fetch_all_nodes already had a max_items guard (default 2000) but fetch_all_edges had no such safeguard, allowing unbounded memory growth on graphs with large numbers of edges. Add _MAX_EDGES = 5000 constant and mirror the same loop-guard pattern from fetch_all_nodes: cap the result list, emit a warning log, and break pagination once the limit is reached. |
||
|---|---|---|
| .. | ||
| __init__.py | ||
| file_parser.py | ||
| llm_client.py | ||
| logger.py | ||
| retry.py | ||
| zep_paging.py | ||