Fixed :py:class:aiohttp.TCPConnector doing blocking I/O in the event loop to create the SSLContext -- by :user:bdraco.
The blocking I/O would only happen once per verify mode. However, it could cause the event loop to block for a long time if the SSLContext creation is slow, which is more likely during startup when the disk cache is not yet present.
Improved performance of :py:meth:~aiohttp.ClientWebSocketResponse.receive and :py:meth:~aiohttp.web.WebSocketResponse.receive when there is no timeout. -- by :user:bdraco.
The timeout context manager is now avoided when there is no timeout as it accounted for up to 50% of the time spent in the :py:meth:~aiohttp.ClientWebSocketResponse.receive and :py:meth:~aiohttp.web.WebSocketResponse.receive methods.
Improved performance of HTTP keep-alive checks -- by :user:bdraco.
Previously, when processing a request for a keep-alive connection, the keep-alive check would happen every second; the check is now rescheduled if it fires too early instead.
Fixed multipart reading when stream buffer splits the boundary over several read() calls -- by :user:Dreamsorcerer.
Related issues and pull requests on GitHub:
:issue:8653.
Fixed :py:class:aiohttp.TCPConnector doing blocking I/O in the event loop to create the SSLContext -- by :user:bdraco.
The blocking I/O would only happen once per verify mode. However, it could cause the event loop to block for a long time if the SSLContext creation is slow, which is more likely during startup when the disk cache is not yet present.
Related issues and pull requests on GitHub:
:issue:8672.
Miscellaneous internal changes
Improved performance of :py:meth:~aiohttp.ClientWebSocketResponse.receive and :py:meth:~aiohttp.web.WebSocketResponse.receive when there is no timeout. -- by :user:bdraco.
The timeout context manager is now avoided when there is no timeout as it accounted for up to 50% of the time spent in the :py:meth:~aiohttp.ClientWebSocketResponse.receive and :py:meth:~aiohttp.web.WebSocketResponse.receive methods.
Related issues and pull requests on GitHub:
:issue:8660.
Improved performance of starting request handlers with Python 3.12+ -- by :user:bdraco.
Related issues and pull requests on GitHub:
:issue:8661.
Improved performance of HTTP keep-alive checks -- by :user:bdraco.
Previously, when processing a request for a keep-alive connection, the keep-alive check would happen every second; the check is now rescheduled if it fires too early instead.
langchain[patch]: Release 0.2.12 (#24954)
langchain[minor]: Upgrade ambiguous root_validator to @pre_init (#24895)
langchain[patch]: Update unit tests to workaround a pydantic 2 issue (#24886)
langchain[patch]: Add missing type annotations (#24889)
langchain[patch]: fix ollama in init_chat_model (#24832)
langchain[patch]: init_chat_model() to import ChatOllama from langchain-ollama and fallback on langchain-community (#24821)
langchain: revert "init_chat_model() to support ChatOllama from langchain-ollama" (#24819)
langchain: init_chat_model() to support ChatOllama from langchain-ollama (#24818)
langchain: Make OutputFixingParser.from_llm() create a useable retry chain (#24687)
community[minor]: Self query retriever for HANA Cloud Vector Engine (#24494)
langchain: fix TrajectoryEvalChain.prep_inputs (#19959)
langchain[patch]: init_chat_model support ChatBedrockConverse (#24564)
core[patch]: Release 0.2.29 (#25126)
core[patch]: base language model disable_streaming (#25070)
core[patch]: Relax rate limit unit tests in terms of timing (#25140)
core[patch]: Sort Config attributes (#25127)
core[patch]: Release 0.2.29rc1 (#25115)
core[patch], openai[patch]: enable strict tool calling (#25111)
core[minor]: Add support for DocumentIndex in the index api (#25100)
