Add job cancellation to the stats dashboard

- Active jobs on /stats get a cancel button wired to the existing
  download/match cancel endpoints, showing "cancelling..." and an inline
  error when the task already finished.
- Cancelling now sets the status immediately, so a task whose runner died
  in a restart stops showing as "downloading".
- Download streams use a bounded read timeout (10 s connect / 60 s read):
  a stalled socket fails within a minute (previously it could block
  forever), and a task cancelled while stalled is marked cancelled rather
  than error.
- The stats job list reaps stale download/match tasks, so phantom jobs
  never appear on the dashboard.

backend/apps/library/downloads.py

@@ -144,13 +144,25 @@ def run_download_task(task_id):
         )
     except Exception as exc:  # noqa: BLE001 - report background failures
         destination.unlink(missing_ok=True)
-        logger.exception("Download task %s failed", task_id)
-        DownloadTask.objects.filter(id=task_id).update(
-            status=DownloadTask.STATUS_ERROR,
-            error=str(exc),
-            speed=None,
-            updated_at=timezone.now(),
-        )
+        if DownloadTask.objects.filter(id=task_id, cancelled=True).exists():
+            # Cancel was requested while the socket was stalled; the read
+            # timeout is what breaks the worker out of it.
+            DownloadTask.objects.filter(id=task_id).update(
+                status=DownloadTask.STATUS_CANCELLED,
+                progress=0,
+                downloaded=0,
+                speed=None,
+                error="",
+                updated_at=timezone.now(),
+            )
+        else:
+            logger.exception("Download task %s failed", task_id)
+            DownloadTask.objects.filter(id=task_id).update(
+                status=DownloadTask.STATUS_ERROR,
+                error=str(exc),
+                speed=None,
+                updated_at=timezone.now(),
+            )
     finally:
         # Background threads hold their own DB connection; release it so
         # Gunicorn workers do not leak connections when threads finish.