Update app.py

app.py

@@ -295,9 +295,6 @@ def analyze_outfit(input_img):
         
         drip_len = len(style_prompts['drippy'])
         mid_len = len(style_prompts['mid'])
-        drip_score = np.mean(all_probs[0 : drip_len])
-        mid_score = np.mean(all_probs[drip_len : drip_len + mid_len])
-        not_score = np.mean(all_probs[drip_len + mid_len : style_prompts_end_index])
 
         # Determine overall style category AND DEFINE score_label
         score_label = "Style Score" # Initialize with a default/fallback
@@ -334,9 +331,20 @@ def analyze_outfit(input_img):
         )
         
         # Clip the score between 0 and 100 for safety
+        # Weighted score formula:
+        # - Drip is worth full points (x1)
+        # - Mid is worth 2/3 points (~66%)
+        # - Trash is worth 1/3 points (~33%)
+        
+        final_score = (drip_score * 1.0) + (mid_score * 0.66) + (not_score * 0.33)
+        
+        # Now map it to a 0–100 scale
+        final_score = final_score * 100
+        
+        # Clip it just in case (though it shouldn't exceed 100 now)
         final_score = min(max(final_score, 0), 100)
         
-        # Assign category based on final_score ranges
+        # Now classify the category based on the final_score
         if final_score >= 75:
             category_key = 'drippy'
             score_label = "Drip Score"
@@ -351,6 +359,7 @@ def analyze_outfit(input_img):
         percentage_score_str = f"{final_score:.0f}%"
 
 
+
       
         print(f"Style analysis: Category={category_label}, Score = {score_label}={percentage_score_str} (Raw Score: {final_score:.4f})")
 
@@ -417,7 +426,7 @@ def analyze_outfit(input_img):
         category_html = f"""
             <div class='results-container'>
                 <h2 class='result-category'>RATING: {category_label.upper()}</h2>
-                <p class='result-score'>{score_label}: {percentage_score}</p>
+                <p class='result-score'>{score_label}: {percentage_score_str}</p>
             </div>
         """
         return category_html, tts_path, response_text