Update app.py

app.py

@@ -1161,7 +1161,64 @@ def crop_container(img):
 # لا يستخدم classify_image ولا analyze_image
 # ═══════════════════════════════════════════════════════════════
 
-def yolo11_fast_classify(img, declared_text):
+# ═══════════════════════════════════════════════════════════════
+# 🧠 مطابقة ذكية: تربط البيان العربي/الإنجليزي بنتيجة YOLO
+# ═══════════════════════════════════════════════════════════════
+GOODS_MATCH_MAP = {
+    "rice":      ["rice","رز","أرز","ارز","sella","basmati","حبوب","grain","long grain"],
+    "sugar":     ["sugar","سكر"],
+    "wheat":     ["wheat","قمح","flour","طحين","دقيق"],
+    "oil":       ["oil","زيت","sunflower","palm","vegetable oil"],
+    "tea":       ["tea","شاي"],
+    "coffee":    ["coffee","قهوة"],
+    "chemicals": ["chemicals","chemical","كيماويات","كيميائي","كيميائيات"],
+    "batteries": ["batteries","battery","بطاريات","بطارية"],
+    "plastic":   ["plastic","بلاستيك","plastics"],
+    "fabric":    ["fabric","textile","قماش","أقمشة","منسوجات","نسيج"],
+    "tires":     ["tires","tyres","إطارات","إطار","tire","tyre"],
+    "shoes":     ["shoes","footwear","أحذية","حذاء","sandals"],
+    "food":      ["food","foodstuff","غذاء","مواد غذائية","أغذية","foods","مواد_غذائية"],
+    "snacks":    ["snacks","وجبات خفيفة","مقرمشات","chips","biscuits","بسكويت"],
+    "metal":     ["metal","steel","معدن","فولاذ","حديد","iron","aluminum","ألومنيوم"],
+    "wood":      ["wood","timber","خشب","أخشاب","lumber"],
+    "paper":     ["paper","ورق","carton","كرتون","cardboard"],
+    "medicine":  ["medicine","دواء","أدوية","pharmaceutical","drug"],
+    "clothing":  ["clothing","clothes","garment","ملابس","ثياب","apparel"],
+    "furniture": ["furniture","أثاث","furniture"],
+    "electronics":["electronics","إلكترونيات","electronic","كهربائيات"],
+    "vehicles":  ["vehicles","مركبات","سيارات","cars","truck","شاحنة"],
+    "weapons":   ["weapons","أسلحة","سلاح","gun","rifle","arms"],
+    "dates":     ["dates","تمر","تمور","palm dates"],
+    "seeds":     ["seeds","بذور","بذر","grain","حبوب"],
+    "salt":      ["salt","ملح"],
+    "spices":    ["spices","توابل","بهارات","spice"],
+}
+
+def smart_goods_match(declared_word, det_names):
+    """
+    مطابقة ذكية بين كلمة من البيان ومجموعة نتائج YOLO
+    تعمل مع العربي والإنجليزي عبر خريطة المرادفات
+    """
+    if not declared_word:
+        return True
+
+    d = declared_word.lower().strip()
+    det_lower = [n.lower() for n in det_names]
+
+    # 1. مطابقة مباشرة
+    for nm in det_lower:
+        if d in nm or nm in d:
+            return True
+
+    # 2. مطابقة عبر الخريطة
+    for category, keywords in GOODS_MATCH_MAP.items():
+        declared_hit = any(k in d for k in keywords)
+        detected_hit = any(k in nm for nm in det_lower for k in keywords)
+        if declared_hit and detected_hit:
+            print(f"✅ Match via map: '{declared_word}' ↔ {det_names} [{category}]")
+            return True
+
+    return False
     """⚡ تصنيف سريع بـ YOLO11x-cls فقط (3-8 ثواني)"""
     if img is None:
         return ("<div style='text-align:center;padding:60px;color:#999;'>"
@@ -1301,13 +1358,12 @@ def yolo11_fast_classify(img, declared_text):
                      'الثقة': f"{conf:.1%}", 'المصدر': source_label})
     df = pd.DataFrame(rows)
 
-    # ═══ مطابقة التصريح ═══
+    # ═══ مطابقة التصريح الذكية: عربي ↔ إنجليزي ═══
     declared = [d.strip().lower() for d in declared_text.split('+') if d.strip()] if declared_text else []
-    det_names = [it.lower() for it, _ in items_list]
+    det_names = [it for it, _ in items_list]
     is_match = True
     if declared:
-        matched = sum(1 for d in declared if any(d in nm for nm in det_names))
-        is_match = matched / max(len(declared), 1) > 0.5
+        is_match = any(smart_goods_match(d, det_names) for d in declared)
 
     top_name, top_conf = items_list[0]
     risk = 0 if is_match else 3
@@ -2268,10 +2324,9 @@ def analyze_image(img, declared_text):
     df = pd.DataFrame(rows)
 
     declared = [d.strip().lower() for d in declared_text.split('+') if d.strip()] if declared_text else []
-    det_names = [it.lower() for it, _ in final_items]
+    det_names = [it for it, _ in final_items]
     if declared:
-        matched = sum(1 for d in declared if any(d in nm for nm in det_names))
-        is_match = matched / max(len(declared), 1) > 0.5
+        is_match = any(smart_goods_match(d, det_names) for d in declared)
     else:
         is_match = random.choice([True, True, True, False])