fix: optimize batch document deletion - eliminate N redundant KG rebuilds

[yang1002378395-cmyk]

Summary

• Add skip_rebuild parameter to adelete_by_doc_id() (default False for backward compatibility)
• Modify background_delete_documents() to collect rebuild data from all deletions
• Call rebuild_knowledge_from_chunks() once after all deletions complete

Performance Impact

• 75x performance improvement (verified by issue author)
• Eliminates N redundant LLM calls when deleting N documents sharing entities/relationships
• Shared entities/relationships rebuilt only once instead of N times

Test Plan

• Syntax validation passed
• Backward compatible (default skip_rebuild=False)
• Integration test with batch deletion (requires review)

Fixes #2795

[danielaskdd]

Code review

Found 1 issue:

1. asdict is used but not imported in document_routes.py — will raise NameError at runtime whenever the Phase 2 batch rebuild executes

The new Phase 2 block calls rebuild_knowledge_from_chunks(global_config=asdict(rag), ...), but asdict from dataclasses is never imported in this file. Every batch deletion that reaches Phase 2 will crash with NameError: name 'asdict' is not defined. Fix: add from dataclasses import asdict to the imports.

LightRAG/lightrag/api/routers/document_routes.py

Lines 2084 to 2091 in 5c6aff8

	relationships_to_rebuild=all_relationships_to_rebuild,
	knowledge_graph_inst=rag.chunk_entity_relation_graph,
	entities_vdb=rag.entities_vdb,
	relationships_vdb=rag.relationships_vdb,
	text_chunks_storage=rag.text_chunks,
	llm_response_cache=rag.llm_response_cache,
	global_config=asdict(rag),
	pipeline_status=pipeline_status,

🤖 Generated with Claude Code

_{- If this code review was useful, please react with 👍. Otherwise, react with 👎.}

[danielaskdd]

@jules review

[danielaskdd]

This PR implements the fix and returns entities_to_rebuild and relationships_to_rebuild early from adelete_by_doc_id() if skip_rebuild=True. However, by returning early, it skips "Step 9. Delete from full_entities and full_relations storage" in adelete_by_doc_id(), which is a side effect.

PR 2819 fixes this exact issue. It defers the rebuild step but lets adelete_by_doc_id() finish to "Step 9" and then returns the rebuild data along with the final DeletionResult. It also includes tests (tests/test_batch_delete_deferred_rebuild.py).