Update utils/bubble_utils.py

utils/bubble_utils.py

@@ -10,7 +10,7 @@ from utils.polygon_utils import (
     draw_translated_text_convex,
     shrink_or_expand_polygon,
     FONT_PATH,
-    correct_ocr_polygons_with_bubbles,
+    match_polygon_to_bubble_by_overlap,
     render_translated_chunk,
 )
 from utils.bubble_detect import detect_speech_bubbles_robust
@@ -70,6 +70,21 @@ def extract_and_translate_with_masks(full_img, interior_polygons):
 
     return translations
 
+def match_translations_to_bubbles(translations, bubble_polygons, min_overlap=0.10):
+    """
+    Add matched_bubble_idx to each OCR translation.
+    OCR polygon is NOT corrected.
+    """
+    for t in translations:
+        ocr_poly = t.get("polygon")
+        if not ocr_poly:
+            continue
+
+        idx = match_polygon_to_bubble_by_overlap(ocr_poly, bubble_polygons, min_overlap)
+        t["matched_bubble_idx"] = idx
+
+    return translations
+
 # ======================== Debug Visualization ============================
 
 def visualize_all_debug(img, translations, bubble_polygons, step_name="debug", prefix="debug"):
@@ -164,9 +179,8 @@ def bubble_pipeline_single(file_obj, num_chunks=1, polygon_strategy="hybrid", de
     # 4. Match OCR polygons to bubble polygons
     # -------------------------------------------------------
     if bubble_polygons:
-        translations = correct_ocr_polygons_with_bubbles(
-            translations, bubble_polygons, strategy=polygon_strategy
-        )
+        print("🔄 Matching OCR → Bubble regions...")
+        translations = match_translations_to_bubbles(translations, bubble_polygons)
     else:
         print("⚠️ No bubble polygons → skip polygon correction")