v1.3.3

Hotfix

• Resolved special character handling issue in PostgreSQL graph storage
• Properly handling of backslashes and quotation marks in entity and relation names

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

v1.3.1

New features for core and server

• Refactoring PostgreSQL AGE graph storage implementation, which introducing breaking changes that are incompatible with existing graph data by @danielaskdd in #1237
• Use startup folder's .env to support multiple instances LightRAG server @ParisNeo in #1185
• Dynamic LLM Selection via QueryParam for Optimized Performance by @omdivyatej in #1167
• Add TokenTracker to track token usage for LLM calls by @choizhang in #1207
• RedisKVStorage: Add Connection Pooling, Error Handling, and Batch Processing by @shmily1012 in #1258
• Add context manager support to simplify token tracking by @choizhang in #1229
• Add client configuration support to OpenAI integration by @shane-lil in #1208
• Optimize config management by @lcjqyml in #1233
• Add FAISS_USE_GPU env var to control which version of FAISS to be installed by @danielaskdd in #1237
• Remove graph storage which lack of support: Gremlin、TiDB graph、MongoDB graph by @danielaskdd in #1237
• Remove Oracle storage completely by @danielaskdd in #1237

New features for WebUI

• Change graph to an undirected one by @danielaskdd in #1284
• Add graph edge thickness by @choizhang in #1242
• Add graph nodes of the same type represented by the same color by @choizhang in #1265
• Add document status filter feature, support filtering documents by status by @ericshao in
• Add file name display in WebUI by @danielaskdd in #1187
• Add sortable columns to document manager by @danielaskdd in #1202
• Add Pipeline status to WebUI by @danielaskdd in #1192
• Add drop all documents in WebUI by @danielaskdd in #1237
• Improve user experience of multiple files upload by @danielaskdd in #1193
• Improve file upload error handling by @danielaskdd in #1214
• Change the label of the edge from type to keyword by @choizhang in #1228
• Add customize title to WebUI @danielaskdd in #1269
• Refactoring LIghtRAG Server setting dialogue by @danielaskdd in #1269

Bugs fixed

• Added validation to prevent unnecessary keyword extraction when explicit keywords are provided by @OxidBurn in #1255
• Fix: auth_mode error in health check end point by @danielaskdd in #1201
• Fix: unknown filepath errors by @JoramMillenaar in #1243
• Fix citation by @LarFii in #1213
• Fix(postgres_impl): added file_path to queries by @Keviddle in #1268
• Fix search label display problem on graph view by @danielaskdd in #1278 #1279
• Fix: tooltip flickering problem by @danielaskdd in #1212
• Remove the comments at the end of the environment variable lines in .env file by @danielaskdd in #1225
• Fix guest login expiration problem by @danielaskdd in #1270

New Contributors

• @omdivyatej made their first contribution in #1167
• @shane-lil made their first contribution in #1208
• @Keviddle made their first contribution in #1268
• @shmily1012 made their first contribution in #1258

Full Changelog: v1.3.0...v1.3.1

v1.3.0

What's Changed

• refactor(SiteHeader): Hide logout button in guest mode by @choizhang in #1161
• Refactor auth and version checks for improved reliability by @danielaskdd in #1162
• Refactor LightRAG Server documentation for clarity and structure. by @danielaskdd in #1166
• Fix mongo doc status by @Gardelll in #1160
• Refactor authentication logic and Swagger Auth config by @danielaskdd in #1171
• Fix API-Key authentication failure while login is enabled by @danielaskdd in #1171
• Fix in-memory db hanged on persistence in some cases by @danielaskdd in #1171

New Contributors

• @Gardelll made their first contribution in #1160

Full Changelog: v1.2.8...v1.3.0

v1.2.8

What's Changed

• Optimize parallel file processing in folder scan by @danielaskdd in #1145
• Fix dead lock problem in Gunicorn multi workers mode by @danielaskdd in #1145
• Fix empty graph not display correctly after database dropped by @danielaskdd in #1150
• Added anthropic binding and few localization to other languages by @ParisNeo in #1148
• Add Arabic and French selection to app settings by @ParisNeo and @danielaskdd in #1152
• Implemented version display in SiteHeader of webui by @danielaskdd in #1153
• Improve node size calculation logic for node expansion to prevent oversize by @danielaskdd in #1156

v1.2.7

What's Changed

• Added camera rotation to graph panel and merge login feature by @danielaskdd in #1108
• Added loginPage by @choizhang in #1049
• Fixed get_by_id type error in PSQL impl by @JoramMillenaar in #1107
• Hot fix: Server can not startup with legacy data (json_doc_status_impl.py) by @danielaskdd in #1115
• Improve login flow by @danielaskdd in #1124
• Improve graph reload by @danielaskdd in #1126
• Fix settings popover toggle logic by @danielaskdd in #1128
• Fix dark mode styles for node property buttons by @danielaskdd in #1132
• Fix: Resolve graph data reloading problem after login token expire by @danielaskdd in #1133
• Update .gitattributes for webui files by @danielaskdd in #1134
• Implemented simple PSQL doc-status get_by_ids method by @JoramMillenaar in #1129
• fix: correct chunk_ids as array type and remove incorrect filepath type conversion by @Howe829 in #1120
• Fix ERROR: Failed to process document doc-xxx: 'metadata' by @LarFii in #1138
• Fixed the issue where editing entities or relationships caused query failures. by @LarFii in #1141
• fix hf_embed torch device use MPS or CPU when CUDA is not available -… by @mvignieri in #1139
• Add ability to conditionally disable file logging by @spo0nman in #1111

