Commit Graph

14468 Commits

Author SHA1 Message Date
Chester Curme
0e404adc60 🦍 langchain-core==1.0.0a5 2025-09-30 11:49:40 -04:00
Chester Curme
ffaedf7cfc fix 2025-09-30 11:29:39 -04:00
Chester Curme
b6ecc0b040 infra(fix): temporarily skip some release checks 2025-09-30 11:28:40 -04:00
ccurme
4d581000ad fix(infra): temporarily skip pre-release checks for alpha branch (#33160)
We have deliberately introduced a breaking change in
https://github.com/langchain-ai/langchain/pull/33021.

Will revert when we have compatible pre-releases for tested packages.
2025-09-30 11:10:06 -04:00
ccurme
06ddc57c7a release(core): 1.0.0a5 (#33158) 2025-09-30 10:37:06 -04:00
Chester Curme
49704ffc19 Merge branch 'master' into wip-v1.0
# Conflicts:
#	libs/partners/anthropic/langchain_anthropic/chat_models.py
#	libs/partners/anthropic/pyproject.toml
#	libs/partners/anthropic/uv.lock
2025-09-30 09:25:19 -04:00
Sydney Runkle
eed0f6c289 feat(langchain): todo middleware (#33152)
Porting the [planning
middleware](39c0138d0f/src/deepagents/middleware.py (L21))
over from deepagents.

Also adding the ability to configure:
* System prompt
* Tool description

```py
from langchain.agents.middleware.planning import PlanningMiddleware
from langchain.agents import create_agent

agent = create_agent("openai:gpt-4o", middleware=[PlanningMiddleware()])

result = await agent.invoke({"messages": [HumanMessage("Help me refactor my codebase")]})

print(result["todos"])  # Array of todo items with status tracking
```
2025-09-30 02:23:26 +00:00
Mason Daugherty
8926986483 chore: standardize translator named params 2025-09-29 16:42:59 -04:00
ccurme
729637a347 docs(anthropic): document support for memory tool and context management (#33149) 2025-09-29 16:38:01 -04:00
Mason Daugherty
3325196be1 fix(langchain): handle gpt-5 model name in init_chat_model (#33148)
expand to match any `gpt-*` model to openai
2025-09-29 16:16:17 -04:00
Mason Daugherty
f402fdcea3 fix(langchain): add context_management to Anthropic chat model init (#33150) 2025-09-29 16:13:47 -04:00
ccurme
ca9217c02d release(anthropic): 0.3.21 (#33147) langchain-anthropic==0.3.21 2025-09-29 19:56:28 +00:00
ccurme
f9bae40475 feat(anthropic): support memory and context management features (#33146)
https://docs.claude.com/en/docs/build-with-claude/context-editing

---------

Co-authored-by: Mason Daugherty <mason@langchain.dev>
2025-09-29 15:42:38 -04:00
Mason Daugherty
7f757cf37d fix(core): extras handling, add test 2025-09-29 15:41:49 -04:00
ccurme
c20bd07f16 Merge branch 'master' into wip-v1.0 2025-09-29 12:31:29 -04:00
ccurme
839a18e112 fix(openai): remove __future__.annotations import from test files (#33144)
Breaks schema conversion in places.
2025-09-29 16:23:32 +00:00
Mohammad Mohtashim
33a6def762 fix(core): Support of 'reasoning' type in 'convert_to_openai_messages' (#33050) langchain==1.0.0a10 2025-09-29 09:17:05 -04:00
nhuang-lc
c456c8ae51 fix(langchain): fix response action for HITL (#33131)
Multiple improvements to HITL flow:

* On a `response` type resume, we should still append the tool call to
the last AIMessage (otherwise we have a ToolResult without a
corresponding ToolCall)
* When all interrupts have `response` types (so there's no pending tool
calls), we should jump back to the first node (instead of end) as we
enforced in the previous `post_model_hook_router`
* Added comments to `model_to_tools` router so clarify all of the
potential exit conditions

Additionally:
* Lockfile update to use latest LG alpha release
* Added test for `jump_to` behaving ephemerally, this was fixed in LG
but surfaced as a bug w/ `jump_to`.
* Bump version to v1.0.0a10 to prep for alpha release

