Optimize GetArtists: parallel artist processing, batch album upserts, batch download lookups, retry on metadata calls

internal/database/album_repository.go

@@ -5,6 +5,7 @@ import (
 	"fmt"
 	"time"
 
+	"github.com/jackc/pgx/v5"
 	"github.com/jackc/pgx/v5/pgxpool"
 )
 
@@ -60,6 +61,46 @@ func (r *AlbumRepository) Create(ctx context.Context, a *Album) error {
 	return nil
 }
 
+func (r *AlbumRepository) CreateBatch(ctx context.Context, albums []*Album) error {
+	if len(albums) == 0 {
+		return nil
+	}
+
+	batch := &pgx.Batch{}
+	for _, a := range albums {
+		batch.Queue(
+			`INSERT INTO albums (external_id, artist_id, title, album_type, release_date, total_tracks, total_discs, label, genres, cover_url, monitor_state)
+			 VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11)
+			 ON CONFLICT (external_id) DO UPDATE SET
+			   title = EXCLUDED.title,
+			   album_type = EXCLUDED.album_type,
+			   release_date = EXCLUDED.release_date,
+			   total_tracks = EXCLUDED.total_tracks,
+			   total_discs = EXCLUDED.total_discs,
+			   label = EXCLUDED.label,
+			   genres = EXCLUDED.genres,
+			   cover_url = EXCLUDED.cover_url,
+			   monitor_state = CASE
+			     WHEN albums.monitor_state = 'excluded' THEN albums.monitor_state
+			     WHEN albums.monitor_state = 'monitored' THEN albums.monitor_state
+			     ELSE EXCLUDED.monitor_state
+			   END,
+			   updated_at = NOW()`,
+			a.ExternalID, a.ArtistID, a.Title, a.AlbumType, a.ReleaseDate, a.TotalTracks, a.TotalDiscs, a.Label, a.Genres, a.CoverURL, a.MonitorState,
+		)
+	}
+
+	results := r.pool.SendBatch(ctx, batch)
+	defer results.Close()
+
+	for range albums {
+		if _, err := results.Exec(); err != nil {
+			return fmt.Errorf("batch creating album: %w", err)
+		}
+	}
+	return nil
+}
+
 func (r *AlbumRepository) GetByExternalID(ctx context.Context, externalID string) (*Album, error) {
 	a := &Album{}
 	err := r.pool.QueryRow(ctx,