check for doubles

This commit is contained in:
2026-05-12 10:16:55 +02:00
parent 09afe40f04
commit c69b4c392a
2 changed files with 23 additions and 8 deletions
+4 -2
View File
@@ -13,7 +13,7 @@ import db
import sftp as sftp_module
import sync
import uploader
from uploader import delete_book, fetch_all_books, find_duplicate_groups
from uploader import CalibreClient, delete_book, fetch_all_books, find_duplicate_groups
logging.basicConfig(level=logging.INFO, format="%(asctime)s %(levelname)s %(name)s%(message)s")
log = logging.getLogger(__name__)
@@ -222,13 +222,15 @@ def _run_dedup():
try:
cfg = config.load()
log.info("Dedup: fetching all books ...")
client = CalibreClient(cfg.calibre)
client._ensure_auth()
books = fetch_all_books(cfg.calibre)
groups = find_duplicate_groups(books)
to_delete = [b for group in groups for b in sorted(group, key=lambda x: x.get("id", 0))[1:]]
_dedup_state.update({"total": len(to_delete), "deleted": 0, "failed": 0})
log.info("Dedup: %d duplicate(s) to delete across %d group(s)", len(to_delete), len(groups))
for book in to_delete:
ok, msg = delete_book(cfg.calibre, book["id"])
ok, msg = delete_book(cfg.calibre, book["id"], client)
if ok:
_dedup_state["deleted"] += 1
else:
+19 -6
View File
@@ -174,7 +174,10 @@ def fetch_all_books(cfg: CalibreConfig) -> list[dict]:
data = resp.json()
# 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") or data.get("totalNotFiltered") or 0
total = (
data.get("recordsTotal") or data.get("total_count") or
data.get("total") or data.get("totalNotFiltered") or 0
)
all_books.extend(rows)
log.info("Books fetched: %d / %d", len(all_books), total)
if not rows or len(all_books) >= total:
@@ -183,13 +186,23 @@ def fetch_all_books(cfg: CalibreConfig) -> list[dict]:
return all_books
def delete_book(cfg: CalibreConfig, book_id: int) -> tuple[bool, str]:
"""Delete a book from Calibre-Web by ID."""
client = CalibreClient(cfg)
client._ensure_auth()
def delete_book(cfg: CalibreConfig, book_id: int, client: "CalibreClient | None" = None) -> tuple[bool, str]:
"""Delete a book from Calibre-Web by ID. Pass a pre-authenticated client to avoid re-auth overhead."""
if client is None:
client = CalibreClient(cfg)
client._ensure_auth()
csrf = client._upload_csrf
if not csrf:
# Try to fetch a CSRF token from the book detail page
try:
page = client._session.get(f"{cfg.url}/book/{book_id}", timeout=15)
csrf = _extract_csrf(page.text)
client._upload_csrf = csrf
except Exception:
pass
resp = client._session.post(
f"{cfg.url}/delete/{book_id}",
data={"csrf_token": client._upload_csrf} if client._upload_csrf else {},
data={"csrf_token": csrf} if csrf else {},
timeout=30,
)
if resp.ok: