split duplicate grouping and hash helpers

internal/service/duplicate.go

@@ -16,13 +16,7 @@ package service
 
 import (
 	"context"
-	"crypto/sha256"
-	"encoding/hex"
-	"errors"
-	"fmt"
-	"io"
 	"os"
-	"sort"
 
 	"go.uber.org/zap"
 
@@ -30,8 +24,6 @@ import (
 	"github.com/ShukeBta/MediaStationGo/internal/repository"
 )
 
-const sampleSize = 1 << 20 // 1 MiB per sample window
-
 // DuplicateService is the entry point for the duplicate finder.
 type DuplicateService struct {
 	log  *zap.Logger
@@ -149,72 +141,6 @@ func (d *DuplicateService) Detect(ctx context.Context, libraryID string) (*Repor
 	return rep, nil
 }
 
-func (d *DuplicateService) markDuplicateGroup(ctx context.Context, rep *Report, key string, group []model.Media, markedIDs map[string]struct{}) {
-	primary := pickPrimary(group)
-	dupes := make([]model.Media, 0, len(group)-1)
-	for _, m := range group {
-		if m.ID == primary.ID || m.DuplicateOf == primary.ID {
-			continue
-		}
-		if _, ok := markedIDs[m.ID]; ok {
-			continue
-		}
-		dupes = append(dupes, m)
-		if err := d.repo.DB.WithContext(ctx).
-			Model(&model.Media{}).
-			Where("id = ?", m.ID).
-			Updates(map[string]any{
-				"is_duplicate": true,
-				"duplicate_of": primary.ID,
-			}).Error; err != nil {
-			d.log.Warn("dup mark failed", zap.Error(err))
-			continue
-		}
-		markedIDs[m.ID] = struct{}{}
-		rep.ItemsMarked++
-	}
-	if len(dupes) == 0 {
-		return
-	}
-	rep.Groups = append(rep.Groups, Group{
-		Hash:       key,
-		Primary:    primary,
-		Duplicates: dupes,
-	})
-}
-
-func groupByExternalIdentity(rows []model.Media) map[string][]model.Media {
-	groups := map[string][]model.Media{}
-	for _, row := range rows {
-		key := mediaExternalIdentityKey(row)
-		if key == "" {
-			continue
-		}
-		groups[key] = append(groups[key], row)
-	}
-	return groups
-}
-
-func mediaExternalIdentityKey(row model.Media) string {
-	var key string
-	switch {
-	case row.TMDbID > 0:
-		key = fmt.Sprintf("tmdb:%d", row.TMDbID)
-	case row.BangumiID > 0:
-		key = fmt.Sprintf("bangumi:%d", row.BangumiID)
-	case row.DoubanID != "":
-		key = "douban:" + row.DoubanID
-	case row.TheTVDBID != "":
-		key = "thetvdb:" + row.TheTVDBID
-	default:
-		return ""
-	}
-	if row.SeasonNum > 0 || row.EpisodeNum > 0 {
-		key += fmt.Sprintf(":s%d:e%d", row.SeasonNum, row.EpisodeNum)
-	}
-	return key
-}
-
 // Current returns duplicate groups already marked in the database. It keeps
 // the UI useful after a prior scan and avoids requiring POST on page load.
 func (d *DuplicateService) Current(ctx context.Context, libraryID string) (*Report, error) {
@@ -288,67 +214,3 @@ func (d *DuplicateService) Unmark(ctx context.Context, libraryID string) (int64,
 	res := q.Updates(map[string]any{"is_duplicate": false, "duplicate_of": ""})
 	return res.RowsAffected, res.Error
 }
-
-// pickPrimary picks the "best" media row to keep: prefer scraped > size > id.
-func pickPrimary(group []model.Media) model.Media {
-	sort.SliceStable(group, func(i, j int) bool {
-		ai, aj := group[i].ScrapeStatus == "matched", group[j].ScrapeStatus == "matched"
-		if ai != aj {
-			return ai
-		}
-		if group[i].SizeBytes != group[j].SizeBytes {
-			return group[i].SizeBytes > group[j].SizeBytes
-		}
-		return group[i].ID < group[j].ID
-	})
-	return group[0]
-}
-
-// SparseFileHash computes the head+mid+tail SHA-256 of a file, suffixed with
-// the file size so two files that happen to collide on the sample window
-// but differ in length are still distinguishable.
-func SparseFileHash(path string) (string, error) {
-	if path == "" {
-		return "", errors.New("empty path")
-	}
-	f, err := os.Open(path) // #nosec G304 -- path is selected from configured media library files for duplicate detection.
-	if err != nil {
-		return "", err
-	}
-	defer f.Close()
-	st, err := f.Stat()
-	if err != nil {
-		return "", err
-	}
-	size := st.Size()
-	h := sha256.New()
-	if size <= int64(sampleSize)*3 {
-		if _, err := io.Copy(h, f); err != nil {
-			return "", err
-		}
-		return fmt.Sprintf("%s-%d", hex.EncodeToString(h.Sum(nil)), size), nil
-	}
-	buf := make([]byte, sampleSize)
-	// head
-	if _, err := io.ReadFull(f, buf); err != nil {
-		return "", err
-	}
-	h.Write(buf)
-	// middle
-	if _, err := f.Seek(size/2-int64(sampleSize)/2, io.SeekStart); err != nil {
-		return "", err
-	}
-	if _, err := io.ReadFull(f, buf); err != nil {
-		return "", err
-	}
-	h.Write(buf)
-	// tail
-	if _, err := f.Seek(size-int64(sampleSize), io.SeekStart); err != nil {
-		return "", err
-	}
-	if _, err := io.ReadFull(f, buf); err != nil {
-		return "", err
-	}
-	h.Write(buf)
-	return fmt.Sprintf("%s-%d", hex.EncodeToString(h.Sum(nil)), size), nil
-}

