feat: duplicates resolve themselves where Lidarr says it is safe (M498)

The duplicate sweep proposed 4,197 groups and every one waited for the
operator. Most are safe to settle, and Lidarr defines what safe means: it
maps one file to each track of the release it monitors and downloads any
mapped file that disappears. Deleting a mapped copy opens exactly the hole
the operator saw Lidarr fill.

Classify (#5435)
- Migration 0075: duplicate_groups.class (same_release, cross_release,
  mismatch, review), resolve_note, resolved_automatically;
  duplicate_group_members.lidarr_state (tracked, unmapped);
  fingerprint_settings.auto_resolve; notification kind
  duplicates_resolved with both kind CHECKs swapped (rule 36).
- library.ClassifyDuplicateGroup, with MatchTitleKey dropping featuring
  credits, remaster notes and video-rip markers, and keeping live, demo,
  remix and instrumental. The rip markers move from api to library.

Choose the copy to keep (#5436)
- ProposeSurvivor ranks the copy Lidarr maps first, then tag fit (a
  clash-free track number, no rip marker in the name, an MBID), then the
  quality rules. File size picked the wrong Humanz copy in 6 of 21 groups.

Act (#5437)
- An hourly resolver pass reads Lidarr's unmapped files, matched by the
  last three path components, and records each copy's state.
- Same album, with at most one copy mapped: merged into the mapped copy.
  The merge is guarded, so a mapped copy can never be removed
  (MergeDuplicateGroupGuarded, ErrCopyTrackedByLidarr).
- Same album, every copy mapped: the monitored release lists the song
  twice (Humanz's 14x12" box set). The pass moves Lidarr to the release
  that lists each song once and best covers what is on disk. It never
  picks one covering less, and is capped at 10 albums per pass.
  - Fixed point (lesson #4183): the chosen release no longer repeats.
  - The album is left alone for 24h while Lidarr rescans, so "every copy
    unmapped" mid-rescan is never read as licence to merge.
- Both actions are audited with no actor and summarised to admins. The
  operator can switch them off in the Fingerprinting card (rule 25).
- Manual merges use the same guard: 409 copy_tracked_by_lidarr, or 503
  lidarr_unavailable when Lidarr cannot say.

Web
- Duplicates gets tabs: Needs review, Across releases, Resolved
  automatically. Each loads as you scroll (rule 172), replacing the
  pager.
- Each copy says whether Lidarr uses it.
- The resolver's note shows on each group.
- The merge confirm blocks, before sending, a merge that would remove
  the copy Lidarr uses.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

internal/db/dbq/duplicates.sql.go

@@ -62,6 +62,29 @@ func (q *Queries) CountPendingDuplicateGroups(ctx context.Context) (int64, error
 	return column_1, err
 }
 
+const countPendingDuplicateGroupsInView = `-- name: CountPendingDuplicateGroupsInView :one
+SELECT count(*)::bigint
+  FROM duplicate_groups g
+ WHERE g.status = 'pending'
+   AND (SELECT count(*) FROM duplicate_group_members m WHERE m.group_id = g.id) >= 2
+   AND CASE $1::text
+         WHEN 'cross_release' THEN g.class = 'cross_release'
+         ELSE g.class IS DISTINCT FROM 'cross_release'
+       END
+`
+
+// CountPendingDuplicateGroups narrowed to one view of the report (M498):
+//
+//	review         what the operator still has to decide: every class but
+//	               cross_release, and groups not yet classified
+//	cross_release  the same song on different releases, which is kept, not merged
+func (q *Queries) CountPendingDuplicateGroupsInView(ctx context.Context, view string) (int64, error) {
+	row := q.db.QueryRow(ctx, countPendingDuplicateGroupsInView, view)
+	var column_1 int64
+	err := row.Scan(&column_1)
+	return column_1, err
+}
+
 const deleteStalePendingDuplicateGroups = `-- name: DeleteStalePendingDuplicateGroups :execrows
 DELETE FROM duplicate_groups g
  WHERE g.status = 'pending'
@@ -187,6 +210,32 @@ func (q *Queries) GetLatestFingerprintComputedAt(ctx context.Context) (pgtype.Ti
 	return latest, err
 }
 
