Changes since 0.3.34
- version bumps and locks
- test + bug fix
- Support multiple resume values with
Command.resume(#4406) - skip yielding interrupt if input was a map
- lint
- update loop to append mapped tasks to task specific values
- no upper bounds for python
- Merge branch 'main' into multi-resumes
- revert debugging note" "
- fixing tests
- Improve logic for drawing virtual end node (#4409)
- Fix
- Update all
- Lint
- Improve logic for drawing virtual end node
- add convenient interrupt access to StateSnapshot
- linting etc
- use resume instead of resume_map and deprecate old mapping task_id -> resume logic
- multi hitl with new hash pattern
- [breaking]: Improve interrupt behavior when
stream_mode='values'(#4374) - docs updates
- await
- more test fixes
- fix test to reflect bug fix
- Merge branch 'main' into sr/better-interrupts
- more explicit tests
- goodness, last test fix
- final test fixes, hopefully
- linting and test fixes
- Merge branch 'main' into sr/better-interrupts
- Merge branch 'sr/better-interrupts' of https://github.com/langchain-ai/langgraph into sr/better-interrupts
- revert changes to lockfile
- more linting
- Merge branch 'main' into sr/better-interrupts
- try linting + maybe test fix
- linting
- initial pass - surfacing interrupts for stream_mode='values'