fix: guard series metadata from episode identity pollution

2026-10-03 04:26:36 +08:00 · 2026-06-25 07:22:00 +08:00
parent 3419077b67
commit 087a00b5d0
8 changed files with 338 additions and 13 deletions
@@ -98,6 +98,16 @@ func TestPathHintMetadataDoesNotMarkMediaMatched(t *testing.T) {
 	}
 }

+func TestPathHintMetadataIgnoresEpisodeFileIDForSeries(t *testing.T) {
+	meta, hints := pathHintMetadata("cloud://openlist/国漫/遮天 (2023)/Season 1/遮天.S01E01.{tmdb-4375419}.mkv", true)
+	if meta == nil || meta.Title != "遮天" || meta.Year != 2023 {
+		t.Fatalf("series path hint metadata = %+v", meta)
+	}
+	if hints.TMDbID != 0 || meta.TMDbID != 0 {
+		t.Fatalf("episode filename tmdb id must not become series id: meta=%+v hints=%+v", meta, hints)
+	}
+}
+
 func TestEnrichOneCloudPathHintOverridesStaleTMDbID(t *testing.T) {
 	var requested []string
 	upstream := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -168,6 +178,124 @@ func TestEnrichOneCloudPathHintOverridesStaleTMDbID(t *testing.T) {
 	}
 }

+func TestEnrichOneRejectsStaleEpisodeTMDbIDBySeriesTitle(t *testing.T) {
+	var requested []string
+	upstream := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		requested = append(requested, r.URL.Path)
+		w.Header().Set("Content-Type", "application/json")
+		switch r.URL.Path {
+		case "/tv/220269":
+			_ = json.NewEncoder(w).Encode(map[string]any{
+				"id":             220269,
+				"name":           "错误剧集",
+				"overview":       "stale episode id resolved as another show",
+				"first_air_date": "2024-01-01",
+			})
+		case "/search/tv":
+			if r.URL.Query().Get("query") != "折腰" {
+				_ = json.NewEncoder(w).Encode(map[string]any{"results": []any{}})
+				return
+			}
+			_ = json.NewEncoder(w).Encode(map[string]any{
+				"results": []map[string]any{{
+					"id":             296753,
+					"name":           "折腰",
+					"overview":       "正确的剧集条目",
+					"poster_path":    "/zheyao.jpg",
+					"first_air_date": "2025-05-13",
+					"origin_country": []string{"CN"},
+				}},
+			})
+		case "/tv/296753":
+			_ = json.NewEncoder(w).Encode(map[string]any{
+				"id":             296753,
+				"name":           "折腰",
+				"overview":       "正确的剧集条目",
+				"poster_path":    "/zheyao.jpg",
+				"first_air_date": "2025-05-13",
+				"origin_country": []string{"CN"},
+			})
+		default:
+			http.NotFound(w, r)
+		}
+	}))
+	defer upstream.Close()
+
+	db, err := gorm.Open(sqlite.Open("file::memory:?cache=shared"), &gorm.Config{})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if err := db.AutoMigrate(&model.Library{}, &model.Series{}, &model.Media{}); err != nil {
+		t.Fatal(err)
+	}
+	repos := repository.New(db)
+	cfg := &config.Config{}
+	cfg.Secrets.TMDbAPIKey = "test-key"
+	cfg.Secrets.TMDbAPIProxy = upstream.URL
+	cfg.Secrets.TMDbImageProxy = upstream.URL + "/images"
+	log := zap.NewNop()
+	scraper := NewScraperService(cfg, log, repos, NewTMDbProvider(cfg, log, nil), nil, nil, nil, NewHub(log))
+
+	lib := model.Library{Name: "OpenList · 国产剧", Path: "cloud://openlist/国产剧", Type: "tv", Enabled: true}
+	if err := repos.DB.Create(&lib).Error; err != nil {
+		t.Fatal(err)
+	}
+	media := model.Media{
+		LibraryID:    lib.ID,
+		Title:        "折腰",
+		Path:         "cloud://openlist/国产剧/折腰 (2025)/Season 1/折腰.S01E01.mkv",
+		SeasonNum:    1,
+		EpisodeNum:   1,
+		TMDbID:       220269,
+		ScrapeStatus: "pending",
+	}
+	if err := repos.DB.Create(&media).Error; err != nil {
+		t.Fatal(err)
+	}
+
+	if err := scraper.EnrichOne(t.Context(), &media); err != nil {
+		t.Fatal(err)
+	}
+	var got model.Media
+	if err := repos.DB.First(&got, "id = ?", media.ID).Error; err != nil {
+		t.Fatal(err)
+	}
+	if got.ScrapeStatus != "matched" || got.TMDbID != 296753 || got.Title != "折腰" || got.PosterURL == "" {
+		t.Fatalf("stale tmdb id should be rejected and repaired by title search: status=%q tmdb=%d title=%q poster=%q requests=%v",
+			got.ScrapeStatus, got.TMDbID, got.Title, got.PosterURL, requested)
+	}
+	if firstIndexFunc(requested, func(path string) bool { return path == "/tv/220269" }) < 0 {
+		t.Fatalf("test did not exercise stale id lookup: requests=%v", requested)
+	}
+	if firstIndexFunc(requested, func(path string) bool { return path == "/search/tv" }) < 0 {
+		t.Fatalf("scraper did not fall back to title search after stale id rejection: requests=%v", requested)
+	}
+}
+
+func TestMediaYearHintUsesSeriesFolderYearForEpisodes(t *testing.T) {
+	media := &model.Media{
+		Year:       2026,
+		Path:       "cloud://openlist/国产剧/折腰 (2025)/Season 1/折腰.S01E01.mkv",
+		SeasonNum:  1,
+		EpisodeNum: 1,
+	}
+	if got := mediaYearHint(media); got != 2025 {
+		t.Fatalf("mediaYearHint = %d, want series folder year 2025", got)
+	}
+}
+
+func TestMediaYearHintIgnoresEpisodeRowYearWithoutSeriesFolderYear(t *testing.T) {
+	media := &model.Media{
+		Year:       2026,
+		Path:       "cloud://openlist/综艺/哈哈哈哈哈/Season 6/哈哈哈哈哈 - S06E01.mkv",
+		SeasonNum:  6,
+		EpisodeNum: 1,
+	}
+	if got := mediaYearHint(media); got != 0 {
+		t.Fatalf("mediaYearHint = %d, want no year when only episode row year is available", got)
+	}
+}
+
 func TestEnrichOneUsesLocalPathExternalIDHints(t *testing.T) {
 	var requested []string
 	upstream := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {