is not a reason
-95
View File
@@ -231,98 +231,3 @@ func TestScan_MultipleDirectoriesDoNotCrossContaminate(t *testing.T) {
t.Errorf("Album A and Album B must not share a group_key: %+v", keys)
}
}
// TestScan_OrphanCleanupPreservesPlaylistPhantoms guards #246: a file
// deleted from the library folder *outside* YellowJacket is discovered
// as an orphan by the next scan, and its playlist entry must survive as
// a re-linkable phantom — the same preservation the full rescan and
// stale-retire paths already perform, scoped here to just the orphaned
// file. Before the fix the entry became an empty row (audio_file_id
// NULL and no phantom_file_path), which nothing can ever re-link.
func TestScan_OrphanCleanupPreservesPlaylistPhantoms(t *testing.T) {
t.Parallel()
lib, db := setupTestLibrary(t)
root := t.TempDir()
track := filepath.Join(root, "gone.mp3")
writeTestTrack(t, track, 0)
library, err := db.Queries.CreateLibrary(lib.ctx, sqlcgen.CreateLibraryParams{
Name: "orphans",
Path: root,
})
if err != nil {
t.Fatalf("create library: %v", err)
}
if metrics := lib.scanInternal(library.ID, library.Name, library.Path); metrics == nil {
t.Fatal("first scan returned nil metrics")
}
trackID := queryInt(
t, db, "SELECT id FROM audio_files WHERE file_path = ?", track,
)
if trackID == 0 {
t.Fatal("first scan did not import the track")
}
if _, err := db.ExecContext(
`INSERT INTO playlists (name) VALUES ('keepme')`,
); err != nil {
t.Fatalf("seed playlist: %v", err)
}
playlistID := queryInt(
t, db, "SELECT id FROM playlists WHERE name = 'keepme'",
)
if _, err := db.ExecContext(
`INSERT INTO playlist_tracks (playlist_id, audio_file_id, position)
VALUES (?, ?, 0)`,
playlistID, trackID,
); err != nil {
t.Fatalf("seed playlist_tracks: %v", err)
}
// The file goes away outside the app.
if err := os.Remove(track); err != nil {
t.Fatalf("remove track: %v", err)
}
if metrics := lib.scanInternal(library.ID, library.Name, library.Path); metrics == nil {
t.Fatal("second scan returned nil metrics")
}
if n := queryInt(
t, db, "SELECT COUNT(*) FROM audio_files WHERE file_path = ?", track,
); n != 0 {
t.Fatalf("audio_files still holds the removed path: %d rows", n)
}
if n := queryInt(
t, db,
"SELECT COUNT(*) FROM playlist_tracks WHERE playlist_id = ? "+
"AND audio_file_id IS NULL",
playlistID,
); n != 1 {
t.Fatalf(
"playlist entry did not become a phantom: %d null-id rows, want 1",
n,
)
}
phantomPath := queryString(
t, db,
"SELECT phantom_file_path FROM playlist_tracks WHERE playlist_id = ?",
playlistID,
)
if phantomPath != track {
t.Fatalf(
"phantom_file_path = %q, want %q -- the entry cannot be "+
"re-linked if the file comes back",
phantomPath, track,
)
}
}
-68
View File
@@ -667,74 +667,6 @@ func TestExpiredHTTPCacheJob_TrimsToBudget(t *testing.T) {
}
}
// TestStaleArtistMetadataJob pins the sweep's two keep rules: an owned
// artist's metadata survives, a browsed artist's survives while it still
// holds cached artwork, and everything else goes (#248).
func TestStaleArtistMetadataJob(t *testing.T) {
t.Parallel()
db := database.NewTestDB(t)
const (
ownedMBID = "11111111-1111-1111-1111-111111111111"
browsedMBID = "22222222-2222-2222-2222-222222222222"
staleMBID = "33333333-3333-3333-3333-333333333333"
)
// The owned artist is in the library - which means a *file* says
// so. An artists row on its own is not ownership.
database.InsertTestTrack(t, db, database.TestTrack{
FilePath: "/music/owned.mp3",
Artist: "Owned",
ArtistMBID: ownedMBID,
})
for _, mbid := range []string{ownedMBID, browsedMBID, staleMBID} {
if _, err := db.ExecContext(
`INSERT INTO artist_metadata (mbid, source, data, fetched_at)
VALUES (?, 'wikidata-p18', x'00', CURRENT_TIMESTAMP)`,
mbid,
); err != nil {
t.Fatalf("seed artist_metadata for %s: %v", mbid, err)
}
}
// The browsed artist holds cached artwork, so its metadata is still
// referenced and must survive.