+const listAlbumPresentTrackTitles = `-- name: ListAlbumPresentTrackTitles :many
+SELECT title FROM tracks WHERE album_id = $1 AND missing_since IS NULL
+`
+
+// The titles an album has on disk, for choosing the Lidarr release that covers
+// them best (M498 #5437).
+func (q *Queries) ListAlbumPresentTrackTitles(ctx context.Context, albumID pgtype.UUID) ([]string, error) {
+	rows, err := q.db.Query(ctx, listAlbumPresentTrackTitles, albumID)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	var items []string
+	for rows.Next() {
+		var title string
+		if err := rows.Scan(&title); err != nil {
+			return nil, err
+		}
+		items = append(items, title)
+	}
+	if err := rows.Err(); err != nil {
+		return nil, err
+	}
+	return items, nil
+}
+
 const listDismissedDuplicateMemberSets = `-- name: ListDismissedDuplicateMemberSets :many
 SELECT g.id, array_agg(m.track_id ORDER BY m.track_id)::uuid[] AS track_ids
   FROM duplicate_groups g
@@ -288,6 +337,98 @@ func (q *Queries) ListDuplicateCandidates(ctx context.Context, arg ListDuplicate
 	return items, nil
 }
 
+const listDuplicateGroupsForResolve = `-- name: ListDuplicateGroupsForResolve :many
+SELECT g.id                  AS group_id,
+       g.tier,
+       t.id                  AS track_id,
+       t.album_id,
+       t.title,
+       artists.name          AS artist_name,
+       t.file_path,
+       t.file_format,
+       t.file_size,
+       t.added_at,
+       t.disc_number,
+       t.track_number,
+       (t.mbid IS NOT NULL)::boolean AS has_mbid,
+       albums.release_group_mbid,
+       albums.title          AS album_title,
+       EXISTS (SELECT 1 FROM tracks o
+                WHERE o.album_id = t.album_id AND o.id <> t.id AND o.missing_since IS NULL
+                  AND t.track_number IS NOT NULL
+                  AND o.disc_number IS NOT DISTINCT FROM t.disc_number
+                  AND o.track_number = t.track_number)::boolean AS position_clash
+  FROM duplicate_groups g
+  JOIN duplicate_group_members m ON m.group_id = g.id
+  JOIN tracks t   ON t.id = m.track_id
+  JOIN albums     ON albums.id = t.album_id
+  JOIN artists    ON artists.id = t.artist_id
+ WHERE g.status = 'pending'
+   AND t.missing_since IS NULL
+ ORDER BY g.id, t.id
+`
+
+type ListDuplicateGroupsForResolveRow struct {
+	GroupID          pgtype.UUID
+	Tier             string
+	TrackID          pgtype.UUID
+	AlbumID          pgtype.UUID
+	Title            string
+	ArtistName       string
+	FilePath         string
+	FileFormat       string
+	FileSize         int64
+	AddedAt          pgtype.Timestamptz
+	DiscNumber       *int32
+	TrackNumber      *int32
+	HasMbid          bool
+	ReleaseGroupMbid *string
+	AlbumTitle       string
+	PositionClash    bool
+}
+
+// Every pending group's present copies, for the resolver (M498 #5435): what it
+// needs to classify the group and to choose the copy to keep. One row per copy,
+// grouped by group. position_clash is true when another present track on the
+// same album claims the same disc and track number: tags that do not fit the
+// album, which the survivor rule weighs against a copy.
+func (q *Queries) ListDuplicateGroupsForResolve(ctx context.Context) ([]ListDuplicateGroupsForResolveRow, error) {
+	rows, err := q.db.Query(ctx, listDuplicateGroupsForResolve)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	var items []ListDuplicateGroupsForResolveRow
+	for rows.Next() {
+		var i ListDuplicateGroupsForResolveRow
+		if err := rows.Scan(
+			&i.GroupID,
+			&i.Tier,
+			&i.TrackID,
+			&i.AlbumID,
+			&i.Title,
+			&i.ArtistName,
+			&i.FilePath,
+			&i.FileFormat,
+			&i.FileSize,
+			&i.AddedAt,
+			&i.DiscNumber,
+			&i.TrackNumber,
+			&i.HasMbid,
+			&i.ReleaseGroupMbid,
+			&i.AlbumTitle,
+			&i.PositionClash,
+		); err != nil {
+			return nil, err
+		}
+		items = append(items, i)
+	}
+	if err := rows.Err(); err != nil {
+		return nil, err
+	}
+	return items, nil
+}
+
 const listExactDuplicateHashes = `-- name: ListExactDuplicateHashes :many
 SELECT f.audio_stream_sha256,
        array_agg(t.id ORDER BY t.id)::uuid[] AS track_ids
@@ -329,17 +470,32 @@ func (q *Queries) ListExactDuplicateHashes(ctx context.Context, currentVersion i
 
