TypeError: ForwardRef._evaluate() missing 1 required keyword-only argument: 'recursive_guard'

[midhun-soffy]

Initial Checks

• I have searched GitHub for a duplicate issue and I'm sure this is something new
• I have searched Google & StackOverflow for a solution and couldn't find anything
• I have read and followed the docs and still think this is a bug
• I am confident that the issue is with pydantic (not my code, or another library in the ecosystem like FastAPI or mypy)

Description

I am using for my build
pydantic-2.7.3
pydantic-core-2.18.4
python runtime - 3.12.4

The issue is that ForwardRef._evaluate() has added a new parameter 'recursive_guard' and this parameter is missing in the code pydantic/v1/typing.py (line 66).

This could be solved by the following code snippet starting line 56

if sys.version_info < (3, 9):

def evaluate_forwardref(type_: ForwardRef, globalns: Any, localns: Any) -> Any:
    return type_._evaluate(globalns, localns)

elif sys.version_info < (3, 12, 4):

def evaluate_forwardref(type_: ForwardRef, globalns: Any, localns: Any) -> Any:
    # Even though it is the right signature for python 3.9, mypy complains with
    # `error: Too many arguments for "_evaluate" of "ForwardRef"` hence the cast...
    return cast(Any, type_)._evaluate(globalns, localns, set())

else:
def evaluate_forwardref(type_: ForwardRef, globalns: Any, localns: Any) -> Any:
# fixing error: TypeError: ForwardRef._evaluate() missing 1 required # keyword-only argument: 'recursive_guard'
return cast(Any, type_)._evaluate(globalns, localns, set(), recursive_guard=set())

Example Code

from langchain_google_vertexai import VertexAI, HarmBlockThreshold, HarmCategory

def read():
    llm = VertexAI(model_name="gemini-1.0-pro", safety_settings=safety_settings())

def safety_settings() -> dict:
    return
    {
        HarmCategory.HARM_CATEGORY_UNSPECIFIED: HarmBlockThreshold.BLOCK_NONE,
        HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: HarmBlockThreshold.BLOCK_NONE,
        HarmCategory.HARM_CATEGORY_HATE_SPEECH: HarmBlockThreshold.BLOCK_NONE,
        HarmCategory.HARM_CATEGORY_HARASSMENT: HarmBlockThreshold.BLOCK_NONE,
        HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT: HarmBlockThreshold.BLOCK_NONE,
    }

Python, Pydantic & OS Version

pydantic version: 2.7.2
 pydantic-core version: 2.18.3
 pydantic-core build: profile=release pgo=false
 python version: 3.12.4

Affected Components

• Compatibility between releases
• Data validation/parsing
• Data serialization - .model_dump() and .model_dump_json()
• JSON Schema
• Dataclasses
• Model Config
• Field Types - adding or changing a particular data type
• Function validation decorator
• Generic Models
• Other Model behaviour - model_construct(), pickling, private attributes, ORM mode
• Plugins and integration with other tools - mypy, FastAPI, python-devtools, Hypothesis, VS Code, PyCharm, etc.

[ellyx13]

I am using fastapi and pydantic v1 and got the same error.
Pydantic version: 1.10.11
Fastapi version: 0.111.0

I checked and found it was similar to this issue:
python/cpython#118418

And this commit in python 3.12 do not add the default None value to the recursive_guard parameters
sobolevn/cpython@d7483de

[ellyx13]

My solution was to downgrade python to version 3.12.3 and it worked properly

[midhun-soffy]

My solution was to downgrade python to version 3.12.3 and it worked properly

I have 3.12.2 on my local. However trying to test the deployment locally in docker the google alpha function deploy local always looks for the latest python environment even if I specify a lower runtime (their alpha function bug).

Thanks for fixing this

[sydney-runkle]

Closed by #9612