v0.7.0

infiniflow/ragflowv0.7.0May 30, 2024by JinHai-CN

AI Summary

Released on May 30, 2024, this release integrates BCE and BGE reranker models, Jina embedding models, and support for Baichuan and VolcanoArk LLMs. It also introduces RAPTOR support and ARM platform support.

Key Highlights

  • Integrates BCE and BGE reranker models.
  • Integrates Jina embedding and reranker models.
  • Supports LLM Baichuan and VolcanoArk.
  • Supports RAPTOR for better text retrieval.
  • Adds a new API for removing documents.

New Features

  • Supports a new file format: HTML.
  • Supports the ARM platform.
  • Fixes for PDF parser and rerank models.

Full Release Notes

Major updates:

- Integrates [BCE](https://github.com/netease-youdao/BCEmbedding) and [BGE](https://github.com/FlagOpen/FlagEmbedding) reranker models.
- Integrates [Jina](https://jina.ai/embeddings/) embedding and reranker models.
- Supports LLM Baichuan and VolcanoArk.
- Supports [RAPTOR](https://arxiv.org/html/2401.18059v1) for better text retrieval.
- Adds a new API for removing documents.
- Supports a new file format: HTML.
- Supports the ARM platform.

## What's Changed
* Fix missing docker image version prefix v. by @CamusGao in https://github.com/infiniflow/ragflow/pull/855
* Updated Launch service from source by @writinwaters in https://github.com/infiniflow/ragflow/pull/856
* fix: the site domain name in the Chat Bot API is hardcoded. #776 by @cike8899 in https://github.com/infiniflow/ragflow/pull/859
* Reorganized docs for docusaurus publish by @writinwaters in https://github.com/infiniflow/ragflow/pull/860
* fix: display specific error message when previewing file error #868 by @cike8899 in https://github.com/infiniflow/ragflow/pull/869
* fixed a format issue for docusaurus publication by @writinwaters in https://github.com/infiniflow/ragflow/pull/871
* Fixed a format issue by @writinwaters in https://github.com/infiniflow/ragflow/pull/872
* Add api/list_kb_docs function and modify api/list_chunks by @guoyuhao2330 in https://github.com/infiniflow/ragflow/pull/874
* fixed: The choices may be empty. by @dashi6174 in https://github.com/infiniflow/ragflow/pull/876
* fix import error by @yonghuili1 in https://github.com/infiniflow/ragflow/pull/877
* Add api.md:/api/list_kb_docs/ description by @guoyuhao2330 in https://github.com/infiniflow/ragflow/pull/881
* Implements RAPTOR for better chunking #882 by @cike8899 in https://github.com/infiniflow/ragflow/pull/883
* feat: remove the space before promptText by @cike8899 in https://github.com/infiniflow/ragflow/pull/886
* Add /api/document/rm function by @guoyuhao2330 in https://github.com/infiniflow/ragflow/pull/887
* Add support for VolcEngine - the current version supports SDK2 by @yungongzi in https://github.com/infiniflow/ragflow/pull/885
* feat: RAPTOR is not displayed when the parsing method is picture. by @cike8899 in https://github.com/infiniflow/ragflow/pull/897
* fixbug for computing 'not concating feature' by @xinzhuang in https://github.com/infiniflow/ragflow/pull/896
* add raptor by @KevinHuSh in https://github.com/infiniflow/ragflow/pull/899
* Update api_md document/rm by @guoyuhao2330 in https://github.com/infiniflow/ragflow/pull/894
* feat: fixed issue with threshold translation #882 and add NodeContextMenu by @cike8899 in https://github.com/infiniflow/ragflow/pull/906
* Fix some syntax errors, re not import by @dashi6174 in https://github.com/infiniflow/ragflow/pull/904
* Update README by @JinHai-CN in https://github.com/infiniflow/ragflow/pull/901
* Bug fix for volcengine by @yungongzi in https://github.com/infiniflow/ragflow/pull/909
* The fasttext library is missing, and it is used in the operators.py file. by @dashi6174 in https://github.com/infiniflow/ragflow/pull/925
* syntax error by @dashi6174 in https://github.com/infiniflow/ragflow/pull/924
* feat: build react flow nodes and edges from mock data #918 by @cike8899 in https://github.com/infiniflow/ragflow/pull/919
