fix(ocr-pipeline): re-populate row.words for word-lookup in Step 5

The row_result stored in DB excludes words to keep payload small.
When Step 5 reconstructs RowGeometry from DB, words were empty,
causing word-lookup to find nothing and return blank cells.

Now re-populates row.words from cached _word_dicts (or re-runs
detect_column_geometry if cache is cold) before cell grid building.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

klausur-service/backend/ocr_pipeline_api.py

@@ -1084,6 +1084,36 @@ async def detect_words(
         for r in row_result["rows"]
     ]
 
+    # Re-populate row.words from cached full-page Tesseract words.
+    # Word-lookup in _ocr_single_cell needs these to avoid re-running OCR.
+    word_dicts = cached.get("_word_dicts")
+    if word_dicts is None:
+        ocr_img_tmp = create_ocr_image(dewarped_bgr)
+        geo_result = detect_column_geometry(ocr_img_tmp, dewarped_bgr)
+        if geo_result is not None:
+            _geoms, left_x, right_x, top_y, bottom_y, word_dicts, inv = geo_result
+            cached["_word_dicts"] = word_dicts
+            cached["_inv"] = inv
+            cached["_content_bounds"] = (left_x, right_x, top_y, bottom_y)
+
+    if word_dicts:
+        # words['top'] is relative to content-ROI top_y.
+        # row.y is absolute. Convert: row_y_rel = row.y - top_y.
+        content_bounds = cached.get("_content_bounds")
+        if content_bounds:
+            _lx, _rx, top_y, _by = content_bounds
+        else:
+            top_y = min(r.y for r in row_geoms) if row_geoms else 0
+
+        for row in row_geoms:
+            row_y_rel = row.y - top_y
+            row_bottom_rel = row_y_rel + row.height
+            row.words = [
+                w for w in word_dicts
+                if row_y_rel <= w['top'] + w['height'] / 2 < row_bottom_rel
+            ]
+            row.word_count = len(row.words)
+
     if stream:
         return StreamingResponse(
             _word_stream_generator(