Hugging Face's logo

Spaces:
Scribbler310
/
gs-port
Sleeping

App Files Community
gs-port / node_modules /@langchain /openai /CHANGELOG.md
Scribbler310's picture
Scribbler310
feat: enhance dashboard
c2b7eb3 verified
preview code
|
raw
history blame contribute delete
26.9 kB

@langchain/openai

1.4.5

Patch Changes

  • #10749 dc20c0e Thanks @open-swe! - fix(openai): add gpt-5.5 and gpt-5.5-pro to profiles, default gpt-5.5-pro to Responses API, bump openai sdk

  • #10776 20a9abe Thanks @hntrl! - fix(deps): remediate uuid vulnerability by removing direct uuid usage

  • Updated dependencies [20a9abe]:

1.4.4

Patch Changes

  • #10681 2301260 Thanks @hntrl! - fix(openai): add index to streaming reasoning content blocks for proper chunk merging

1.4.3

Patch Changes

1.4.2

Patch Changes

1.4.1

Patch Changes

1.4.0

Minor Changes

  • #10509 5552999 Thanks @hntrl! - feat(openai): add support for phase parameter on Responses API messages
    • Extract phase from message output items and surface it on text content blocks
    • Support phase in streaming via response.output_item.added events
    • Round-trip phase through both raw provider and standard content paths
    • Move phase into extras dict in the core standard content translator

Patch Changes

1.3.1

Patch Changes

  • #10481 478652c Thanks @hnustwjj! - fix(openai): detect DeepSeek context overflow errors as ContextOverflowError

    DeepSeek returns maximum context length in 400 error messages when the context limit is exceeded. These are now recognized by wrapOpenAIClientError, so downstream code (e.g. summarization middleware fallback) can handle them correctly.

  • #10507 52e501b Thanks @App-arently! - fix(openai): guard JSON.parse in streaming json_schema when text is empty

  • Updated dependencies [bbbfea1]:

1.3.0

Minor Changes

  • #10317 af9bbd3 Thanks @hntrl! - feat(openai): support tool search and defer_loading
    • Add tools.toolSearch() factory for the OpenAI Responses API tool_search built-in tool, supporting both server-executed and client-executed modes.
    • Propagate defer_loading from LangChain tool extras through bindTools() and into the Responses API payload, enabling on-demand tool discovery.
    • Handle tool_search_call and tool_search_output response items in both streaming and non-streaming converters.
    • Add core block translator support to convert tool_search_call → server_tool_call and tool_search_output → server_tool_call_result.

Patch Changes

1.2.13

Patch Changes

1.2.12

Patch Changes

1.2.11

Patch Changes

  • #10106 9f30267 Thanks @hntrl! - Add package version metadata to runnable traces. Each package now stamps its version in this.metadata.versions at construction time, making version info available in LangSmith trace metadata.

  • #10151 f298a9b Thanks @hntrl! - Bump openai SDK to ^6.24.0, fix ChatCompletionTool type narrowing for new union type, add file input converter tests for newly supported document types (docx, pptx, xlsx, csv)

  • Updated dependencies [9f30267, 403a99f, 3b1fd54, 77bd982]:

1.2.10

Patch Changes

  • #10143 62ba83e Thanks @topliceanurazvan! - fix(openai): emit handleLLMNewToken callback for usage chunk in Completions API streaming

    The final usage chunk in _streamResponseChunks was only yielded via the async generator but did not call runManager.handleLLMNewToken(). This meant callback-based consumers (e.g. LangGraph's StreamMessagesHandler) never received the usage_metadata chunk. Added the missing handleLLMNewToken call to match the behavior of the main streaming loop.

  • Updated dependencies [10a876c, b46d96a]:

1.2.9

Patch Changes

  • #10080 b583729 Thanks @hntrl! - Add string-model constructor overloads for chat models (with supporting tests where applicable).

  • Updated dependencies [fb2226e]:

1.2.8

Patch Changes

  • #10077 05396f7 Thanks @christian-bromann! - feat(core): add ContextOverflowError, raise in anthropic and openai

  • #10081 5a6f26b Thanks @hntrl! - feat(core): add namespace-based symbol branding for error class hierarchies

    Introduces createNamespace utility for hierarchical symbol-based branding of class hierarchies. All LangChain error classes now use this pattern, replacing hand-rolled duck-type isInstance checks with reliable cross-realm Symbol.for-based identity.