* fix raptor bugs by @KevinHuSh in https://github.com/infiniflow/ragflow/pull/928
* Expanded list of locally deployed embedding models by @writinwaters in https://github.com/infiniflow/ragflow/pull/930
* Remove duplicated FROM. by @CamusGao in https://github.com/infiniflow/ragflow/pull/935
* Updated RESTful API Reference by @writinwaters in https://github.com/infiniflow/ragflow/pull/908
* Update docnm_kwd to doc_name  by @guoyuhao2330 in https://github.com/infiniflow/ragflow/pull/939
* Add Dockerfile and requirements.txt for arm by @aopstudio in https://github.com/infiniflow/ragflow/pull/936
* feat: test buildNodesAndEdgesFromDSLComponents by @cike8899 in https://github.com/infiniflow/ragflow/pull/940
* minor editorial updates for clarity by @writinwaters in https://github.com/infiniflow/ragflow/pull/941
* Added support for Baichuan LLM by @yungongzi in https://github.com/infiniflow/ragflow/pull/934
* fix #917 #915 by @KevinHuSh in https://github.com/infiniflow/ragflow/pull/946
* Add umap-learn, fasttext and volcengine in requirements_arm.txt by @aopstudio in https://github.com/infiniflow/ragflow/pull/945
* fix: fixed the issue of 404 error in the user settings page of the demo site by @cike8899 in https://github.com/infiniflow/ragflow/pull/948
* Fixed a docusaurus display issue by @writinwaters in https://github.com/infiniflow/ragflow/pull/954
* Update README by @JinHai-CN in https://github.com/infiniflow/ragflow/pull/956
* Expanded the supported LLM list by @writinwaters in https://github.com/infiniflow/ragflow/pull/960
* feat: add FlowHeader and delete edge by @cike8899 in https://github.com/infiniflow/ragflow/pull/959
* limit the system context length of conversation messages. by @KevinHuSh in https://github.com/infiniflow/ragflow/pull/962
* feat: add rerank models to the project #724 #162 by @cike8899 in https://github.com/infiniflow/ragflow/pull/966
* feat: add jina by @cike8899 in https://github.com/infiniflow/ragflow/pull/967
* add rerank model by @KevinHuSh in https://github.com/infiniflow/ragflow/pull/969
* fix: if the conversation name is too long, it will overflow the current item. #607 by @cike8899 in https://github.com/infiniflow/ragflow/pull/972
* fix jina adding issure and term weight refinement by @KevinHuSh in https://github.com/infiniflow/ragflow/pull/974
* Update README.md by @writinwaters in https://github.com/infiniflow/ragflow/pull/976
* Update README.md by @writinwaters in https://github.com/infiniflow/ragflow/pull/978
* Fix:After being idle for a while, new tasks need to be cancel and redo by @cklogic in https://github.com/infiniflow/ragflow/pull/958
* Add support for HTML file by @aopstudio in https://github.com/infiniflow/ragflow/pull/973
* fix create dialog bug by @KevinHuSh in https://github.com/infiniflow/ragflow/pull/982
* fix: fixed the issue of error reporting when saving chat configuration #965 by @cike8899 in https://github.com/infiniflow/ragflow/pull/984
* set ollama keep_alive by @KevinHuSh in https://github.com/infiniflow/ragflow/pull/985
* fix bug in pdf parser by @KevinHuSh in https://github.com/infiniflow/ragflow/pull/986
* fix bge rerank normalize issue by @KevinHuSh in https://github.com/infiniflow/ragflow/pull/988
* fix tk_count undefine issue by @KevinHuSh in https://github.com/infiniflow/ragflow/pull/996
* Update README by @JinHai-CN in https://github.com/infiniflow/ragflow/pull/998

## New Contributors
* @CamusGao made their first contribution in https://github.com/infiniflow/ragflow/pull/855
* @yungongzi made their first contribution in https://github.com/infiniflow/ragflow/pull/885
* @xinzhuang made their first contribution in https://github.com/infiniflow/ragflow/pull/896
* @aopstudio made their first contribution in https://github.com/infiniflow/ragflow/pull/936

**Full Changelog**: https://github.com/infiniflow/ragflow/compare/v0.6.0...v0.7.0