 const listPendingDuplicateGroupMembers = `-- name: ListPendingDuplicateGroupMembers :many
 WITH page AS (
-    SELECT g.id, g.tier, g.worst_bit_error_rate, g.detected_at
+    SELECT g.id, g.tier, g.worst_bit_error_rate, g.detected_at, g.class, g.resolve_note
       FROM duplicate_groups g
      WHERE g.status = 'pending'
        AND (SELECT count(*) FROM duplicate_group_members m WHERE m.group_id = g.id) >= 2
+       AND CASE $1::text
+             WHEN 'cross_release' THEN g.class = 'cross_release'
+             ELSE g.class IS DISTINCT FROM 'cross_release'
+           END
      ORDER BY g.detected_at DESC, g.id
-     LIMIT $2 OFFSET $1
+     LIMIT $3 OFFSET $2
 )
 SELECT p.id                  AS group_id,
        p.tier,
        p.worst_bit_error_rate,
        p.detected_at,
+       p.class,
+       p.resolve_note,
+       m.lidarr_state,
+       t.disc_number,
+       t.track_number,
+       (t.mbid IS NOT NULL)::boolean AS has_mbid,
+       EXISTS (SELECT 1 FROM tracks o
+                WHERE o.album_id = t.album_id AND o.id <> t.id AND o.missing_since IS NULL
+                  AND t.track_number IS NOT NULL
+                  AND o.disc_number IS NOT DISTINCT FROM t.disc_number
+                  AND o.track_number = t.track_number)::boolean AS position_clash,
        t.id                  AS track_id,
        t.title,
        artists.name          AS artist_name,
@@ -361,6 +517,7 @@ SELECT p.id                  AS group_id,
 `
 
 type ListPendingDuplicateGroupMembersParams struct {
+	View       string
 	PageOffset int32
 	PageLimit  int32
 }
@@ -370,6 +527,13 @@ type ListPendingDuplicateGroupMembersRow struct {
 	Tier              string
 	WorstBitErrorRate *float32
 	DetectedAt        pgtype.Timestamptz
+	Class             *string
+	ResolveNote       *string
+	LidarrState       *string
+	DiscNumber        *int32
+	TrackNumber       *int32
+	HasMbid           bool
+	PositionClash     bool
 	TrackID           pgtype.UUID
 	Title             string
 	ArtistName        string
@@ -384,12 +548,14 @@ type ListPendingDuplicateGroupMembersRow struct {
 	PlayCount         int64
 }
 
