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

Pydantic populate_by_name not respected in StateGraph input validation #1042

New Issue
Closed
opened 2026-02-20 17:42:51 -05:00 by yindo · 1 comment
yindo commented 2026-02-20 17:42:51 -05:00
Owner
Copy Link
Copy Source

Originally created by @smart-cau on GitHub (Nov 6, 2025).

Bug Description

Pydantic's populate_by_name=True configuration is not respected during LangGraph's StateGraph input validation, causing validation errors when using field aliases (camelCase) instead of original field names (snake_case).

Environment

  • LangGraph version: 1.0.1
  • Python version: 3.13.5
  • Pydantic version: 2.x

Steps to Reproduce

from pydantic import BaseModel, ConfigDict, Field
from langgraph.graph import StateGraph, START, END

def to_camel(string: str) -> str:
    """Convert snake_case to camelCase."""
    return "".join(
        word.capitalize() if i != 0 else word
        for i, word in enumerate(string.split("_"))
    )

class BaseStateConfig(BaseModel):
    model_config = ConfigDict(
        alias_generator=to_camel,
        populate_by_name=True,  # Should allow both snake_case and camelCase
        serialize_by_alias=True,
    )

class InputState(BaseStateConfig):
    user_name: str = Field(description="User name")
    user_age: int | None = Field(default=None, description="User age")

class OutputState(BaseStateConfig):
    result: str | None = Field(default=None, description="Result")

class GraphState(InputState, OutputState):
    pass

def process_node(state: GraphState) -> dict:
    return {"result": f"Hello, {state.user_name}!"}

# Create graph
workflow = StateGraph(GraphState, input_schema=InputState, output_schema=OutputState)
workflow.add_node("process", process_node)
workflow.add_edge(START, "process")
workflow.add_edge("process", END)
graph = workflow.compile()

# Test 1: snake_case input (works)
result1 = graph.invoke({"user_name": "Alice", "user_age": 30})
print("✅ Test 1 passed:", result1)

# Test 2: camelCase input (fails, but should work due to populate_by_name=True)
result2 = graph.invoke({"userName": "Bob", "userAge": 25})
print("✅ Test 2 passed:", result2)

Expected Behavior

Both Test 1 (snake_case) and Test 2 (camelCase) should succeed, because:

  1. alias_generator=to_camel creates camelCase aliases for all fields
  2. populate_by_name=True allows both original names (snake_case) and aliases (camelCase) during validation

Actual Behavior

Test 2 fails with:

pydantic_core._pydantic_core.ValidationError: 1 validation error for GraphState
userName
  Field required [type=missing, input_value={'userName': 'Bob', 'userAge': 25}, input_type=dict]

The camelCase fields are not recognized, even though Pydantic's populate_by_name=True should enable this.

Root Cause Analysis

The issue appears to be in langgraph/graph/state.py around line 1237 in the _coerce_state function:

def _coerce_state(schema, input):
    return schema(**input)

LangGraph's state coercion logic may be bypassing Pydantic's normal validation flow, which respects populate_by_name. The direct field mapping or optimization in LangGraph's state handling might not honor Pydantic's model_config settings.

Verification: Pydantic Works Correctly

Direct Pydantic validation works as expected:

# This works perfectly:
state1 = InputState(user_name="Alice", user_age=30)  # snake_case
state2 = InputState(userName="Bob", userAge=25)      # camelCase
# Both succeed ✅

The issue only occurs when going through LangGraph's StateGraph.invoke().

Suggested Fix

LangGraph's state validation should respect Pydantic's model_config settings, particularly populate_by_name. This would allow users to:

  1. Define schemas with snake_case (Python convention)
  2. Use camelCase aliases for external APIs (JavaScript/JSON convention)
  3. Accept both formats during validation (thanks to populate_by_name=True)

Workaround

Currently, users must manually convert camelCase to snake_case before passing input to LangGraph:

# Manual conversion required
def camel_to_snake(data: dict) -> dict:
    # Convert all keys from camelCase to snake_case
    ...

graph.invoke(camel_to_snake(input_data))

This workaround defeats the purpose of Pydantic's populate_by_name feature.

Impact

This issue affects any LangGraph application that:

  • Receives input from external systems using different naming conventions (e.g., JavaScript/TypeScript frontends)
  • Wants to use Pydantic's alias features for field name mapping
  • Expects populate_by_name=True to work as documented

Thank you for your consideration!

