From 0230054f92fc9138b90594f558641a512c99128d Mon Sep 17 00:00:00 2001
From: vndangkhoa <vndangkhoa@users.noreply.github.com>
Date: Fri, 20 Feb 2026 20:15:59 +0700
Subject: [PATCH] Fix missing Episode 1 and duplicate search results
 (Unicode-aware dedup)

---
 backend/go.mod                           |  2 +-
 backend/internal/api/handlers.go         | 31 ++++++++++++++++++++++++
 backend/internal/scraper/ophim.go        |  7 ++++--
 backend/internal/scraper/phimmoichill.go | 11 +++++----
 4 files changed, 43 insertions(+), 8 deletions(-)

diff --git a/backend/go.mod b/backend/go.mod
index b48e243..fc82cc7 100644
--- a/backend/go.mod
+++ b/backend/go.mod
@@ -8,6 +8,7 @@ require (
 	github.com/go-chi/chi/v5 v5.2.4
 	github.com/go-chi/cors v1.2.2
 	golang.org/x/image v0.35.0
+	golang.org/x/text v0.33.0
 	gorm.io/gorm v1.31.1
 )
 
@@ -22,7 +23,6 @@ require (
 	github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec // indirect
 	golang.org/x/net v0.49.0 // indirect
 	golang.org/x/sys v0.40.0 // indirect
-	golang.org/x/text v0.33.0 // indirect
 	modernc.org/libc v1.22.5 // indirect
 	modernc.org/mathutil v1.5.0 // indirect
 	modernc.org/memory v1.5.0 // indirect
diff --git a/backend/internal/api/handlers.go b/backend/internal/api/handlers.go
index b8afe24..fd43ff1 100644
--- a/backend/internal/api/handlers.go
+++ b/backend/internal/api/handlers.go
@@ -11,6 +11,7 @@ import (
 	"strconv"
 	"strings"
 	"sync"
+	"unicode"
 
 	"streamflow-backend/internal/database"
 	"streamflow-backend/internal/models"
@@ -18,6 +19,9 @@ import (
 	"streamflow-backend/internal/service"
 
 	"github.com/go-chi/chi/v5"
+	"golang.org/x/text/runes"
+	"golang.org/x/text/transform"
+	"golang.org/x/text/unicode/norm"
 )
 
 const (
@@ -128,6 +132,7 @@ func (h *Handler) fetchAndMergeMovies(fetch movieFetcher) []models.RophimMovie {
 func (h *Handler) mergeMovies(providerResults [][]models.RophimMovie, maxLen int) []models.RophimMovie {
 	var allMovies []models.RophimMovie
 	seenID := make(map[string]int)
+	seenSlug := make(map[string]int)
 	seenTitle := make(map[string]int)
 
 	for i := 0; i < maxLen; i++ {
@@ -135,11 +140,22 @@ func (h *Handler) mergeMovies(providerResults [][]models.RophimMovie, maxLen int
 			if i < len(movies) {
 				movie := movies[i]
 
+				// Check 1: Exact ID match
 				if idx, found := seenID[movie.ID]; found {
 					h.mergeMovieMetadata(&allMovies[idx], &movie)
 					continue
 				}
 
+				// Check 2: Slug match (e.g. "vu-tru-cua-doi-ta" from both providers)
+				slugKey := normalizeKey(movie.Slug)
+				if slugKey != "" {
+					if idx, found := seenSlug[slugKey]; found {
+						h.mergeMovieMetadata(&allMovies[idx], &movie)
+						continue
+					}
+				}
+
+				// Check 3: Normalized title match
 				titleKey := normalizeKey(movie.OriginalTitle)
 				if titleKey == "" {
 					titleKey = normalizeKey(movie.Title)
@@ -152,6 +168,9 @@ func (h *Handler) mergeMovies(providerResults [][]models.RophimMovie, maxLen int
 				allMovies = append(allMovies, movie)
 				currIdx := len(allMovies) - 1
 				seenID[movie.ID] = currIdx
+				if slugKey != "" {
+					seenSlug[slugKey] = currIdx
+				}
 				if titleKey != "" {
 					seenTitle[titleKey] = currIdx
 				}
@@ -418,7 +437,19 @@ func (h *Handler) mergeMovieMetadata(existing, new *models.RophimMovie) {
 }
 
 func normalizeKey(s string) string {
+	if s == "" {
+		return ""
+	}
 	s = strings.ToLower(s)
+	// Strip Vietnamese diacritics: Vũ Trụ Của Đôi Ta → vu tru cua doi ta
+	t := transform.Chain(norm.NFD, runes.Remove(runes.In(unicode.Mn)), norm.NFC)
+	result, _, err := transform.String(t, s)
+	if err == nil {
+		s = result
+	}
+	// Replace đ/Đ which NFD doesn't decompose
+	s = strings.ReplaceAll(s, "đ", "d")
+	// Keep only alphanumeric
 	reg := regexp.MustCompile("[^a-z0-9]+")
 	return reg.ReplaceAllString(s, "")
 }
diff --git a/backend/internal/scraper/ophim.go b/backend/internal/scraper/ophim.go
index 4e0dcac..235f796 100644
--- a/backend/internal/scraper/ophim.go
+++ b/backend/internal/scraper/ophim.go
@@ -303,8 +303,11 @@ func (s *OphimScraper) GetMovieDetail(slug string) (*models.RophimMovie, error)
 				var n int
 				if _, err := fmt.Sscanf(ep.Name, "Tap %d", &n); err == nil {
 					epNum = n
-				} else {
-					epNum = 1
+				}
+				// If still 0 (e.g. "Full", "Trailer"), skip — don't default to 1
+				// as that would collide with real Episode 1 during dedup
+				if epNum == 0 {
+					continue
 				}
 			}
 
diff --git a/backend/internal/scraper/phimmoichill.go b/backend/internal/scraper/phimmoichill.go
index a5c72b2..2626821 100644
--- a/backend/internal/scraper/phimmoichill.go
+++ b/backend/internal/scraper/phimmoichill.go
@@ -204,14 +204,15 @@ func (s *PhimMoiChillScraper) GetMovieDetail(slug string) (*models.RophimMovie,
 
 		epNum := 0
 		if strings.EqualFold(epName, "Full") {
-			epNum = 1
-		} else {
-			// Try "Tập 1", "Tập 2"
-			fmt.Sscanf(epName, "Tập %d", &epNum)
+			// Single-movie "Full" — will be handled by the fallback below
+			// Don't assign epNum=1 as it collides with real Episode 1 in series
+			return
 		}
+		// Try "Tập 1", "Tập 2"
+		fmt.Sscanf(epName, "Tập %d", &epNum)
 
 		if epNum == 0 {
-			// Try to extract from title if current text is just "Tap X"
+			// Try plain number
 			fmt.Sscanf(epName, "%d", &epNum)
 		}