What's Changed
- experimental[patch]: Release 0.0.56 by @baskaryan in #19840
- docs: remove unnecessary args from the pip install by @cyai in #19823
- Update cross_encoder_reranker.ipynb by @eltociear in #19846
- core: generate mermaid syntax and render visual graph by @angeligareta in #19599
- ai21[patch]: release 0.1.3 by @efriis in #19867
- š„ Update LangChain people data by @jacoblee93 in #19858
- community[patch]: Revert " Fix the bug that Chroma does not specify `eā¦ by @baskaryan in #19866
- openai[patch]: fix azure embedding length check by @efriis in #19870
- Partially Revert "openai[patch]: Update openai chat model to new base class interface" by @nfcampos in #19871
- ai21[patch]: fix core dep by @efriis in #19874
- community[patch]: Release 0.0.31 by @baskaryan in #19873
- community: Update ChatZhipuAI to support GLM-4 model by @zhangch9 in #16695
- openai[patch]: remove openai chunk size validation by @efriis in #19878
- Add OpenVINO rerank model support by @OpenVINO-dev-contest in #19791
- robocorp[patch]: Fix nested arguments descriptors and tool names by @mkorpela in #19707
- robocorp[patch]: fix core min version by @efriis in #19879
- community: Add Dria retriever by @anilaltuner in #17098
- docs[patch]: Revert quarto update by @bracesproul in #19880
- docs: Fix link in Unstructured notebook by @northern-64bit in #19851
- cli[minor]: Add version to integration package template by @eyurtsev in #19876
- langchain: Adding a new section aware splitter to langchain by @msetbar in #16526
- docs: Add docs for RunnableConfigurableFields by @spike-spiegel-21 in #19849
- feat(partners): support request timeout in BaseCohere by @mspronesti in #19641
- docs[patch]: Hide google from function calling docs by @bracesproul in #19887
- community: add Layerup Security integration by @JamsheedMistri in #19787
- Add remove_comments option (default True): do not extract html comments by @petervandenabeele in #13259
- core: Assign missing message ids in BaseChatModel by @nfcampos in #19863
- Core[major]: Base Tracer to propagate raw output from tool for on_tool_end by @keenborder786 in #18932
- core[patch]: Release 0.1.38 by @baskaryan in #19895
- langchain: fix ElasticsearchStore reference for self query by @maxjakob in #19907
- Cohere: Add multihop tool agent by @harry-cohere in #19919
- cohere[patch]: release 0.1.0 by @efriis in #19924
- cohere, docs: update imports and installs to langchain_cohere by @billytrend-cohere in #19918
- cohere: simplify integration test by @harry-cohere in #19928
- cohere: Improve integration test stability, fix documents bug by @harry-cohere in #19929
- docs: mention caveats with CacheBackedEmbeddings.embed_query by @jokester in #19926
- pinecone[patch]: source tag by @efriis in #19739
- core[patch]: remove requests by @efriis in #19891
- deprecating integrations moved to langchain_google_community by @lkuligin in #19841
- docs: update cohere documentation by @sepiatone in #19700
- core: BaseChatModel modify chat message before passing to run_manager by @nfcampos in #19939
- core[patch]: Release 0.1.39 by @baskaryan in #19940
- core: fix return of draw_mermaid_png and change to not save image by default by @angeligareta in #19950
- core[minor]: Add aload to document loader by @eyurtsev in #19936
- langchain-postgres: Initial package with postgres chat history implementation by @eyurtsev in #19884
- Support weight only quantization with intel-extension-for-transformers. by @PenghuiCheng in #14504
- comunity: Implement delete method and all async methods in opensearch_vector_search by @2jimoo in #17321
- Update metadata filtering examples of documents by @tomasonjo in #19963
- core: 0.1.40, fix try_load_from_hub for older langchain versions load_chain by @efriis in #19964
- docs: Custom Document Loaders by @eyurtsev in #19935
- cli[minor]: Add disable sockets in unit tests by @eyurtsev in #19877
- langchain_groq[feat]: Add tool calling support by @gradenr in #19971
- groq: release 0.1.0 by @efriis in #19975
- groq: fix core version by @efriis in #19976
- groq: handle streaming tool call case by @efriis in #19978
- Jacob/docs new by @jacoblee93 in #19765
