chore(docs): document ScoreExtractor tiling and refactor debug scripts (#563)

scripts/debug/patch_tracker.py

@@ -0,0 +1,80 @@
+import sys
+
+with open('youtube_tab_to_pdf.py', 'r', encoding='utf-8') as f:
+    lines = f.readlines()
+
+new_lines = []
+skip = False
+
+import_added = False
+
+for line in lines:
+    if line.startswith('import cv2') and not import_added:
+        new_lines.append(line)
+        new_lines.append('from video_cv_tracker import TemporalTracker\n')
+        import_added = True
+        continue
+        
+    if line.startswith('def extract_unique_scroll(frames:'):
+        skip = True
+        new_lines.append('''def extract_unique_scroll(frames: List[np.ndarray], threshold: float = SIMILARITY_THRESHOLD) -> List[np.ndarray]:
+    print(f"[4/5] 스크롤형 Tab 시계열 추적 추출 중...")
+    
+    strip_tops, strip_bottoms = [], []
+    for frame in frames[:50]:
+        strip = _find_white_tab_strip(frame)
+        if strip:
+            strip_tops.append(strip[0])
+            strip_bottoms.append(strip[1])
+            
+    if not strip_tops:
+        return []
+        
+    median_top = int(np.median(strip_tops))
+    median_bottom = int(np.median(strip_bottoms))
+    
+    tracker = TemporalTracker()
+    
+    for frame in frames:
+        h = frame.shape[0]
+        tab_crop = frame[max(0, median_top):min(h, median_bottom), :]
+        if not _has_tab_content(tab_crop): 
+            continue
+        tracker.process_frame(tab_crop)
+
+    panorama = tracker.get_final_panorama()
+    if panorama is None:
+        return []
+        
+    print(f"  -> 생성된 파노라마 길이: {panorama.shape[1]}px")
+    
+    chunk_width = 1280
+    final_chunks = []
+    
+    w = panorama.shape[1]
+    start_x = 0
+    
+    while start_x < w:
+        chunk = panorama[:, start_x:min(w, start_x + chunk_width)]
+        if chunk.shape[1] < chunk_width:
+            pad = np.full((chunk.shape[0], chunk_width - chunk.shape[1], 3), 255, dtype=np.uint8)
+            chunk = np.hstack([chunk, pad])
+        final_chunks.append(chunk)
+        start_x += chunk_width
+        
+    print(f"  -> A4 분할 컷: {len(final_chunks)}개")
+    return final_chunks
+
+''')
+        continue
+        
+    if skip and line.startswith('def extract_unique_overlay('):
+        skip = False
+
+    if not skip:
+        new_lines.append(line)
+
+with open('youtube_tab_to_pdf.py', 'w', encoding='utf-8') as f:
+    f.writelines(new_lines)
+    
+print("Patched youtube_tab_to_pdf.py successfully.")