Originally created by @smart-cau on GitHub (Nov 6, 2025). ## Bug Description Pydantic's `populate_by_name=True` configuration is not respected during LangGraph's `StateGraph` input validation, causing validation errors when using field aliases (camelCase) instead of original field names (snake_case). ## Environment - **LangGraph version**: 1.0.1 - **Python version**: 3.13.5 - **Pydantic version**: 2.x ## Steps to Reproduce ```python from pydantic import BaseModel, ConfigDict, Field from langgraph.graph import StateGraph, START, END def to_camel(string: str) -> str: """Convert snake_case to camelCase.""" return "".join( word.capitalize() if i != 0 else word for i, word in enumerate(string.split("_")) ) class BaseStateConfig(BaseModel): model_config = ConfigDict( alias_generator=to_camel, populate_by_name=True, # Should allow both snake_case and camelCase serialize_by_alias=True, ) class InputState(BaseStateConfig): user_name: str = Field(description="User name") user_age: int | None = Field(default=None, description="User age") class OutputState(BaseStateConfig): result: str | None = Field(default=None, description="Result") class GraphState(InputState, OutputState): pass def process_node(state: GraphState) -> dict: return {"result": f"Hello, {state.user_name}!"} # Create graph workflow = StateGraph(GraphState, input_schema=InputState, output_schema=OutputState) workflow.add_node("process", process_node) workflow.add_edge(START, "process") workflow.add_edge("process", END) graph = workflow.compile() # Test 1: snake_case input (works) result1 = graph.invoke({"user_name": "Alice", "user_age": 30}) print("✅ Test 1 passed:", result1) # Test 2: camelCase input (fails, but should work due to populate_by_name=True) result2 = graph.invoke({"userName": "Bob", "userAge": 25}) print("✅ Test 2 passed:", result2) ``` ## Expected Behavior Both Test 1 (snake_case) and Test 2 (camelCase) should succeed, because: 1. `alias_generator=to_camel` creates camelCase aliases for all fields 2. `populate_by_name=True` allows **both** original names (snake_case) **and** aliases (camelCase) during validation ## Actual Behavior Test 2 fails with: ``` pydantic_core._pydantic_core.ValidationError: 1 validation error for GraphState userName Field required [type=missing, input_value={'userName': 'Bob', 'userAge': 25}, input_type=dict] ``` The camelCase fields are not recognized, even though Pydantic's `populate_by_name=True` should enable this. ## Root Cause Analysis The issue appears to be in `langgraph/graph/state.py` around line 1237 in the `_coerce_state` function: ```python def _coerce_state(schema, input): return schema(**input) ``` LangGraph's state coercion logic may be bypassing Pydantic's normal validation flow, which respects `populate_by_name`. The direct field mapping or optimization in LangGraph's state handling might not honor Pydantic's `model_config` settings. ## Verification: Pydantic Works Correctly Direct Pydantic validation works as expected: ```python # This works perfectly: state1 = InputState(user_name="Alice", user_age=30) # snake_case state2 = InputState(userName="Bob", userAge=25) # camelCase # Both succeed ✅ ``` The issue only occurs when going through LangGraph's `StateGraph.invoke()`. ## Suggested Fix LangGraph's state validation should respect Pydantic's `model_config` settings, particularly `populate_by_name`. This would allow users to: 1. Define schemas with snake_case (Python convention) 2. Use camelCase aliases for external APIs (JavaScript/JSON convention) 3. Accept both formats during validation (thanks to `populate_by_name=True`) ## Workaround Currently, users must manually convert camelCase to snake_case before passing input to LangGraph: ```python # Manual conversion required def camel_to_snake(data: dict) -> dict: # Convert all keys from camelCase to snake_case ... graph.invoke(camel_to_snake(input_data)) ``` This workaround defeats the purpose of Pydantic's `populate_by_name` feature. ## Impact This issue affects any LangGraph application that: - Receives input from external systems using different naming conventions (e.g., JavaScript/TypeScript frontends) - Wants to use Pydantic's alias features for field name mapping - Expects `populate_by_name=True` to work as documented Thank you for your consideration!
yindo closed this issue 2026-02-20 17:42:51 -05:00
yindo commented 2026-02-20 17:42:52 -05:00
Author
Owner
Copy Link
Copy Source

@sydney-runkle commented on GitHub (Nov 7, 2025):

Hiya, this is a dupe of https://github.com/langchain-ai/langgraph/issues/2555, closing as such

@sydney-runkle commented on GitHub (Nov 7, 2025): Hiya, this is a dupe of https://github.com/langchain-ai/langgraph/issues/2555, closing as such
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
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#1042
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?