- cohere: Add citations to agent, flexibility to tool parsing, fix SDK issue by @harry-cohere in #19965
- core[Patch]: mypy ignore fixes #17048 by @UtkarshaGupte in #19931
- core[minor]: Add aformat to FewShotPromptTemplate by @cbornet in #19652
- core: support pydantic V2 for JSONOutputParser, allow for other sources of JSON schemas by @jnis23 in #19716
- [docs][minor]: Fix typo in Custom Document Loader doc by @cbornet in #20003
- community: Implement Async OpenSearch
afrom_texts&afrom_embeddingsby @crispyricepc in #20009 - community[minor]: Add metadata filtering support for neo4j vector by @tomasonjo in #20001
- langchain: enhance
LocalFileStoreto allow directory/file permissions to be specified by @chrispy-snps in #18857 - docs[patch]: Make Docusaurus and Vercel add trailing slashes when navigating by default by @jacoblee93 in #20014
- community[minor]: added missed class to all by @leo-gan in #19888
- anthropic[minor]: tool use by @baskaryan in #20016
- anthropic[patch]: bump core dep by @baskaryan in #20019
- Add cookbook for Anthropic .with_structured_output() by @rlancemartin in #20017
- anthropic[patch]: fix experimental tests by @baskaryan in #20021
- docs:
graphsupdate by @leo-gan in #19675 - docs
integrations/providersupdate 10 by @leo-gan in #19970 - anthropic[patch]: use anthropic 0.23 by @baskaryan in #20022
- anthropic[patch]: Release 0.1.6 by @baskaryan in #20026
- docs:
integrations/providers/unstructuredupdate by @leo-gan in #19892 - docs: mark anthropic tools wrapper as deprecated by @baskaryan in #20024
- docs:
integrations/providersupdate 9 by @leo-gan in #19941 - Update example cookbook for Anthropic tool use by @rlancemartin in #20029
- docs: hide experimental anthropic by @baskaryan in #20030
- docs: fixing typo in argument name by @0ssamaak0 in #20028
- docs[patch]: Fix Model I/O quickstart by @jacoblee93 in #20031
- docs: fix together model tab by @baskaryan in #20032
- docs: weaviate docs by @efriis in #20042
- Docs: Update custom chat model by @eyurtsev in #19967
- docs: fix title cap by @efriis in #20048
- core: Implement aformat_messages for ChatMessagePromptTemplate by @cbornet in #20038
- core: Add async aformat_document method by @cbornet in #20037
- core: Implement aformat_prompt and ainvoke in BasePromptTemplate by @cbornet in #20035
- core[patch]: Document BaseCache abstraction in code by @eyurtsev in #20046
- langchain-core[minor]: Allow passing local cache to language models by @liugddx in #19331
- community[patch]: Improve import callbacks to make it IDE friendly by @eyurtsev in #20050
- docs[patch]: Add missing redirects by @jacoblee93 in #20076
- cohere: move package to external repo by @efriis in #20081
- docs: anthropic tool docstring by @baskaryan in #20091
- template: add rag azure search template by @kristapratico in #18143
- partners[anthropic]: fix anthropic chat model message type lookup keys by @maximeperrindev in #19034
- templates: migrate to langchain_anthropic package to support Claude 3 models by @donbr in #19393
- pinecone[patch]: release 0.1.0 by @efriis in #20109
- Documentation: Fixed the typo of Discord -> Telegram by @TAAGECH9 in #20008
- [core] fix: manually specifying run_id for chat models.invoke() and .ainvoke() by @hinthornw in #20082
- postgres[minor]: add postgres checkpoint implementation by @eyurtsev in #20025
- postgres[minor]: Add pgvector community as is by @eyurtsev in #20096
- community[minor]: Add support for Pebblo cloud_api_key in PebbloSafeLoader by @rahul-trip in #19855
- Community: Updating Azure Retriever and Docs to be Azure AI Search instead of Azure Cognitive Search by @marlenezw in #19925
- community: Add PHP language parser to document_loaders by @david02871 in #19850
- docs: use standard openai params by @baskaryan in #20160
- docs: standardize fireworks params by @baskaryan in #20162
- mistralai[patch]: standardize model params by @baskaryan in #20163
- anthropic[patch]: standardize init args by @baskaryan in #20161
- community: extend Predibase integration to support fine-tuned LLM adapters by @alexsherstinsky in #19979
- langchain: fix pinecone upsert when async_req is set to False by @harryhaibojiang in #19793
