check for doubles

uploader.py

@@ -167,16 +167,19 @@ def fetch_all_books(cfg: CalibreConfig) -> list[dict]:
     while True:
         resp = client._session.get(
             f"{cfg.url}/ajax/listbooks",
-            params={"start": start, "length": page_size, "sort": "title", "order": "asc"},
+            params={"draw": 1, "start": start, "length": page_size, "sort": "title", "order": "asc"},
             timeout=30,
         )
         resp.raise_for_status()
         data = resp.json()
-        rows = data.get("rows", [])
+        # Calibre-Web uses DataTables format: "data"/"recordsTotal", older versions use "rows"/"total_count"
+        rows = data.get("rows") or data.get("data") or []
+        total = data.get("total_count") or data.get("recordsTotal") or data.get("recordsFiltered") or 0
         all_books.extend(rows)
-        if start + page_size >= data.get("total_count", 0):
+        log.info("Books fetched: %d so far (page start=%d, page_size=%d, total=%d)", len(all_books), start, len(rows), total)
+        if not rows or len(all_books) >= total:
             break
-        start += page_size
+        start += len(rows)
     return all_books
 
 
@@ -199,8 +202,10 @@ def find_duplicate_groups(books: list[dict]) -> list[list[dict]]:
     from collections import defaultdict
     groups: dict[str, list[dict]] = defaultdict(list)
     for book in books:
-        words = _normalize_words(book.get("title", ""))
+        title = book.get("title", "")
-        key = " ".join(sorted(words))
+        # Normalise: lowercase, strip punctuation and extra whitespace — no word removal
+        key = re.sub(r"[^\w\s]", " ", title.lower())
+        key = re.sub(r"\s+", " ", key).strip()
         if key:
             groups[key].append(book)
     return sorted(