langchain-ai/langgraph
1
0
Fork 0
mirror of https://github.com/langchain-ai/langgraph.git synced 2026-07-15 12:35:43 -04:00
Code Issues 314 Packages Projects Releases 17 Wiki Activity

[PR #5296] langgraph[fix]: remove deprecated pydantic logic + fix schema gen behavior for typed dicts #4351

New Issue
Closed
opened 2026-02-20 17:50:03 -05:00 by yindo · 0 comments
yindo commented 2026-02-20 17:50:03 -05:00
Owner
Copy Link
Copy Source

Original Pull Request: https://github.com/langchain-ai/langgraph/pull/5296

State: closed
Merged: Yes

Fixes https://github.com/langchain-ai/langgraph/issues/4360

  1. Remove deprecated use of __fields__
  2. Remove deprecated use of .schema()
  3. Remove deprecated access of model_fields on an instance (instead access on class)
  4. Remove definitions of get_input_schema and get_output_schema from compiled state graph. Instead, migrate to direct json schema gen.
  5. Updated lockfile to use latest pydantic

More on 4:
Previously, if a user used a TypedDict input/output/state schema, we decomposed the typed dict and generated a pydantic model instead. This doesn't work with things like Required / NotRequired which are typed dict specific.

Now, we're directly generating json schema from either pydantic models of type adapters of said typed dicts.

Perf wise, might want to add a cache. Let's see.

Example of new and improved typed dict behavior:

from typing_extensions import TypedDict, Required, NotRequired
from pydantic import TypeAdapter
from pprint import pprint


class Base(TypedDict):
    a: str
    b: str | None

class Sub(Base, total=False):
    c: str
    d: str | None
    e: Required[str]
    f: NotRequired[str]

ta = TypeAdapter(Sub)
pprint(ta.json_schema())

"""
{'properties': {'a': {'title': 'A', 'type': 'string'},
                'b': {'anyOf': [{'type': 'string'}, {'type': 'null'}],
                      'title': 'B'},
                'c': {'title': 'C', 'type': 'string'},
                'd': {'anyOf': [{'type': 'string'}, {'type': 'null'}],
                      'title': 'D'},
                'e': {'title': 'E', 'type': 'string'},
                'f': {'title': 'F', 'type': 'string'}},
 'required': ['a', 'b', 'e'],
 'title': 'Sub',
 'type': 'object'}
"""

In this case:

  • a is required, Base has total=True by default
  • b is required, Base has total=True by default and | None aka Optional doesn't imply NotRequired
  • c is not required bc Sub has total=False
  • d is not required bc Sub has total=False
  • e is required bc of Required
  • f is not required bc of NotRequired
**Original Pull Request:** https://github.com/langchain-ai/langgraph/pull/5296 **State:** closed **Merged:** Yes --- Fixes https://github.com/langchain-ai/langgraph/issues/4360 1. Remove deprecated use of `__fields__` 2. Remove deprecated use of `.schema()` 3. Remove deprecated access of `model_fields` on an instance (instead access on class) 4. Remove definitions of `get_input_schema` and `get_output_schema` from compiled state graph. Instead, migrate to direct json schema gen. 5. Updated lockfile to use latest pydantic More on 4: Previously, if a user used a `TypedDict` input/output/state schema, we decomposed the typed dict and generated a pydantic model instead. This doesn't work with things like `Required` / `NotRequired` which are typed dict specific. Now, we're directly generating json schema from either pydantic models of type adapters of said typed dicts. Perf wise, might want to add a cache. Let's see. Example of new and improved typed dict behavior: ```py from typing_extensions import TypedDict, Required, NotRequired from pydantic import TypeAdapter from pprint import pprint class Base(TypedDict): a: str b: str | None class Sub(Base, total=False): c: str d: str | None e: Required[str] f: NotRequired[str] ta = TypeAdapter(Sub) pprint(ta.json_schema()) """ {'properties': {'a': {'title': 'A', 'type': 'string'}, 'b': {'anyOf': [{'type': 'string'}, {'type': 'null'}], 'title': 'B'}, 'c': {'title': 'C', 'type': 'string'}, 'd': {'anyOf': [{'type': 'string'}, {'type': 'null'}], 'title': 'D'}, 'e': {'title': 'E', 'type': 'string'}, 'f': {'title': 'F', 'type': 'string'}}, 'required': ['a', 'b', 'e'], 'title': 'Sub', 'type': 'object'} """ ``` In this case: * `a` is required, `Base` has `total=True` by default * `b` is required, `Base` has `total=True` by default and `| None` aka `Optional` doesn't imply `NotRequired` * `c` is not required bc `Sub` has `total=False` * `d` is not required bc `Sub` has `total=False` * `e` is required bc of `Required` * `f` is not required bc of `NotRequired`
yindo added the pull-request label 2026-02-20 17:50:03 -05:00
yindo closed this issue 2026-02-20 17:50:03 -05:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
agents
awaiting author revision
bug
change
chat models
checkpointer
ci
code quality
code quality
dependencies
documentation
drawing
duplicate
enhancement
good first issue
help wanted
invalid
maintainer
meta
missing information
not maintained by langchain
open-swe-auto
open-swe-max
open-swe-max-auto
pending
performance
pull-request
Mirrored from GitHub Pull Request
 question
requires design
sdk
server
stale
studio
tests
type checking
usage question
v0.5
v1.1
No Label pull-request
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langchain-ai/langgraph#4351
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?