- pinecone[patch]: fix core min version by @efriis in #20177
- Adding MongoDB Cookbook for Chat history and semantic cache by @RichmondAlake in #19998
- community: add request_timeout and max_retries to ChatAnthropic by @kaijietti in #19402
- docs: add vertexai to structured output by @baskaryan in #20171
- community:
cross_encodersflatten namespaces by @leo-gan in #20183 - docs: TogetherAI as a drop-in replacement for OpenAI by @sepiatone in #19900
- Community: Add support for MLX models (chat & llm) by @Blaizzy in #18152
- community: add bedrock anthropic callback for token usage counting by @Sukitly in #19864
- Fix pr 19772 by @3coins in #20047
- baichuan[patch]: standardize init args by @liugddx in #20209
- docs: Fix the class links in openai_tools and openai_functions description in output parser documentations by @Haris-Ali007 in #20197
- community[patch]: pass through sql agent kwargs by @baskaryan in #19962
- community: Enhance Tencent Cloud VectorDB, langchain: make Tencent Cloud VectorDB self query retrieve compatible by @jeffkit in #19651
- GCSDirectoryLoader bugfix by @timothywong731 in #20005
- doc:(sharememory&bittensor) Get rid of ZeroShotAgent and use create_react_agent instead by @liugddx in #20157
- community: add allow_dangerous_requests for OpenAPI toolkits by @daviddwlee84 in #19493
- docs: Add documentation of
ElasticsearchStore.BM25RetrievalStrategyby @g-votte in #20098 - openai: wrap stream code in context manager blocks by @snopoke in #18013
- Remove postgres package by @eyurtsev in #20207
- langchain-postgres: Remove remaining README.md file by @eyurtsev in #20221
- together: release 0.1.0 by @efriis in #20225
- standard-tests: a standard unit and integration test set by @efriis in #20182
- core: Implement aformat for FewShotPromptWithTemplates by @cbornet in #20039
- core: Implement aformat_messages for _StringImageMessagePromptTemplate by @cbornet in #20036
- community: switch to falkordb python client by @efriis in #20229
- community[patch]: OpenLLM Async Client Fixes and Timeout Parameter by @charlod in #20007
- langchain[patch]: make BooleanOutputParser check words not substrings by @casperdcl in #20064
- langchain[patch]: Update unit test by @eyurtsev in #20228
- docs: Fix typo in citations example by @sjnarmstrong in #20218
- community: fixed multithreading returning List[List[Documents]] instead of List[Documents] by @chip-davis in #20230
- core[patch]: Release 0.1.41 by @baskaryan in #20233
- community[patch]: Release 0.0.32 by @baskaryan in #20236
- langchain[patch]: Release 0.1.15 by @baskaryan in #20237
- mistralai[patch]: Release 0.1.1 by @baskaryan in #20239
- anthropic[patch]: Release 0.1.7 by @baskaryan in #20240
New Contributors
- @angeligareta made their first contribution in #19599
- @zhangch9 made their first contribution in #16695
- @mkorpela made their first contribution in #19707
- @anilaltuner made their first contribution in #17098
- @spike-spiegel-21 made their first contribution in #19849
- @JamsheedMistri made their first contribution in #19787
- @jokester made their first contribution in #19926
- @PenghuiCheng made their first contribution in #14504
- @UtkarshaGupte made their first contribution in #19931
- @crispyricepc made their first contribution in #20009
- @0ssamaak0 made their first contribution in #20028
- @donbr made their first contribution in #19393
- @TAAGECH9 made their first contribution in #20008
- @rahul-trip made their first contribution in #19855
- @david02871 made their first contribution in #19850
- @harryhaibojiang made their first contribution in #19793
- @RichmondAlake made their first contribution in #19998
- @Blaizzy made their first contribution in #18152
- @Haris-Ali007 made their first contribution in #20197
- @jeffkit made their first contribution in #19651
- @daviddwlee84 made their first contribution in #19493
- @snopoke made their first contribution in #18013
- @charlod made their first contribution in #20007
- @casperdcl made their first contribution in #20064
- @sjnarmstrong made their first contribution in #20218
Full Changelog: v0.1.14...v0.1.15