-// One page of proposals, newest first, flattened to one row per member so the
-// handler folds them without a query per group. What each copy carries — likes
-// and plays from every user — is here because it is what the operator weighs
-// when deciding which copy to keep.
+// One page of proposals in one view (see CountPendingDuplicateGroupsInView),
+// newest first, flattened to one row per member so the handler folds them
+// without a query per group. What each copy carries — likes and plays from
+// every user — is here because it is what the operator weighs when deciding
+// which copy to keep; Lidarr's view of it, because a copy Lidarr tracks cannot
+// be removed without Lidarr downloading it again.
 func (q *Queries) ListPendingDuplicateGroupMembers(ctx context.Context, arg ListPendingDuplicateGroupMembersParams) ([]ListPendingDuplicateGroupMembersRow, error) {
-	rows, err := q.db.Query(ctx, listPendingDuplicateGroupMembers, arg.PageOffset, arg.PageLimit)
+	rows, err := q.db.Query(ctx, listPendingDuplicateGroupMembers, arg.View, arg.PageOffset, arg.PageLimit)
 	if err != nil {
 		return nil, err
 	}
@@ -402,6 +568,13 @@ func (q *Queries) ListPendingDuplicateGroupMembers(ctx context.Context, arg List
 			&i.Tier,
 			&i.WorstBitErrorRate,
 			&i.DetectedAt,
+			&i.Class,
+			&i.ResolveNote,
+			&i.LidarrState,
+			&i.DiscNumber,
+			&i.TrackNumber,
+			&i.HasMbid,
+			&i.PositionClash,
 			&i.TrackID,
 			&i.Title,
 			&i.ArtistName,
@@ -425,6 +598,64 @@ func (q *Queries) ListPendingDuplicateGroupMembers(ctx context.Context, arg List
 	return items, nil
 }
 
+const markDuplicateGroupResolvedAutomatically = `-- name: MarkDuplicateGroupResolvedAutomatically :exec
+UPDATE duplicate_groups SET resolved_automatically = true WHERE id = $1
+`
+
+func (q *Queries) MarkDuplicateGroupResolvedAutomatically(ctx context.Context, id pgtype.UUID) error {
+	_, err := q.db.Exec(ctx, markDuplicateGroupResolvedAutomatically, id)
+	return err
+}
+
+const setDuplicateGroupClasses = `-- name: SetDuplicateGroupClasses :exec
+UPDATE duplicate_groups g
+   SET class = v.class, resolve_note = NULLIF(v.note, '')
+  FROM (SELECT unnest($1::uuid[]) AS id,
+               unnest($2::text[])   AS class,
+               unnest($3::text[])     AS note) v
+ WHERE g.id = v.id
+   AND g.status = 'pending'
+   AND (g.class IS DISTINCT FROM v.class OR g.resolve_note IS DISTINCT FROM NULLIF(v.note, ''))
+`
+
+type SetDuplicateGroupClassesParams struct {
+	GroupIds []pgtype.UUID
+	Classes  []string
+	Notes    []string
+}
+
+// The resolver's verdicts, written in one statement. Only rows whose class or
+// note actually changed are touched, so an hourly pass over thousands of
+// unchanged groups writes nothing. An empty note clears it.
+func (q *Queries) SetDuplicateGroupClasses(ctx context.Context, arg SetDuplicateGroupClassesParams) error {
+	_, err := q.db.Exec(ctx, setDuplicateGroupClasses, arg.GroupIds, arg.Classes, arg.Notes)
+	return err
+}
+
+const setDuplicateMemberLidarrStates = `-- name: SetDuplicateMemberLidarrStates :exec
+UPDATE duplicate_group_members m
+   SET lidarr_state = NULLIF(v.state, '')
+  FROM (SELECT unnest($1::uuid[]) AS group_id,
+               unnest($2::uuid[]) AS track_id,
+               unnest($3::text[])    AS state) v
+ WHERE m.group_id = v.group_id
+   AND m.track_id = v.track_id
+   AND m.lidarr_state IS DISTINCT FROM NULLIF(v.state, '')
+`
+
+type SetDuplicateMemberLidarrStatesParams struct {
+	GroupIds []pgtype.UUID
+	TrackIds []pgtype.UUID
+	States   []string
+}
+
+// Lidarr's view of each copy as of this pass; an empty state clears it (Lidarr
+// not consulted). Unchanged rows are left alone.
+func (q *Queries) SetDuplicateMemberLidarrStates(ctx context.Context, arg SetDuplicateMemberLidarrStatesParams) error {
+	_, err := q.db.Exec(ctx, setDuplicateMemberLidarrStates, arg.GroupIds, arg.TrackIds, arg.States)
+	return err
+}
+
 const startDuplicateSweep = `-- name: StartDuplicateSweep :one
 INSERT INTO duplicate_sweeps DEFAULT VALUES RETURNING id, started_at
 `