New Contributors

• @choizhang made their first contribution in #1049
• @Howe829 made their first contribution in #1120
• @mvignieri made their first contribution in #1139

Full Changelog: v1.2.6...v1.2.7

v1.2.6

What's Changed

• Resolve the issue with making API calls to Azure OpenAI service by @zhicwu in #1043
• force keywords_extraction output as JSON by @zhenya-zhu in #1045
• Improve Entity Extraction Robustness for Truncated LLM Responses by @danielaskdd in #1046
• Filter by ID during Query for Postgres VDB by @ArindamRoy23 in #1032
• Refactor LightRAG for better code organization by @LarFii in #1051
• Fix the merge bug with Neo4j by @LarFii in #1053
• Standardizing Vector Database Interfaces by @LarFii in #1052
• The WebUI has added multi-language support, currently covering Chines… by @baoheping in #1028
• Fix optional query item in all DBs by @ArindamRoy23 in #1060
• Refactor Neo4J storage to use entity_id for node identification by @danielaskdd in #1050
• Add PipelineStatusResponse model for API endpoint by @danielaskdd in #1068
• Fix get single node subgrap problem by @danielaskdd in #1073
• Update .gitignore by @danielaskdd in #1074
• Explicitly set API docs and schema URLs. by @danielaskdd in #1078
• Fixed lint and Added new imports at the top of the file by @RushiChaganti in #1061
• Consolidate the two log filters into a single utility and relocate it to the utils module. by @danielaskdd in #1080
• Fix graspologic build error in docker container by @coredevorg in #1058
• fix stream by @lybtt in #1070
• Implemented several enhancements to the Graph UI by @danielaskdd in #1064
• Replace unreasonable parameter names by @pengjunfeng11 in #1079
• Fix PGDocStatusStorage content extraction for DocProcessingStatus by @ericshao in #1062
• Updated PSQL's chunk_id field to be a TEXT field by @JoramMillenaar in #1091
• Adjust concurrency limits more LLM friendly settings for new comers by @danielaskdd in #1097
• Fix pipeline bactch process problem by @danielaskdd in #1098
• Postgresql - refactor the store entity/relation data by @OxidBurn in #1085
• Fixed Postgres query parsing issues by @JoramMillenaar in #1087
• fix continue prompt format error by @LarFii in #1102
• Update postgres_impl.py by @pengjunfeng11 in #1100
• Adding support for handling file paths in document processing. by @LarFii in #1103

New Contributors

• @ArindamRoy23 made their first contribution in #1032
• @baoheping made their first contribution in #1028
• @RushiChaganti made their first contribution in #1061
• @coredevorg made their first contribution in #1058
• @lybtt made their first contribution in #1070
• @pengjunfeng11 made their first contribution in #1079
• @ericshao made their first contribution in #1062
• @JoramMillenaar made their first contribution in #1091
• @OxidBurn made their first contribution in #1085

Full Changelog: v1.2.5...v1.2.6

v1.2.5

What's Changed

• Add summary language setting by env by @danielaskdd in #989
• Completed the missing methods in various databases to support modifications for the Graph. by @LarFii in #990
• Made .env file loaded from the current folder instead of the script folder by @ParisNeo in #995
• [EVO] Add language configuration to environment and argument parsing by @Brocowlee in #994
• Improved task handling and error management in LightRAG by @danielaskdd in #1002
• Revert "[EVO] Add language configuration to environment and argument … by @LarFii in #1005
• [Bug]Fix the issue where editing entity names does not properly update the corresponding relationship in the vdb. by @LarFii in #1006
• Added entity merging functionality by @LarFii in #1009
• feat(api): Add user authentication functionality by @lcjqyml in #1000
• Feat: Added minimum degree filter for graph queries by @danielaskdd in #1003
• Update dev by @LarFii in #1015
• Fix edit entity and relation bugs by @LarFii in #1018
• fix the postgres get all labels and get knowledge graph by @ShanGor in #1024
• Fixed entites_section_list comma error by @da-luggas in #1021
• Add LLM cache control env for API Server by @danielaskdd in #1025
• Fix qdrant payload id by @dixyes in #1027
• Fixed and improved Iteration extraction by @LarFii in #1030
• Added docling option to load files by @ParisNeo in #1011
• Refactoring Neo4j implementation and fixing storage init problem for Gunicorn by @danielaskdd in #1036

New Contributors

• @Brocowlee made their first contribution in #994
• @lcjqyml made their first contribution in #1000
• @dixyes made their first contribution in #1027

Full Changelog: v1.2.4...v1.2.5

v1.2.4

What's Changed

• Add multi workers support for API Server by @danielaskdd in #969
• adding full_doc_id to insert by @abylikhsanov in #971
• Improved caching check by @MdNazishArmanShorthillsAI in #978
• Add max nodes limit for graph retrieval by @danielaskdd in #974
• Providing setup_logger for standalone LightRAG logger initialization by @danielaskdd in #982
• Fix package name checks for docx and pptx modules. by @danielaskdd in #986
• fix: api server installation missing MANIFEST.in file by @danielaskdd in #987
• Improved file handling and validation for document processing by @danielaskdd in #976

Full Changelog: v1.2.3...v1.2.4

v1.2.3

What's Changed

• Use consistent format as in examples by @zhicwu in #953
• feat: fix delete by document id by @FeHuynhVI in #960

New Contributors

• @zhicwu made their first contribution in #953

Full Changelog: v1.2.2...v1.2.3