internal/service/duplicate_groups.go

@@ -0,0 +1,93 @@
+// Package service — duplicate grouping helpers.
+package service
+
+import (
+	"context"
+	"fmt"
+	"sort"
+
+	"go.uber.org/zap"
+
+	"github.com/ShukeBta/MediaStationGo/internal/model"
+)
+
+func (d *DuplicateService) markDuplicateGroup(ctx context.Context, rep *Report, key string, group []model.Media, markedIDs map[string]struct{}) {
+	primary := pickPrimary(group)
+	dupes := make([]model.Media, 0, len(group)-1)
+	for _, m := range group {
+		if m.ID == primary.ID || m.DuplicateOf == primary.ID {
+			continue
+		}
+		if _, ok := markedIDs[m.ID]; ok {
+			continue
+		}
+		dupes = append(dupes, m)
+		if err := d.repo.DB.WithContext(ctx).
+			Model(&model.Media{}).
+			Where("id = ?", m.ID).
+			Updates(map[string]any{
+				"is_duplicate": true,
+				"duplicate_of": primary.ID,
+			}).Error; err != nil {
+			d.log.Warn("dup mark failed", zap.Error(err))
+			continue
+		}
+		markedIDs[m.ID] = struct{}{}
+		rep.ItemsMarked++
+	}
+	if len(dupes) == 0 {
+		return
+	}
+	rep.Groups = append(rep.Groups, Group{
+		Hash:       key,
+		Primary:    primary,
+		Duplicates: dupes,
+	})
+}
+
+func groupByExternalIdentity(rows []model.Media) map[string][]model.Media {
+	groups := map[string][]model.Media{}
+	for _, row := range rows {
+		key := mediaExternalIdentityKey(row)
+		if key == "" {
+			continue
+		}
+		groups[key] = append(groups[key], row)
+	}
+	return groups
+}
+
+func mediaExternalIdentityKey(row model.Media) string {
+	var key string
+	switch {
+	case row.TMDbID > 0:
+		key = fmt.Sprintf("tmdb:%d", row.TMDbID)
+	case row.BangumiID > 0:
+		key = fmt.Sprintf("bangumi:%d", row.BangumiID)
+	case row.DoubanID != "":
+		key = "douban:" + row.DoubanID
+	case row.TheTVDBID != "":
+		key = "thetvdb:" + row.TheTVDBID
+	default:
+		return ""
+	}
+	if row.SeasonNum > 0 || row.EpisodeNum > 0 {
+		key += fmt.Sprintf(":s%d:e%d", row.SeasonNum, row.EpisodeNum)
+	}
+	return key
+}
+
+// pickPrimary picks the "best" media row to keep: prefer scraped > size > id.
+func pickPrimary(group []model.Media) model.Media {
+	sort.SliceStable(group, func(i, j int) bool {
+		ai, aj := group[i].ScrapeStatus == "matched", group[j].ScrapeStatus == "matched"
+		if ai != aj {
+			return ai
+		}
+		if group[i].SizeBytes != group[j].SizeBytes {
+			return group[i].SizeBytes > group[j].SizeBytes
+		}
+		return group[i].ID < group[j].ID
+	})
+	return group[0]
+}

internal/service/duplicate_hash.go

@@ -0,0 +1,62 @@
+// Package service — duplicate sparse file hashing.
+package service
+
+import (
+	"crypto/sha256"
+	"encoding/hex"
+	"errors"
+	"fmt"
+	"io"
+	"os"
+)
+
+const sampleSize = 1 << 20 // 1 MiB per sample window
+
+// SparseFileHash computes the head+mid+tail SHA-256 of a file, suffixed with
+// the file size so two files that happen to collide on the sample window
+// but differ in length are still distinguishable.
+func SparseFileHash(path string) (string, error) {
+	if path == "" {
+		return "", errors.New("empty path")
+	}
+	f, err := os.Open(path) // #nosec G304 -- path is selected from configured media library files for duplicate detection.
+	if err != nil {
+		return "", err
+	}
+	defer f.Close()
+	st, err := f.Stat()
+	if err != nil {
+		return "", err
+	}
+	size := st.Size()
+	h := sha256.New()
+	if size <= int64(sampleSize)*3 {
+		if _, err := io.Copy(h, f); err != nil {
+			return "", err
+		}
+		return fmt.Sprintf("%s-%d", hex.EncodeToString(h.Sum(nil)), size), nil
+	}
+	buf := make([]byte, sampleSize)
+	// head
+	if _, err := io.ReadFull(f, buf); err != nil {
+		return "", err
+	}
+	h.Write(buf)
+	// middle
+	if _, err := f.Seek(size/2-int64(sampleSize)/2, io.SeekStart); err != nil {
+		return "", err
+	}
+	if _, err := io.ReadFull(f, buf); err != nil {
+		return "", err
+	}
+	h.Write(buf)
+	// tail
+	if _, err := f.Seek(size-int64(sampleSize), io.SeekStart); err != nil {
+		return "", err
+	}
+	if _, err := io.ReadFull(f, buf); err != nil {
+		return "", err
+	}
+	h.Write(buf)
+	return fmt.Sprintf("%s-%d", hex.EncodeToString(h.Sum(nil)), size), nil
+}