langchain-xai==0.2.0

Changes since langchain-xai==0.1.1

xai[minor]: release 0.2 (#29132)
openai[minor]: release 0.3 (#29100)
multiple: disable socket for unit tests (#29080)
Revert "integrations[patch]: remove non-required chat param defaults" (#29048)
integrations[patch]: remove non-required chat param defaults (#26730)

langchain-tests==0.3.8

Changes since langchain-tests==0.3.7

tests[patch]: release 0.3.8 (#29141)
openai[minor]: release 0.3 (#29100)
json mode standard test (#25497)

langchain-pinecone==0.2.2

Changes since langchain-pinecone==0.2.1

pinecone[patch]: release 0.2.2 (#29139)
partner: Update aiohttp in langchain pinecone. (#28863)
multiple: disable socket for unit tests (#29080)

langchain-openai==0.3.0

langchain-openai==0.3 implements two breaking changes:

Structured output

We update the default method parameter for ChatOpenAI(...).with_structured_output(method=<method>) from method="function_calling" to method="json_schema".

• "json_schema" (0.3 default) uses OpenAI's dedicated structured output feature to get a structured response
• "function_calling" (0.2 default) uses function calling to get a structured response

For schemas specified via TypedDict or JSON schema, strict schema validation is disabled by default but can be enabled by specifying strict=True.

Note: conceptually there is a difference between forcing a tool call and forcing a response format. Tool calls may have more concise arguments versus generating content adhering to a schema. Prompts may need to be adjusted to recover desired behavior.

How to retain the 0.2 with_structured_output behavior after upgrading to 0.3

To change this behavior back, you can pass method="function_calling" to your with_structured_output calls that you want to switch the behavior back.

Expected errors

1. Models that don’t support method="json_schema" (e.g., gpt-4 and gpt-3.5-turbo, currently the default model for ChatOpenAI) will raise an error unless method is explicitly specified. To recover the previous default, pass method="function_calling" into with_structured_output.

2. Schemas specified via Pydantic BaseModel that have fields with non-null defaults or metadata (like min/max constraints) will raise an error. To recover the previous default, pass method="function_calling" into with_structured_output. See OpenAI's docs for supported schemas.

Optional parameters

We no longer implement non-null defaults for temperature, max_retires, and n, which are optional fields. In particular, we no longer specify a default temperature of 0.7.

The previous defaults can be set by specifying:

• temperature=0.7
• max_retries=2
• n=1

langchain-chroma==0.2.0

Changes since langchain-chroma==0.1.4

multiple: disable socket for unit tests (#29080)
chroma: fix persistence if client_settings is passed in (#25199)
chroma[patch]: Update logic for assigning ids
multiple: combine sync/async vector store standard test suites (#28580)
[VectorStore] Improvement: Improve chroma vector store (#28524)
chroma[patch]: fix bug (#28538)
chroma[patch]: (nit) simplify test (#28517)
chroma[patch]: add get_by_ids and fix bug (#28516)
langchain_chroma: added document.id support (#27995)
partners[chroma]: add retrieval of embedding vectors (#28290)
chroma[patch]: test python 3.13 in CI (#28387)
langchain_chroma: fixed integration tests (#27968)
community: chroma error patch(attribute changed on chroma) (#27827)
chroma[minor]: release 0.2.0 (#27840)
multiple: rely on asyncio_mode auto in tests (#27200)
chroma: Deprecating versions 0.5.7 thru 0.5.12 (#27305)
langchain_chroma[patch]: updated example for get documents with where clause (#26767)
langchain_chroma: Pass through kwargs to Chroma collection.delete (#25970)

langchain-text-splitters==0.3.5

Changes since langchain-text-splitters==0.3.4

text-splitters[patch]: release 0.3.5 (#29054)
infra: speed up unit tests (#28974)
text_splitters: Add HTMLSemanticPreservingSplitter (#25911)
text-splitters: fix state persistence issue in ExperimentalMarkdownSyntaxTextSplitter (#28373)

langchain-pinecone==0.2.1

Changes since langchain-pinecone==0.2.0

pinecone: bump core version (#29077)
pinecone: Review pinecone tests (#29073)
pinecone: release 0.2.1 (version sequence) (#28485)
partners: update version header for Pinecone integration (#28481)
partners/pinecone: release 0.2.2 (#28466)
pinecone: update pinecone client (#28320)
multiple: push deprecation removals to 1.0 (#28236)
many: use core 0.3.15 (#27834)
multiple: rely on asyncio_mode auto in tests (#27200)

langchain-groq==0.2.3

Changes since langchain-groq==0.2.2

partners/groq: release 0.2.3 (#29081)
groq: user agent (#29079)
Revert "integrations[patch]: remove non-required chat param defaults" (#29048)
integrations[patch]: remove non-required chat param defaults (#26730)
partners: fix default value for stop_sequences in ChatGroq (#28924)

langchain-voyageai==0.1.4

Changes since langchain-voyageai==0.1.3

voyageai[patch]: release 0.1.4 (#29046)
partners/voyageai: enable setting output dimension (#28740)

langchain==0.3.14

Changes since langchain==0.3.13

langchain[patch]: release 0.3.14 (#29018)
langchain.agents: corrected documentation (#28986)
infra: speed up unit tests (#28974)
langchain[patch]: fix test following update to langchain-openai (#28838)
langchain: add support for Google Anthropic Vertex AI model garden provider in init_chat_model (#28177)