• Add null checks for envFrom fields
• Support both secrets and configmaps
• Build envFrom list dynamically
• Only render envFrom when entries exist
• Fix template indentation issues
- Changed create_entity to use rag.acreate_entity() instead of direct graph manipulation
- Changed create_relation to use rag.acreate_relation() instead of direct graph manipulation
- This ensures vector embeddings are created and entities/relations are searchable
- Adds proper concurrency locks and metadata population
- Added optional page tracking fields (start_page, end_page, pages) to TextChunkSchema.
- Updated LightRAG class to handle page metadata during document processing.
- Implemented validation for LLM responses to ensure only valid reference IDs are used.
- Enhanced chunking functions to include page data for better context management.
- Improved reference generation to include page ranges for citations.
- Added PDF extraction methods to capture page-level data using PyPDF2 and Docling.
Added three new REST API endpoints for direct knowledge graph manipulation:
- POST /graph/entity/create: Create new entities in the knowledge graph
- POST /graph/relation/create: Create relationships between entities
- POST /graph/entities/merge: Merge duplicate/misspelled entities while preserving relationships
The merge endpoint is particularly useful for consolidating entities discovered after document processing, fixing spelling errors, and cleaning up the knowledge graph. All relationships from source entities are transferred to the target entity, with intelligent handling of duplicate relationships.
Updated API documentation in lightrag/api/README.md with usage examples for all three endpoints.
Prepared statement caching is disabled by setting
`statement_cache_size=0` in the `asyncpg` connection pool parameters.
This is necessary to prevent
`asyncpg.exceptions.InvalidSQLStatementNameError` when using
transaction-level connection poolers like Supabase Supavisor or
pgbouncer, which do not support prepared statements.
- Store file_path in full_docs storage
- Update PostgreSQL implementation by map file_path to doc_name
- Other storage implementation automatically handles the new field
• Add track ID to tooltip display
• Remove JSON braces from metadata
• Reorder tooltip content layout
• Clean up metadata indentation
• Show track ID before metadata