Set Context in RunnableSequence & RunnableParallel (#25073)
docs: runnable docstring space (#25106)
core[patch]: Include dependencies in sys_info (#25076)
docs: update numbering of items in method docs (#25093)
core[patch]: Remove default BaseModel init docstring (#25009)
core[patch]: Add pydantic metadata to subset model (#25032)
docs: change for optional variables in chatprompt (#25017)
core[patch]: exclude special pydantic init params (#25084)
core[patch]: Relax time constraints on rate limit test (#25071)
core[minor]: Introduce DocumentIndex abstraction (#25062)
docs: show beta directive (#25013)
docs: fix kwargs docstring (#25010)
docs: show deprecation warning first in api ref (#25001)
langchain-core==0.2.29rc1
Changes since langchain-core==0.2.28
core[patch]: Release 0.2.29rc1 (#25115)
core[patch], openai[patch]: enable strict tool calling (#25111)
core[minor]: Add support for DocumentIndex in the index api (#25100)
Set Context in RunnableSequence & RunnableParallel (#25073)
docs: runnable docstring space (#25106)
core[patch]: Include dependencies in sys_info (#25076)
docs: update numbering of items in method docs (#25093)
core[patch]: Remove default BaseModel init docstring (#25009)
core[patch]: Add pydantic metadata to subset model (#25032)
docs: change for optional variables in chatprompt (#25017)
core[patch]: exclude special pydantic init params (#25084)
core[patch]: Relax time constraints on rate limit test (#25071)
core[minor]: Introduce DocumentIndex abstraction (#25062)
docs: show beta directive (#25013)
docs: fix kwargs docstring (#25010)
docs: show deprecation warning first in api ref (#25001)
openai[patch]: Release 0.1.21 (#25269)
partners: fix of issue #24880 (#25229)
infra: temp skip oai embeddings test (#25148)
openai[patch]: Release 0.1.21rc2 (#25146)
openai[patch]: ChatOpenAI.with_structured_output json_schema support (#25123)
openai[patch]: Release 0.1.21rc1 (#25116)
core[patch], openai[patch]: enable strict tool calling (#25111)
patch[Partners] Unified fix of incorrect variable declarations in all check_imports (#25014)
langchain-openai==0.1.21rc2
Changes since langchain-openai==0.1.20
openai[patch]: Release 0.1.21rc2 (#25146)
openai[patch]: ChatOpenAI.with_structured_output json_schema support (#25123)
openai[patch]: Release 0.1.21rc1 (#25116)
core[patch], openai[patch]: enable strict tool calling (#25111)
patch[Partners] Unified fix of incorrect variable declarations in all check_imports (#25014)
langchain-openai==0.1.21rc1
Changes since langchain-openai==0.1.20
openai[patch]: Release 0.1.21rc1 (#25116)
core[patch], openai[patch]: enable strict tool calling (#25111)
patch[Partners] Unified fix of incorrect variable declarations in all check_imports (#25014)
langchain-openai==0.1.20
Changes since langchain-openai==0.1.19
integrations[patch]: release model packages (#24900)
core[patch], integrations[patch]: convert TypedDict to tool schema support (#24641)
openai[patch]: openai proxy added to base embeddings (#24539)
langchain-openai==0.1.19
Changes since langchain-openai==0.1.17
integration releases (#24725)
openai[small]: Add the new model: gpt-4o-mini (#24594)
openai[patch]: move test (#24552)
standard-tests: add override check (#24407)
all: add release notes to pypi (#24519)
core,groq,openai,mistralai,robocorp,fireworks,anthropic[patch]: Update BaseModel subclass and instance checks to handle both v1 and proper namespaces (#24417)
openai: release 0.1.18 (#24369)
openai[patch]: use model_name in AzureOpenAI.ls_model_name (#24366)
NumPy 2.0.1 is a maintenance release that fixes bugs and regressions
discovered after the 2.0.0 release. NumPy 2.0.1 is the last planned
release in the 2.0.x series, 2.1.0rc1 should be out shortly.
