Big changes

server/internal/recommend/profile.go

@@ -10,7 +10,9 @@ import (
 	"encoding/json"
 	"math"
 	"sort"
+	"strconv"
 	"strings"
+	"unicode"
 )
 
 // recencyDecay is applied per position down the history list. At 0.94, the 12th item
@@ -26,14 +28,23 @@ const favoriteWeight = 0.6
 // Item is the slice of an Emby item this package reasons about. The raw payload rides
 // along so rows can be emitted without re-fetching or re-encoding.
 type Item struct {
-	ID              string   `json:"Id"`
-	Name            string   `json:"Name"`
-	Type            string   `json:"Type"`
-	SeriesID        string   `json:"SeriesId"`
-	SeriesName      string   `json:"SeriesName"`
-	Genres          []string `json:"Genres"`
-	CommunityRating float64  `json:"CommunityRating"`
-	Studios         []struct {
+	ID                string   `json:"Id"`
+	Name              string   `json:"Name"`
+	Type              string   `json:"Type"`
+	SeriesID          string   `json:"SeriesId"`
+	SeriesName        string   `json:"SeriesName"`
+	ProductionYear    int      `json:"ProductionYear"`
+	Genres            []string `json:"Genres"`
+	CommunityRating   float64  `json:"CommunityRating"`
+	RunTimeTicks      int64    `json:"RunTimeTicks"`
+	IndexNumber       int      `json:"IndexNumber"`
+	ParentIndexNumber int      `json:"ParentIndexNumber"`
+	Container         string   `json:"Container"`
+	MediaStreams      []struct {
+		Type  string `json:"Type"`
+		Codec string `json:"Codec"`
+	} `json:"MediaStreams"`
+	Studios []struct {
 		Name string `json:"Name"`
 	} `json:"Studios"`
 	UserData struct {
@@ -46,6 +57,38 @@ type Item struct {
 	Raw json.RawMessage `json:"-"`
 }
 
+func (i Item) RuntimeMinutes() int {
+	if i.RunTimeTicks <= 0 {
+		return 0
+	}
+	return int(i.RunTimeTicks / 600_000_000)
+}
+
+func (i Item) TitleKey() string {
+	value := i.Name
+	if i.Type == "Episode" && strings.TrimSpace(i.SeriesName) != "" {
+		value = i.SeriesName
+	}
+	var b strings.Builder
+	for _, r := range strings.ToLower(value) {
+		if unicode.IsLetter(r) || unicode.IsDigit(r) {
+			b.WriteRune(r)
+		}
+	}
+	return b.String()
+}
+
+// SeenKey is a title-level fallback for imported catalogue records, whose payloads
+// deliberately contain no per-user UserData. Movies include their year so watching an
+// older film does not hide a remake with the same name; episodes collapse to series.
+func (i Item) SeenKey() string {
+	key := i.TitleKey()
+	if key != "" && strings.EqualFold(i.Type, "Movie") && i.ProductionYear > 0 {
+		return key + "|" + strconv.Itoa(i.ProductionYear)
+	}
+	return key
+}
+
 // Seed is a title recent enough to anchor a "Because you watched …" row.
 type Seed struct {
 	ID   string
@@ -58,8 +101,9 @@ type Profile struct {
 	StudioWeights map[string]float64
 	// Seen holds item ids *and* series ids already watched or in progress, so a
 	// recommendation never suggests something the user is already partway through.
-	Seen  map[string]bool
-	Seeds []Seed
+	Seen       map[string]bool
+	SeenTitles map[string]bool
+	Seeds      []Seed
 }
 
 func (p Profile) IsEmpty() bool { return len(p.GenreWeights) == 0 && len(p.Seeds) == 0 }
@@ -87,12 +131,25 @@ func BuildProfile(history, favorites []Item) Profile {
 		GenreWeights:  map[string]float64{},
 		StudioWeights: map[string]float64{},
 		Seen:          map[string]bool{},
+		SeenTitles:    map[string]bool{},
 	}
 
 	seedSeen := map[string]bool{}
+	tasteSeen := map[string]bool{}
 	for i, item := range history {
-		weight := math.Pow(recencyDecay, float64(i))
-		profile.absorb(item, weight)
+		profile.markSeen(item)
+
+		// Several episodes of one series are evidence for one taste, not several
+		// independent tastes. Keep the newest occurrence's recency weight and still
+		// mark every item/series identifier as seen.
+		tasteID := item.ID
+		if item.SeriesID != "" {
+			tasteID = item.SeriesID
+		}
+		if !tasteSeen[tasteID] {
+			tasteSeen[tasteID] = true
+			profile.absorbTaste(item, math.Pow(recencyDecay, float64(i)))
+		}
 
 		// An episode seeds its series, not itself: "Because you watched Severance"
 		// reads better than "Because you watched Good News".
@@ -113,12 +170,25 @@ func BuildProfile(history, favorites []Item) Profile {
 }
 
 func (p *Profile) absorb(item Item, weight float64) {
+	p.markSeen(item)
+	p.absorbTaste(item, weight)
+}
+
+func (p *Profile) markSeen(item Item) {
 	if item.ID != "" {
 		p.Seen[item.ID] = true
 	}
 	if item.SeriesID != "" {
 		p.Seen[item.SeriesID] = true
 	}
+	if key := item.SeenKey(); key != "" {
+		p.SeenTitles[key] = true
+	}
+}
+
+// absorbTaste learns affinity without marking the item watched. This is used for
+// browsing signals: lingering on a card is meaningful, but must not hide that card.
+func (p *Profile) absorbTaste(item Item, weight float64) {
 	for _, genre := range item.Genres {
 		if g := strings.TrimSpace(genre); g != "" {
 			p.GenreWeights[g] += weight
@@ -168,6 +238,9 @@ func (p Profile) Score(candidate Item) float64 {
 	if candidate.SeriesID != "" && p.Seen[candidate.SeriesID] {
 		return -1
 	}
+	if p.SeenTitles[candidate.SeenKey()] {
+		return -1
+	}
 	if candidate.UserData.Played || candidate.UserData.PlaybackPositionTicks > 0 {
 		return -1
 	}