v1.3.2

What's New

Key improvements: Enhanced document indexing speed. Significantly improved the query performance of Neo4j and PostgreSQL AGE graph storage.

• Node/Edge name and description can be edited with WebUi by @choizhang in #1364
• Improve Neo4j performance by batching with UNWIND by @frederikhendrix in #1341
• Improve PostgreSQL AGE performance by @danielaskdd in #1383
• Fix node color by node type by @danielaskdd in #1308
• Add /context query prefix to Ollama model simulation for LightRAG Server by @Exploding-Soda in #1305
• Add ENABLE_LLM_CACHE env fix mix_kg_vector_query return value error when only_need_context is enabled by @danielaskdd in #1318
• Add FORCE_LLM_SUMMARY_ON_MERGE env by @danielaskdd in #1334
• Add OPENAI_API_BASE env by @QodiCat in #1330
• Add middle-content matching for graph label search by @danielaskdd in #1322
• Only merge new entities/edges during gleaning stage by @danielaskdd in #1337
• Add streaming response support for query test in WebUI by @choizhang in #1331
• Add bypass query mode by @choizhang in #1338
• Optimize: Use python-docx for better parsing. by @zhudongwork in #1345
• Locales: add traditional Chinese language support by @iLuJack in #1359
• Cancel pending tasks when any chunk processing fails by @danielaskdd

Bug fixed

• Fix mix_kg_vector_query return value error when only_need_context is enabled by @danielaskdd in #1318
• Fixed null value handling and ensure exceptions are avoided by @IcySugar000 in #1315
• Fix document list sorting problem while filter is off by @danielaskdd in #1326
• Fix: LLM cache now work for nodes and edges merging by @danielaskdd in #1328
• Fixes a bug where file_path was not present dictionary by @FeHuynhVI in #1304
• Fix error when running caused by global_args.timeout being None by @HyuPete in #1333
• Fix grammar and typo in LightRAG Server REAEME.md by @earayu in #1329
• Fix history turns not working for query test in WebUI by @choizhang in #1332
• Fix the issue where PostgresSQL AGE cannot handle content containing backslashes by @danielaskdd in #1360
• Add graph_db_lock to ensure consistency across multiple processes for node and edge edition jobs by @danielaskdd in #1367
• Fix: LLM cache not saving anymore while enable_llm_cache is disabled by @danielaskdd in #1377
• Fix PostgreSQL subquery in FROM missing alias by @superhuahua in #1375

New Contributors

• @IcySugar000 made their first contribution in #1298
• @Exploding-Soda made their first contribution in #1305
• @HyuPete made their first contribution in #1333
• @earayu made their first contribution in #1329
• @QodiCat made their first contribution in #1330
• @zhudongwork made their first contribution in #1345
• @frederikhendrix made their first contribution in #1341
• @iLuJack made their first contribution in #1359
• @superhuahua made their first contribution in #1375

Full Changelog: v1.3.1...v1.3.2