feat(backend): duplicate pairs, dismissals, and merge resolution

Adds the duplicate-detection backend on top of perceptual hashing:

- Two tables (edited into the original migrations): data.duplicate_pairs holds
  precomputed near-duplicate candidates (rebuilt wholesale by the rescan), and
  data.duplicate_dismissals is a global "not a duplicate" overlay that survives
  rescans. New audit actions file_merge / duplicate_dismiss.
- DuplicateService:
  - Rescan builds every pair within DUPLICATE_HASH_THRESHOLD via a BK-tree over
    the perceptual hashes and replaces the pairs table. This is the only thing
    that populates pairs, so GET never compares all-vs-all (scales to 110k+).
  - Clusters reads the precomputed pairs (ACL-filtered, non-trashed, non-
    dismissed), groups them into connected components via union-find, and
    paginates whole clusters.
  - Resolve merges a pair field-by-field: each scalar from keep or discard,
    metadata keep/discard/shallow-merge, tags/pools keep or union; then trashes
    the discarded file. Enforces edit ACL on both.
  - Dismiss records a canonical pair (view ACL on both).
- Endpoints under /files: GET /files/duplicates, POST /files/duplicates/dismiss,
  POST /files/duplicates/resolve (registered before /:id to avoid collision).
  Plain delete reuses /files/bulk/delete.
- Repo support: ListMissingPHash, ListAllPHashes, CopyPoolMemberships, plus the
  DuplicatePairRepo (ReplaceAll via COPY, ListVisible) and DismissalRepo.

Unit tests cover the BK-tree pairing, union-find clustering, metadata merge and
field validation; an integration test covers rescan -> list -> merge -> dismiss
(including that a dismissal survives a re-rescan).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

backend/internal/port/repository.go

@@ -52,6 +52,15 @@ type FileRepo interface {
 	SetNeedsReview(ctx context.Context, ids []uuid.UUID, value bool) error
 	// SetPHash sets (or clears, when nil) the perceptual hash of a file.
 	SetPHash(ctx context.Context, id uuid.UUID, phash *int64) error
+	// ListMissingPHash returns live image/video files that have no perceptual
+	// hash yet (the dedup backfill work list).
+	ListMissingPHash(ctx context.Context) ([]domain.File, error)
+	// ListAllPHashes returns the id and perceptual hash of every live, hashed
+	// file (the global input to the dedup rescan; not ACL-filtered).
+	ListAllPHashes(ctx context.Context) ([]domain.PHashEntry, error)
+	// CopyPoolMemberships adds targetID to every pool sourceID belongs to,
+	// skipping pools target is already in (used by the duplicate merge).
+	CopyPoolMemberships(ctx context.Context, targetID, sourceID uuid.UUID) error
 	// SoftDelete moves a file to trash (sets is_deleted = true).
 	SoftDelete(ctx context.Context, id uuid.UUID) error
 	// Restore moves a file out of trash (sets is_deleted = false).
@@ -72,6 +81,21 @@ type FileRepo interface {
 	RecordTagUses(ctx context.Context, userID int16, filterDSL string) error
 }
 
+// DuplicatePairRepo persists the precomputed near-duplicate candidate pairs.
+type DuplicatePairRepo interface {
+	// ReplaceAll atomically replaces the whole pairs table (used by the rescan).
+	ReplaceAll(ctx context.Context, pairs []domain.DuplicatePair) error
+	// ListVisible returns pairs whose both files are live, not dismissed, and
+	// (for non-admins) visible to the viewer.
+	ListVisible(ctx context.Context, viewerID int16, isAdmin bool) ([]domain.DuplicatePair, error)
+}
+
+// DismissalRepo persists "not a duplicate" decisions.
+type DismissalRepo interface {
+	// Add records a pair as dismissed (canonical order, idempotent).
+	Add(ctx context.Context, a, b uuid.UUID, userID int16) error
+}
+
 // TagRepo is the persistence interface for tags.
 type TagRepo interface {
 	List(ctx context.Context, params OffsetParams) (*domain.TagOffsetPage, error)