Changes since langchain-community==0.3.21

community: release 0.3.22 (#30962) community: fix cost calculations for 4.1 and o4 in OpenAI callback (#30899) Community: Add bind variable support for oracle adb docloader (#30937) langchain_community: OpenAIEmbeddings not respecting chunk_size argument (#30946) community: Add deprecation decorator to SingleStore community integrations (#30846) community: support usage_metadata for litellm streaming calls (#30683) community: change metadata in opensearch mmr (#30921) community: redis tool typos fix (#30811) community: Add ruff rule PGH003 (#30812) community[patch]: fix cost calculations for o3 in OpenAI callback (#30807) Community: Add Managed Identity support for Azure AI Search (#30730) feat(community): add oauth2 support for Jira toolkit (#30684) community[patch]: simplify cache logic (#30760) community: deprecate AzureCosmosDBNoSqlVectorSearch in favor of langchain-azure-ai implementation (#30756) community: Google Vertex AI Search now returns the website title as part of the document metadata (#30688) Clean up numpy dependencies and speed up 3.13 CI with numpy>=2.1.0 (#30714) docs: Update couchbase vector store docs (#30710) ChatTongyi reasoning_content fix (#30694) community[patch]: Fix empty producer in PDF Parsers (#30620) community[patch]: Redundant Parser checker for Webbaseloader (#30632) feature: removed pandas dataframe dependency for similary_search when using DuckDB as vector store (#30445) community: Add support for custom runtimes to Riza tools (#30664)