Releases: langchain-ai/langgraph
Releases · langchain-ai/langgraph
sdk==0.1.45
Changes since sdk==0.1.44
- [SDK] Add auth types
cli==0.1.63
Changes since cli==0.1.62
- [CLI] Add openapi param to config
cli==0.1.62
Changes since cli==0.1.61
- Update dev
- [CLI] Add auth env var
- Add Memory Store conceptual doc section (#2624)
sdk==0.1.44
Changes since sdk==0.1.43
- sdk-py 0.1.44
- sdk-py: Strip out unused keys in command parameter
checkpoint==2.0.9
Changes since checkpoint==2.0.8
- checkpoint: release 2.0.9 (#2744)
- all lines
- lint
- checkpoint: set strict_map_key=False in serde
- Fix
- lib: Performance improvements
- Add sync support for the AsyncPostgresStore (#2673)
- Nicer item repr (#2655)
- Add langgraph.json snippet to concept doc (#2630)
- Clarify behavior in docstring (#2628)
- Add Memory Store conceptual doc section (#2624)
- Fix ref doc formatting (#2623)
- Update docstrings for store classes (#2616)
0.2.59
0.2.58
Changes since 0.2.57
- 0.2.58
- Fix
- lib: Performance improvements
- Update test
- reformat
- x
- qxqx
- Fix
- Update again
- Update test
- lib: Add support for invoke(Command(goto=))
- Fix
- x
- Merge pull request #2695 from langchain-ai/nc/10dec/multistep-plan
- Merge pull request #2691 from langchain-ai/dqbd/enhanced-config-type-extraction
- lib: Add unit test for multistep planner graph
- Fix lint
- Handle PydanticUndefined, add tests
- Move to langgraph.utils.fields
- fix(config): extract default values, description from pydantic models, typeddict and dataclass
0.2.57
Changes since 0.2.56
- Revert "langgraph: release 0.2.58 (#2692)"
- langgraph: release 0.2.58 (#2692)
- langgraph: allow tools to return Command in tool node (#2656)
- lib: Treat Command as "resuming" signal (#2682)
- Merge pull request #2679 from langchain-ai/nc/9dec/imperative-generator
- lib: imperative api: Generators use yield to publish stream_mode=custom events
- Add sync support for the AsyncPostgresStore (#2673)
- 0.2.57
- Merge pull request #2378 from langchain-ai/nc/8nov/send-future
- Fix
- Add tests, missing return stmt
- lib: Support returning mixed list of commands and state updates
- Merge pull request #2658 from langchain-ai/nc/5dec/return-multiple-commands
- Lint
- fix docstring (#2660)
- Lint
- lib: Add support for returning multiple commands from a node
- Merge pull request #2468 from cab938/issue2159
- Merge pull request #2502 from langchain-ai/vb/fix-annotation
- Fix
- Merge branch 'main' into issue2159
- Update
- Update
- Comment
- Wait until next tick to start send task
- Fix
- Fix writes for task being saved against next checkpoint id
- Add test for cancellation
- Lint
- Lint
- Lint
- Lint
- Fix get_state
- Wire up retry policy
- Rename
- Lint
- Fix stream order
- Fix async
- Comment
- Fix
- Lint
- Lint
- Comments
- Get it working with interrupt (sync)
- WIP
- Finish impl
- Add
call
function to call a node and get a future - langgraph: fix issue w/ type annotations in tools_condition
- lint
- feat: Make CompiledGraph displayable in Juypyter with display() (#2159)
0.2.56
Changes since 0.2.55
- lib 0.2.56
- Merge pull request #2635 from langchain-ai/vb/add-graph-command-docs
- cr
- Update libs/langgraph/langgraph/types.py
- Update libs/langgraph/langgraph/types.py
- Update libs/langgraph/langgraph/types.py
- langgraph: remove print (#2640)
- update
- Merge branch 'main' into vb/add-graph-command-docs
- cr
- docs: add Command/GraphCommand docs
sdk==0.1.43
Changes since sdk==0.1.42
- sdk-py 0.1.43
- sdk-py: Handle stream(params=)
- lib: Merge GraphCommand and Command
- Fix typos in langgraph_sdk client. (#2621)