Releases: HKUDS/LightRAG
Releases · HKUDS/LightRAG
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
Contributors
OxidBurn, ericshao, and 11 other contributors
Assets 2
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
Contributors
ParisNeo, dixyes, and 6 other contributors
Assets 2
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
Contributors
abylikhsanov, danielaskdd, and MdNazishArmanShorthillsAI
Assets 2
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
Full Changelog: v1.2.2...v1.2.3
Contributors
zhicwu and FeHuynhVI
Assets 2
v1.2.2
What's Changed
- fix AttributeError: 'NoneType' object has no attribute 'dim' by @konrad-woj in #931
- Add retrieval message code highlighting and copy button by @ArnoChenFx in #933
- Enhance Graph Visualization with Configurable Depth and Layout Iterations by @ArnoChenFx in #946
- fix this event loop is already running by @FeHuynhVI in #940
- add support for the single document and custom chunks method by @cnjack in #950
- Version 1.2.2 by @YanSte in #952
New Contributors
- @FeHuynhVI made their first contribution in #940
- @cnjack made their first contribution in #950
Full Changelog: v1.2.1...v1.2.2
Contributors
cnjack, YanSte, and 3 other contributors
Assets 2
v1.2.1
What's Changed
- Added more file types support to the webui by @ParisNeo in #897
- Simplify dockerfile and docker-compose files by @ArnoChenFx in #905
- add support of providing ids for documents insert by @PiochU19 in #892
- fix insert_custom_chunks skipping every new doc by @konrad-woj in #906
- Refactor api server by split it to smaller files by @danielaskdd in #880
- Fix the demo issue of PG to cater with new LightRag changes by @ShanGor in #914
- Add support to namespace_prefix for postgres_impl by @ShanGor in #916
- fix apipeline_enqueue_documents by @enriquecatala in #923
- Simplify cli arguments by @danielaskdd in #925
- HotFix: fix document indexing error API Server, fix PostgreSQL error while handle doc with null bytes. by @danielaskdd in #908
- Implement Knowledge Graph API for NetworkX Storage by @danielaskdd in #928
New Contributors
- @PiochU19 made their first contribution in #892
- @konrad-woj made their first contribution in #906
Full Changelog: v1.1.12...v1.2.1
Contributors
ParisNeo, konrad-woj, and 5 other contributors
Assets 2
v1.1.12
v1.1.11
v1.1.10
What's Changed
- Template github by @YanSte in #874
- updated template by @YanSte in #875
- Updated templates by @YanSte in #876
- Removed all useless try except around imports as pipmaster already handles that by @ParisNeo in #873
- cleaned code by @YanSte in #882
- Optimised returns by @YanSte in #883
- Cleanup of code by @YanSte in #885
- removed get_knowledge_graph by @YanSte in #886
- Improved Paralles by @YanSte in #887
- Parelle by @YanSte in #888
- Multi batches by @YanSte in #889
- fixed the behaviour Multi paralle by @YanSte in #890
- added lock by @YanSte in #891
- Add LlamaIndex LLM implementation module by @spo0nman in #864
- Database Cleanup by @YanSte in #898
- Database Cleanup by @YanSte in #899
- Database Cleanup by @YanSte in #900
Full Changelog: v1.1.9...v1.1.10
Contributors
ParisNeo, spo0nman, and YanSte