    • New LangChainError base class that all LangChain errors extend
    • New createNamespace / Namespace API in @langchain/core/utils/namespace
    • Refactored ModelAbortError, ContextOverflowError to use namespace branding
    • Added ContextOverflowError.fromError() static factory method
    • Deprecated addLangChainErrorFields in favor of LangChainError subclasses
    • Migrated Google provider errors (GoogleError, ConfigurationError, etc.) to namespace branding
    • Updated Anthropic and OpenAI providers to use ContextOverflowError.fromError()

  • #10078 7be50a7 Thanks @christian-bromann! - chore(*): update model profiles

1.2.7

Patch Changes

1.2.6

Patch Changes

  • #9972 16d691c Thanks @hntrl! - fix(openai): drop Anthropic tool_use content blocks when converting messages for OpenAI

    When messages originating from Anthropic (e.g. via ChatAnthropic) are passed to ChatOpenAI, Anthropic-native tool_use blocks in message.content are now filtered out during conversion. These blocks are already represented in message.tool_calls and would cause an OpenAI API error if passed through.

  • #9940 1058574 Thanks @saakshigupta2002! - fix(openai): correctly convert annotations back to OpenAI format in Responses API multi-turn conversations

1.2.5

Patch Changes

  • #9743 0870ca0 Thanks @d2201! - fix(openai): include encrypted reasoning in ZDR responses input

  • #9934 cf46089 Thanks @hntrl! - feat(openai): update openai SDK to ^6.18.0

    • Adds support for codex 5.3
    • Added action option to image generation tool (generate, edit, auto)
    • Removed @ts-expect-error for gpt-image-1.5 model (now in SDK types)
    • Auto-route codex models (codex-mini-latest, gpt-5-codex, gpt-5.1-codex, etc.) to Responses API
    • Added shell_call and local_shell_call to streaming converter and input reconstruction
    • Added unit tests for isReasoningModel and _modelPrefersResponsesAPI

1.2.4

Patch Changes

  • #9887 1fa865b Thanks @Muhammad-Kamran-Khan! - Fix validation to allow file_url and file_id without filename metadata in Responses API, and prevent sending filename when not allowed.

  • #9873 28efb57 Thanks @hntrl! - Add reasoningEffort call option as a convenience shorthand for reasoning.effort

    • Adds reasoningEffort to BaseChatOpenAICallOptions for easier configuration of reasoning models
    • Automatically coalesces reasoningEffort into reasoning.effort when calling reasoning models (o1, o3, etc.)
    • If both reasoningEffort and reasoning.effort are provided, reasoning.effort takes precedence
    • Marked as @deprecated to encourage use of the full reasoning.effort option

  • #9876 4e42452 Thanks @sflanker! - fix(openai): pass runManager to responses._generate function in ChatOpenAI

  • #9900 a9b5059 Thanks @hntrl! - Improved abort signal handling for chat models:

    • Added ModelAbortError class in @langchain/core/errors that contains partial output when a model invocation is aborted mid-stream
    • invoke() now throws ModelAbortError with accumulated partialOutput when aborted during streaming (when using streaming callback handlers)
    • stream() throws a regular AbortError when aborted (since chunks are already yielded to the caller)
    • All provider implementations now properly check and propagate abort signals in both _generate() and _streamResponseChunks() methods
    • Added standard tests for abort signal behavior

  • #9900 a9b5059 Thanks @hntrl! - fix(providers): add proper abort signal handling for invoke and stream operations

    • Added early abort check (signal.throwIfAborted()) at the start of _generate methods to immediately throw when signal is already aborted
    • Added abort signal checks inside streaming loops in _streamResponseChunks to return early when signal is aborted
    • Propagated abort signals to underlying SDK calls where applicable (Google GenAI, Google Common/VertexAI, Cohere)
    • Added standard tests for abort signal behavior in @langchain/standard-tests

    This enables proper cancellation behavior for both invoke and streaming operations, and allows fallback chains to correctly proceed to the next runnable when the previous one is aborted.

1.2.3

Patch Changes

1.2.2

Patch Changes

1.2.1

Patch Changes

1.2.0

Minor Changes

Patch Changes

1.1.3

Patch Changes

1.1.2

Patch Changes

1.1.1

Patch Changes

1.1.0

Minor Changes

  • 8319201: hoist message/tool conversion utilities from classes

Patch Changes

  • 4906522: fix(openai): pair reasoning with function_call id

1.0.0

This release updates the package for compatibility with LangChain v1.0. See the v1.0 release notes for details on what's new.

0.6.16

Patch Changes

  • b8ffc1e: fix(openai): Remove raw OpenAI fields from token usage

0.6.15

Patch Changes

  • e63c7cc: fix(openai): Convert OpenAI responses API usage to tracing format

0.6.14

Patch Changes

  • d38e9d6: fix(openai): fix streaming in openai

0.6.12

Patch Changes

  • 41bd944: support base64 embeddings format
  • 707a768: handle undefined disableStreaming to restore streaming functionality

0.6.11

Patch Changes

  • 65459e3: use proper casing for reasoning effort param

0.6.10

Patch Changes

  • 4a3f5af: add verbosity to json schema response format (#8754)
  • 424360b: re-add reasoning_effort param