Fix entity not becoming unavailable on server shutdown

Trigger async_request_refresh() on WebSocket disconnect to restart
the polling loop. Without this, the coordinator's polling stays
stopped and last_update_success is never set to False.

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

custom_components/remote_media_player/media_player.py

@@ -199,6 +199,10 @@ class MediaPlayerCoordinator(DataUpdateCoordinator[dict[str, Any]]):
         # Re-enable polling as fallback
         self.update_interval = timedelta(seconds=self._poll_interval)
         _LOGGER.warning("WebSocket disconnected, falling back to polling")
+        # Trigger an immediate refresh to restart the polling loop.
+        # Without this, the polling loop stays stopped (it was disabled when
+        # WebSocket was active) and the entity never becomes unavailable.
+        self.hass.async_create_task(self.async_request_refresh())
         # Schedule reconnect attempt
         self._schedule_reconnect()