The Python versions supported by this release are 3.9-3.12.
NOTE: Do not use the GitHub generated "Source code" files listed in the "Assets", they are garbage.
Improvements
np.quantile with method closest_observation chooses nearest even order statistic
This changes the definition of nearest for border cases from the nearest
odd order statistic to nearest even order statistic. The numpy
implementation now matches other reference implementations.
Bumps the python-requirements group with 16 updates in the / directory:
3.9.53.10.323.2.024.2.00.2.100.2.120.2.90.2.110.2.220.2.290.1.170.1.210.1.930.1.991.26.42.0.11.36.11.40.33.10.63.10.72.20.12.21.06.0.16.0.22024.5.152024.7.242.0.312.0.328.5.09.0.04.66.44.66.5Updates
aiohttpfrom 3.9.5 to 3.10.3Release notes
Sourced from aiohttp's releases.
... (truncated)
Changelog
Sourced from aiohttp's changelog.
... (truncated)
Commits
ef20502Release 3.10.3 (#8675)73d17d4[PR #8676/2915102 backport][3.10] Fix type ignore in SSLContext creation conn...f3fcba4[PR #8672/c3219bf backport][3.10] Fix TCPConnector doing blocking I/O in the ...f96182a[PR #8662/be23d16f backport][3.10] Improve performance of keepalive reschedul...dbcdb16[PR #8667/406cd2c7 backport][3.10] Improve performance of generating random W...b4ad882[PR #8661/4d604ea backport][3.10] Improve performance of starting request han...3a9de0c[PR #8660/14d5295 backport][3.10] Improve performance of WebSockets when ther...1bc8d53[PR #8657/6c6ecfaf backport][3.10] Fix multipart reading with split boundary ...491106eRelease 3.10.2 (#8655)ce2e975[PR #8652/b0536ae6 backport][3.10] Do not follow symlinks for compressed file...Updates
attrsfrom 23.2.0 to 24.2.0Commits
Updates
langchainfrom 0.2.10 to 0.2.12Release notes
Sourced from langchain's releases.
Commits
7b08de8langchain[patch]: Release 0.2.12 (#24954)245cb5acore[patch]: Release 0.2.27 (#24952)199e9c5core[patch]: Fix tool args schema inherited field parsing (#24936)fba65bainfra: test core on py 3.9, 10, 11 (#24951)4092876core: docstrings `BaseCallbackHandler update (#24948)6e45dbadocs: fix redirect (#24950)ad16eedcore[patch]: runnable config ensure_config deep copy from var_child_runnable…...3ab09d8docs[patch]: Adds components for prereqs, compatibility, fix chat model tab i...9cb69a8docs: update retriever template, add arxiv retriever (#24947)db3ceb4partners/mongodb: Improved search index commands (#24745)Updates
langchain-communityfrom 0.2.9 to 0.2.11Release notes
Sourced from langchain-community's releases.
... (truncated)
Commits
8e2316bcommunity[patch]: Release 0.2.11 (#24989)c2538e7experimental[patch]: bump min versions of core and community (#24996)acba38adocs: update toolkit guides (#24992)22c1a40community[patch]: support named arguments in github toolkit (#24986)4797b80experimental[patch]: release 0.0.64 (#24990)7061869Add relik graph transformer (#24982)98c22e9docs: feature table component (#24985)c04d95bstandard-tests: set integration test parameters independent of unit test (#24...54e9ea4fix: Modify the order of init_chat_model import ollama package. (#24977)fe1820cdocs: add wikipedia integration docs (#24932)Updates
langchain-corefrom 0.2.22 to 0.2.29Release notes
Sourced from langchain-core's releases.