---------

Co-authored-by: Sydney Runkle <sydneymarierunkle@gmail.com>
Co-authored-by: Sydney Runkle <54324534+sydney-runkle@users.noreply.github.com>
2025-09-29 13:08:18 +00:00
Mason Daugherty
8f60946d5a more docstrings 2025-09-29 00:52:14 -04:00
Mason Daugherty
790c4a8e43 Merge branch 'master' into wip-v1.0 2025-09-28 23:53:54 -04:00
Mason Daugherty
b83d45bff7 fix: add continue when guarding against v0 blocks 2025-09-28 23:53:11 -04:00
Mason Daugherty
ba663defc2 fix: anthropic and converse docstrings 2025-09-28 23:46:17 -04:00
Mason Daugherty
0633974eb0 conversion logic - add missing non-id cases 2025-09-28 23:46:07 -04:00
Mason Daugherty
4d39cf39ff update docstring/comments 2025-09-28 23:44:28 -04:00
Mason Daugherty
77e52a6c9c add docstring 2025-09-28 23:43:44 -04:00
Mason Daugherty
2af1cb6ca3 update conversion docstrings 2025-09-28 23:43:27 -04:00
Mason Daugherty
54f3a6d9cf add note about PlainTextContentBlock and v0/v1 compat 2025-09-28 22:51:17 -04:00
Eugene Yurtsev
54ea62050b chore(langchain_v1): move tool node to tools namespace (#33132)
* Move ToolNode to tools namespace
* Expose injected variable as well in tools namespace
* Update doc-strings throughout
2025-09-26 15:23:57 -04:00
Mason Daugherty
87e5be1097 feat(core): parse reasoning_content from additional_kwargs 2025-09-26 10:50:12 -04:00
Mason Daugherty
fcd8fdd748 revert: remove accidental symlink 2025-09-25 23:48:26 -04:00
Mason Daugherty
370010d195 Merge branch 'master' into wip-v1.0 2025-09-25 20:57:34 -04:00
Mason Daugherty
986302322f docs: more standardization (#33124) 2025-09-25 20:46:20 -04:00
Mason Daugherty
adc941d1dc Merge branch 'master' into wip-v1.0 2025-09-25 20:36:59 -04:00
Mason Daugherty
d15514d571 integrations: delete deprecated items 2025-09-25 18:02:08 -04:00
Mason Daugherty
a5137b0a3e refactor(langchain): resolve pydantic deprecation warnings (#33125) 2025-09-25 17:33:18 -04:00
Mason Daugherty
5bea28393d docs: standardize .. code-block directive usage (#33122)
and fix typos
2025-09-25 16:49:56 -04:00
Mason Daugherty
c3fed20940 docs: correct ported over directives (#33121)
Match rest of repo
2025-09-25 15:54:54 -04:00
Mason Daugherty
6d418ba983 test(mistralai): add xfail for structured output test (#33119)
In rare cases (difficult to reproduce), Mistral's API fails to return
valid bodies, leading to failures from `PydanticToolsParser`
2025-09-25 13:05:31 -04:00
Mason Daugherty
12daba63ff test(openai): raise token limit for o1 test (#33118)
`test_o1[False-False]` was sometimes failing because the OpenAI o1 model
was hitting a token limit with only 100 tokens
2025-09-25 12:57:33 -04:00
Christophe Bornet
eaf8dce7c2 chore: bump ruff version to 0.13 (#33043)
Co-authored-by: Mason Daugherty <mason@langchain.dev>
2025-09-25 12:27:39 -04:00
Mason Daugherty
ee94f9567a Merge branch 'master' into wip-v1.0 2025-09-25 01:47:00 -04:00
Mason Daugherty
f82de1a8d7 chore: bump locks (#33114) 2025-09-25 01:46:01 -04:00
Mason Daugherty
f3f5b93be6 Merge branch 'master' into wip-v1.0 2025-09-25 01:36:00 -04:00
Mason Daugherty
e3efd1e891 test(text-splitters): capture beta warnings (#33113) 2025-09-25 01:30:20 -04:00
Mason Daugherty
d6769cf032 test(text-splitters): resolve pytest marker warning (#33112)
#33111
2025-09-25 01:29:42 -04:00
Mason Daugherty
00565d7bf6 Merge branch 'master' into wip-v1.0 2025-09-25 01:09:44 -04:00
Mason Daugherty
7ab2e0dd3b test(core): resolve pytest marker warning (#33111)
Remove redundant/outdated `@pytest.mark.requires("jinja2")` decorator

Pytest marks (like `@pytest.mark.requires(...)`) applied directly to
fixtures have no effect and are deprecated.
2025-09-25 01:08:54 -04:00
Mason Daugherty
81319ad3f0 test(core): resolve pydantic_v1 deprecation warning (#33110)
Excluded pydantic_v1 module from import testing

Acceptable since this pydantic_v1 is explicitly deprecated. Testing its
importability at this stage serves little purpose since users should
migrate away from it.
2025-09-25 01:08:03 -04:00
Mason Daugherty
e3f3c13b75 refactor(core): use aadd_documents in vectorstores unit tests (#33109)
Don't use the deprecated `upsert()` and `aupsert()`

Instead use the recommended alternatives
2025-09-25 00:57:08 -04:00
Mason Daugherty
c30844fce4 fix(core): use version agnostic get_fields (#33108)
Resolves a warning
2025-09-25 00:54:29 -04:00