Skip to content

Conversation

jmaillefaud
Copy link

Description

Currently, the _get_model function to fetch the model for structured output binding does not support DynamicRunnable. Using the provided prepare method, we can simply fetch to configured runnable which can then be processed.

Tests

Added three more asserts to test_get_model.

bajajku and others added 4 commits August 16, 2025 13:53
- Add efficient PyArrow IPC serialization for pandas DataFrames and Series
- Implement intelligent fallback to pickle for complex object types (lists, dicts)
- Support both msgpack and JSON serialization modes
- Preserve all pandas features: MultiIndex, dtypes, structure
- Add comprehensive test coverage (122 tests)
- Add pyarrow>=20.0.0 dependency to langgraph-checkpoint

Performance benefits:
- ~45% smaller serialized size
- Sub-millisecond serialization/deserialization
- Better cross-language compatibility

Handles pandas nuances:
- MultiIndex (both row and column)
- All pandas dtypes including nullable, categorical, datetime
- Arbitrary objects in cells with graceful fallback
- Complete structure preservation
Copy link

vercel bot commented Aug 19, 2025

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Preview Comments Updated (UTC)
langgraph-docs-preview Ready Ready Preview Comment Aug 20, 2025 2:55pm

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants