fix: Use hashed request_id as cache key in shared RQ client reclaim_request

The reclaim_request method was using request.unique_key directly as the
cache key, while all other methods use unique_key_to_request_id() to
compute a hashed request_id. This inconsistency caused orphaned cache
entries since cached requests were stored under a different key than
what other methods would look up.

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

Author: vdusek

src/apify/storage_clients/_apify/_request_queue_shared_client.py

@@ -267,9 +267,9 @@ async def reclaim_request(
                     self.metadata.pending_request_count += 1
 
                 # Update the cache
-                cache_key = request.unique_key
+                request_id = unique_key_to_request_id(request.unique_key)
                 self._cache_request(
-                    cache_key,
+                    request_id,
                     processed_request,
                     hydrated_request=request,
                 )