... (truncated)
Commits
b4c1234core[patch]: Release 0.2.29 (#25126)dff83cccore[patch]: base language model disable_streaming (#25070)130e80bdocs: rag.ipynb - fixing typo (#25142)09fbce1openai[patch]: ChatOpenAI.with_structured_output json_schema support (#25123)0ba125cdocs: Standardize QianfanLLMEndpoint LLM (#25139)28e0958core[patch]: Relax rate limit unit tests in terms of timing (#25140)a2e9910Documentation Update for Upstash Semantic Caching (#25114)7e7fcf5community: Fix ValidationError on creating GPT4AllEmbeddings with no gpt4all_...04dd8d3Update google_search.ipynb (#25135)63d84e9patch[doc] Fix word spelling error (#25128)Updates
langchain-openaifrom 0.1.17 to 0.1.21Release notes
Sourced from langchain-openai's releases.
Commits
fd54619openai[patch]: Release 0.1.21 (#25269)6dd9f05core[patch]: Deprecating beta upsert APIs in vectorstore (#25069)ca9dceestandard-tests[patch]: test ToolMessage.status="error" (#25210)dadb6f1cli[patch]: Update integration template for embedding models (#25248)b6f0174community[patch],core[patch]: Update EdenaiTool root_validator and add unit t...c3ced4ccore[patch]: use time.monotonic() instead time.time() in InMemoryRateLimiterbd6c316community[patch]: Remove more@allow_reuse=True validators (#25236)6e57aa7community[patch]: Remove usage of@root_validator(allow_reuse=True) (#25235)a2b4c33community[patch]: FAISS: ValueError mentions normalize_score_fn isntead of re...4825dc0langchain[patch]: add deprecations (#24792)Updates
langsmithfrom 0.1.93 to 0.1.99Release notes
Sourced from langsmith's releases.
Commits
d722293chore: bump version (#917)77904d2chore: bump version73aa21efix: pushing new manifest without metadata update (#907)41bf6a4Merge branch 'main' of github.com:langchain-ai/langsmith-sdk into maddy/patch...0b26e82merge50d3d92Accept EvaluationResults with dict (#915)6e09e09prettier281043dtest248bb56change negation to anyce0c8f4just falseUpdates
numpyfrom 1.26.4 to 2.0.1Release notes
Sourced from numpy's releases.
... (truncated)
Commits
4c9f431Merge pull request #27000 from charris/prepare-2.0.10e70e00REL: Prepare for the NumPy 2.0.1 release [wheel build]4d10ffcMerge pull request #26995 from charris/backport-26985764b667BUG: Add object cast to avoid warning with limited API9be6ad6Merge pull request #26971 from charris/backport-269356d950e9BUG: fix f2py tests to work with v2 API89630c0Merge pull request #26962 from charris/backport-2691988fa840TST: Apply test suggestion by Nathan for rlstrip fixesa9da01eBUG,MAINT: Fix utf-8 character stripping memory access6afbbf8Merge pull request #26963 from charris/backport-26930Updates
openaifrom 1.36.1 to 1.40.3Release notes
Sourced from openai's releases.
... (truncated)
Changelog
Sourced from openai's changelog.
... (truncated)
Commits
1a388a1release: 1.40.31d88db0chore(internal): ensure package is importable in lint cmd (#1631)2a87736chore(ci): bump prism mock server version (#1630)4373893chore(ci): codeowners file (#1627)646fff0release: 1.40.2a7f4ba7chore(internal): updates (#1624)ef859c8fix(json schema): support recursive BaseModels in Pydantic v1 (#1623)a9b8aa8chore(internal): format some docstrings5f52e47fix(client): raise helpful error message for response_format misuse631a2a7release: 1.40.1Updates
orjsonfrom 3.10.6 to 3.10.7Release notes
Sourced from orjson's releases.
Changelog
Sourced from orjson's changelog.
Commits
5b951e83.10.7d18f26csse2 stable implbd9dca5cargo update, misc4c05bfaUpdate benchmarks for 3.10.6Updates
pydantic-corefrom 2.20.1 to 2.21.0Release notes
Sourced from pydantic-core's releases.