Update scoring_calculation_system.py

scoring_calculation_system.py

@@ -1317,82 +1317,107 @@ def calculate_compatibility_score(breed_info: dict, user_prefs: UserPreferences)
         }
 
         # 2. 優化權重配置 - 根據使用者情況動態調整權重
-        base_weights = {
-            'space': 0.28,      
-            'exercise': 0.18,   
-            'grooming': 0.12,   
-            'experience': 0.22, 
-            'health': 0.12,     
-            'noise': 0.08      
-        }
-
-        # 根據特殊情況調整權重
-        weights = base_weights.copy()
-        
-        # 有孩童時的權重調整
-        if user_prefs.has_children:
-            if user_prefs.children_age == 'toddler':
-                weights['experience'] *= 1.4  # 幼童需要更有經驗的配對
-                weights['noise'] *= 1.3       # 噪音影響更重要
-                weights['health'] *= 1.2      # 健康因素更關鍵
-            elif user_prefs.children_age == 'school_age':
-                weights['experience'] *= 1.2
-                weights['noise'] *= 1.2
+        def get_adjusted_weights(user_prefs, scores):
+            """
+            根據使用者條件動態調整各項評分的權重。
+            確保權重調整合理，避免過度偏移。
+            """
+            weights = {
+                'space': 0.28,      
+                'exercise': 0.18,   
+                'grooming': 0.12,   
+                'experience': 0.22, 
+                'health': 0.12,     
+                'noise': 0.08      
+            }
             
-        # 居住環境的權重調整
-        if user_prefs.living_space == 'apartment':
-            weights['space'] *= 1.3           # 空間限制更重要
-            weights['noise'] *= 1.2           # 噪音影響更顯著
-        
-        # 重新正規化權重
-        total_weight = sum(weights.values())
-        weights = {k: v/total_weight for k, v in weights.items()}
+            # 公寓住戶需要更注重空間和噪音
+            if user_prefs.living_space == 'apartment':
+                if scores['space'] < 0.6:  # 空間評分不理想時更重視
+                    weights['space'] *= 1.25
+                    weights['noise'] *= 1.15
+                
+            # 新手飼主需要更注重經驗要求
+            if user_prefs.experience_level == 'beginner':
+                if scores['experience'] < 0.5:  # 經驗需求較高時更重視
+                    weights['experience'] *= 1.3
+            
+            # 有孩童時的特殊考量
+            if user_prefs.has_children:
+                child_age_weights = {
+                    'toddler': {'experience': 1.3, 'health': 1.2, 'noise': 1.2},
+                    'school_age': {'experience': 1.2, 'health': 1.1, 'noise': 1.1},
+                    'teenager': {'experience': 1.1, 'health': 1.05, 'noise': 1.05}
+                }
+                
+                age_adjustments = child_age_weights.get(user_prefs.children_age, 
+                                                      child_age_weights['school_age'])
+                for key, mult in age_adjustments.items():
+                    weights[key] *= mult
 
-        # 3. 計算加權總分
-        weighted_score = sum(score * weights[category] for category, score in scores.items())
+            # 重新正規化權重總和為1
+            total = sum(weights.values())
+            return {k: v/total for k, v in weights.items()}
+
+        # 3. 套用調整後的權重
+        weights = get_adjusted_weights(user_prefs, scores)
         
-        # 4. 改進的分數放大函數
-        def amplify_score(score):
+        # 4. 計算初始加權分數
+        weighted_score = sum(score * weights[category] for category, score in scores.items())
+
+        # 5. 改進的分數放大函數
+        def amplify_score(raw_score, scores, user_prefs):
             """
-            改進的分數放大函數，提供更大的分數差異。
-            
-            主要改進：
-            1. 調整基礎計算參數以產生更明顯的差異
-            2. 根據分數區間使用不同的放大策略
-            3. 擴大最終分數範圍
+            優化的分數放大函數，確保分數分布合理且有意義。
+            考慮原始分數的分布和關鍵條件的影響。
             """
-            # 基礎調整，擴大差異
-            adjusted = (score - 0.4) * 2.0  # 從0.35調整到0.4，倍數從1.8提高到2.0
+            # 基礎轉換：將分數範圍從[0.1-1.0]映射到[0-1]
+            normalized = (raw_score - 0.1) / 0.9
             
-            # 使用更陡峭的曲線，從3.2降到2.8使曲線不會過於極端
-            amplified = pow(adjusted, 2.8) / 4.5 + score
+            # 使用S型曲線轉換，使中間範圍的差異更明顯
+            transformed = 1 / (1 + math.exp(-6 * (normalized - 0.5)))
             
-            # 分數區間處理
-            if score < 0.65:  # 低分區間
-                amplified *= 0.85  # 進一步降低不適合的配對
-            elif score > 0.85:  # 高分區間
-                amplified = 0.85 + (amplified - 0.85) * 0.4  # 高分區間的緩和壓縮
+            # 將分數映射回目標範圍[0.6-0.95]
+            score = 0.6 + transformed * 0.35
             
-            # 擴大分數範圍到0.45-0.95，原本是0.55-0.95
-            final_score = max(0.45, min(0.95, amplified))
+            # 關鍵條件檢查和調整
+            critical_conditions = [
+                (scores['experience'] < 0.4 and user_prefs.has_children, 0.85),
+                (scores['noise'] < 0.3 and user_prefs.living_space == 'apartment', 0.85),
+                (scores['health'] < 0.3, 0.9),
+                (scores['space'] < 0.3 and user_prefs.living_space == 'apartment', 0.85)
+            ]
             
-            # 四捨五入到小數點後第三位
-            return round(final_score, 3)
+            # 應用關鍵條件調整
+            for condition, factor in critical_conditions:
+                if condition:
+                    score *= factor
+            
+            return round(max(0.6, min(0.95, score)), 4)
 
-        # 5. 計算最終分數並應用關鍵條件檢查
-        final_score = amplify_score(weighted_score)
+        # 6. 計算最終分數
+        final_score = amplify_score(weighted_score, scores, user_prefs)
         
-        # 針對特殊情況的最終調整
-        if user_prefs.has_children and scores['experience'] < 0.4:
-            final_score *= 0.8  # 有孩童但經驗分數過低時大幅降低
-        if user_prefs.living_space == 'apartment' and scores['noise'] < 0.3:
-            final_score *= 0.75  # 住公寓但噪音分數極低時顯著降低
+        # 7. 確保分數差異性
+        def adjust_final_scores(base_scores, final_score):
+            """
+            確保最終分數能反映品種間的實際差異。
+            """
+            scores = base_scores.copy()
+            scores['overall'] = final_score
+            
+            # 檢查是否有極端低分項目
+            min_score = min(v for k, v in base_scores.items())
+            if min_score < 0.4:  # 如果有特別低的分數
+                scores['overall'] *= 0.92  # 適度降低整體評分
+            
+            return scores
 
-        # 6. 準備返回結果
-        scores = {k: round(v, 4) for k, v in scores.items()}
-        scores['overall'] = round(final_score, 4)
+        # 8. 準備最終結果
+        final_scores = adjust_final_scores(scores, final_score)
         
-        return scores
+        # 四捨五入所有分數
+        return {k: round(v, 4) for k, v in final_scores.items()}
 
     except Exception as e:
         print(f"Error details: {str(e)}")