refactor(hash-conflicts): one usage-counts reader for the API and SSR surfaces

The pending-hash-conflict walk existed twice (JSON admin API emitting
numeric counts, SSR page emitting a formatted summary) — deliberately
different shapes, but both duplicated the counts-or-zero lookup and
the five-field total. MediaItemUsageCountsTotal is the shared read;
the per-surface shaping stays local on purpose.

internal/handlers/hash_conflicts.go

@@ -2,6 +2,7 @@ package handlers
 
 import (
 	"bookhoard/internal/database"
+	"context"
 	"fmt"
 	"net/http"
 	"time"
@@ -48,6 +49,20 @@ type HashConflictResponse struct {
 	Items       []HashConflictItem `json:"items"`
 }
 
+// MediaItemUsageCountsTotal reads an item's usage tally for the
+// hash-conflict surfaces, treating a lookup failure as "no usage"
+// (these surfaces only use it to warn about data loss), and returns
+// the counts with their precomputed total.
+func MediaItemUsageCountsTotal(ctx context.Context, db *database.Queries, id pgtype.UUID) (database.GetMediaItemUsageCountsRow, int) {
+	counts, err := db.GetMediaItemUsageCounts(ctx, id)
+	if err != nil {
+		counts = database.GetMediaItemUsageCountsRow{}
+	}
+	total := counts.ProgressCount + counts.HighlightsCount + counts.BookmarksCount +
+		counts.NotesCount + counts.CollectionsCount
+	return counts, int(total)
+}
+
 // ListHashConflicts returns all pending hash conflict groups with their member
 // items and usage counts.
 // GET /api/admin/hash-conflicts
@@ -80,10 +95,7 @@ func (h *HashConflictsHandler) ListHashConflicts(c *echo.Context) error {
 		}
 
 		for _, mi := range items {
-			counts, err := h.db.GetMediaItemUsageCounts(ctx, mi.ID)
-			if err != nil {
-				counts = database.GetMediaItemUsageCountsRow{}
-			}
+			counts, _ := MediaItemUsageCountsTotal(ctx, h.db, mi.ID)
 			resp.Items = append(resp.Items, HashConflictItem{
 				ID:              uuid.UUID(mi.ID.Bytes),
 				Title:           mi.Title,