- Refactor the trigger condition for LLM-based summarization of entities and relations. Instead of relying on character length, the summary is now triggered when the number of merged description fragments exceeds a configured threshold. This provides a more robust and logical condition for consolidation.
- Introduce the `OLLAMA_NUM_CTX` environment variable to explicitly configure the context window size (`num_ctx`) for Ollama models. This decouples the model's context length from the `MAX_TOKENS` parameter, which is now specifically used to limit input for summary generation, making the configuration clearer and more flexible.
- Updated `README` files, `env.example`, and default values to reflect these changes.
Move knowledge_graph_inst.has_node check inside get_storage_keyed_lock
in _merge_edges_then_upsert to ensure atomic check-then-act operations
and prevent duplicate node creation during concurrent updates.
Refactored the `KeyedUnifiedLock` to be generic and support dynamic namespaces. This decouples the locking mechanism from a specific "GraphDB" implementation, allowing it to be reused across different components and workspaces safely.
Key changes:
- `KeyedUnifiedLock` now takes a `namespace` parameter on lock acquisition.
- Renamed `_graph_db_lock_keyed` to a more generic _storage_keyed_lock`
- Replaced `get_graph_db_lock_keyed` with get_storage_keyed_lock` to support namespaces
Added validation checks in entity and relationship extraction functions to filter out entities that become empty strings after normalize_extracted_info processing. This prevents empty labels from appearing in get_all_labels() results and maintains knowledge graph data integrity.
Refactored the LLM cache to a flat Key-Value (KV) structure, replacing the previous nested format. The old structure used the 'mode' as a key and stored specific cache content as JSON nested under it. This change significantly enhances cache recall efficiency.
- Logs detailed progress, including warnings and failures, to the pipeline status.
- Adds counters to report the total number of successfully rebuilt entities and relationships upon completion.
- To enhance performance during document deletion, new batch-get methods, `get_nodes_by_chunk_ids` and `get_edges_by_chunk_ids`, have been added to the graph storage layer (`BaseGraphStorage` and its implementations). The [`adelete_by_doc_id`](lightrag/lightrag.py:1681) function now leverages these methods to avoid unnecessary iteration over the entire knowledge graph, significantly improving efficiency.
- Graph storage updated: Networkx, Neo4j, Postgres AGE
- Add validation in _find_most_related_edges_from_entities and _get_edge_data function during edge data construction
- Add warning logs when 'weight' attribute is missing and set default value of 0.0
