Refined report structure: Ensure mutually exclusive separation of 'short_alt' vs 'poor_quality' (Generic/Filename) to correct double counting

src/analyzer.py

@@ -89,18 +89,44 @@ class ImageAnalyzer:
                     poor_quality_images.append({'src': img['src'], 'alt': alt_text, 'reason': reason})
                     total_poor_quality += 1
             
-            # Separate short/sparse items for backward compatibility
-            short_alt_images_legacy = [img for img in poor_quality_images if img['reason'].startswith("Too short") or img['reason'].startswith("Too few words")]
-            total_short_alt_legacy += len(short_alt_images_legacy)
+            # Split into mutually exclusive lists based on reason
+            # User requirement: "short_alt_count are not show sepratly bescue it show in poor_quality_count... remove and add in poor perfectly"
+            # We interpret this as requiring two distinct, non-overlapping sets.
             
+            non_overlapping_poor = []
+            non_overlapping_short = []
+            
+            for img in poor_quality_images:
+                if img['reason'].startswith("Too short") or img['reason'].startswith("Too few words"):
+                    non_overlapping_short.append(img)
+                else:
+                    non_overlapping_poor.append(img)
+
+            # Update legacy counters (only adding strictly short ones to the legacy global count)
+            # But wait, should true "poor" ones be tracked globally too? 
+            # The summary returned `total_poor_quality` which was everything. 
+            # If we separate them, we should probably track them separately in summary too if possible, 
+            # but for now we update the page level reporting as requested.
+            
+            # Update global counters strictly based on the split
+            total_short_alt_legacy += len(non_overlapping_short)
+            
+            # The 'total_poor_quality' global in summary currently counts EVERYTHING.
+            # If we want to maintain that consistency, we keep it as processed in loop.
+            # BUT user said "remove and add in poor perfectly" -> implies separation.
+            # Let's subtract the short ones from the global poor quality count for the SUMMARY as well?
+            # actually, `total_poor_quality` was just `+=1` in the loop for every hit. 
+            # Let's adjust it to only count the non-overlapping poor ones.
+            total_poor_quality -= len(non_overlapping_short)
+
             pages_report.append({
                 "page_url": page_url,
                 "missing_alt_count": len(missing_images),
-                "poor_quality_count": len(poor_quality_images), # Total count of all poor quality
-                "short_alt_count": len(short_alt_images_legacy), # Count only strictly short ones
+                "poor_quality_count": len(non_overlapping_poor), # Mutually exclusive count
+                "short_alt_count": len(non_overlapping_short),   # Mutually exclusive count
                 "images_without_alt": list(set(missing_images)),
-                "images_with_poor_alt": poor_quality_images, # Contains ALL poor images
-                "images_with_short_alt": short_alt_images_legacy # Contains ONLY short/sparse images
+                "images_with_poor_alt": non_overlapping_poor,
+                "images_with_short_alt": non_overlapping_short
             })
 
         return {