if _, err := db.ExecContext(
`INSERT INTO artist_images
(artist_mbid, source, source_url, file_path)
VALUES (?, 'test', 'http://x', '/art/primary.jpg')`,
browsedMBID,
); err != nil {
t.Fatalf("seed artist_images: %v", err)
}
if _, err := StaleArtistMetadataJob(db).Run(context.Background()); err != nil {
t.Fatalf("run job: %v", err)
}
for _, tc := range []struct {
mbid string
want int
}{
{ownedMBID, 1},
{browsedMBID, 1},
{staleMBID, 0},
} {
var n int
if err := db.QueryRowWriter(
"SELECT COUNT(*) FROM artist_metadata WHERE mbid = ?", tc.mbid,
).Scan(&n); err != nil {
t.Fatalf("count %s: %v", tc.mbid, err)
}
if n != tc.want {
t.Errorf("artist_metadata rows for %s = %d, want %d", tc.mbid, n, tc.want)
}
}
}
// TestStaleSearchClicksJob deletes only the clicks old enough to have
// left the retention window (#249).
func TestStaleSearchClicksJob(t *testing.T) {
-34
View File
@@ -629,40 +629,6 @@ func dirSize(dir string) (bytes, files int64) {
return bytes, files
}
// StaleArtistMetadataJob evicts long-lived artist metadata (bios, wiki
// leads, relationships) for artists the user no longer has any reason
// to keep around: not owned and holding no cached artwork.
//
// artist_metadata has no TTL by design — entity data changes rarely and
// re-fetching spends someone else's rate limit — so without a sweep it
// grows for the life of the install. This is the "swept when the
// artist is no longer referenced" contract the datamap always declared
// for it and nothing ever performed (#248).
func StaleArtistMetadataJob(db *database.DB) Job {
return Job{
Name: "artist-metadata-sweep",
MinInterval: dailyInterval,
Run: func(_ context.Context) (Result, error) {
res, err := db.ExecContext(
`DELETE FROM artist_metadata
WHERE mbid NOT IN (` + ownedArtistMBIDs + `)
AND mbid NOT IN (
SELECT artist_mbid FROM artist_images
)`,
)
if err != nil {
return Result{}, fmt.Errorf(
"delete stale artist_metadata rows: %w", err,
)
}
rows, _ := res.RowsAffected()
return Result{RowsDeleted: rows}, nil
},
}
}
// searchClicksRetention is how long a search-click ranking signal stays
// useful. search_clicks is authored behavioural data — nothing that
// owns a row ever drops it — so age is the ceiling that keeps the table
@@ -121,12 +121,6 @@ export interface CandidateFile {
"bitrate"?: number;
"isAudio": boolean;
/**
* LengthMillis is the file's duration as the source reports it, or
* 0 when it does not. Soulseek reports it for most audio files.
*/
"lengthMillis"?: number;
/**
* expected track position
*/
@@ -459,19 +453,10 @@ export interface MatchScore {
"albumFit": number;
/**
* aligned tracks vs expected count
* audio files vs expected count
*/
"completeness": number;
/**
* DurationFit is how well the aligned files' lengths agree with the
* expected tracks', and DurationKnown whether enough of them stated
* a length for that to count. When it does not, the score is the
* four text signals alone, exactly as before durations were read.
*/
"durationFit": number;
"durationKnown": boolean;
/**
* Anchored records whether an MBID drove this score. Unanchored
* matches are capped, because there is nothing to be right about.
@@ -56,16 +56,6 @@ export function CycleRepeat(): $CancellablePromise<void> {
return $Call.ByID(3510519482);
}
/**
* DropSourceForPlaylist clears the queue's "Playing from" label when
* its source playlist is deleted. A link back to a playlist that no
* longer exists is worse than none, and the label otherwise survives
* the deletion until the next SetQueue (#249).
*/
export function DropSourceForPlaylist(playlistID: number): $CancellablePromise<void> {
return $Call.ByID(1435106374, playlistID);
}
/**
* EmitCurrentState emits the current queue state to the frontend.
* This is called after the frontend DOM is ready.